Encyclopedia of Remote Sensing (Encyclopedia of Earth Sciences Series) (2014)

ENCYCLOPEDIA of REMOTE SENSING

Encyclopedia of Earth Sciences Series ENCYCLOPEDIA OF REMOTE SENSING Volume Editor

Eni G. Njoku is a Senior Research Scientist at the Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California, USA. He has a B.A. from the University of Cambridge, and S.M. and Ph.D. from the Massachusetts Institute of Technology. His research focuses on spaceborne microwave sensing with application to land surface hydrology and the global water cycle. Amongst his awards are the NASA Exceptional Service Medal (1985) and Fellow of the Institute of Electrical and Electronics Engineers (1995).

Section Editors

Michael J. Abrams Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA

Vincent V. Salomonson Department of Geography University of Utah Salt Lake City, UT 84112 USA

Ghassem R. Asrar World Climate Research Programme World Meteorological Organization 1211 Geneva Switzerland

Vernon H. Singhroy Canada Centre for Remote Sensing Ottawa Ontario K1A 0Y7 Canada

Frank S. Marzano Department of Information Engineering Sapienza University of Rome 00184 Rome, Italy and Centre of Excellence CETEMPS University of L'Aquila 67100 L'Aquila Italy

F. Joseph Turk Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA

Peter J. Minnett Meteorology and Physical Oceanography Rosenstiel School of Marine and Atmospheric Science University of Miami Miami, FL 33149 USA Aims of the Series The Encyclopedia of Earth Sciences Series provides comprehensive and authoritative coverage of all the main areas in the Earth Sciences. Each volume comprises a focused and carefully chosen collection of contributions from leading names in the subject, with copious illustrations and reference lists. These books represent one of the world’s leading resources for the Earth Sciences community. Previous volumes are being updated and new works published so that the volumes will continue to be essential reading for all professional earth scientists, geologists, geophysicists, climatologists, and oceanographers as well as for teachers and students. See the back of this volume for a current list of titles in the Encyclopedia of Earth Sciences Series. Go to http://www.springerlink.com/reference-works/ and springerreference.com to visit the “Earth Sciences Series” on-line. About the Series Editor Professor Charles W. Finkl has edited and/or contributed to more than eight volumes in the Encyclopedia of Earth Sciences Series. For the past 25 years he has been the Executive Director of the Coastal Education & Research Foundation and Editor-in-Chief of the international Journal of Coastal Research. In addition to these duties, he is Professor at Florida Atlantic University in Boca Raton, Florida, USA. He is a graduate of the University of Western Australia (Perth) and previously worked for a wholly owned Australian subsidiary of the International Nickel Company of Canada (INCO). During his career, he acquired field experience in Australia; the Caribbean; South America; SW Pacific islands; southern Africa; Western Europe; and the Pacific Northwest, Midwest, and Southeast USA. Founding Series Editor Professor Rhodes W. Fairbridge (deceased) has edited more than 24 Encyclopedias in the Earth Sciences Series. During his career he has worked as a petroleum geologist in the Middle East, been a WW II intelligence officer in the SW Pacific and led expeditions to the Sahara, Arctic Canada, Arctic Scandinavia, Brazil and New Guinea. He was Emeritus Professor of Geology at Columbia University and was affiliated with the Goddard Institute for Space Studies.

ENCYCLOPEDIA OF EARTH SCIENCES SERIES

ENCYCLOPEDIA of REMOTE SENSING edited by

ENI G. NJOKU Jet Propulsion Laboratory California Institute of Technology Pasadena, California USA

Library of Congress Control Number: 2013953424

ISBN 978-0-387-36698-2 This publication is available also as: Electronic publication under ISBN 978-0-387-36699-9 and Print and electronic bundle under ISBN 978-0-387-36700-2

Springer New York, Heidelberg, Dordrecht, London

Printed on acid-free paper

Cover photo: Cloud formations over the western Aleuthian Islands, taken by Landsat 7, 1 June 2000. Credit: US Geological Survey, Earth Resources Observation and Science (EROS) Center.

Every effort has been made to contact the copyright holders of the figures and tables which have been reproduced from other sources. Anyone who has not been properly credited is requested to contact the publishers, so that due acknowledgment may be made in subsequent editions.

All rights reserved for the contributions Aerosols; Air Pollution; Atmospheric General Circulation Models; Calibration and Validation; Calibration, Optical/Infrared Passive Sensors; Calibration, Synthetic Aperture Radars; Cloud Properties; Data Processing, SAR Sensors; Earth System Models; Emerging Technologies; Emerging Technologies, Free-Space Optical Communications; Emerging Technologies, Radar; Emerging Technologies, Radiometer; Geodesy; Geomorphology; Geophysical Retrieval, Forward Models in Remote Sensing; Geophysical Retrieval, Inverse Problems in Remote Sensing; Geophysical Retrieval, Overview; GPS, Occultation Systems; Ionospheric Effects on the Propagation of Electromagnetic Waves; Irrigation Management; Land Surface Roughness; Land-Atmosphere Interactions, Evapotranspiration; Lidar Systems; Limb Sounding, Atmospheric; Madden-Julian Oscillation (MJO); Mission Costs of Earth-Observing Satellite; Ocean Surface Topography; Ocean-Atmosphere Water Flux and Evaporation; Precision Agriculture; Reflected Solar Radiation Sensors, Multiangle Imaging; Reflected Solar Radiation Sensors, Polarimetric; Sea Level Rise; Sea Surface Wind/Stress Vector; Solid Earth Mass Transport; Stratospheric Ozone; Terrestrial Snow; Thermal Radiation Sensors (Emitted); Trace Gases, Stratosphere, and Mesosphere; Urban Environments, Beijing Case Study; Volcanism; Water Vapor © Springer Science+Business Media New York 2014 No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work.

Contents

Contributors Preface

xi xxiii

Calibration, Optical/Infrared Passive Sensors Carol Bruegge

47

Calibration, Synthetic Aperture Radars Anthony Freeman

51

Calibration, Scatterometers David Long

54

Climate Data Records Eric F. Wood

58

Climate Monitoring and Prediction Mathew R. P. Sapiano

58

Cloud Liquid Water Fuzhong Weng

68

Cloud Properties Matthew Lebsock and Steve Cooper

70

Coastal Ecosystems Xiaojun Yang

73

Commercial Remote Sensing William Gail

78

Cosmic-Ray Hydrometeorology Darin Desilets and Marek Zreda

83

Cost Benefit Assessment Molly Macauley

86

Acknowledgments

xxv

Acoustic Radiation Alain Weill

1

Acoustic Tomography, Ocean Brian Dushaw

4

Acoustic Waves, Propagation Alain Weill

11

Acoustic Waves, Scattering Alain Weill

13

Aerosols Ralph Kahn

16

Agricultural Expansion and Abandonment Jiaguo Qi

20

Agriculture and Remote Sensing Jerry Hatfield and Susan Moran

22

Air Pollution Annmarie Eldering

32

Atmospheric General Circulation Models Joao Teixeira, Mark Taylor, Anders Persson and Georgios Matheou

35

Calibration and Validation Andreas Colliander

39

Crop Stress Susan Moran

88

Calibration, Microwave Radiometers Christopher Ruf

46

Cryosphere and Polar Region Observing System Mark Drinkwater

91

vi

CONTENTS

Cryosphere, Climate Change Effects Aixue Hu

98

Emerging Technologies, Radiometer Todd Gaier

186

Cryosphere, Climate Change Feedbacks Peter J. Minnett

101

Emerging Technologies, Sensor Web Mahta Moghaddam, Agnelo Silva and Mingyan Liu

190

Cryosphere, Measurements and Applications Roger Barry

104

Environmental Treaties Alexander de Sherbinin

196

Data Access Ron Weaver

119

Fields and Radiation Frank S. Marzano

201

Data Archival and Distribution Mark A. Parsons

121

Fisheries Cara Wilson

202

Data Archives and Repositories Ruth Duerr

127

Forestry Dar Roberts

210

Data Assimilation Dennis McLaughlin

131

Gamma and X-Radiation Enrico Costa and Fabio Muleri

219

Data Policies Ray Harris

134

Geodesy Calvin Klatt

228

Data Processing, SAR Sensors Jakob van Zyl

136

Decision Fusion, Classification of Multisource Data Björn Waske and Jón Atli Benediktsson

Geological Mapping Using Earth’s Magnetic Field Vernon H. Singhroy and Mark Pilkington

140

Geomorphology David Pieri

Earth Radiation Budget, Top-of-Atmosphere Radiation Bing Lin

145

Geophysical Retrieval, Forward Models in Remote Sensing Eugene Ustinov

241

Geophysical Retrieval, Inverse Problems in Remote Sensing Eugene Ustinov

247

Earth System Models Andrea Donnellan

146

Electromagnetic Theory and Wave Propagation Yang Du

150

Emerging Applications William Gail Emerging Technologies Jason Hyon Emerging Technologies, Free-Space Optical Communications Hamid Hemmati

232 237

Geophysical Retrieval, Overview Eugene Ustinov

251

159

Global Climate Observing System Jean-Louis Fellous

254

162

Global Earth Observation System of Systems (GEOSS) Steffen Fritz

257

163

Global Land Observing System Johannes A. Dolman

261

Emerging Technologies, Lidar David M. Tratt

177

Global Programs, Operational Systems Mary Kicza

263

Emerging Technologies, Radar Alina Moussessian

185

GPS, Occultation Systems Chi O. Ao

264

CONTENTS

vii

Ice Sheets and Ice Volume Robert Thomas

269

Microwave Dielectric Properties of Materials Martti Hallikainen

364

Icebergs Donald L. Murphy

281

Microwave Horn Antennas Yahya Rahmat-Samii

375

International Collaboration Lisa Robock Shaffer

284

Microwave Radiometers Niels Skou

382

286

Microwave Radiometers, Conventional Niels Skou

386 389

Irrigation Management Steven R. Evett, Paul D. Colaizzi, Susan A. O’Shaughnessy, Douglas J. Hunsaker and Robert G. Evans

291

Microwave Radiometers, Correlation Christopher Ruf Microwave Radiometers, Interferometers Manuel Martin-Neira

390

Land Surface Emissivity Alan Gillespie

303

Microwave Radiometers, Polarimeters David Kunkee

395

Land Surface Roughness Thomas Farr

311

Microwave Subsurface Propagation and Scattering Alexander Yarovoy

398

Land Surface Temperature Alan Gillespie

314

Microwave Surface Scattering and Emission David R. Lyzenga

403

Land Surface Topography G. Bryan Bailey

320

Mission Costs of Earth-Observing Satellites Randall Friedl and Stacey Boland

405

Land-Atmosphere Interactions, Evapotranspiration Joshua B. Fisher

325

Mission Operations, Science Applications/Requirements David L. Glackin

Landslides Vernon H. Singhroy

328

Observational Platforms, Aircraft, and UAVs Jeffrey Myers

409

Law of Remote Sensing Joanne Irene Gabrynowicz

332

Observational Systems, Satellite David L. Glackin

412

Lidar Systems Robert Menzies

334

Ocean Applications of Interferometric SAR Roland Romeiser

426

Lightning Rachel I. Albrecht, Daniel J. Cecil and Steven J. Goodman

339

Ocean Data Telemetry Michael R. Prior-Jones

429

Limb Sounding, Atmospheric Nathaniel Livesey

344

Ocean Internal Waves Werner Alpers

433

Madden-Julian Oscillation (MJO) Baijun Tian and Duane Waliser

349

Ocean Measurements and Applications, Ocean Color Samantha Lavender

Magnetic Field Nils Olsen

358

Ocean Modeling and Data Assimilation Detlef Stammer

446

Media, Electromagnetic Characteristics Yang Du

362

Ocean Surface Topography Lee-Lueng Fu

455

Ionospheric Effects on the Propagation of Electromagnetic Waves Attila Komjathy

407

437

viii

CONTENTS

Ocean Surface Velocity Bertrand Chapron, Johnny Johannessen and Fabrice Collard

461

Radiation (Natural) Within the Earth’s Environment Anthony England

Ocean, Measurements and Applications Ian Robinson

469

Radiation Sources (Natural) and Characteristics Anthony England

574

Ocean-Atmosphere Water Flux and Evaporation W. Timothy Liu and Xiaosu Xie

480

Radiation, Electromagnetic Frank S. Marzano

576

Operational Transition Richard Anthes

489

Radiation, Galactic, and Cosmic Background David M. Le Vine

581

492

Radiation, Multiple Scattering Frank S. Marzano

585

495

Radiation, Polarization, and Coherence Yang Du

588

Radiation, Solar and Lunar David M. Le Vine

591

Radiation, Volume Scattering Leung Tsang and Kung-Hau Ding

595

Radiative Transfer, Solution Techniques Rodolfo Guzzi

606

Radiative Transfer, Theory Frank S. Marzano

624

Optical/Infrared, Atmospheric Absorption/ Transmission, and Media Spectral Properties Gian Luigi Liberti Optical/Infrared, Radiative Transfer Knut Stamnes Optical/Infrared, Scattering by Aerosols and Hydrometeors Gian Luigi Liberti

498

558

Pattern Recognition and Classification Björn Waske and Jón Atli Benediktsson

503

Polar Ice Dynamics James Maslanik

509

Polar Ocean Navigation Lawson Brigham

512

Policies and Economics Roberta Balstad

515

Radio-Frequency Interference (RFI) in Passive Microwave Sensing David Kunkee

Precision Agriculture Kelly Thorp

515

Rainfall Ralph Ferraro

640

Processing Levels Ron Weaver

517

Rangelands and Grazing Hunt E. Raymond, Jr.

653

Public-Private Partnerships William Gail

520

Radar, Altimeters Keith Raney

525

Reflected Solar Radiation Sensors, Polarimetric David J. Diner

663

Radar, Scatterometers David Long

532

Reflector Antennas Yahya Rahmat-Samii

668

Radar, Synthetic Aperture Keith Raney

536

Remote Sensing and Geologic Structure Vernon H. Singhroy and Paul Lowman

681

Radars Keith Raney

547

Remote Sensing, Historical Perspective Vincent V. Salomonson

684

Reflected Solar Radiation Sensors, Multiangle Imaging David J. Diner

634

658

CONTENTS

ix

Terrestrial Snow Son V. Nghiem, Dorothy K. Hall, James L. Foster and Gregory Neumann

821

Thermal Radiation Sensors (Emitted) Simon Hook

830

719

Trace Gases, Stratosphere, and Mesosphere Nathaniel Livesey

834

Sea Ice Albedo Donald Perovich

722

Trace Gases, Troposphere - Detection from Space 838 Pieternel F. Levelt, J. P. Veefkind and K. F. Boersma

Sea Ice Concentration and Extent Josefino C. Comiso

727

Trafficability of Desert Terrains Charles Hibbitts

846

Sea Level Rise Josh Willis

743

Tropospheric Winds Chris Velden

849

Sea Surface Salinity Gary Lagerloef

747

Ultraviolet Remote Sensing Arlin Krueger

853

Sea Surface Temperature Peter J. Minnett

754

Ultraviolet Sensors Arlin Krueger

860

Sea Surface Wind/Stress Vector W. Timothy Liu and Xiaosu Xie

759

Urban Environments, Beijing Case Study 869 Son V. Nghiem, Alessandro Sorichetta, Christopher D. Elvidge, Christopher Small, Deborah Balk, Uwe Deichmann and Gregory Neumann

Severe Storms Charles A. III Doswell

767

Snowfall Ralf Bennartz

Remote Sensing, Physics and Techniques David L. Glackin

691

Resource Exploration Fred A. Kruse and Sandra L. Perry

702

SAR-Based Bathymetry Han Wensink and Werner Alpers

Urban Heat Island Lela Prashad

878

780

Vegetation Indices Alfredo Huete

883

Soil Moisture Yann Kerr

783

Vegetation Phenology John Kimball

886

Soil Properties Alfredo Huete

788

Volcanism Michael J. Abrams

890

Solid Earth Mass Transport Erik Ivins

791

Water and Energy Cycles Taikan Oki and Pat J.-F. Yeh

895

Stratospheric Ozone Michelle Santee

796

Water Resources Taikan Oki and Pat J.-F. Yeh

903

Subsidence Stuart Marsh and Martin Culshaw

800

Water Vapor Eric Fetzer

909

Surface Radiative Fluxes Rachel T. Pinker

806

Weather Prediction Peter Bauer

912 921

Surface Truth Christopher Ruf

815

Wetlands John Melack Author Index

923

Surface Water Michael Durand

816

Subject Index

925

Contributors

Michael J. Abrams Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Rachel I. Albrecht Divisão de Satélites e Sistemas Ambientais (DSA/CPTEC), Instituto Nacional de Pesquisas Espaciais (INPE) 12630-000 Cachoeira Paulista, SP Brazil [email protected] Werner Alpers Institute of Oceanography University of Hamburg 20146 Hamburg Germany [email protected] Richard Anthes University Corporation for Atmospheric Research Boulder, CO 80301 USA [email protected] Chi O. Ao Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected]

G. Bryan Bailey USGS Earth Resources Observation and Science Center Sioux Falls, SD 57198 USA [email protected] Deborah Balk School of Public Affairs, Baruch College City University of New York New York, NY USA [email protected] Roberta Balstad CIESIN Columbia University Palisades, NY 10964 USA [email protected] Roger Barry National Snow and Ice Data Center NSIDC 449 UCB University of Colorado Boulder, CO 80309-0449 USA [email protected] Peter Bauer European Centre for Medium-Range Weather Forecasts (ECMWF) Shinfield Park Reading RG2 9AX UK [email protected]

xii

CONTRIBUTORS

Jón Atli Benediktsson Faculty of Electrical and Computer Engineering University of Iceland 107 Reykjavik Iceland [email protected] Ralf Bennartz Atmospheric and Oceanic Sciences Department University of Wisconsin-Madison Madison, WI 53706-1481 USA [email protected] K. F. Boersma Koninklijk Nederlands Meteorologisch Instituut (KNMI) 3732 GK, De Bilt The Netherlands and Technical University Eindhoven (TUE) 5612 AZ, Eindhoven The Netherlands [email protected] Stacey Boland Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA Lawson Brigham University of Alaska Fairbanks, AK 99775-5840 USA [email protected] Carol Bruegge Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Daniel J. Cecil Marshall Space Flight Center (MSFC), National Aeronautics and Space Administration (NASA) Huntsville, AL 35805 USA [email protected] Bertrand Chapron Satellite Oceanography Laboratory, IFREMER Plouzané 29280 France [email protected]

Paul D. Colaizzi USDA-ARS Conservation and Production Research Laboratory Bushland, TX 79012 USA [email protected] Fabrice Collard CLS, Division Radar Plouzané 29280 France [email protected] Andreas Colliander Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Josefino C. Comiso Cryospheric Sciences Laboratory, Code 615 Earth Sciences Division, NASA Goddard Space Flight Center Greenbelt, MD 20771 USA [email protected] Steve Cooper Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA Enrico Costa Istituto di Astrofisica e Planetologia Spaziali, INAF 00133 Rome Italy [email protected] Martin Culshaw Honorary Research Associate, British Geological Survey, Keyworth Nottingham NG12 1AE UK and Honorary Visiting Professor, School of Civil Engineering, University of Birmingham, Edgbaston Birmingham B15 2TT UK Alexander de Sherbinin Center for International Earth Science Information Network (CIESIN) Columbia University Palisades, NY 10964 USA [email protected]

CONTRIBUTORS

Uwe Deichmann Development Research Group, The World Bank Washington, DC USA [email protected] Darin Desilets Hydroinnova LLC Albuquerque, NM 87106 USA [email protected] David J. Diner Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Kung-Hau Ding Air Force Research Laboratory Wright-Patterson AFB Dayton, OH 45433 USA Johannes A. Dolman Department of Earth Sciences VU University Amsterdam 1081 Amsterdam The Netherlands [email protected] Andrea Donnellan Science Division Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Charles A. III Doswell Doswell Scientific Consulting Norman, OK 73071 USA [email protected] Mark Drinkwater Mission Science Division European Space Agency, ESA/ESTEC 2201 AZ Noordwijk ZH The Netherlands [email protected] Yang Du Zhejiang University 310027 Hangzhou People’s Republic of China [email protected]

Ruth Duerr National Snow and Ice Data Center, CIRES 449 UCB, University of Colorado Boulder, CO 80309 USA [email protected] Michael Durand School of Earth Sciences The Ohio State University 275 Mendenhall Laboratory Columbus, OH 43210 USA [email protected] Brian Dushaw Applied Physics Laboratory University of Washington Seattle, WA 98105-6698 USA [email protected] Annmarie Eldering Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Christopher D. Elvidge Earth Observation Group, NOAA-NESDIS National Geophysical Data Center E/GC2 Boulder, CO USA [email protected] Anthony England College of Engineering University of Michigan Ann Arbor, MI 48109 USA [email protected] Robert G. Evans USDA-ARS Sidney, MT 59270 USA [email protected] Steven R. Evett USDA-ARS Conservation and Production Research Laboratory Bushland, TX 79012 USA [email protected]

xiii

xiv

CONTRIBUTORS

Thomas Farr Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Jean-Louis Fellous Committee on Space Research (COSPAR) Secretariat c/o CNES-2, place Maurice Quentin 75039 Paris France [email protected] Ralph Ferraro NOAA/NESDIS, ESSIC/CICS College Park, MD 20740 USA [email protected] Eric Fetzer Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Joshua B. Fisher Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] James L. Foster Hydrological Sciences Laboratory, Code 617 National Aeronautics and Space Administration Goddard Space Flight Center Greenbelt, MD USA [email protected] Anthony Freeman Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Randall Friedl Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected]

Steffen Fritz International Institute for Applied Systems Analysis 2361 Laxenburg Austria [email protected] Lee-Lueng Fu Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Joanne Irene Gabrynowicz National Center for Remote Sensing, Air, and Space Law The University of Mississippi School of Law Missisippi, MS 38677-1848 USA [email protected] Todd Gaier Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] William Gail Global Weather Corporation Boulder, CO 80303 USA [email protected] Alan Gillespie Department of Earth and Space Sciences University of Washington Seattle, WA 98195 USA [email protected] David L. Glackin Los Angeles, CA USA Steven J. Goodman National Environmental Satellite, Data, and Information Service (NESDIS), National Oceanic and Atmospheric Administration (NOAA) Silver Spring, MD 20910 USA [email protected] Rodolfo Guzzi Agenzia Spaziale Italiana ASI 00133 Roma Italy [email protected]

CONTRIBUTORS

Dorothy K. Hall Cryospheric Sciences Laboratory, Code 615 NASA/Goddard Space Flight Center Greenbelt, MD 20771 USA [email protected]

Alfredo Huete Plant Functional Biology and Climate Change Cluster Faculty of Science University of Technology 2007 Sydney, NSW Australia [email protected]

Martti Hallikainen Aalto University 00076 Aalto Espoo Finland [email protected]

Douglas J. Hunsaker USDA-ARS Maricopa, AZ 85138 USA [email protected]

Ray Harris Department of Geography University College London London WC1E 6BT UK [email protected] Jerry Hatfield National Laboratory for Agriculture and the Environment Ames, IA 50011 USA [email protected] Hamid Hemmati Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Charles Hibbitts Applied Physics Laboratory Laurel, MD 20723 USA [email protected] Simon Hook Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Aixue Hu Climate and Global Dynamics Division National Center for Atmospheric Research Boulder, CO 80305 USA [email protected]

xv

E. Raymond Hunt, Jr. USDA-ARS Hydrology and Remote Sensing Laboratory Beltsville, MD 20705 USA [email protected] Jason Hyon Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Erik Ivins Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 9109 USA [email protected] Johnny Johannessen Nansen Environmental and Remote Sensing Center 5006 Bergen Norway [email protected] Ralph Kahn NASA Goddard Space Flight Center Greenbelt, MD 20771 USA [email protected] Yann Kerr CNES/CESBIO 31401 Toulouse France [email protected] Mary Kicza National Oceanic and Atmospheric Administration (NOAA) Washington, DC 20230 USA [email protected]

xvi

CONTRIBUTORS

John Kimball Flathead Lake Biological Station University of Montana Polson, MT 59860-6815 USA [email protected]

David M. Le Vine Code 615, Cryopsheric Sciences Branch NASA/Goddard Space Flight Center Greenbelt, MD 20771 USA [email protected]

Calvin Klatt Geodetic Survey Division Natural Resources Canada Ottawa, ON K1A 0E9 Canada [email protected]

Matthew Lebsock Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected]

Attila Komjathy Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Arlin Krueger Atmospheric Chemistry and Dynamics Laboratory (Code 614) NASA/Goddard Space Flight Center Greenbelt, MD 20771 USA [email protected] Fred A. Kruse Physics Department and Remote Sensing Center Naval Postgraduate School Monterey, CA 93943 USA [email protected] David Kunkee The Aerospace Corporation Los Angeles, CA 90009 USA [email protected] Gary Lagerloef ESR Seattle, WA 98121 USA [email protected] Samantha Lavender Pixalytics Ltd Plymouth, Devon PL6 8BX UK [email protected]

Pieternel F. Levelt Koninklijk Nederlands Meteorologisch Instituut (KNMI) 3730 AE De Bilt The Netherlands and Delft University of Technology 5612 AE Eindhoven The Netherlands [email protected] Gian Luigi Liberti CNR/ISAC 00133 Rome Italy [email protected] Bing Lin NASA Langley Research Center, MS 420 Hampton, VA 23681-2199 USA [email protected] Mingyan Liu Electrical and Computer Engineering University of Michigan Ann Arbor, MI 48109 USA W. Timothy Liu Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Nathaniel Livesey Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected]

CONTRIBUTORS

David Long Department of Electrical and Computer Engineering BYU Center for Remote Sensing Brigham Young University Provo, UT 84602 USA [email protected] Paul Lowman NASA Goddard, Code 698.0 Greenbelt, MD 20771 USA [email protected] David R. Lyzenga College of Engineering, Naval Architecture and Marine Engineering University of Michigan Ann Arbor, MI 48109-2145 USA [email protected] Molly Macauley Resources for the Future Washington, DC 202-328-5043 USA [email protected] Stuart Marsh Nottingham Geospatial Institute The University of Nottingham Nottingham Geospatial Building, Triumph Road Nottingham NG7 2TU UK [email protected] Manuel Martin-Neira European Space Agency (ESA-ESTEC) Keplerlaan 1 2200 Noordwijk The Netherlands [email protected] Frank S. Marzano Department of Information Engineering Sapienza University of Rome 00184 Rome Italy and Centre of Excellence CETEMPS University of L'Aquila 67100 L'Aquila Italy [email protected]

xvii

James Maslanik Department of Aerospace Engineering Sciences University of Colorado, CCAR Boulder, CO 80309 USA [email protected] Georgios Matheou Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Dennis McLaughlin Department of Civil and Environmental Engineering Massachusetts Institute of Technology Cambridge, MA 02139 USA [email protected] John Melack Department of Ecology, Evolution and Marine Biology University of California Santa Barbara, CA 93106 USA [email protected] Robert Menzies Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Peter J. Minnett Rosenstiel School of Marine and Atmospheric Science University of Miami Miami, FL 33149 USA [email protected] Mahta Moghaddam Electrical Engineering – Electrophysics University of Southern California Los Angeles, CA 0089 USA [email protected] Susan Moran USDA ARS Southwest Watershed Research Center Tuscon, AZ 85719 USA [email protected]

xviii

CONTRIBUTORS

Alina Moussessian Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected]

Susan A. O’Shaughnessy USDA-ARS Conservation and Production Research Laboratory Bushland, TX 79012 USA [email protected]

Fabio Muleri Istituto di Astrofisica e Planetologia Spaziali, INAF 00133 Rome Italy [email protected]

Mark A. Parsons Center for a Digital Society Rensselaer Polytechnic Institute Troy, NY 12180 USA [email protected]

Donald L. Murphy International Ice Patrol, US Coast Guard New London, CT 06320 USA [email protected] Jeffrey Myers NASA/Ames Research Center, Airborne Science and Technology Laboratory University of California, Santa Cruz MS244-15 Moffett Field, CA 94035 USA [email protected] Gregory Neumann Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Son V. Nghiem Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected]

Donald Perovich USACE Cold Regions Research and Engineering Laboratory Hanover, NH 03755-1250 USA [email protected] Sandra L. Perry Perry Remote Sensing, LLC Denver, CO 80231 USA [email protected] Anders Persson United Kingdom Meteorological Office Exeter Devon, EX1 3PB UK David Pieri Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected]

Taikan Oki Institute of Industrial Science, University of Tokyo 153-8505 Tokyo Japan [email protected]

Mark Pilkington Geological Survey of Canada Ottawa, ON K1A 0E9 Canada [email protected]

Nils Olsen DTU Space Technical University of Denmark 2800 Kgs. Lyngby Denmark [email protected]

Rachel T. Pinker Department of Atmospheric and Oceanic Science University of Maryland College Park, MD 20742 USA [email protected]

CONTRIBUTORS

xix

Lela Prashad School of Earth and Space Exploration, 100 Cities Project Arizona State University Tempe, AZ 85287-1404 USA [email protected]

Christopher Ruf Department of Atmospheric, Oceanic and Space Sciences University of Michigan Ann Arbor, MI 48109 USA [email protected]

Michael R. Prior-Jones British Antarctic Survey Cambridge CB3 OET UK [email protected]

Vincent V. Salomonson Department of Geography, University of Utah South Jordan, UT 84095 USA [email protected]

Jiaguo Qi Department of Geography/CGCEO Michigan State University East Lansing, MI 48823 USA [email protected]

Michelle Santee Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected]

Yahya Rahmat-Samii Department of Electrical Engineering University of California at Los Angeles Los Angeles, CA 90095 USA [email protected]

Mathew R. P. Sapiano University of Maryland - College Park College Park, MD 20742 USA [email protected]

Keith Raney Applied Physics Laboratory Johns Hopkins University Laurel, MD 20723 USA [email protected]

Lisa Robock Shaffer MC 0553 Rady School of Management University of California, San Diego La Jolla, CA 92093-0553 USA [email protected]

Dar Roberts Department of Geography University of California Santa Barbara, CA 93106 USA [email protected]

Agnelo Silva Electrical Engineering – Electrophysics University of Southern California Los Angeles, CA 0089 USA

Ian Robinson Ocean and Earth Science University of Southampton, at National Oceanography Centre Southampton SO14 3ZH UK [email protected]

Vernon H. Singhroy Applications Development Section Natural Resources Canada Canada Centre for Remote Sensing Ottawa, ON K1A 0Y7 Canada [email protected]

Roland Romeiser Rosenstiel School of Marine and Atmospheric Science University of Miami Miami, FL 33149-1031 USA [email protected]

Niels Skou National Space Institute Technical University of Denmark 2800 Lyngby Denmark [email protected]

xx

CONTRIBUTORS

Christopher Small Lamont Doherty Earth Observatory Marine Geology and Geophysics Columbia University Palisades, NY USA [email protected] Alessandro Sorichetta Dipartimento di Scienze della Terra “A. Desio” Universita' degli Studi di Milano 20122 Milan Italy [email protected] Detlef Stammer Institut für Meereskunde, Zentrum für Marine und Atmosphärische Wissenschaften Universität Hamburg 20146 Hamburg Germany [email protected] Knut Stamnes Stevens Institute of Technology, Castle Point on Hudson Hoboken, NJ 07030-5991 USA [email protected] Mark Taylor Sandia National Laboratory Albuquerque, New Mexico 91109 USA [email protected] Joao Teixeira Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected]

Baijun Tian Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] David M. Tratt The Aerospace Corporation Los Angeles, CA 90009-2957 USA [email protected] Leung Tsang Paul Allen Center Department of Electrical Engineering University of Washington Seattle, WA 98195-2500 USA [email protected] Eugene Ustinov Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Jakob van Zyl Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] J. P. Veefkind Koninklijk Nederlands Meteorologisch Instituut (KNMI) 3732 GK, De Bilt The Netherlands and Eindhoven University of Technology 5612 AE Eindhoven The Netherlands [email protected]

Robert Thomas Sigma Space 66-400 Gorzow Wlkp Poland [email protected]

Chris Velden University of Wisconsin, CIMSS Madison, WI 53706 USA [email protected]

Kelly Thorp USDA-ARS U.S. Arid-Land Agricultural Research Center Maricopa, AZ 85138 USA [email protected]

Duane Waliser Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 9109 USA [email protected]

CONTRIBUTORS

Björn Waske Institute of Geodesy and Geoinformation University of Bonn 53115 Bonn Germany [email protected] Ron Weaver National Snow and Ice Data Center, Cooperative Institute for Research in Environmental Sciences University of Colorado Boulder, CO 80309-0449 USA [email protected] Alain Weill Bur. Jussieu LATMOS, Laboratoire Atmosphere Milieux Observations Spatiales 75005 Paris France [email protected] Fuzhong Weng Center for Satellite Applications and Research (STAR) National Oceanic and Atmospheric Administration College Park, MD 20740 USA [email protected] Han Wensink ARGOSS BV 8325ZH Vollenhove The Netherlands [email protected] Josh Willis Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected]

Cara Wilson Southwest Fisheries Science Center NOAA/NMFS, Environmental Research Division Pacific Grove, CA 93950-2097 USA [email protected] Eric F. Wood Department of Civil and Environmental Engineering Princeton University Princeton, NJ 08544 USA [email protected] Xiaosu Xie Jet Propulsion Laboratory California Institute of Technology Pasadena, CA 91109 USA [email protected] Xiaojun Yang Department of Geography Florida State University Tallahassee, FL 32306-2190 USA [email protected] Alexander Yarovoy Delft University of Technology 2628 CN Delft The Netherlands [email protected] Pat J.-F. Yeh Department of Civil and Environmental Engineering National University of Singapore 117576 Singapore Singapore [email protected] Marek Zreda Department of Hydrology and Water Resources University of Arizona Tucson, AZ 85721 USA [email protected]

xxi

Preface

During the past few decades, the emergence of remote sensing as a discipline – its science, instruments, missions, and applications – has inspired new and comprehensive studies of the Earth. Detailed observations of Earth’s land, ocean and atmospheric processes, and measurements of hitherto unexplored geophysical phenomena have been made possible by remote sensing instruments on groundbased, airborne, and spaceborne platforms. In particular, the unique vantage point of space provides spatially extensive and global perspectives of Earth. Frequent measurements, made hourly, daily, or weekly, over extended periods of years to decades, depending on the observing system and its configuration, have enabled comprehensive studies of Earth’s global system. Remote sensing has thus profoundly altered our understanding of the world in which we live, and has revolutionized the approaches we use to study our environment. Each year the growing number of Earth observing satellites, and the increasingly huge amounts of data and information provided, yield new knowledge and greater appreciation of the changes occurring on our planet, with important implications for future generations of Earth inhabitants. This encyclopedia is a comprehensive reference work on Earth remote sensing that presents the foundations, principles, and state of the art of remote sensing and describes the diverse applications it serves. It covers the concepts, techniques, instrumentation, data analysis, interpretation, and applications of remote sensing. This volume is part of the Encyclopedia of Earth Science series and is organized in the same style as other volumes in the series. The scientific disciplines covered by the series have all benefited in one way or another from the new understanding and discoveries afforded by remote sensing. It is thus timely for publication of an encyclopedia that can link these disciplines and the remote sensing techniques relevant to them in an integrated framework. The focus of the encyclopedia is on remote sensing of Earth – its atmosphere, oceans, cryosphere, and land

surface and subsurface. Some of the techniques described in this volume have their origins in the disciplines of astronomy and astrophysics, and the study of the stars and planets for which, until recently, remote sensing was the only means of obtaining observational scientific data. When applied to Earth, these techniques have blossomed into a remarkably diverse and increasingly sophisticated set of scientific, technological, and computational approaches that all fall under the umbrella of remote sensing. The rapid growth of remote sensing as a discipline is evidenced by the large number of scientific journals now devoted to this field, and the number of courses and degree programs offered at universities around the world. The measurement and interpretation of radiation scattered and emitted by Earth’s atmosphere, surface, and subsurface is what we generally mean when we speak of Earth remote sensing. These measurements are obtained by instruments on remote platforms that include satellites, aircraft, balloons, drones, trucks, and stationary towers. Remote sensing instruments take many forms and are designed to measure electromagnetic radiation in specific wavelength regions of the broad electromagnetic spectrum; some instruments use other forms of radiation such as acoustic radiation. Measurements from the wide array of instruments, operating on the variety of available platforms available, can be processed and analyzed to extract characteristic information about Earth and its constituent biological, chemical, and physical structures, at resolutions from centimeters to thousands of kilometers. This remotely sensed information can be used on its own or combined with direct or ‘in situ’ measurements and geophysical models to give a more comprehensive understanding of the diversity of Earth science phenomena, some of which would be very limited without the unique perspective brought by remote sensing. It is clear that an attempt to fully cover the breadth and depth of topics in remote sensing is a daunting task. Nevertheless, the need for a compendium that can be used as a reference work for this field, as a living document that

xxiv

PREFACE

can be updated periodically to capture new advances, is a pressing one. It is with this aim in mind that the Springer Encyclopedia of Remote Sensing was conceived. Both this print version of the encyclopedia, which can be updated with revisions once every several years, and an online version, which can be updated on a more frequent basis by authors of individual entries, are provided. The online version can accommodate introduction of new entries as the need for new topics or treatments emerges. The encyclopedia entries cover topics that include broad introductory surveys as well as more in-depth treatment of some subjects. The entries treat topics of the physical principles of remote sensing in different wavelength regimes, propagation and scattering of radiation, geophysical models, remote sensing instrumentation, retrieval methods, remote sensing platforms and observational configurations. The models and retrieval methods are described with reference to specific applications in atmosphere, ocean, cryosphere, land, and solid earth geophysics. These applications include human impacts of climate change, and the enabling interdisciplinary science, as well as applications of direct societal benefit such as human health, food security, and prediction and mitigation of natural hazards. Earth remote sensing from space has flourished in the past few decades, and has become a truly global enterprise through development of international collaborations and partnerships, with investments from an increasing number of countries in building and operating satellite observational systems. Several entries in this volume have been devoted to describing these programs, and associated international policies and principles. This encyclopedia is designed to support the needs of students, teachers, and professionals across a broad

spectrum of science, technology, and societal applications related to Earth remote sensing. The intended audience includes those with observational interests in the fields of oceanography, atmospheric sciences, meteorology, climate, cryospheric studies, hydrology, geology, solid earth geophysics, ecology, agronomy, forestry, environmental pollution, geography, land use and social studies, among others. The target audience also includes those with interests in remote sensing theory and practice, electromagnetic propagation, radiative transfer modeling, remote sensing instruments, spacecraft systems and orbits, environmental policy and decisionmaking, resource planning, and monitoring and forecasting of extreme events and natural hazards. In the commercial sector, economists, legal and insurance companies, and commercial and industrial concerns relying on the production, marketing and availability of value-added remote sensing products will also find the encyclopedia a valuable resource. The entries are presented in alphabetical order with titles that are designed to aid searches for specific topics. Crossreferencing using keywords to related entries is also provided to support efficient searches for information of interest to readers. The entries provide bibliographies for further in-depth reading. In summary, though it cannot be claimed that this encyclopedia represents an exhaustive treatment or complete coverage of the field of Earth remote sensing, it is hoped that the volume will serve as a comprehensive and dynamic introduction, and initial entry point, to inspire further reading and study of this exciting and rapidly developing field. September 2013

Eni G. Njoku

Acknowledgments

A work the size of this encyclopedia inevitably relies on the help and cooperation of a large number of people, only some of whom can be individually identified and thanked here. My particular thanks go to the Board of Section Editors, a group of diverse and highly respected remote sensing scientists. To Mike Abrams, Ghassem Asrar, Frank Marzano, Peter Minnett, Vince Salomonson, Vern Singhroy, and Joe Turk, thank you for keeping this project on course by helping to choose the topics that form the entries, suggesting high-quality authors, reviewing the initial manuscripts, and finally checking proofs with your selected groups of authors as well as writing important contributions yourselves. I also wish to acknowledge the great help of Roberta Balstad, Farouk El-Baz, Moustafa Chahine, Jean-Marie Dubois, A.J. Chen, Robert Gurney, Jim Smith, and Guido Visconti who contributed to the early selection of topics and authors for the encyclopedia. My thanks are due also to Tom Farr who assisted me with the editorial duties during a critical stage in the project. This leads to the largest group I wish to acknowledge, the authors of the 170 entries that range in size from several hundred words up to major contributions of several thousand words. Many authors took on more than one entry within their specialty area. Due to the length of time required to produce a volume of this type many

authors who submitted entries early had to wait a considerable amount of time before their entries were finally published. To these authors I especially wish to express my thanks for their patience and dedication to the completion of the project, and I hope they find the final volume worth the wait. At the production end of the project has been the staff of Springer. Their help, understanding, and cooperation, especially when problems needed to be overcome, is something that cannot be appreciated enough. Their patient discussions and exchanges with the authors and board members did much to maintain the smooth progress of the project. Special acknowledgment should go to Petra van Steenbergen and Sylvia Blago who provided me with encouragement through many difficult periods, and to Simone Giesler, all of who were involved with the encyclopedia from start to finish. I must also acknowledge the rewarding environment of my host institution and colleagues, all of whom provided a rich source of motivation and ideas in the field of remote sensing to inspire a publication of this type. My final appreciation goes to my wife Mary whose patience and support over the years has been a major factor in my ability to undertake this task, and to my son Eni Jr. who reminds me constantly of the power of positive thinking.

A

In a gas,

ACOUSTIC RADIATION Alain Weill Bur. Jussieu, LATMOS, Laboratoire Atmosphere Milieux Observations Spatiales, Paris, France

Definition Acoustic. One branch of physics which studies sound. The word acoustic comes from the Greek word akoustikos. “which is related to hearing.” Sound. It comes from the Latin word sonum: “which is related to the hearing sensation created by perturbation of the material medium (elastic, fluid, solid).” In physics, it is a vibration, generally in a gas, created by expansion and compression of gas molecules. Sound waves propagate in the fluid medium and do not propagate in the vacuum. Sounds can be produced in the atmosphere and oceans by living animals or by structures through interaction with the wind, as, for example, trees murmuring, mountains roaring, river sounds, and waves breaking and can be created by various instruments such as music instruments, microphones, speakers, and transducers and also by instruments developed for remote sensing such as SONAR (Sound Navigation and Ranging), ADCP (Acoustic Doppler Current Profiler), and SODAR (Sound Detection and Ranging or what are called echo sounders for atmosphere and ocean). A sound propagating in a medium is characterized by its speed c: c2 ¼ @P=@r

(1)

where P is the pressure and r the density, and @ is derived.

c ¼ ðgP=rÞ1=2

(2)

where g is the heat capacity ratio. Notice that sound speed in the air for standard conditions of temperature and pressure near the surface is close to 340 m/s, while at the ocean surface it is close to 1,500 m/s, which is faster. This will have an incidence on different ways for acoustic signal processing to be done in the ocean and atmosphere. Sound or rather a sound wave is a mechanical pressure oscillation, which is generally longitudinally propagating. Period T. It is the signal duration corresponding to the time when the sound wave is reproduced identically. Frequency. f ¼ 1/T (T in s and f in hertz). Frequency audio spectrum (distribution of acoustic energy as function of frequencies can be divided in four zones related to human hearing power: 0–20 Hz infrasound (not audible), 20–300 Hz is low-pitched, 300–6,000 Hz is medium range, 6,000–20,000 Hz is high-pitched, more than 20,000 Hz are ultrasonic sounds (not audible). Sound amplitude. It corresponds to acoustic pressure fluctuation of the medium Dp (amount of energy in the sound wave) measured at one point of a surface S. It is the ratio of pressure P by the surface element S. I ¼ P=S in W/m2 For a spherical acoustic source, the intensity at distance r is
Ir ¼ P ðsound power of the sourceÞ 4p r2 Radiation. It is the way acoustic wave energy radiates and concerns acoustic rays from the acoustic sources through

E.G. Njoku (ed.), Encyclopedia of Remote Sensing, DOI 10.1007/978-0-387-36699-9, © Springer Science+Business Media New York 2014

2

ACOUSTIC RADIATION

the concerned medium. For example, from a microphone radiating along different directions, we are interested in the radiation diagram corresponding to the knowledge of rays (analogy with optical rays) along different directions.

Introduction We shall begin to analyze (sound) or (acoustic) waves and the wave equation from which we are able to describe energy propagation and rays. If one compares them to electromagnetic waves, acoustic waves are simpler and can be described by the velocity potential, which is a scalar. We shall apply the principles of acoustic wave radiation to different types of acoustic sources used such as monopolar, dipolar sources and the response of the medium at different distances showing that the acoustic field will change of as the characteristics do so. These considerations are really fundamental for remote sensing with acoustic sounders and they are similar to electromagnetic waves, though propagation equations are different. They give information about distances when field characteristics will present some useful organization properties. Wave equation and acoustic waves in flows It corresponds to the study of a plane wave transmitted in a flow by a vibrating plane. It can be, for example, the diaphragm of a microphone vibrating along an axis x. A plane acoustic wave is a general concept from the physics of waves. It corresponds to a wave where wave fronts (surfaces of the same phase) are infinite planes, perpendicular to the same direction of propagation. The equation of pressure variation p (or wave propagation equation) is


(3) @ 2 p @x2 ¼ 1 c2 @ 2 p @t 2 where, as already said, c is the sound speed. For a sinusoidal wave, the solution of the equation is pðx; tÞ ¼ p0 cos½ð2p=T Þðt  x=cÞ

(4)

with the wave number k ¼ 2p=Tc ¼ 2p=l and l is the wave length. We get pðx; tÞ ¼ p0 cosð2p t =T  kxÞ

(5)

where 2p/T is the pulsation o. Generally, one uses the complex notation: pðx; tÞ ¼ p0 expðot  kxÞ

(6)

A linear relation between the displacement gradient (compressibility) gives p ¼ k @z/@x (where k is a coefficient of compressibility) from which we have, solving the sound equation: @z=@t ¼ ð1=ðr0 cÞ expð jðot  kxÞÞ

(7)

This means that particle velocity is in phase with acoustic pressure.

Other derived definitions are useful parameters for remote sensing techniques as the acoustic impedance Z, which is the ratio between pressure and the complex amplitude of the particle. To illustrate the interest in using acoustic impedance, in a project for the Titan satellite sounding (Weill and Blanc, 1987), it was suggested to use acoustic impedance from the satellite’s surface to discriminate, by acoustic remote sensing, between solid and liquid surface just before a possible crash of the rocket at the satellite’s surface. For the flow, it is equal to Zc ¼ r0c. For the plane wave (or progressive wave) 8 x, we have Z(x) ¼ Zc. The condensation of the wave is the spatial derivative @z/@x, which corresponds to a relative change of density.

Monopoles, dipoles, and pulsing sphere If wave propagation Eq. 3 is satisfied, we can work with harmonic solutions of the equation and use wave superposition in the Fourier space. Let us consider one source S of strength q radiating at radial distance r and the solution of the equation for particles radial velocities is
(8) v ¼ A ðexp ðjkrÞ r2 þ jk exp ðjkrÞÞ=r A is a constant and the boundary conditions are such that the solution vanishes at infinity. It is important to notice that radiation behavior is different as function of distance. The acoustic flux Fa of the radial velocity v across a sphere of radius is id v*4pr2 (sphere considered at the distance r) is Fa ¼ 4pA ðexpðjkrÞ þ jkr expðjkrÞÞ

(9)

Therefore, when kr is small, the first term of (Eq. 9) predominates and the conditions correspond to what is called the near field (velocity in phase with the source), and when kr is large, we are in the far field conditions (velocity 90 in advance with the source). These very simple statements are very important in acoustic remote sensing, if (for example) active sea foam, which is an acoustic transmitter at the sea surface, has to be modeled; see Vagle and Farmer (1992) to understand acoustic noise below the surface and bubble sound emission. A more general representation of acoustic sources corresponds to dipolar sources constituting two radiating sources of strengths or magnitudes q+ and q separated by a distance a and such that m ¼ qa is the dipolar momentum (as considered in electromagnetism). Solving the wave equation for the two monopoles with y the angle between r and the direction of the dipole gives two velocity components: (a) one
for the near field (small r): mð1 þ jkrÞ expðjkrÞ cosðyÞ 4pr3 , to which is added a transverse component varying as sin y (orthogonal component), and (b) one for the far field (large r):  k 2 m cosðyÞ expðjkrÞ=4pr, which is typically a radial component. Dipolar sources or a combination of dipolar sources are very useful to simulate acoustic antennas and

ACOUSTIC RADIATION

3

Acoustic Radiation, Figure 1 Realization of a three offset antennas for a 6 kHz minisounder. One antenna is vertical and two others are slanting. One distinguishes compression chambers as acoustic sources and the horns and antennas parts (parabolic portion + circular aperture covered by acoustic foam).

acoustic sources. Moreover, field analysis considering the distance where the field can be considered as far is very important to interpret, for example, signals coming from active systems (which transmit acoustic waves) or passive systems (which only receive acoustic waves as acoustic radiometers by analogy with electromagnetic and optical radiometers. Close to monopoles and dipoles, which are theoretical concepts, is the pulsating sphere of radius a. With the condition that ka 0.98. Many soils also have a high emissivity, >0.95. Lyon (1965) pointed out that, for most rocks, the emissivity at
10 mm is
0.95. Once e has been estimated, LST is found by c2

(8) Tm ¼ c1 lIn eðlÞBðl;T 5 þ 1 Þpl

Image classification has been used to determine surface composition, from which e can be inferred. For example, van de Griend and Owe, (1993) showed that in the 8 – 14 mm spectral range, e is highly correlated with NDVI, the normalized difference vegetation index, for different vegetation types: e  a þ b In NDVI; NDVI ¼

r0:85  r0:65 r0:85 þ r065

(9)

where r is evaluated at near-infrared (l ¼ 0.85 mm) and visible red (l ¼ 0.65 mm) wavelengths. Equation 9 allows e to be estimated pixel by pixel, improving LST estimates especially for agricultural areas with simple geological substrates (homogeneous soils). Empirical coefficients a and b in Equation 9 are dependent on substrate emissivities, but if these can be estimated it is possible to extend greatly the range of surfaces for which accurate model LSTs can be calculated. Model temperatures have been calculated for lava flowing from volcanoes, using Landsat Thematic Mapper (TM) Band 7 (2.25 mm) images and emissivities appropriate for lava (Pieri et al., 1990). The SWIR bands are effective because the peak thermal radiance is near 2 mm, whereas the emitted TIR radiance is too high and exceeds the dynamic range of the TIR band. Care must be taken because the emissivity of the hot lava appears to vary with temperature (Abtahi et al., 2001).

Color temperature, Tc The simplest approach that utilizes the changing shape of the Planck function to estimate LST is a ratio of two spectral bands centered at la and lb, and for which emissivities are known. The ratio is a monotonic function of temperature:

0 1 1 5 exp c ðl T Þ 1 2 b Lðla ; T Þ eðla Þlb @ A:

(10) ¼ Lðlb ; T Þ eðlb Þl5a exp c2 ðla T Þ1  1 Generally, it is assumed that e(la) ¼ e(lb). Equation 10 is sometimes simplified further (Wien’s approximation) by ignoring the (1) term, such that !!1   c2 c2 Lðla ; T Þl5a In  (11) Tc ¼ lb la Lðlb ; T Þl5b For typical LSTs (e.g., 300 K), Tc is best calculated for widely separated central wavelengths, for example, 3 and 10 mm, because the ratio is more sensitive to LST, and therefore measurement precision has less of an effect on the recovered Tc. However, the use of MIR channels is generally limited to night time data because of the need to make large corrections for reflected sunlight.

LAND SURFACE TEMPERATURE

Generalized split-window algorithm Wan and Dozier, (1996) showed that provided the emissivity was known, the SST algorithm could be generalized for use over land. This algorithm is based on spectral radiance differences rather than ratios and has empirical coefficients a and b that permit the elimination of atmospheric effects. The form is Ts ¼ T4 þ

1 b ðT4  T5 Þ  a1 a1

(12)

T4 and T5 are the brightness temperatures for AVHRR (Advanced Very High Resolution Radiometer) bands 4 and 5 (10.8 and 11.9 mm) or MODIS bands 31 and 32. Equation 12 can be generalized further to account for directional effects, but is still of the form of a difference equation.

Summary and conclusions Temperature and emissivity are generally both unknown for “geological” surfaces of rock and soil and must be solved for simultaneously using inversion of Planck’s equation. These algorithms thus go beyond just finding LST and are discussed in the entry “Land Surface Emissivity.” Because of the underdetermined nature of the modified Planck equation (Equation 2), no single solution for LST has been found that satisfies all analysts. Therefore, literally dozens of algorithms have been proposed, tested, and applied. This entry summarized only a few, representative of the basic approaches. Bibliography Abtahi, A. A., Kahle, A. B., Abbott, E. A., Gillespie, A. R., Sabol, D., Yamada, G., and Pieri, D., 2002. Emissivity changes in basalt cooling after eruption from Pu’u O’o, Kilauea, Hawaii. Eos Transactions of the American Geophysical Union, 83(47), Fall Meeting Supplement, Abstract V71A–1263. Anding, D., and Kauth, R., 1970. Estimation of sea surface temperature from space. Remote Sensing of Environment, 1, 217–220. Berk, A., Anderson, G. P., Acharya, P. K., Bernstein, L. S., Muratov, L., Lee, J., Fox, M., Adler-Golden, S. M., Chetwynd, J. H., Hoke, M. L., Lockwood, R. B., Gardner, J. A., Cooley, T. W., and Lewis, P. E., 2005. MODTRAN5: a reformulated atmospheric band model with auxiliary species and practical multiple scattering options. In Proceedings of Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery, SPIE, 5806: 662; DOI:10.1117/ 12.606026. Conel, J. E., Green, R. O., Carrere, V., Margolis, J. S., Alley, R. E., Vane, G., Bruegge, C. J., and Gary, B. L., 1988. Atmospheric water mapping with the airborne visible/infrared imaging spectrometer (AVIRIS), Mountain Pass, CA. In Vane, G. (ed.), Proceedings of the AVIRIS Performance Evaluation Workshop. JPL Publication 88–38, Pasadena: Jet Propulsion Lab., pp. 21–26. Gao, B.-C., and Goetz, A. F. H., 1990. Column atmospheric water vapor and vegetation liquid water retrievals from airborne imaging spectrometer data. Journal of Geophysical ResearchAtmospheres, 95, 3549–3564. Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., Iredell, M., Saha, S., White, G., Woollen, J., Zhu, Y., Chelliah, M., Ebsuzaki, W., Higgins, W., Janowiak, J., Mo,

319

K. C., Ropelewski, C., Wang, J., Leetma, A., Reynolds, R., Jenne, R., and Joseph, D., 1996. The NCEP/NCAR 40-year reanalysis project. Bulletin of the American Meteorological Society, 77, 437–470. Lyon, R. J. P., 1965. Analysis of rocks by spectral infrared emission (8 to 25 microns). Economic Geology, 60(4), 715–736. Mushkin, A., and Gillespie, A. R., 2005. Estimating sub-pixel surface roughness using remotely sensed stereoscopic data. Remote Sensing of Environment, 99, 75–83. Norman, J. M., and Becker, F., 1995. Terminology in thermal infrared remote sensing of natural surfaces. Remote Sensing Reviews, 12, 159–173. Pieri, D. C., Glaze, L. S., and Abrams, M. J., 1990. Thermal radiance observations of an active lava flow during the June 1984 eruption of Mt. Etna. Geology, 18, 1018–1022. Seeman, S. W., Borbas, E. E., Li, J., Menzel, W. P., and Gumley, L. E., 2006. MODIS atmospheric profile retrieval algorithm theoretical basis document, version 6. Available from http:// modis.gsfc.nasa.gov/data/atbd/atbd_mod07.pdf. Accessed 7 June, 2013. Snyder, W. C., and Wan, Z.-M., 1998. BRDF models to predict spectral reflectance and emissivity in the thermal infrared. IEEE Transactions on Geoscience and Remote Sensing, 36, 214–255. Sobrino, J. A., Jiménez-Muñoz, J. C., and Paolini, L., 2004. Land surface temperature retrieval from Landsat TM 5. Remote Sensing of Environment, 90, 434–440. Tonooka, H., 2005. Accurate atmospheric correction of ASTER thermal infrared imagery using the water vapor scaling method. IEEE Transactions on Geoscience and Remote Sensing, 43(12), 2778–2792. Tonooka, H., and Palluconi, F. D., 2005. Validation of ASTER/TIR standard atmospheric correction using water surfaces. IEEE Transactions on Geoscience and Remote Sensing, 43(12), 2769–2777. van de Griend, A. A., and Owe, M., 1993. On the relationship between thermal emissivity and the normalized difference vegetation index for natural surface. International Journal of Remote Sensing, 14(6), 1119–1131. Wan, Z., 1999. MODIS land-surface temperature algorithm theoretical basis document (LST ATBD), Version 3.3. Contract NAS5-31370. Wan, Z., and Dozier, J., 1992. Effects of temperature-dependent molecular absorption coefficients on the thermal infrared remote sensing of the earth surface. In Proceedings IGARSS’92. pp. 1242–1245. Wan, Z., and Dozier, J., 1996. A generalized split-window algorithm for retrieving land-surface temperature from space. IEEE Transactions on Geoscience and Remote Sensing, 34(4), 892–905. Young, S. J., Johnson, R., and Hackwell, J. A., 2002. An in-scene method for atmospheric compensation of thermal hyperspectral data. Journal of Geophysical Research, 107(24), 4774, doi:10.1029/2001JD001266.

Cross-references Crop Stress Cryosphere and Polar Region Observing System Fields and Radiation Global Programs, Operational Systems Land Surface Emissivity Ocean, Measurements and Applications Optical/Infrared, Radiative Transfer Sea Surface Temperature Thermal Radiation Sensors (Emitted) Volcanism Water and Energy Cycles Water Vapor

320

LAND SURFACE TOPOGRAPHY

LAND SURFACE TOPOGRAPHY G. Bryan Bailey USGS Earth Resources Observation and Science Center, Sioux Falls, SD, USA

Synonyms Elevation; Landscape; Relief; Terrain Definition Topography of the Land Surface. The three-dimensional arrangement of physical attributes (such as shape, height, and depth) of a land surface in a place or region. Physical features that make up the topography of an area include mountains, valleys, plains, and bodies of water. Humanmade features such as roads, railroads, and landfills are also often considered part of a region’s topography (American Heritage Science Dictionary, 2005). Importance of topographic information The topography of the land surface is one of the most fundamental geophysical measurements of the Earth, and it is a dominant controlling factor in virtually all physical processes that occur on the land surface. Topography of the land surface also significantly controls processes within the overlying atmosphere, and it reflects the processes within the underlying lithosphere. Consequently, topographic information is important across the full spectrum of earth sciences. Precipitation, runoff, soil moisture, incident sunlight, and temperature all vary with topography. Consequently, topography dominantly controls the local and regional distribution and character of vegetation. Erosion and sedimentation, and consequently soil formation and nutrient transport, also are strongly controlled by topography and are key factors in ecological studies. Topography strongly influences the location and magnitude of surface and subsurface water flow. Modeling of water supply and flood potential requires knowledge of the area’s drainage extent, its slopes, and the pattern of the drainage network. Particularly in rugged terrain, topography is commonly the dominant variable in remote sensing imagery. Topographic shading affects the radiance measured at every wavelength and is consequently the statistical principal component of many remotely sensed data sets. Meanwhile, atmospheric optical thickness varies inversely with topographic height, so that topography is an important factor in the atmospheric correction of remotely sensed data. Topographic data are imperative for the orthorectification of satellite imagery. While topography controls many natural processes at and near the Earth’s surface, many natural processes conversely control topography. Consequently, to various G. Bryan Bailey has retired.

degrees, topography records and reveals evidence of current and past natural processes. An obvious example is the development and occurrence of erosional and depositional fluvial landforms. Tectonic, volcanic, glacial, and gravitational processes also produce characteristic landforms that reveal past and ongoing change. Consequently, topographic information is an important tool in the study of such processes (Crippen, 2008).

Describing the topography of the land surface A topographic map (Figure 1) is a planimetric, or twodimensional, representation of the three-dimensional configuration of the land surface where relief, or change in elevation, typically is represented by contour lines. A contour line is a line traced on the map such that all points on that line have the same elevation. That is, contour lines describe continuous points of equal elevation. Until recent years, paper topographic maps were the most common tool used to describe the topography of the land surface. Until about 1940, most topographic maps were made by field crews who used alidades and plane tables to survey and map the topography of the landscape. World War II ushered in the age of aerial photogrammetry as the most common method for making topographic maps. This method, which uses overlapping and nadir-looking aerial photographs and a stereoplotter, revolutionized topographic mapping, resulting in greatly increased map coverage and enhanced map standardization (USGS, 1998). Advances in computer technology brought about the latest great change in topographic mapping, the digital mapping revolution. Perhaps most notable was the replacement of the analog stereoplotter by the computer-assisted analytical stereoplotter. Not only has computer-assisted map production made it easier to make new paper maps and revise old ones, computer technology has accelerated demand for topographic data and other map information in digital form for use with the ever-growing number of computer-based mapping applications (McGlone, 2007). A digital elevation model or DEM (Figure 2) is the generic term used most frequently to denote digital topographic data in all their various forms. The word “model” is applied because computers can use such data to model and automatically analyze the Earth’s topography in three dimensions, thus avoiding much timeconsuming human interpretation (Maune et al., 2007). Digital terrain model (DTM) and digital surface model (DSM) are two other common terms with similar meaning to DEM. A DEM is a digital file consisting of terrain elevations for ground positions at regularly spaced horizontal intervals. The shorter those intervals are, the higher the spatial resolution of the DEM. The U.S. Geological Survey National Elevation Dataset (NED) typically has elevation data spaced at 30 m intervals, and it is thus said to have 30 m postings. DEMs are referenced to a vertical

LAND SURFACE TOPOGRAPHY

321

Land Surface Topography, Figure 1 Part of USGS 1:250,000 topographic map of the Drum Mts., Utah, area.

datum, such as the WGS84/EGM96 geoid, and to a geographic coordinate system, such as Universal Transverse Mercator (UTM).

Determining land surface topography from remotely sensed digital data Many DEMs have been, and continue to be, generated by digitizing topographic maps produced largely by aerial photogrammetric techniques. However, today most DEMs of the Earth’s land surface are being generated, using a variety of automated processes, directly from digital data acquired by a rather large variety of airborne sensors and land surface-imaging satellite systems.

Most DEMs produced today from remotely sensed digital data are derived from one of three primary sources: optical imaging systems, interferometric synthetic aperture radar systems, and lidar systems (McGlone, 2007).

Optical imaging systems Three types of optical imaging sensors are used for photogrammetric production of DEMs: airborne film cameras, airborne digital sensors, and digital sensors onboard satellites. Film mapping cameras, for decades the staple of aerial photogrammetric mapping, continue to be important sources of stereo images for DEM generation. Aerial film is scanned by high-resolution scanners to produce digital

322

LAND SURFACE TOPOGRAPHY

Land Surface Topography, Figure 2 DEM intensity image (left) and DEM shaded relief image (right) of the Drum Mts., Utah.

images that can be processed by a softcopy stereoplotter or by one of the many available DEM generation software systems (McGlone, 2007). Airborne and spaceborne digital sensors capable of collecting imagery useful in the generation of DEMs have certain similarities, but they also have important differences. Airborne systems are essentially digital mapping cameras capable of acquiring very high-spatial-resolution images, but typically with fewer spectral bands than spaceborne optical imaging systems. Stereo acquisition by airborne digital sensors typically is achieved by overlap of successive images acquired along the flight path of the aircraft, similar to film mapping cameras. Optimal stereo coverage for DEM generation by satellite sensors is achieved along the orbital track of the satellite by using two (or more) sensors. One of the sensors is nadir looking, while the other points at some fixed angle along the orbital track fore and/or aft of the spacecraft. Some satellite systems acquire stereo imagery from adjacent orbits by pointing across track, and some are able to acquire limited stereo coverage along the same orbit using a single sensor that looks forward to image an area from one angle and then is pointed backward to image the same area from a different angle. Generating DEMs from imagery acquired by airborne or spaceborne digital sensors is accomplished with the aid of a softcopy stereoplotter or one of the many available DEM generation software systems. The process may or may not employ the use of ground control points (GCPs), and it typically involves a sequence of steps that include selecting tie points in each of the stereo pair, co-registration of the stereo images, stereo correlation for parallax difference measurement (image matching), and calculation of elevation values. In the co-registered stereo

images, any positional differences between common points parallel to the direction of travel (parallax differences) are attributed to displacements caused by relief. Relative ground elevations are determined by measuring the parallax differences in the registered images, which then are converted to elevation (Lang and Welch, 1999).

Interferometric SAR systems Synthetic aperture radars (SAR) illuminate the Earth’s surface with microwave pulses, and they receive and record the return signals with respect to the magnitude and phase of those sine wave pulses (Bamler, 1997). While it is possible to generate accurate DEMs from stereo radar images using techniques similar to those described for optical imaging systems, DEMs are generated more commonly from interferometric synthetic aperture radar (InSAR). InSAR exploits the phase of SAR signals to measure stereo parallaxes to an accuracy of a fraction of a wavelength. The phase of the return radar wave depends on the distance to the ground, so it is possible to accurately determine land surface elevation on a pixel-by-pixel basis from the phase information. To generate a DEM, InSAR uses two SAR images of the same land surface area taken from slightly different positions and determines phase differences between them, producing an image called an interferogram. Further processing of the interferogram results in the generation of a DEM of the land surface imaged by the two sensors (Henslely et al., 2007). Lidar systems Lidar stands for Light Detection and Ranging, and like radar, it is an active remote sensing system. Lidars use

LAND SURFACE TOPOGRAPHY

laser technology to measure distances to specific points by transmitting pulses or continuous waves of light, amplifying the light that is scattered back, and recording the precise time the transmitted pulse takes to travel to the target and back (Fowler et al., 2007). Most lidar systems used today to produce DEMs are airborne systems, and they employ a variety of different beam steering or scanning strategies. They also operate at a wide range of altitudes above the land surface, depending on the resolution requirements for the DEM to be produced. DEMs produced from lidar data typically have significantly greater spatial detail and better accuracy than DEMs produced from optical imaging systems or InSAR. Lidar is a complex remote sensing technology, and the data processing required to convert raw lidar data to DEMs also is complex. Typically, such processing is done by the company or agency that collects that data, because the processing software has been developed specifically for the lidar system that collected the data (Fowler et al., 2007). The lidar product most commonly associated with topography is a DEM known as the bare earth model,

Land Surface Topography, Figure 3 Example of bare earth model DEM produced from lidar data collected over North Carolina.

323

which is a product from which the processing has removed virtually all returns not associated with the bare land surface (Figure 3).

Satellite systems that produce topographic data This section briefly examines the system and data characteristics of selected Earth-observing satellite systems that acquire or have acquired data from which DEMs of the global land surface can be or have been generated. Airborne systems provide data from which highly accurate DEMs can be produced, but there is not one program or system that can provide such data for anywhere on the Earth’s surface. The international remote sensing and earth science communities need access to quality DEMs for the entire global land surface, hence the emphasis here on satellite systems. Radar satellite systems Beginning in 1991, a number of polar-orbiting SAR satellite systems have been in operation, providing a continuous collective capability to acquire InSAR data from which DEMs can be generated for virtually any place on the global land surface. However, it was not until the 2010 launch of TanDEM-X, which operates in tandem formation with TerraSAR-X, that systematic efforts to produce a consistent global DEM from InSAR data were undertaken. Table 1 lists the SAR satellite systems that have contributed to DEM generation over the past two decades. Shuttle radar topography mission (SRTM) In 2000, the US National Aeronautics and Space Administration (NASA) and National Geospatial-Intelligence Agency (NGA) cooperated with the German Space Agency (DLR) to fly a Space Shuttle mission dedicated to acquiring digital topographic data for more than 80 % of the Earth’s land surface. NASA flew a C-band SAR and DLR flew an X-band SAR, both of which were configured with one antenna in the bay of the space shuttle and the other at the end of a 60 m collapsible mast. The mission lasted 11 days, and NASA’s C-band SAR collected complete InSAR coverage of the global landmass between 60  N to 56  S latitude. The X-band SAR was not designed for uninterrupted coverage.

Land Surface Topography, Table 1 International SAR systems from which DEMs can be produced from InSAR data Country/agency

Satellite/sensor

Band

Launch date

Still active

European Space Agency (ESA) ESA Canadian Space Agency (CSA) CSA Japan (JAXA) ESA Japan (JAXA) Germany (DLR) Germany (DLR)

ERS-1 ERS-2 RADARSAT-1 RADARSAT-2 JERS-1 Envisat ASAR ALOS PALSAR TerraSAR-X TanDEM-X

C-band C-band C-band C-band L-band C-band L-band X-band X-band

1991 1995 1995 2007 1992 2002 2006 2007 2010

No No No Yes No No No Yes Yes

324

LAND SURFACE TOPOGRAPHY

The SRTM DEM data set is available from NASA and from the U.S. Geological Survey at no charge to users. Data are at 1 arc-second (30 m) for the United States and its territories and 3 arc-seconds (90 m) for all other covered areas. Vertical accuracy specification for the 1 arc-second data was 16 m (Rabus et al., 2003), but the data frequently have vertical accuracies better than 10 m (at 90 % confidence).

SPOT satellites The French Système Pour l’Observation de la Terre (SPOT) satellites have been capable of acquiring crosstrack stereo digital imagery from which DEMs can be generated since their initial launch in 1984. Spatial resolution of SPOT sensors has increased over that time, so the spatial details of DEMs that can be generated from those data likewise have increased. Currently, SPOT Image offers for sale DEMs produced from SPOT optical image data that cover most of the global land surface. The DEMs have 30 m postings and a vertical accuracy of less than 10 m where the slope of terrain is less than 20 %. ASTER The Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) was built by the Japanese Ministry of Economy, Trade and Industry (METI), and it flies onboard NASA’s Terra satellite. ASTER collects along-track stereo optical data with 15 m spatial resolution from which DEMs with 30 m postings are routinely produced as standard data products without the need for GCPs. The accuracy of ASTER DEMs varies some depending on terrain and other conditions, but they routinely have vertical accuracies better than 15 m (Fujisada et al., 2005). ASTER 60 km by 60 km DEMs are available for purchase to the general user from METI’s Earth Resources Data Analysis Center and NASA’s Land Processes Distributed Active Archive Center. Since ASTER was launched in late 1999, more than two million scenes have been acquired of the global land surface. NASA and METI cooperated to produce a global DEM from these ASTER data. The ASTER Global DEM, which covers the Earth’s landmasses from 83  N to 83  S latitude with nominal accuracies of 20 m vertical and 30 m horizontal at 95 % confidence, has 30 m postings. The ASTER Global DEM was contributed to the Global Earth Observing System of Systems (GEOSS) by NASA and METI, and thus it is available at no cost to users worldwide. Cartosat and prism Two optical imaging satellite systems that were designed to acquire data for generation of DEMs are the Indian Space Research Organisation’s (ISRO) Cartosats and the Japanese Aerospace Exploration Agency’s (JAXA) Prism sensor that flew onboard the Advanced Land Observation Satellite (ALOS). All acquire stereo image data along the orbital track. Cartosat-1 has a spatial resolution of 2.5 m

with a 30 km swath, and Cartosat-2 has a spatial resolution of less than 1 m with a 9.6 km swath. Prism has a 2.5 m spatial resolution and a 35 km swath. DEMs are not offered as standard products by either ISRO or JAXA, but the image data are available for purchase.

Summary and conclusions The topography of the land surface is one of the most fundamental geophysical measurements of the Earth, and it is a dominant controlling factor in virtually all natural physical processes that occur on the land surface. The topographic map, which is a planimetric representation of the three-dimensional land, has been the most common tool used to describe the topography of the land surface until recently. Now, digital topographic data, in the form of a DEMs, are the tools of choice for many who wish to characterize the topography of the land surface. DEMs are most frequently generated by automated computer techniques from digital data acquired by airborne and spaceborne sensors. Locally to regionally, airborne film and electro-optical systems and lidar systems provide users with high-quality and very accurate DEMs. For global studies, spaceborne optical systems capable of acquiring stereo imagery and InSAR systems offer the opportunity to produce DEMs with improving quality and accuracy worldwide. The SRTM DEM that cover 80 % of the global land surface and the ASTER Global DEM that covers virtually all of it are examples of two recent contributions by land remote sensing systems to better characterize the global land surface topography. Almost certainly, even greater advancements will be achieved in the next few years. Bibliography Bamler, R., 1997. Digital terrain models from radar interferometry. In Fritsch, D., and Hobbie, D. (eds.), Photogrammetric Week 1997. Heidelberg: Wichmann Verlag, pp. 93–105. Crippen, R. E., 2008. Global topographic exploration and analysis with the SRTM and ASTER elevation models. In Elevation Models for Geoscience, Special Publication. London: Geological Society (in press). Fowler, R. A., Samberg, A., Flood, M. J., and Greaves, T. J., 2007. Topographic and terrestrial lidar. In Maune, D. F. (ed.), Digital Elevation Model Technologies and Applications: The DEM Users Manuel, 2nd edn. Bethesda: American Society of Photogrammetry and Remote Sensing, pp. 199–252. Fujisada, H., Bailey, G. B., Kelly, G., Hara, S., and Abrams, M., 2005. ASTER DEM performance. IEEE Transactions on Geoscience and Remote Sensing, 43, 2715–2724. Henslely, S., Munjy, R., and Rosen, P., 2007. Interferometric synthetic aperture radar (IFSAR). In Maune, D. F. (ed.), Digital Elevation Model Technologies and Applications: The DEM Users Manuel, 2nd edn. Bethesda: American Society of Photogrammetry and Remote Sensing, pp. 141–198. Lang, H. R., and Welch, R., 1999. ATBD-AST-08 Algorithm Theoretical Basis Document for ASTER Digital Elevation Models (Standard Product AST14). Washington, DC: National Aeronautics and Space Administration/Earth Observing System Program, p. 69. Maune, D. F., Kopp, S. M., Crawford, C. A., and Zervas, C. E., 2007. Introduction. In Maune, D. F. (ed.), Digital Elevation Model Technologies and Applications: The DEM Users Manuel,

LAND-ATMOSPHERE INTERACTIONS, EVAPOTRANSPIRATION

2nd edn. Bethesda: American Society of Photogrammetry and Remote Sensing, pp. 1–36. McGlone, J. C., 2007. Photogrammetry. In Maune, D. F. (ed.), Digital Elevation Model Technologies and Applications: The DEM Users Manuel, 2nd edn. Bethesda: American Society of Photogrammetry and Remote Sensing, pp. 119–140. Rabus, B., Eineder, M., Roth, A., and Bamler, R., 2003. The shuttle radar topography mission: a new class of digital elevation models acquired by spaceborne radar. ISPRS Journal of Photogrammetry and Remote Sensing, 57, 241–262. The American Heritage® Science Dictionary Copyright © 2005 by Houghton Mifflin Company. Published by Houghton Mifflin Company. USGS, 1998. Topographic Mapping. Reston: U.S. Geological Survey. (Out of print; online version: http://erg.usgs.gov/isb/ pubs/booklets/topo/topo.html).

Cross-references Geodesy Geomorphology Land Surface Roughness Lidar Systems Radar, Synthetic Aperture Reflected Solar Radiation Sensors, Multiangle Imaging

LAND-ATMOSPHERE INTERACTIONS, EVAPOTRANSPIRATION Joshua B. Fisher Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA

Synonyms Evaporation; Water flux Definition Evapotranspiration (ET). The transfer of liquid water from open water and through plant transpiration to the atmosphere as water vapor. Transpiration. The loss of water vapor through plant pores called stomata on leaves/needles or stems. Basics of evapotranspiration Evapotranspiration (ET) is the movement and transfer (i.e., flux) of water as a liquid at the Earth’s surface to the atmosphere as a gas. ET is a combination of open water evaporation and plant transpiration. (Sublimation, which is the transition of solid water (i.e., ice, snow) to vapor due to low atmospheric pressure (i.e., high altitude), dry air, and high sunlight, is generally considered separate from ET.) Sources of open water evaporation could include oceans, seas, lakes, rivers, ponds, puddles, and water on objects such as plants, buildings, rocks, the soil surface (including movement of water vertically through the soil to the surface), or in the context of measuring devices such as a pan. Plants take up water from the soil through their roots, transferring that water through

325

their stems via conduits called xylem to their leaves, where it is used in the process of photosynthesis. The photosynthetic machinery in leaves (e.g., chlorophyll) takes in CO2 from the atmosphere through stomatal pores and combines it with water and energy (i.e., light) to create sugars used to maintain and grow plant tissue and functions. While stomata are open, plants may lose water from their leaves to the atmosphere – this water loss is called transpiration. Plants regulate the opening and closing of their stomata to minimize water loss (closed), yet maximize CO2 absorption (open). Energy is required to break the strong bonds that hold water molecules together as a liquid – when those bonds break, the individual water molecules may enter the surrounding atmosphere as vapor. Energy may be in the form of heat, radiation, or pressure. Regardless of the availability of energy, water molecules may not be able to enter the atmosphere if the atmosphere is already saturated with moisture (humidity) or if there is no wind to facilitate the transfer of the molecules from the water source to the atmosphere. The wind itself may be differentially influenced by friction as it passes over smooth versus rough surfaces. Therefore, solar radiation (or, indirectly, air temperature), air humidity, and wind speed are the main climate influences on ET. The main vegetative controls include leaf and canopy structures, regulation of stomata, and rooting dynamics. Finally, soil characteristics control soil moisture retention of precipitation inputs. All of these potential controls vary in influence depending on the system in question, as well as the associated spatial and temporal scales of analysis (Fisher et al., 2011).

Remote sensing of ET ET can be measured “remotely” with instruments attached to towers extending over vegetation using the eddy covariance technique (e.g., FLUXNET: Baldocchi et al., 2001). These same instruments may be attached to airplanes for regional measurements. However, ET cannot be measured directly from satellite remote sensing, so it must be inferred from a model or the residual of other measurements. There are three orders of complexity in spacebased estimation of ET:

Simple: Empirical, semiempirical

Intermediate: Water balance, energy balance

Complex: Land surface/Earth system models Empirical, semiempirical approaches One of the simplest approaches to estimating ET is to take another closely related variable that is measureable and convert that to ET using a statistical relationship. The statistical relationship (e.g., linear regression) may be developed from studies where both the other variable and ET were measured and then used to extrapolate beyond the site. One commonly used variable is the Normalized Difference Vegetation Index (NDVI), as well as related “greenness” indices, constructed from

326

LAND-ATMOSPHERE INTERACTIONS, EVAPOTRANSPIRATION

measurements primarily in the red and near-infrared (NIR) wavelengths, and which is indicative of plant productivity. Where there is plenty of water and energy, there will be both high NDVI and ET; where there is no water and energy, one would not expect much NDVI and ET. However, this relationship may fall apart, for example, under deforestation or nutrient limitation (high ET, low NDVI), or diurnal/seasonal water stress (low ET, high NDVI). NDVI may be obtained from satellite instruments such as the Advanced Very High Resolution Radiometer (AVHRR), the Moderate resolution Imaging Spectroradiometer (MODIS), or the Visible Infrared Imager Radiometer Suite (VIIRS). Another commonly used variable is Land Surface Temperature (LST), constructed from thermal-infrared (TIR) measurements. A given surface may be cooled (lower LST) when evaporating and hotter when there is less ET. However, other forces may change the temperature of the surface, including advecting warm/cool/dry/ moist air. LST may be obtained from satellite instruments such as MODIS, the Atmospheric Infrared Sounder (AIRS), or Landsat. LST may be combined with NDVI for a somewhat more complex empirical approach. One of the leading empirical approaches comes from the MPI-BGC product, which is constructed from a machine learning technique and model tree ensemble that developed statistical relationships between measured ET and globally available ancillary data at over 250 FLUXNET sites (Jung et al., 2009). Finally, many agriculturalists use semiempirical algorithms to estimate ET, using physics-based equations for potential ET (PET), then converting or downscaling PET to actual ET (AET) using an empirical scalar multiplier, called a crop coefficient, developed for their specific crop and location.

Water balance ET may be calculated as the residual of known measurements in the water balance equation: P ¼ dS þ Q þ ET

(1)

where P is precipitation (rainfall and snow); dS is the change in stored standing water (e.g., lakes, ponds, or in/on plants), soil moisture, and groundwater; and Q is runoff. From a remote sensing standpoint, rainfall is measured from a variety of satellites including the Tropical Rainfall Measuring Mission (TRMM) (the Global Precipitation Mission (GPM) is currently in development as the next major multi-satellite precipitation-measuring mission) and snow from MODIS. dS is measureable at large spatial scales from the Gravity Recovery And Climate Experiment (GRACE). Q is not yet measureable from space, (The proposed Surface Water Ocean Topography (SWOT) mission currently in development would measure river discharge from space.) but is readily obtained from river discharge measurements, though many rivers are sparsely instrumented, for

example, in developing nations. Equation 1 may be rearranged to solve for ET given the known measurements of the three other variables in the equation.

Energy balance ET may also be considered an energy (water fluxes such as precipitation and ET are usually given in units of depth per time (i.e., mm · day1); the units are consistent when they are in volume per area per time (i.e., m3 · ha1 · day1). 1 m3 is equal to 1,000 l. Water can also be expressed in units of mass – 1 kg of water is equal to 1 mm of water spread over 1 m2. ET, like Rn, can be expressed in units of energy too. Because it requires 2.45 MJ to vaporize 1 kg of water (at 20 C), 1 kg of water is therefore equivalent to 2.45 MJ; 1 mm of water is thus equal to 2.45 MJ · m2) variable, called the latent heat of evaporation (LE), as it requires a certain amount of energy to convert a given quantity of liquid water to gas. Energy coming from the sun less any radiation that gets reflected back to the atmosphere – or net radiation (Rn) – is energy available to drive ET. Any Rn that does not drive ET either gets converted to sensible heat (H) or stored in the soil or other objects (G): Rn ¼ ET þ H þ G

(2)

A few space-based Rn are available, including those from the Surface Radiation Budget (SRB), the Clouds and Earth’s Radiant Energy System (CERES), the International Satellite Cloud Climatology Project (ISCCP), and MODIS. H and, to a lesser extent, G are not remotely measureable and are the focus of models such as the Surface Energy Balance System (SEBS), the Atmosphere-Land Exchange Inverse (ALEXI), the Surface Energy Balance Algorithm for Land (SEBAL), and Mapping EvapoTranspiration at high Resolution with Internalized Calibration (METRIC), all of which rely particularly on remotely sensed LST (Li et al., 2009).

Direct approaches ET may also be calculated “directly” from the physics that control ET, as outlined earlier in the “Basics of Evapotranspiration” subsection. The most widely used equation for determining ET comes in the form of the Penman-Monteith equation: ET ¼

cp rVPD ra

þ g rras

DRn þ Dþg

(3)

where D is the slope of the saturation-to-vapor pressure curve, cp is the specific heat of water, r is air density, VPD is vapor pressure deficit, ra is aerodynamic resistance, g is the psychrometric constant, and rs is surface resistance. Equation 3 forms the foundation of the algorithm for the official MODIS ET product (MOD16) (Mu et al., 2011), which relies on MODIS-based leaf area index (LAI), fraction of absorbed photosynthetically active radiation (fAPAR), land cover, and a general

LAND-ATMOSPHERE INTERACTIONS, EVAPOTRANSPIRATION

327

Land-Atmosphere Interactions, Evapotranspiration, Figure 1 Mean monthly ET for 2004 from the PT-JPL product.

biome-specific lookup table to parameterize the resistances. Rn, VPD, and air temperature (Ta; i.e., included in D) in MOD16 are derived from the NASA/ GMAO Modern Era Retrospective Analysis (MERRA). The PT-JPL product (Figure 1: Fisher et al., 2008) is based on the PET formulation of the Priestley-Taylor equation, which is a reduced version of the PenmanMonteith equation, eliminating the need to parameterize the stomatal and aerodynamic resistances, leaving only equilibrium evaporation multiplied by a constant (1.26) called the a coefficient: PET ¼ a

D Rn Dþg

(4)

PET is reduced to AET using ecophysiological constraint functions (f-functions, unitless multipliers, 0–1) based on atmospheric moisture (VPD and relative humidity, RH) and vegetation indices (NDVI and SAVI): ET ¼ ETs þ ETc þ ETi ETc ¼ ð1  fwet Þfg fT fM a

(5) D Rnc Dþg

ETs ¼ ðfwet þ fSM Þð1  fwet ÞÞa D Rnc ETi ¼ fwet a Dþg

D ðRnc  GÞ Dþg

(6) (7) (8)

where ETs, ETc, and ETi are evaporation from the soil, canopy, and intercepted water, respectively, each calculated explicitly. fwet is relative surface wetness (RH 4), fg is green canopy fraction ( fAPAR/fIPAR),

fT is a plant temperature constraint (exp(((Tmax  Topt)/Topt)2)), fM is a plant moisture constraint ( fAPAR/fAPARmax) and fSM is a soil moisture constraint, (RHVPD). fAPAR is absorbed photosynthetically active radiation (PAR), fIPAR is intercepted PAR, Tmax is maximum air temperature, Topt is Tmax at max(RnTmaxSAVI/ VPD), and G is the soil heat flux.

Land surface models/Earth system models The most complex approach to estimating ET is through full Land Surface Models (LSMs) or Earth System Models (ESMs). These models are typically driven by meteorological data and aim to simulate all of the relevant biogeochemical processes and states governing the exchange of energy, water, and carbon throughout the entire land surface or complete Earth system, including ocean and atmosphere. Some of these models assimilate any relevant observation from both space and in situ to constrain the complexity of linkages and feedbacks. While the estimate of ET from LSMs and ESMs is subject to potentially greater uncertainty relative to the previously described approaches due to increased complexity and degrees of freedom, LSMs and ESMs allow more realistic feedbacks to and from ET given changes in the Earth system or climate (Mueller et al., 2011). Summary Remote sensing of ET is currently a high-level research and science priority, especially as ET is central to connecting the water, energy, and carbon cycles; a modulator of regional rainfall; a significant factor in flood and drought processes and models; the primary climatic predictor of biodiversity; and critical for the agricultural industry. In situ measurement of ET requires

328

LANDSLIDES

cost-constraining equipment; as such, major international efforts, such as the Global Energy and Water Cycle Experiment (GEWEX), have focused on determination of ET from existing remote sensing assets (Jiménez et al., 2011; Vinukollu et al., 2011). The techniques described here provide an overview of how the scientific community estimates ET from remote sensing.

Acknowledgment This research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the NASA. Bibliography Baldocchi, D., et al., 2001. FLUXNET: a new tool to study the temporal and spatial variability of ecosystem-scale carbon dioxide, water vapor, and energy flux densities. Bulletin of the American Meteorological Society, 82, 2415–2434. Fisher, J. B., Tu, K., and Baldocchi, D. D., 2008. Global estimates of the land-atmosphere water flux based on monthly AVHRR and ISLSCP-II data, validated at 16 FLUXNET sites. Remote Sensing of Environment, 112, 901–919. Fisher, J. B., Whittaker, R. H., and Malhi, Y., 2011. ET come home: a critical evaluation of the use of evapotranspiration in geographical ecology. Global Ecology and Biogeography, 20, 1–18. Jiménez, C., et al., 2011. Global inter-comparison of 12 land surface heat flux estimates. Journal of Geophysical Research, 116, 1–27, doi:10.1029/2010JD014545. Jung, M., Reichstein, M., and Bondeau, A., 2009. Towards global empirical upscaling of FLUXNET eddy covariance observations: validation of a model tree ensemble approach using a biosphere model. Biogeosciences, 6, 2001–2013. Li, Z.-L., et al., 2009. A review of current methodologies for regional evapotranspiration estimation from remotely sensed data. Sensors, 9, 3801–3853. Mu, Q., Zhao, M., and Running, S. W., 2011. Improvements to a MODIS global terrestrial evapotranspiration algorithm. Remote Sensing of Environment, 111, 519–536. Mueller, B., et al., 2011. Evaluation of global observations-based evapotranspiration datasets and IPCC AR4 simulations. Geophysical Research Letters, 38, 1–7, doi:10.1029/2010GL046230. Vinukollu, R. K., Wood, E. F., Ferguson, C. R., and Fisher, J. B., 2011. Global estimates of evapotranspiration for climate studies using multi-sensor remote sensing data: evaluation of three process-based approaches. Remote Sensing of Environment, 115, 801–823.

LANDSLIDES Vernon H. Singhroy Applications Development Section, Natural Resources Canada, Canada Centre for Remote Sensing, Ottawa, ON, Canada

Synonyms Debris flow; Mass movement; Mudslide; Rock avalanche

Definition Landslide is used to describe the downslope movement of soil and rock under the effects of gravity. In some cases, other terms such as mass movements and slope failure are used interchangeably with landslides. The most common triggers of landslides are earthquakes, heavy rains, thawing of frozen ground, river and coastal erosion, and frequent infrastructure and building development. Many types of landslides are usually associated with specific mechanics of slope failure and the properties and characteristics of failure type. Figure 1 shows a simple illustration of a rotational landslide which illustrates the commonly used labels for the parts of a landslide. Introduction Landslides are among one of the serious geological hazards which threaten and influence the socioeconomic conditions of many countries (Schuster, 1996). They are the manifestation of slope instability. An example of their destructive nature is shown in Figure 2. The La Conchita landslide in California killed ten persons. Geologists and engineers have long tried to identify the conditions of slope failure and mitigate their risk to infrastructure and populated areas. There are various techniques used to map and evaluate landslides. Large-scale, stereo aerial photographs are one of these tools that have been extensively used in landslide investigations. Because of their three-dimensional capability, they provide essential geologic and geomorphic information necessary for landslide inventory mapping. Recently, there has been increasing uses of high-resolution satellite images (1–5 m) for various landslide investigations. The recent advances in radar interferometry are providing valuable insights in monitoring slow-moving landslides. The following section briefly discusses the uses of these techniques in landslide investigations. Landslide mapping Stereo aerial photographs are used extensively to produce landslide inventory maps. They allow the identification of geomorphic, geologic, and related land use features related to landslides (Mollard and Janes, 1993). Geological and geomorphologic units related to landslide inventories can be interpreted on the basis of morphological, textural, and structural characteristics using stereo aerial photos and high-resolution satellite images. Landslide inventory maps are usually published at various scales, such as national (1; 1,000,000), regional (1:100,000), medium (1:25,000–50,000), and large scales (>1:15,000). For instance, detail inventory requires detail aerial photos and high-resolution satellite images to assist the interpreter to make conclusions on types and causes of the landslide. Recently, the high-resolution satellite Google images provided a cheap and valuable source of

LANDSLIDES

329

Crown cracks Crown Original ground surface Minor scarp

k an

Transverse cracks

gh

l tf

Ri

Ma

in s

He

car

p

ad

Transverse ridges

Radial cracks

Surface of rupture Toe Main body Foot

Toe of surface of rupture

Surface of separation

Landslides, Figure 1 Parts of a landslide (Modified from Varnes (1978), by Highland and Bobrowsky (2008)).

Landslides, Figure 2 This landslide occurred at La Conchita, California, USA, in 2005. Ten people were killed (USGS Photo).

locating landslides before and after the devastating earthquake in Haiti in January 2010 (Figure 3). Other high-resolution optical systems (1–5 m) such as IKONOS, Quickbird, and IRS images and the stereo capability of SPOT 5 are useful for landslide recognition

and related land use mapping. Whenever possible, the highest-resolution images should be used to identify and interpret the geomorphic and associated features shown in Figure 1. Large landslides are easily recognized from medium resolution 30 m Landsat TM images.

330

LANDSLIDES

Landslides, Figure 3 Landslides near Port-au-Prince, before and after Haiti earthquake.

Recent research has shown that high-resolution stereo radar and optical images, combined with topographic and geological information, have assisted in mapping the geomorphic characteristics of deep-seated landslides needed to produce of landslide inventory maps (Singhroy et al., 1998; Singhroy, 2005). The multi-incidence angle, stereo, and high-resolution capabilities of the various radar satellites are particularly useful in providing terrain and geomorphic information needed to produce landslide inventory maps. Currently, damage assessment related to landslides and other disasters in support of relief efforts uses aerial photography, videos, high-resolution satellite images, and ground checks.

Landslide monitoring using InSAR Landslides usually resulted in extreme economic and societal costs, despite our increased understanding of the mechanisms of failure and large ground deformation. Current state of the art in real-time monitoring of active slopes developed for early warning of landslides is very expensive. Satellite radar interferometry is used increasingly to complement real-time monitoring such as GPS and in situ field measurements (Singhroy, 2008). Interferometric Synthetic Aperture Radar (InSAR) techniques are being used to measure small millimeter displacement on slow-moving landslides. An interferometric image represents the phase differences between the backscatter signals in two SAR images obtained from similar positions in space. In case of spaceborne SAR, the images are acquired

from repeat-pass orbits. The phase differences between two repeat-pass images result from topography and from changes in the line-of-sight distance (range) to the radar due to displacement of the surface or change in the atmospheric propagation path length. For a nonmoving target, the phase differences can be converted into a digital elevation map if very precise satellite orbit data are available. Typical scales for SAR interferometry application to landslide movements are millimeters to centimeters per orbit cycle of the radar satellite. This orbit cycle can range from 44 days for ALOS and 10 days for TerraSAR-X. Constellation missions such as Cosmo-SkyMed, Sentinel, and RADARSAT constellation mission are reducing the orbits to 1–4 days. It is clear that InSAR techniques can be used to monitor landslide motion under specific conditions, provided coherence is maintained over the respective orbit cycle. Using data pairs with short perpendicular baselines, short time intervals between acquisitions, and correcting for the effect of topography and atmospheric effects, reliable measurements of surface displacement can be achieved. The InSAR deformation maps provide linear motion at the line of site. Although this is very useful information, landslide motion is very complex with nonlinear vectors. Therefore, InSAR techniques do not provide the complete 3d motion Vegetation decorrelates the radar signals. Therefore, stable coherent targets such as installed corner reflectors or man-made constructions such as houses, roads, and bridges are used to calculate the landslide motion. The uses of installed field corner reflectors are increasing on remote vegetated sites. Acquiring about 30 InSAR images on coherent targets over long periods are analyzed by a Permanent

LANDSLIDES

331

Instrumentation Legend Tiltmeters Crackmeters Extensometers Differential GPS

RADARSAT-1 100 50 Millimeters

0 −50 −100 −150 −200 −250 2000

Vertical displacement (millimeters)

RADARSAT 2

2001

2002

2003 2 2004 Time (Ye (Years)

2005

2006

2007

30 0 -30 -60 -90 -120 -150 2008

2009

2010

Time (Years)

2011

30mm/y

Landslides, Figure 4 Frank Slide, Canada: monitoring landslide motion from RADARSAT InSAR (2000–2010) (Modified from Mei et al. 2007).

Scatterer Technique PSInSAR™ developed by Ferretti et al. (2001). With the Permanent Scatterer Technique, the movement of small objects (down to about 1 m2) can be monitored. This method has been applied to map subsidence and slow-moving landslides and many parts of the world. The Frank rock avalanche is provided to demonstrate the capability of InSAR to monitor gradual motion on large rock avalanche in the Canadian Rockies. The Frank Slide, a 30 106 m3 rock avalanche of Paleozoic limestone, occurred in April 1903 on the east face of Turtle Mountain in southern Alberta, Canada. Seventy fatalities were recorded. This slide is still active. Several factors contribute to this rock avalanche. These include the geological structure of the mountain, subsidence from coal mining at the toe of the mountain, blast-induced seismicity, above-average precipitation in years prior to the slide, and freeze-thaw cycles (Cruden and Hungr, 1996). GPS stations and several in situ monitors are installed to monitor post-slide activity at specific

locations (Figure 4). Current InSAR monitoring is complementing the in situ measurements. The fact that the rock covering the rock avalanche is bare and dry leads to the high coherence and identification of more than 95 % of the coherent target monitoring targets for the Frank Slide area. Due to their great density and excellent coverage, the coherent target measurements of this area are a reliable reflection of current deformation pattern. The most recent InSAR results (Figure 4) have shown that during a period from April 2004 to October 2006, the foot of the eastern slope of Turtle Mountain, the ground surface above the coal mine, was found to subside at an average rate of about 3.1 mm per year supporting the speculation that underground coal mining triggered the Frank landslide (Mei et al., 2007). The above examples show that satellite images are providing reliable complementary techniques to landslide mapping and monitoring, and therefore, its uses are increasing in landslide investigation and mitigation.

332

LAW OF REMOTE SENSING

Bibliography Cruden, D. M., and Hungr, O., 1996, 1986. The debris of Frank Slide and theories of rockslide-avalanche mobility. Canadian Journal of Earth Sciences, 23, 425–432. Ferretti, A., Prati, C., and Rocca, F., 2001. Permanent scatterers in SAR interferometry. IEEE Transactions on Geoscience and Remote Sensing, 39, 8–20. Highland, L. M., and Bobrowsky, P., 2008. The Landslide Handbook – A Guide to Understanding Landslides. Reston: U.S. Geological Survey Circular, Vol. 1325, p. 129. Mei, S., Poncos, V., and Froese, C., 2007. InSAR Mapping of Millimetre-scale Ground Deformation over Frank Slide, Turtle Mountain, Alberta. Alberta Energy and Utilities Board, EUB/ AGS Earth Science Report 2007, 1–62 pp. Mollard, J. D., and Janes, J. R., 1993. Airphoto Interpretation of the Canadian Landscape. Ottawa: Energy, Mines and Resources, Canada, p. 415p. Schuster, R. L., 1996. Socio- economic significance of landslides. In Turner, A. K., and Schuster, R. L. (eds.), Landslides: Investigation and Mitigation. Washington, DC: National Academy Press. Report 247, Transportation Research Board, NRC, pp. 12–35. Singhroy, V., 2005. Remote sensing for landslide assessment: chapter 16. In Glade, T., Anderson, M., and Crozier, M. J. (eds.), Landslides Hazard and Risk. Chichester/Hoboken: Wiley, pp. 469–549. Singhroy, V., 2008. Satellite remote sensing applications for landslide detection and monitoring (chap. 7). In Sassa, K., and Canuti, P. (eds.), Landslide Disaster Risk Reduction. Berlin: Springer, pp. 143–158. Singhroy, V., Mattar, K. E., and Gray, A. L., 1998. Landslide characteristics in Canada using interferometric SAR and combined SAR and TM images. Advances in Space Research, 3, 465–476. Varnes, D. J., 1978. Slope movement, types and processes. In Schuster, R. L., and Krizek, R. J. (eds.), Landslides-Analysis and Control. Washington, DC: National Research Council. Transportation Research Board Special Report 176, pp. 11–23.

LAW OF REMOTE SENSING Joanne Irene Gabrynowicz National Center for Remote Sensing, Air, and Space Law, The University of Mississippi School of Law, Missisippi, MS, USA

Overview All of the international space law began with the Declaration of Legal Principles Governing the Activities of States in the Exploration and Use of Outer Space (Declaration), adopted in 1962 by the United Nations General Assembly. National space laws, like that of the United States, were influenced by international space law and developed in tandem. The Declaration is the foundation for the Treaty on Principles Governing the Activities of States in the Exploration and Use of Outer Space, including the Moon and Other Celestial Bodies (Outer Space Treaty). The Outer Space Treaty entered into force at the height of the Cold War on October 10, 1967. It contains fundamental space

law principles that are directly applicable to remote sensing activities, such as all nations have the nonexclusive right to use space. In less than a decade, four more treaties followed, some of which also have legal principles applicable to remote sensing.

Legislative history in the United Nations In 1970, Prof. A.A. Cocca of Argentina first introduced remote sensing as a specific legal topic in a paper to the Legal Subcommittee of the U.N. Committee on the Peaceful Uses of Outer Space (UNCOPUOS). In 1971, a Working Group was formed in the Legal Subcommittee to consider the paper, and in 1973, the Scientific and Technical Subcommittee of UNCOPUOS issued its first report containing a section on remote sensing. In 1974, the General Assembly adopted a resolution recommending that the LSC should consider the question of the legal implications of remote sensing of the Earth from space. In 1975–1976, the first discussions concerning the legal implications of remote sensing began. Initially, the participating Nation-States organized themselves into three groups, or blocs: the Soviet Union/East Europe; the Group of 77 (G-77), consisting of developing nations in Africa, Asia, and Latin America; and the Western Group. The primary issues were whether or not sensing States had to obtain the consent of a sensed State prior to acquiring data from space or prior to distributing data to a third party. The Soviet Union and France submitted a proposal to require sensing States to obtain the prior consent of sensed States before data could be made available to other entities. The G-77 initially opposed both remote sensing itself and the distribution of data. The United States, then the only sensing state, advocated a free flow of data and therefore opposed prior constraints. In 1975, the General Assembly recommended that the Legal Subcommittee continue consideration of remote sensing from space as a high priority. It specifically pointed to the use of remote sensing regarding the Earth’s natural resources and environment. It also recommended drafting principles regarding points on which States agreed. The General Assembly noted that the Scientific and Technical Subcommittee had examined operational and experimental questions and now recommended that studies be conducted on organizational and financial matters. It also endorsed an international remote sensing training center for personnel from developing nations. In 1976, this work slowed down because the Soviet Union attempted to link some remote sensing issues to the developing Moon Treaty. By late 1976, progress was made on formulating some draft principles. The General Assembly noted in a resolution that the Legal Subcommittee formulated five draft principles and identified three new common elements identified by States. Main issues From 1977 to 1979, the Working Group focused on three main issues: whether or not “should” or “shall” ought to

LAW OF REMOTE SENSING

be used in the principles, providing consultation and dispute resolution procedures, and the Soviet Union’s proposal to limit image gathering to 50 m spatial resolution. Cold War politics drove the legal debate. A specific attempt to codify the 50 m limit was made on May 19, 1978, when Cuba, Czechoslovakia, the German Democratic Republic, Hungary, Mongolia, Poland, Romania, and the Union of Soviet Socialist Republics signed the Convention on Transfer and Use of Data of Remote Sensing of the Earth from Outer Space in Moscow. It provided that if a Contracting Party was in possession of data with resolution higher than 50 m, it was forbidden from making the data available to anyone without the explicit consent of the sensed State. This did not significantly influence the Legal Subcommittee negotiations. In a 1979 resolution, the General Assembly again recommended that the Legal Subcommittee continue work on the draft principles on a priority basis. The evolution and use of the US Landsat system catalyzed addressing many general and specific remote sensing issues in legal terms. In 1982 and 1983, as a cost recovery method, the United States raised the access fee for a Landsat ground station from $200,000 (US) to $600,000 (US), and the cost of computer-compatible tapes increased. The United States announced its intention to commercialize the Landsat system. Identifying principles regarding data access, among others, became more pressing. In 1981, 1982, 1983, and 1984, the General Assembly adopted resolutions that noted each year’s progress toward developing principles and continued to urge the Legal Subcommittee to develop the legal implications of remote sensing on a priority basis. In 1984, the US Congress passed the Land Remote Sensing Commercialization Act (Commercialization Act). It adopted the nondiscriminatory access policy forged in the Legal Subcommittee and provided for a three-phased process to establish commercial remote sensing. The first phase was to award a contract for existing Landsat operations to a private sector operator. The second phase was to be a transition period in which both the government and the private sector would operate satellites, with government activities phasing out. The third phase was to be a fully private, commercial environment. The US contract award process had begun and the negotiators in the United Nations took note. Conversely, the United States needed to have the legitimacy of commercial remote sensing activities accepted. France was preparing to launch its first remote sensing satellite, SPOT-1, which it did in 1986. In 1984, France proposed alternate language for the draft data access principle, and negotiations on the remote sensing principles were revitalized. In 1985, significant portions of the draft principles were still not agreed upon. According to at least one report, the Soviet Union/Eastern bloc did not participate in discussions of Article XII, the data access principle, out of concern that there would be no acceptable solutions for the G-77 nations. Some nations, including Mexico and

333

Vietnam, believed there was insufficient time to consider the draft text. The General Assembly again adopted a resolution endorsing the Legal Subcommittee’s work and added that it should finalize the draft set of principles. The Chair of the Working Group was Austria, and it offered an alternate text based on consultations regarding the French proposal. Discussions on the Austrian text were held. No changes were made in 1986. On December 3, 1986, The Principles Relating to Remote Sensing of the Earth from Outer Space (Principles) were adopted by the General Assembly.

The principles relating to remote sensing of the Earth from outer space The Principles embody the view that Outer Space is a resource for all humanity and should be used for the general benefit of all nations. They encourage international cooperation and address access and distribution of data and information generated by national civilian remote sensing systems. Primary data are the raw data delivered in the form of electromagnetic signals, photographic film, magnetic tape, or any other means. Processed data are the products resulting from processing primary data, and analyzed information means information resulting from interpreting processed data. Remote sensing activities include operations, data collection, storage, processing, interpretation, and dissemination. The Principles set a standard of international cooperation among sensing and sensed States while attempting to achieve a balance between their rights and interests. Needs of developing nations are given special regard. The Principles specifically promote protection of the Earth’s environment and of humanity from natural disasters. States that possess remotely sensed information useful for averting harmful phenomena are required to disclose the information to concerned States. If the potential harm threatens people, the obligation to disclose requires promptness and extends to processed data and analyzed information. The rights and responsibilities of sensed and sensing States are particularly addressed in Articles IV and XII. Article IV sets a legal standard for behavior among sensed and sensing States, and Article XII is a data dissemination statute. Together, they provide a fluid legal regime that obliges sensing States to avoid harm to sensed States and to provide them with access to primary data and processed data concerning their own territory on a nondiscriminatory basis. This was the compromise between terrestrial sovereignty and the freedom to use space. The legitimacy of space-based remote sensing was accepted by ensuring that a sensed State would have access to the imagery of its territory. Analyzed information available to sensing States is also to be available to the sensed States on the same basis and terms. In turn, sensed States are to meet reasonable cost terms and do not have access to analyzed information legally unavailable to the sensed States, for example, proprietary information.

334

LIDAR SYSTEMS

The Principles were contained in the first major resolution to emerge from COPUOS in more than a decade and provide a foundation for the continued evolution of international remote sensing law. The question of whether or not the remote sensing principles ought to become a treaty continues to be raised in the COPUOS Legal Subcommittee.

US law: The oldest national remote sensing law On October 28, 1992, the US Congress passed the Land Remote Sensing Policy Act of 1992 (Policy Act). It replaced the 1984 Commercialization Act. Congress adopted the nondiscriminatory access policy for second time. The Policy Act’s focus is long-term remote sensing policy and its numerous facets. Specific matters addressed include program management, Landsat 7 procurement, Landsat 4 through 7 data policy, transfer of Landsat 6 program responsibilities, regulatory authority and administration of public and private remote sensing systems, federal research and development, advanced technology demonstration, Landsat 7 successor systems, data availability and archiving, and the continued prohibition of weather satellite commercialization. The legislation features a focus on the value of remote sensing in conducting global change research and other public sector applications, a recasting of remote sensing activities, and provisions for the future evolution of remote sensing policy. In 2008, efforts were made to replace the 1992 law with a new statute titled, the National Land Imaging Program. This bill was intended to embody the US new national remote sensing policy to implement a long-term operational land imaging program. It was not made into law. New law and policy of remote sensing nations India, the United Kingdom, and some other remote sensing nations have policies rather than laws. However, increasingly, major remote sensing nations are promulgating national laws. In 1999, Canada announced a policy that, in 2005, received Royal Assent and which came into force in April of 2007 as the Remote Sensing Space Systems Act. In 2007, the German Act on Satellite Data Security entered into force. These laws address the commercial availability of high-resolution imagery, and both seek to ensure national security interests within a commercial context. In 2008, Japan and France each passed a comprehensive national space law that includes sections on remote sensing. Court decisions in France, Germany, and the United States regarding the intellectual property and other aspects of remote sensing are also adding to the overall corpus of law. Globalizing Earth observations In the 1990s, the trend to internationalize Earth observation satellite operations began, and important new agreements were formulated. On November 19, 1998, the US National Oceanic and Atmospheric Administration and

the European Organisation for the Exploitation of Meteorological Satellites entered into an agreement on a joint polar-orbiting operational system, and a second agreement was entered into on June 24, 2003. The International Charter on Space and Major Disasters became operational on November 1, 2000. Some nations, like Belgium, do not have indigenous remote sensing capabilities but, nonetheless, are developing national remote sensing laws because they are participating in remote sensing consortia. It is clear that remote sensing law will continue to develop.

Bibliography Christol, C. Q., 1982. The Modern International Law of Outer Space. New York: Pergamon Press. Gabrynowicz, J., 2002a. The United Nations Principles Relating to Remote Sensing of the Earth from Space: A legislative History – Interviews of Members of the United States Delegation. Oxford, MS: The National Center for Remote Sensing, Air, and Space Law. ISBN 0-9720432-1-7. Gabrynowicz, J., 2002b. Proceedings, The First International Conference on the State of Remote Sensing Law. Oxford, MS: The National Center for Remote Sensing, Air, and Space Law. ISBN 0–9720432-3–3. Gabrynowicz, J., 2005. The perils of landsat from grassroots to globalization: a comprehensive review of U.S. remote sensing law with a few thoughts for the future. Chicago Journal of International Law, 6, 72. Gabrynowicz, J., 2008. The second international conference on the state of remote sensing law. Journal of Space Law, 34(1), ISSN 0095–7577 - 9720432–3–3, The National Center for Remote Sensing, Air, and Space Law: Oxford, MS. Graham, J. F., and Gabrynowicz, J., 2002. Landsat 7: Past, Present and Future. Oxford, MS: The National Center for Remote Sensing, Air, and Space Law. ISBN 0-9720432-0-9. The Land Remote Sensing Laws and Policies of National Governments: A Global Survey, The National Center for Remote Sensing, Air, and Space Law, Oxford, MS, 2007. Available from http://www.spacelaw.olemiss.edu/resources/pdfs/noaa.pdf. Last accessed 21 June 2012. United Nations Office for Outer Space Affairs. United Nations Office for Outer Space Affairs, http://www.unoosa.org/oosa/en/ SpaceLaw/index.html. Last accessed 21 June 2012.

LIDAR SYSTEMS Robert Menzies Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA

Synonyms Ladar; Laser radar Definitions Lidar. LIght Detection And Ranging Ladar. LAser Detection And Ranging

LIDAR SYSTEMS

335

Introduction A lidar system in the strictly defined sense of the acronym measures range to a “target” that provides a signal that can be detected. Thus, the lidar system includes both a transmitter and a receiver. Ranging is accomplished using time-of-flight methods. The target can be a “hard” target that is essentially opaque to the lidar wavelength, not allowing measureable penetration beyond its range, or a “diffuse” scattering medium that allows penetration and range gating. Examples of the former are the surfaces of the Earth and other planets, or man-made objects. Examples of the latter are atmospheric aerosols and gases. In reality, these are terms that are commonly used but do not have strict, universally accepted definitions. In fact, the term “lidar” itself is commonly applied to systems that contain transmitters and receivers but do not have inherent range measurement capability. The lidar community is inclusive in this regard. Following David Tratt’s introductory entry (“Emerging Technologies: Lidar”), which describes lidar basics and various classes or categories of lidar, we provide here a summary of the current capabilities in these various lidar applications areas. Our lidar categories are altimetry and mapping systems, backscatter systems, Doppler systems, and differential absorption systems. Comments on emerging technologies and methods are included. Lidar/ladar applications cover a wide range of activities and interests. The 3D imaging applications are a growth area with strong support from the defense community. System developments in this area are included only in brief overview mode. The balance in this entry is tilted more toward systems developed for scientific investigations.

a challenge, however. The advent of avalanche photodiode (APD) arrays and photon-counting receivers (e.g., Aull et al., 2002; Albota et al., 2002), combined with optical methods for simultaneous transmission of multiple beams, have greatly increased the mapping efficiencies of airborne and space systems. The use of statistical methods in a photon-counting mode has allowed the use of compact, high pulse-repetition frequency (prf), low pulse energy laser transmitters in various imaging and mapping systems (Degnan et al., 2008; Steinvall et al., 2008). The Lunar Orbiter Laser Altimeter (LOLA) instrument, scheduled to launched in June, 2009, uses a Diffractive Optical Element (DOE) to produce a 5-beam pattern for provision of more spatial coverage than with prior space laser altimeters (Ramos-Izquierdo et al., 2009; Smith et al., 2010). The DOEs have found use in various airborne laser 3D mappers. An alternative to the use of scanners or elements such as DOEs, matched with APD arrays, is a flash lidar/ladar. The images in this type of system record the intensity reflected by the scene when flood-illuminated by the laser transmitter pulse. The laser transmitter irradiates the entire field of view of the receiver camera pixel array, and each pulse generates an entire frame of data (Stettner et al., 2005). The array elements are high-speed detectors that are periodically sampled in time at nanosecond timescales. The advances in hybridizing the focal planes with silicon CMOS read-out integrated circuits (ROICs), utilizing steady improvements in high-speed circuitry, provide the potential for growth with this approach. Laser sources can include semiconductor lasers and fiber lasers mated to power amplifiers.

Altimetry and mapping systems Laser altimetry is relatively mature, with heritage in aircraft instruments, followed by Earth-orbiting, Mars-orbiting, and Lunar-orbiting systems. The early altimeter/mapping instruments used a form of threshold detection to trigger a circuit that enabled range measurement to a “first return” scattering surface. The implementation of fast waveform recovery, or multistop detection circuits, increases data rates but provides structure information in the line-of-sight dimension. The Geoscience Laser Altimeter System (GLAS) on the Earthorbiting ICESat (Abshire et al., 2005) provided structure detail in the time domain, a capability that is essential for future use of laser 3D mappers in obtaining global estimates of biomass. High-resolution 3D imaging with very high depth resolution (
1 mm) can be achieved at km distances using fiber lasers and high bandwidth waveform encoding and decoding techniques (Buck et al., 2007). The current and next-generation systems combine multi-beam transmitter patterns with structural detail in the range dimension. The laser altimetric observational method provides line-of-sight detail that complements radar methods as well as higher spatial resolution in the cross dimensions. Spatial coverage is

Backscatter lidars Here we include elastic backscatter lidars and various types of inelastic backscatter lidars (e.g., Raman, fluorescence). The emphasis is on atmospheric studies using these systems. The intensity or energy in the return signal is important with backscatter lidar measurements. Some method of calibration and/or normalization must be used in order to turn the data into useful observations. In the visible, the molecular density, if known sufficiently well, can be used to provide a Rayleigh backscatter intensity that effectively calibrates at least the range dependence of the lidar efficiency factor, or the efficiency factor itself at a particular atmospheric altitude where particle scattering is assumed negligible. This is not a viable technique at longer wavelengths in the infrared, due to the rapid decrease of the Rayleigh scattering cross section with increasing wavelength. Backscatter lidars for cloud and aerosol studies date back to the early years of lidar, when ground-based lidars operating at visible wavelengths probed the stratospheric aerosol layers (e.g., Fiocco and Grams, 1964). The first Earth-orbiting lidar used for atmospheric studies was an elastic backscatter lidar (LITE, launched in 1994). GLAS operated both as an altimeter and an atmospheric lidar

336

LIDAR SYSTEMS

(Spinhirne et al., 2005). Currently the CALIPSO lidar is in Earth orbit, being used for cloud and aerosol studies. The CALIPSO transmitter is a diode-array pumped Nd: YAG laser, by far the most commonly used in backscatter lidars. A wide dynamic range of pulse energies and pulse-repetition frequencies are available in this laser medium. The compact micropulse lidars, which emit pulses in the micro-Joule range, are deployed around the globe in networks such as the MPLNET (Micropulse Lidar Network) (Campbell et al., 2008). Cloud and aerosol detection, characterization, and monitoring algorithms continue to improve for these compact lidars, making them more useful for deployment. The underlying technologies are robust. The vertical profiling capabilities of these lidars cannot be duplicated with passive instruments. Recently, a compact backscatter lidar was deployed on the surface of Mars as part of the Phoenix mission (Whiteway et al., 2008). A variant of the elastic backscatter lidar that is taking center stage in current and future atmospheric investigations is the High Spectral Resolution Lidar or HSRL. Although Doppler lidars are the ultimate high spectral resolution lidars, the term “HSRL” is commonly used in the lidar community to refer to a system that can separate the molecular Rayleigh backscatter signal from the aerosol backscatter signal. This obviates the need to assume a “lidar ratio” (i.e., aerosol extinction-to-backscatter ratio) when interpreting the range-dependent backscatter signals to deduce aerosol optical properties, thereby achieving more robust estimates of aerosol extinction coefficients. A progression in HSRL implementation has gone from early 1980s laser technology such as fragile dye laser systems (Shipley et al., 1983) to more robust solid-state lasers (Grund and Eloranta, 1991). Iodine vapor filters offer simplicity compared with the etalon filters in the HSRL receiver (Hair et al., 2001). More recently, airborne HSRL has been developed, and measurement results have been reported (Hair et al., 2008). The next-generation Earth-orbiting backscatter lidar for cloud and aerosol studies will likely be an HSRL. In fact, the European Space Agency’s Atmospheric Laser Doppler Instrument (ALADIN), a Doppler lidar in the Atmospheric Dynamics Mission (ADM) with atmospheric wind field measurements as its primary objective, is fundamentally an HSRL and will be used for investigations of aerosol optical properties (Ansmann et al., 2007) (see www.esa.int for further information). Resonance fluorescence lidars have been in use for decades to study dynamics and thermal properties of the middle atmosphere, particularly the mesosphere. Lidars built to measure alkalis in the upper mesosphere were also used as Rayleigh backscatter lidars to measure density and temperature profiles in the stratosphere and mesosphere (Hauchecorne and Chanin, 1980). Developments in solid-state laser technology and injection seeding methods have resulted in systems that are more amenable to transportation and operation at remote sites

(e.g., She et al., 2007). Systems that interact with a variety of metals in addition to sodium and the other alkalis are now in development for investigations over a wider range of altitudes (Gardner, 2004). Raman lidars are now commonly used for water vapor profiling and for characterizing the optical and microphysical properties of atmospheric aerosol. The latter method was described 20 years ago (Ansmann et al., 1990) and has continued to evolve into systems that are being used for characterization of major dust plumes that are transported long distances (e.g., Asian dust, Saharan dust) and for calibration/validation exercises (Mona et al., 2007). The former method has a long history and has slowly evolved with the use of improved techniques for minimizing background light, improved algorithms, and improved understanding of sources of bias. The use of Raman lidar for water vapor profiling in the lower atmosphere continues to gain credibility as the level of accuracy continues to improve (Adam and Venable, 2007; Leblanc and McDermid, 2008).

Doppler lidars The atmospheric gas molecules and aerosol particles are in bulk motion in the dynamic atmosphere, and backscattering of laser radiation from the molecules and aerosol particles produces Doppler shifts in frequency. Doppler lidars detect these frequency shifts to deduce wind profiles. Two types of Doppler lidar have received attention over the years: direct detection and coherent detection lidars. The coherent detection lidar is more sensitive and less difficult to implement at relatively longer wavelengths in the infrared, particularly at wavelengths longer than 1.5 mm, the so-called eye-safe region. The ultrahigh spectral resolution that is inherent with these systems makes coherent detection suitable for measuring Doppler-shifted backscatter from the atmospheric aerosol particles. The signal processing has similarities with Doppler radar. The use of rare-earth-doped solid-state laser technologies in the 2 mm wavelength region has been a popular choice for compact coherent detection systems. An example is the NOAA shipborne lidar, which has been used in many field campaigns (Tucker et al., 2009). Airborne systems date back to the mid-1980s when carbon dioxide gas laser transmitters were used (Bilbro et al., 1986). More recent, much more compact systems have also been deployed for measuring wind profiles with high spatial resolution (Hannon et al., 1999). Both the rare-earth-ion-doped solid-state crystal laser technology at 2 mm and the fiber laser technology developed primarily by the telecom industry have been employed in recent ground-based coherent Doppler lidars stationed at airports for airport safety enhancements. These lidar systems are being used for both wake vortex monitoring (e.g., Kopp et al., 2004) and wind shear detection and warning (e.g., Shun and Chan, 2008). Fiber laser

LIDAR SYSTEMS

technologies are being incorporated into current and future systems. Direct detection lidar is the appropriate choice for regions of the atmosphere containing very low aerosol particle concentration in the size range that is useful for optical scattering. The predominant scattering is molecular Rayleigh scattering. An early example was the use of direct detection Rayleigh lidar, modified with the incorporation of twin Fabry-Pérot interferometer filters in the receiver, for measurements of horizontal winds in the middle atmosphere (Chanin et al., 1989). An airborne direct detection Doppler lidar was developed, for tropospheric wind field measurements (Gentry et al., 2007). It is designed for autonomous operation on a high-altitude aircraft. The European Space Agency’s ALADIN lidar is planned for launch in 2010, as the centerpiece instrument in the Atmospheric Dynamics Mission (ADM). ALADIN uses solid-state Nd: YAG laser transmitter technology, frequency-tripled to the 355 nm near-UV wavelength. It contains two receivers, one for the narrow-band Mie scattered radiation from the atmospheric aerosol particles (employing a multichannel Fizeau interferometer) and the other for the Rayleigh scattered radiation from the molecules (employing a double-edge Fabry-Pérot etalon). Accumulation CCD’s are used in both receivers (see www.esa.int for further information).

Differential absorption lidars Differential absorption lidars require typically two carefully selected closely spaced transmit wavelengths and a laser transmitter subsystem that has either discrete or continuous tenability in the desired spectral region to interact with the species of interest. Early systems used dye lasers or nonlinear optics such as optical parametric oscillators to provide tunability. Atmospheric ozone and water vapor have been favorite measurement subjects for decades. More recent systems rely on solid-state laser technologies and modern techniques for generating tunable single-mode radiation with high spectral purity. Airborne systems have progressed in sophistication, with corresponding reductions in mass and dimensions as well. The LASE (Lidar Atmospheric Sensing Experiment) system was demonstrated in the 1990s as an autonomous operation water vapor differential absorption lidar on the high-altitude ER-2 aircraft (Browell et al., 1997). Currently, intercomparison campaigns involving multiple airborne water vapor systems with different designs are being planned and implemented in order to better understand the accuracies of measurement and quantify biases that might exist (Behrendt et al., 2007). Results to date show that measurement accuracies are in good agreement with expectations. Currently, a major challenge for differential absorption lidar is the measurement of atmospheric CO2. Measurements with very high accuracy over regional to global scales would improve understanding of fluxes between

337

atmosphere, land surface, and ocean surface. The influence that increasing carbon dioxide mixing ratio has on climate change has spurred an interest in applying both passive and active remote sensing techniques to address this question. Desired mixing ratio accuracy levels of better than 1 % place great demands on a differential absorption lidar system itself and require the minimization of errors due to imperfect knowledge of the relevant atmospheric parameters (Menzies and Tratt, 2003; Ehret et al., 2008). Demonstrations of CO2 mixing ratio measurement capability using ground-based, coherent detection systems have been reported (Gibert et al., 2008; Koch et al., 2008). Airborne systems are now being tested in flight campaigns, using both the solid-state 2 mm laser technology and the 1.6 mm fiber laser technology (Abshire et al., 2010; Spiers, et al., 2011). Studies of Earth-orbiting lidar systems for CO2 measurements are being conducted under the sponsorship of European, US, and Japanese space agencies (ESA, NASA, and JAXA respectively).

Summary Using an unofficial taxonomy of lidar systems, selected highlights of recent developments and future plans have been provided. Generally speaking, the future applications for altimetry and three-dimensional mapping will motivate increases in coverage within a given available time frame. This will most likely come from increases in total laser transmitter output power, along with optical technology. In other lidar application areas, engineering advances will be critical. For example, advances in compactness, electrical power efficiency, autonomy, and reliability will be essential for further use in hazard detection and monitoring, as well as expansion of regional and global networks for weather, climate, atmospheric composition, and environmental monitoring. Atmospheric greenhouse gas measurements, on a global scale, present high-precision measurement challenges. Nearly 50 years after the first demonstration of the laser, many lidar system applications are still driven by laser technology advances. For example, many applications still await the development of a wider range of laser sources in infrared spectral regions that are presently underutilized. The advent of the quantum cascade laser and other “bandgapengineered” semiconductor laser technologies, as well as fiber laser/amplifier technologies, are good examples of continuing laser technology advances. Acknowledgment This research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the NASA. Bibliography Abshire, J. B., Graham, H. R., Allan, R., Weaver, C. J., Mao, J., Sun, X., Hasselbrack, W. E., Kawa, S. R., and Biraud, S.,

338

LIDAR SYSTEMS

2010. Pulsed airborne lidar measurements of atmospheric CO2 column absorption. Tellus, 62B, 770–783. Abshire, J. B., Sun, X., Riris, H., Sirota, J. M., McGarry, J. F., Palm, S., Yi, D., and Liiva, P., 2005. Geoscience laser altimeter system (GLAS) on the ICESat mission: on-orbit measurement performance. Geophysical Research Letters, 32, L21S02, doi:10.1029/2005GL024028. Adam, M., and Venable, D., 2007. Systematic distortions in water vapor mixing ratio and aerosol scattering ratio from a Raman lidar. In Proceedings of SPIE, Vol. 6750, doi:10.1117/ 12.738205. Albota, M. A., Heinrichs, R. M., Kocher, D. G., Fouche, D. G., Player, B. E., O’Brien, M. E., Aull, B. F., Zayhowski, J. J., Mooney, J., Willard, B. C., and Carlson, R. R., 2002. Threedimensional imaging laser radar with a photon-counting avalanche photodiode array and microchip laser. Applied Optics, 41, 7671, doi:10.1364/AO.41.007671. Ansmann, A., Riebesell, M., and Weitkakmp, C., 1990. Measurements of atmospheric aerosol extinction profiles with a Raman lidar. Optics Letters, 15, 746–748. Ansmann, A., Wandinger, U., Le Rille, O., Lajas, D., and Straume, A. G., 2007. Particle backscatter and extinction profiling with the spaceborne high-spectral-resolution Doppler lidar ALADIN: methodology and simulations. Applied Optics, 46, 6606–6622. Aull, B. F., Loomis, A. H., Young, D. J., Heinrichs, R. M., Felton, B. J., Daniels, P. J., and Landers, D. J., 2002. Geiger-mode avalanche photodiodes for three-dimensional imaging. Lincoln Laboratory Journal, 13, 335–343. Behrendt, A., Wulfmeyer, V., Kiemle, C., Ehret, G., Flamant, C., Schaberl, T., Bauer, H.-S., Kiio, S., Ismail, S., Ferrare, R., Browell, E. V., and Whiteman, D. N., 2007. Intercomparison of water vapor data measured with lidar during IHOP_2002. Part II: airborne-to-airborne systems. Journal of Atmospheric and Oceanic Technology, 24, 22–39, doi:10.1175/ JTECH1925.1. Bilbro, J. W., DiMarzio, C. A., Fitzjarrald, D. E., Johnson, S. C., and Jones, W. D., 1986. Airborne Doppler lidar measurements. Applied Optics, 25, 3952, doi:10.1364/AO.25.003952. Browell, E. V., Ismail, S., Hall, W. M., Moore, A. S., Jr., Kooi, S. A., Brackett, V. G., Clayton, M. B., Barrick, J. D. W., Schmidlin, F. J., Higdon, N. S., Melfi, S. H., and Whiteman, D. N., 1997. LASE validation experiment. In Ansmann, A., Neuber, R., Rairoux, P., and Wandinger, U. (eds.), Advances in Atmospheric Remote Sensing with Lidar. Berlin: Springer, pp. 289–295. Buck, J., Malm, A., Zakel, A., Krause, G., and Tiemann, B., 2007. High-resolution 3D coherent laser radar imaging. In Proceedings of SPIE, Vol. 6550, doi:10-1117/12.719483. Campbell, J. R., Sassen, K., and Welton, E. J., 2008. Elevated cloud and aerosol layer retrievals from micropulse lidar signal profiles. Journal of Atmospheric and Oceanic Technology, 25, 685–700, doi:10.1175/2007JTECHA1034.1. Chanin, M. L., Garnier, A., Hauchecorne, A., and Porteneuve, J., 1989. A Doppler lidar for measuring winds in the middle atmosphere. Geophysical Research Letters, 16, 1273–1276. Degnan, J., Machan, R., Leventhal, E., Lawrence, D., Jodor, G., and Field, C., 2008. In-flight performance of a second-generation, photon counting, 3D imaging lidar. In Proceedings of SPIE, Vol. 6950, Laser Radar Technology and Applications XIII, doi:10-1117/12.784759. Durand, Y., Chinal, E., Endemann, M., Meynart, R., Reitebuch, O., and Treichel, R., 2006. ALADIN airborne demonstrator: a Doppler wind lidar to prepare ESA’s ADM-Aeolus Explorer mission. In Proceedings of SPIE, Vol. 6296, 62961D, doi:10.1117/12.680958. Ehret, G., Kiemle, C., Wirth, M., Amediek, A., Fix, A., and Houweling, S., 2008. Space-borne remote sensing of CO2,

CH4, and N2O by integrated path differential absorption lidar: a sensitivity analysis. Applied Physics B, 90, 593–608, doi:10.1007/s00340-007-2892-3. Esselborn, M., Wirth, M., Fix, A., Tesche, M., and Ehret, G., 2008. Airborne high spectral resolution lidar for measuring aerosol extinction and backscatter coefficients. Applied Optics, 47, 346–358. Fiocco, G., and Grams, G., 1964. Observation of the aerosol layer at 20 km by optical radar. Journal of Atmospheric Science, 21, 323–324. Gardner, C. S., 2004. Performance capabilities of middleatmosphere temperature lidars: comparison of Na, Fe, K, Ca, Ca+, and Rayleigh systems. Applied Optics, 43, 4941–4956. Gentry, B., McGill, M., Schwemmer, G., Hardesty, M., Brewer, A., Wilkerson, T., Atlas, R., Sirota, M., Lindemann, S., and Hovis, F., 2007. Development of an airborne molecular direct detection Doppler lidar for tropospheric wind profiling. In Proceedings of SPIE, Vol. 6681, doi:10.1117/12.739379. Gibert, F., Flamant, P. H., Cuesta, J., and Bruneau, D., 2008. Vertical 2-mm heterodyne differential absorption lidar measurements of mean CO2 mixing ratio in the troposphere. Journal of Atmospheric and Oceanic Technology, 25, 1477–1497, doi:10.1175/ 2008JTECHA1070.1. Grund, C. J., and Eloranta, E. W., 1991. University of Wisconsin high spectral resolution lidar. Optical Engineering, 30, 6–12. Hair, J. W., Caldwell, L. M., Krueger, D. A., and She, C.-Y., 2001. High-spectral-resolution lidar with iodine-vapor filters: measurements of atmospheric-state and aerosol profiles. Applied Optics, 40, 5280–5294. Hair, J. W., Hostetler, C. A., Cook, A. L., Harper, D. B., Ferrare, R. A., Mack, T. L., Welch, W., Ramos-Izquierdo, L., and Hovis, F. E., 2008. Airborne high spectral resolution lidar for profiling aerosol optical properties. Applied Optics, 47, 6734–6752. Hannon, S. M., Bagley, H. R., and Bogue, R. K., 1999. Airborne Doppler lidar turbulence detection: ACLAIM flight test results. In Proceedings of SPIE, Vol. 3707, 234, doi:10.1117/12.351378. Hauchecorne, A., and Chanin, M.-L., 1980. Density and temperature profiles obtained by lidar between 35 and 70 km. Geophysical Research Letters, 7, 565–568. Koch, G. J., Petros, M., Barnes, B., Beyon, J. Y., Amzajerdian, F., Yu, J., Kavaya, M. J., and Singh, U. N., 2004. Validar: a testbed for advanced 2-micron Doppler lidar. In Proceedings of SPIE, Vol. 5412, pp. 87–98, doi:10.1117/12.542116. Koch, G. J., Beyon, J. Y., Gibert, F., Bernes, B. W., Ismail, S., Petros, M., Petzar, P. J., Yu, J., Modlin, E. A., Davis, K. J., and Singh, U. N., 2008. Side-line tunable laser transmitter for differential absorption lidar measurements of CO2: design and application to atmospheric measurements. Applied Optics, 47, 944–956. Kopp, F., Rahm, S., and Smalikho, I. M., 2004. Characterization of aircraft wake vortices by 2-mm pulsed Doppler lidar. Journal Atmospheric and Oceanic Technology, 21, 194–206. Leblanc, T., and McDermid, I. S., 2008. Accuracy of Raman lidar water vapor calibration and its applicability to long-term measurements. Applied Optics, 47, 5592–5603. Menzies, R. T., and Tratt, D. M., 2003. Differential laser absorption spectrometry for global profiling of tropospheric carbon dioxide: selection of optimum sounding frequencies for high-precision measurements. Applied Optics, 42, 6569, doi:10.1364/ AO.42.006569. Mona, L., Amodeo, A., D’Amico, G., and Pappalardo, G., 2007. First comparisons between CNR-IAMM multi-wavelength Raman lidar measurements and CALIPSO measurements. In Proceedings of the SPIE, Vol. 6750, 675010, doi:10.1117/ 12.738011.

LIGHTNING

Ramos-Izquierdo, L., Scott, V. S., Connelly, J., Schmidt, S., Mamakos, W., Guzek, J., Peters, C., Liiva, P., Rodriguez, M., Cavanaugh, H., and Riris, H., 2009. Optical system design and integration of the lunar orbiter laser altimeter. Applied Optics, 44, 3035–3049, doi:10.1364/AO.44.007621. She, C. Y., Vance, J. D., Kawahara, T. D., Williams, B. P., and Wu, Q., 2007. A proposed all-solid-state transportable narrowband sodium lidar for mesopause region temperature and horizontal wind measurement. Canadian Journal of Physics, 85, 111–118. Shipley, S. T., Tracy, D. H., Eloranta, E. W., Trauger, J. T., Sroga, J. T., Roesler, F. L., and Weinman, J. A., 1983. High spectral resolution lidar to measure optical scattering properties of atmospheric aerosols. 1: theory and instrumentation. Applied Optics, 22, 3716–3724. Shun, C. M., and Chan, P. W., 2008. Applications of an infrared Doppler lidar in detection of wind shear. Journal of Atmospheric and Oceanic Technology, 25, 637–655, doi:10.1175/ 2007JTECHA1057.1. Smith, D. E., Zuber, M. T., Neumann, G. A., Lemoine, F. G., Mazarico, E., Torrence, M. H., McGarry, J. F., Rowlands, D. D., Head, III, J. W., Duxbury, T. H., Aharonson, O., Lucey, P. G., Robinson, M. S., Barnouin, O. S., Cavanaugh, J. F., Sun, X., Liiva, P., Mao, D., Smith, J. C., and Bartels, A. E., 2010. Initial observations from the Lunar Orbiter Laser Altimeter (LOLA). Geophysical Research Letters, 37, L18204, doi:10.1029/ 2010gl043751. Spiers, G. D., Menzies, R. T., Jacob, J., Christensen, L. E., Phillips, M. W., Choi, Y., and Browell, E. V., 2011. Atmospheric CO2 measurements with a 2 μm airborne laser absorption spectrometer employing coherent detection. Applied Optics, 50(14). Spinhirne, J. D., Palm, S. P., Hart, W. D., Hlavka, D. L., and Welton, E. J., 2005. Cloud and aerosol measurements from GLAS: overview and initial results. Geophysical Research Letters, 32, L22S03, doi:10.1029/2005GL023507. Steinvall, O., Sjoqvist, L., Henriksson, M., and Jonsson, P., 2008. High resolution ladar using time-correlated single-photon counting. In Proceedings of SPIE, Vol. 6950, doi:10.1117/ 12.778323. Stettner, R., Bailey, H., and Silverman, S., 2005. Large format time-of-flight focal plane detector development. In Proceedings of SPIE, Vol. 5791, pp. 288–292. Tucker, S. C., Brewer, W. A., Banta, R. M., Senff, C. J., Sandverg, S. P., Law, D. C., Weickmann, A. M., and Hardesty, R. M., 2009. Doppler lidar estimation of mixing height using turbulence, shear, and aerosol profiles. Journal of Atmospheric and Oceanic Technology, 26, 673–688, doi:10.1175/ 2008JTECHA1157.1. Werner, C., Flamant, P. H., Reitebuch, O., Köpp, F., Streicher, J., Rahm, S., Nagel, E., Klier, M., and Herrmann, H., 2001. Wind infrared Doppler lidar instrument. Optical Engineering, 40, 115, doi:10.1117/1.1335530. Whiteway, J., Daly, M., Carswell, A., Duck, T., Dickinson, C., Komguem, L., and Cook, C., 2008. Lidar on the Phoenix mission to Mars. Journal of Geophysical Research, 113, E00A08, doi:10.1029/2007JE003002.

Cross-references Cryosphere, Measurements and Applications Ocean, Measurements and Applications Optical/Infrared, Atmospheric Absorption/Transmission, and Media Spectral Properties Optical/Infrared, Radiative Transfer Optical/Infrared, Scattering by Aerosols and Hydrometeors

339

LIGHTNING Rachel I. Albrecht1, Daniel J. Cecil2 and Steven J. Goodman3 1 Divisão de Satélites e Sistemas Ambientais (DSA/CPTEC), Instituto Nacional de Pesquisas Espaciais (INPE), Cachoeira Paulista, SP, Brazil 2 Marshall Space Flight Center (MSFC), National Aeronautics and Space Administration (NASA), Huntsville, AL, USA 3 National Environmental Satellite, Data, and Information Service (NESDIS), National Oceanic and Atmospheric Administration (NOAA), Silver Spring, MD, USA

Synonyms Cloud flash; Ground flash Definition Lightning or lightning discharge. A series of transient and multiple electrical breakdown pulses producing high-current channels (Uman, 1987). Lightning flash. A luminous manifestation accompanying a sudden electrical discharge which takes place from or inside a cloud or, less often, from high structures on the ground or from mountains (WMO, 2011). Introduction A lightning flash is a noncontinuous multi-scale physical process that ranges from the initial breakdown of air to the actual discharge propagation in discrete steps that can occur from cloud to ground (CG) or inside the clouds, i.e., intracloud (IC). In the case of CG lightning, the lightning channel formation is led by stepped leaders (that creates a conducting path between charge centers) and then followed by one or multiple return strokes that traverse the channel moving electric charges and neutralizing the leaders (Rakov and Uman, 2003). These series of return strokes are the lightning flash, and each stroke is guided by the dart leaders that propagate downward on the track of a preceding return stroke. CG flashes are also classified by the polarity of lowered charge: negative and positive. Negative flashes are more common and exhibit several return strokes, while positive flashes have a single or very few return strokes, but higher current than the negative ones. These processes occur too rapidly for the human eye to distinguish, and the flash appears as a single channel lasting for less than a second. Lightning detection networks typically look for the electric field changes associated with such processes. In the case of IC lightning, recoil streamers propagate within the track of positive branches of a bi-leader carrying strong negative charges. The lightning flash is terminated when the electric field is reduced to the point where it cannot sustain the discharge’s propagation anymore.

340

LIGHTNING

The electromagnetic spectrum of lightning The rapid release of electric energy inside the lightning channel generates a shock wave and electromagnetic radiation in a broad spectrum, see the entry Radiation, Electromagnetic. The shock wave rapidly decays into an acoustic wave we know as thunder. The electromagnetic radiation ranges from radio frequencies through visible to X-rays and gamma rays, composing the basis for ground-based lightning location systems (LLS) and remote sensing from satellites. Each lightning component (stepped leaders, return strokes, recoil streamers) emits electromagnetic energy proportional to the charge carried and its derivative in time. Negative stepped leaders are associated with strong negative currents in very short pulses (1 ms) and are detectable in very high frequencies (VHF, 1–200 MHz), as well as the dart leaders and recoil streamers, but with a relatively lower electrical current (MacGorman and Rust, 1998). The return strokes of a CG are high-energy discharges typically of the magnitude 10–100 kA in long pulses and radiate from the very low to high frequency range (1 kHz–10 MHz). Positive CG return strokes usually have continuous high current (>100 kA) and therefore are easily detected by VLF systems. Radio emissions from lightning occur in the form of short pulses by accelerated charges during the fastchanging current steps, while the optical emissions occur from ionized and dissociated gases by thermal radiation of the lightning channel (Goodman et al., 1988). The heating in the channel reaches temperatures above 20,000 K resulting in optical emissions primarily in discrete atomic lines with some continuum at shorter wavelengths. Several measurements of lightning emission in the cloud top have shown strongest emissions at the neutral oxygen (OI(l)) and neutral nitrogen (NI(l)) lines, i.e., 777.4 and 868.3 nm in the near infrared, respectively (Goodman et al., 1988). The radio electromagnetic waves of the lightning processes described above travel through the atmosphere and then are likely to be dissipated, reflected, scattered, refracted, and absorbed. The main effect is the dissipation, reducing the amplitude of the signal inversely proportional to the square of the distance. Ionospheric reflection, where the energy from waves with frequency lower than 5 MHz is trapped in the atmospheric waveguide formed by the ionosphere and the ground, permits long-range propagation of waves from high-energy return strokes. In the optical spectrum, the scattered energy by the cloud particles is observed from satellites as a diffuse light source at cloud top (Christian et al., 1989). Ground-based lightning location systems Several instruments can be used to locate lightning flashes, and more detail can be found at MacGorman and Rust, (1998) and Betz et al., (2009). The main technique consists of a network of sensors that detect IC and/or CG lightning by recording the electromagnetic radiation from

VLF to VHF continuously with time. The radiation detected by each sensor is then compared to other sensors in the network using two main location methods: the timeof-arrival and interferometer techniques. In the time-ofarrival method (TOA), time difference of lightning waveforms from several stations is computed and the location of lightning occurrence is given by the intersection of the hyperbolas for equal time differences. The interferometer method consists of determining the directions of the lightning waveform (azimuth and elevation) by analyzing the phase difference of an incident wave at several stations, and the intersection of these directions gives the location of the lightning source. Today’s operational lightning detection networks usually consist of different sensor types that use one or more location method for redundancy. These networks can be local, regional, or global depending on their operation baseline (distance between the sensors), and their detection efficiency and location accuracy are determined by the density of sensors and radio frequency used (Betz et al., 2009). Table 1 summarizes some of these more widely used lightning networks. The largest regional network is the US National Lightning Detection Network (NLDN) created in 1998, composed by 114 sensors operating in LF that locates mainly CG lightning in North America. Similar regional networks are found in Australia, Brazil, Canada, and Europe. Long-range networks operate in VLF and have been deployed worldwide in an attempt to locate lightning over remote areas like the oceans and the tropics. These networks operate with a sensor baseline of thousands of kilometers, which limits the detection efficiency to the stronger amplitude lightning signals (Cramer and Cummins, 1999). Total lightning (IC + CG lightning) is monitored using VHF and a combination of LF and VLF or VHF and LF. In the USA, total lightning is monitored by several VHF Lightning Mapping Array (LMA) research networks (Table 1) developed by New Mexico Tech (Rison et al., 1999). The individual LMA regional networks consist of 10 or more stations extending
80 km. The LMA measures the TOA of the magnetic peak signals at the different receiving stations to locate the source of impulsive VHF radio signals. Hundreds to thousands of sources per flash can be correlated in space and time, allowing a 3-D or 2-D lightning mapping of the channel over a regional domain of
200 km.

Lightning detection from space Several astronauts reported seeing lightning while looking down from space in the 1960s, describing flashes with hundreds of kilometers in extent and simultaneous flashes occurring between widely separated storms. Lightning was detected in early satellite imagery (Sparrow and Ney, 1971), and in 1981, the space shuttle astronauts recorded lightning in a 16 mm movie camera (Goodman et al., 1993). Although it was not their primary objective, several instruments onboard of the US Air Force DMSP

LIGHTNING

341

Lightning, Table 1 Ground-based lightning location systems operating in the world Frequency Type of discharges used detected Coverage area

Website

NLDN (US National Lightning Detection Network) CLDN (Canadian Lightning Detection Network)

LF

Mainly CG

United States of America

http://www.vaisala.com/

LF

Mainly CG

Canada

EUCLID (EUropean Cooperation for LIghtning Detection) RINDAT (Rede Integrada Nacional de Detecção de Descargas Atmosféricas) LINET (LIghtning location NETwork) LDAR (Lightning Detection and Ranging) LMA (Lightning Mapping Array)

LF

Mainly CG

Europe

http://www.ec.gc.ca/foudrelightning/default.asp? lang¼En&n¼D88E34E8-1 http://www.euclid.org

LF

Mainly CG

South-Southeast Brazil

http://www.rindat.com.br/

VLF, LF

Total lightning (IC + CG) Total lightning (IC + CG) Total lightning (IC + CG)

Europe

http://www.pa.op.dlr.de/linet/

ENTLN (Earth Networks Total Lightning Networks)

ELF-HF

STARNET (Sferics Timing and Ranging NETwork) WWLLN (World Wide Lightning Location Network) Vaisala GLD360 (Global Lightning Dataset 360) GLN (Global Lightning Network) ATDnet (Met Office’s Arrival Time Difference network )

VLF

Total lightning (IC + CG) Mainly CG Mainly CG

VLF

Mainly CG

Globe South America and East Africa Globe

VLF

Mainly CG

Globe

http://www.vaisala.com/

VLF VLF

Mainly CG Mainly CG

Globe Globe

http://www.uspln.com/gln.html http://www.metoffice.gov.uk/

Network

VHF VHF

(Defense Meteorological Satellite Program) satellites have also recorded lightning, providing the first global lightning distribution map as a bonus to the mission (Goodman et al., 1993). The Optical Transient Detector (OTD) onboard of the Microlab-1 (later renamed as OrbView-1) satellite was the first instrument designed to measure lightning from space day and night with storm scale resolution. The OTD operated between 1995 and 2000 in a 70 inclination low Earth orbit (see Low Earth Orbit (LEO)) at an altitude of 740 km. From this altitude, the OTD observed an individual storm for about 3 min. The design concept was based on the earlier research on optical emissions of lightning at cloud top (Christian and Goodman, 1987; Goodman et al., 1988). The OTD detected optical impulses with a 128 128 charge-coupled device (CCD) using a 1 nm narrow-band interference filter centered at 777.4 nm (Christian et al., 2003). Whereas the earlier satellite-based studies were limited to detecting visible lightning flashes during the darkness of night, the near-infrared wavelength combined with the use of spatial and temporal filtering used by OTD also allowed lightning detection during daylight. In 1997, the Lightning Imaging Sensor (LIS) onboard the Tropical Rainfall

Florida, USA

http://branch.nsstc.nasa.gov/ PUBLIC/LDARII/ USA-New Mexico, Oklahoma, http://lightning.nmt.edu/ nmt_lms/ Northern Alabama, Western Texas, Colorado, Atlanta, Washington DC, Spain Australia, Americas, Europe http://www.earthnetworks.com/ http://www.zeus.iag.usp.br/ http://wwlln.net/

Measuring Mission (TRMM) (Kummerow et al., 1998) was launched into a lower orbit inclination of 35 at an altitude of 350 km, later raised to 402 km in August 2001 to extend the mission lifetime. From this altitude, the LIS observed an individual storm for about 90 s. The OTD was a flight qualified engineering model of the LIS, and thus, they share the same basic design heritage. In both OTD and LIS, the signal is read out from the focal plane into a real-time event processor for lightning event detection. The background scene is updated during each frame readout sequence and when a pixel’s brightness compared to the prior background values exceeds a threshold, it is identified as a lightning event. The events are sent to the satellite ground station for geolocation processing in space and time, and an algorithm clusters the events into “flashes” (multiple CCD events grouped into time and space). The flash cannot be distinguished between CG and IC lightning, although in a statistical sense, the fraction of CG and IC flashes might be retrievable from a large sample of flashes (Koshak, 2010). FORTE (Fast On-Orbit Recording of Transient Events) satellite was built by Los Alamos National Laboratory to study lightning signals from space (Jacobson et al., 2000,

342

LIGHTNING

Lightning, Figure 1 Total lightning climatology derived from OTD (1995–2000) and LIS (1998–2010) at 0.5 resolution.

Suszcynsky et al., 2000; Hamlin et al., 2009). Launched in August 1997 at 800 km of altitude with a 70 of inclination and a circular orbit, the optical lightning location system has the same design of OTD and LIS but it also carries a broad band photo diode and the VHF receivers had two broad band channels which are selectable from a grid covering the entire high frequency and to very high frequency (Hamlin et al., 2009), allowing a combined optical and radio frequency observations of lightning. FORTE demonstrated that lightning can be located from space based on multiple-satellite VHF receivers. The “ORAGE” project has also been studying the possibility of locating lightning flashes using VHF-UHF interferometry from a constellation of microsatellites (Bondiou-Clergerie et al., 1999).

OTD and LIS findings The first global distribution of total lightning was derived from 5 years of OTD measurements by Christian et al., (2003), who found that the annual average global flash rate is 44 fl s1, with a maximum of 55 fl s1 in the boreal summer and a minimum of 35 fl s1 in the boreal winter. Recently, Blakeslee et al. (2012) and Cecil et al. (2012) found that these values remained nearly the same combining OTD (1995–2000) and LIS observations (1998–2010). These authors also showed that all continents display a strong diurnal variation with lightning peaking in the late afternoon, while oceans exhibit a

minimal nearly flat diurnal variation, but morning hours are typically slightly enhanced over afternoon. In Figure 1, we present the updated LIS/OTD climatology for 16 years of OTD (1995–2000) and LIS (1999–2010) combined observations of total lightning flash rate density (FRD, fl km2 year1) from Marshall Space Flight Center gridded LIS-OTD climatology product (High Resolution Flash Climatology, HRFC_COM_FR - Cecil et al. 2012). The difference between land and ocean can be clearly observed, with lightning occurring more frequently over continental (> 20 fl km2 year1) regions having greater instability and stronger vertical motion than oceanic environments. However, some coastal regions presented moderate FRD (1–10 fl km2 year1) associated with frequent synoptic scale extratropical cyclones and cold fronts (such as south-southeast coasts of Brazil, South Africa, Australia, and United States), and large-scale convergence zones (such as the South Atlantic, South Pacific, and the Intertropical Convergence Zones). High elevated and complex terrain regions over the tropics can be identified by high thunderstorm activity (> 30 fl km2 year1) at the mountains foot (e.g., Andes, Himalayas, Sierra Madre Occidental, Cameroon Line, and Mitumba Mountains). Congo Basin is dramatically highlighted by its extensive area of large FRD (> 50 fl km2 year1), where the greatest annual number of individual thunderstorms is observed (Zipser et al., 2006). However, higher resolution (0.10 ) LIS climatological maps highlighting topographical features and complex

LIGHTNING

terrain indicate the Congo Basin has the second highest climatological FRD, with 232 fl km2 year1 (at the foothills of Mitumba Mountains), while the total lightning “hot spot” on Earth is observed over Lake Maracaibo with 250 fl km2 year1 (Albrecht et al., 2011). Lake Maracaibo thunderstorm activity is very localized, determined by nocturnal convergence of land-lake and mountain-valley breezes over a warm lake, building the perfect scenario for thunderstorm development of more than 300 days per year (Albrecht et al., 2011). Frequent lightning activity (30–50 fl km2 year1) is also observed over Florida, Cuba, and Indonesia-Malaysia due to land-ocean sea breezes and over the borders of Argentina, Paraguay, and Brazil where the greatest individual flash rates per mesoscale convective systems are observed (Cecil et al., 2005; Zipser et al., 2006). In addition to mapping the lightning distribution, the instrument suite on the TRMM satellite allows more detailed characterization of the thunderstorms producing lightning. The TRMM radar and radiometer (see Microwave Radiometers) also show more intense storms over land (Cecil et al., 2005; Zipser et al., 2006). But for a given radar or radiometer signature, a storm over land is likely to produce more lightning than an otherwise similar storm over ocean (Liu et al., 2011). This suggests differences in the mixed phase microphysics and precipitation, hinted at by lightning but not resolved by the radar or radiometer uniquely by themselves. More information on LIS and OTD can be found at http://thunder.msfc.nasa.gov/.

The future of lightning mapping from space The next generation of NOAA Geostationary Operational Environmental Satellite (GOES-R) series and the EUMETSAT Meteosat Third Generation (MTG) will detect, locate, and measure continuous total lightning activity over their full disk with a nominal resolution of 10 km. GOES-R will carry the Geostationary Lightning Mapper (GLM) and it is scheduled to be launched in late 2015, while the MTG will carry the Lightning Imager (LI) and it is scheduled to be launched in 2018. Both GLM and LI are heritages of OTD and LIS, but GOES-R and MTG are equipped with improved communications systems and much greater telemetry bandwidth to ensure a continuous and reliable flow of the remote sensing products. The GOES-R series will maintain the 2-satellite system over the western hemisphere, with the operational GOES-R satellites at 75  W and 137  W. The GLM and LI together will provide continuous full-disk total lightning for storm warning and nowcasting (e.g., early warnings of tornadic activity, hail, and floods – see Severe Storms) for half of the globe. A geostationary lightning imager (GLI) having more limited coverage of mainland China and adjacent ocean is also planned for the Chinese FY-4 next-generation geostationary satellite series. More information on GOES-R GLM and MTG-LI can be found at http://www.goes-r. gov/ and http://www.eumetsat.int/, respectively.

343

Bibliography Albrecht, R. I., et al., 2011. The 13 years of TRMM lightning imaging sensor: from individual flash characteristics to decadal tendencies. In Proceedings XIV International Conference on Atmospheric Electricity. Rio de Janeiro, 08–12, Aug 2011. Betz, H. D., Schumann, U., and Laroche, P., 2009. Lightning: Principles, Instruments and Applications. Review of Modern Lightning Research. Dordrecht: Springer, p. 641. Blakeslee, R. J., Mach, D. M., Bateman, M. G., and Bailey, J. C., 2012. Seasonal variations in the lightning diurnal cycle and implications for the global electric circuit. Atmospheric Research, http://dx.doi.org/10.1016/j.atmosres.2012.09.023. Bondiou-Clergerie, A., Blanchet, P., Théry, C., Delannoy, A., Lojou, J.Y., Soulage, A., Richard, P., Roux, F., and Chauzy, S., 1999. “ORAGES”: a project for space-borne detection of lightning flashes using interferometry in the VHF-UHF band. In Proceedings of the 11th International Conference on Atmospheric Electricity, Guntersville, Alabama, pp. 184–187. Cecil, D. J., Buechler, D. E., and Blakeslee, R. J., 2012. Gridded lightning climatology from TRMM-LIS and OTD: Dataset description. Atmospheric Research, doi: http://dx.doi. org/10.1016/j.atmosres.2012.06.028. Cecil, D. J., Goodman, S. J., Boccippio, D. J., Zipser, E. J., and Nesbitt, S. W., 2005. Three years of TRMM precipitation features. Part I: radar, radiometric, and lightning characteristics. Monthly Weather Review, 133(3), 543–566. Christian, H. J., and Goodman, S. J., 1987. Optical observations of lightning from a high altitude airplane. Journal of Atmospheric and Oceanic Technology, 4, 701–711. Christian, H. J., Blakeslee, R. J., and Goodman, S. J., 1989. The detection of lightning from geostationary orbit. Journal of Geophysical Research, 94, 13329–13337. Christian, H. J., et al., 2003. Global frequency and distribution of lightning as observed from space by the optical transient detector. Journal of Geophysical Research, 108, 4005, doi:10.1029/ 2002JD002347. Cramer, J. A., and Cummins, K. L., 1999. Long-range and transoceanic lightning detection. In Proceedings 11th International Conference on Atmospheric Electricity. Guntersville, 7–11, June 1999, pp. 250–253. Goodman, S. J., Christian, H. J., and Rust, W. D., 1988. Optical pulse characteristics of intracloud and cloud-to-ground lightning observed from above clouds. Journal of Applied Meteorology, 27, 1369–1381. Goodman, S. J., Christian, H. J., and Rust, W. D., 1993. Global observations of lightning. In Gurney, R. J., Foster, J. L., and Parkinson, C. L. (eds.), Atlas of Satellite Observations Related to Global Change. Cambridge, UK: Cambridge University Press, pp. 191–219. Hamlin, T., Wiens, K. C., Jacobson, A. R., Light, T. E. L., and Eack, K. B., 2009. Space- and ground-based studies of lightning signatures. In Lightning: Principles, Instruments and Applications. Review of Modern Lightning Research. Dordrecht: Springer, pp. 287–2017. Jacobson, A. R., Cummins, K. L., Carter, M., Klingner, P., Dupre, D. R., and Knox, S. O., 2000. FORTE radiofrequency observations of lightning strokes detected by the National Lightning Detection Network. Journal of Geophysical Research, 105, 15653–15662. Koshak, W. J., 2010. Optical characteristics of OTD flashes and the implementations for flash-type discrimination. Journal of Atmospheric and Oceanic Technology, 27, 1822–1838. Kummerow, C., Barnes, W., Kozu, T., Shiue, J., and Simpson, J., 1998. The tropical rainfall measuring mission (TRMM) sensor package. Journal of Atmospheric and Oceanic Technology, 15, 809–817.

344

LIMB SOUNDING, ATMOSPHERIC

Liu, C., Cecil, D., and Zipser, E. J., 2011. Relationships between lightning flash rates and passive microwave brightness temperatures at 85 and 37 GHz over the tropics and subtropics. Journal of Geophysical Research, 116, D23108. MacGorman, D. R., and Rust, W. D., 1998. The Electrical Nature of Storms. New York: Oxford University Press. Rakov, V. A., and Uman, M. A., 2003. Lightning: Physics and Effects. Cambridge, UK: Cambridge University Press, p. 687. Rison, W., Thomas, R. J., Krehbiel, P. R., Hamlin, T., and Harlin, J., 1999. A GPS-based three-dimensional lightning mapping system: initial observations in central New Mexico. Geophysical Research Letters, 26, 3573–3576. Sparrow, J. G., and Ney, E. P., 1971. Lightning observations by satellite. Nature, 232, 540–541. Suszcynsky, D. M., Kirkland, M. W., Jacobson, A. R., Franz, R. C., Knox, S. O., Guillen, J. L. L., and Green, J. L., 2000. FORTE observations of simultaneous VHF and optical emissions from lightning: Basic phenomenology. Journal of Geophysical Research, 10(D2), 2191–2201. Uman, M. A., 1987. The Lightning Discharge. New York: Elsevier, p. 228. World Meteorological Organization (WMO), 2011. METOTERM http://www.wmo.int/pages/prog/lsp/meteoterm_wmo_en.html. Visited on 16, Oct 2011. Zipser, E., Cecil, D., Liu, C., Nesbitt, S. W., and Yorty, S., 2006. Where are the most intense thunderstorms on Earth? Bulletin of the American Meteorological Society, 87, 1057–1071.

Cross-references Microwave Radiometers Optical/Infrared, Radiative Transfer Radiation, Electromagnetic Severe Storms

LIMB SOUNDING, ATMOSPHERIC Nathaniel Livesey Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA

Synonyms Limb profiling; Occultation measurements Definition Limb. The portion of a planetary (or stellar) atmosphere at the outer boundary of the disk, viewed “edge on.” Limb sounding. Atmospheric remote sounding technique involving observing radiation emitted or scattered from the limb. Occultation. Atmospheric remote sounding technique involving observing radiation emitted (or reflected) by a distant body (solar, stellar, lunar, or an orbiting satellite), transmitted along a limb path through an absorbing and/or scattering planetary atmosphere, and detected by a remote observer.

Introduction Limb sounding is a widely used atmospheric remote sounding technique, whereby the atmosphere is viewed “edge on” by a space- or airborne instrument. Limb sounding observations are made from the microwave and infrared – where thermal emission is observed – to the visible and ultraviolet, where observations are typically of sunlight scattered in the limb or of airglow. A wide range of spaceborne limb sounding instruments have been used to observe atmospheric temperature, composition, and dynamics from the upper troposphere (
10 km altitude) to the mid-thermosphere (
450 km). A closely related technique, known as “occultation,” involves observing the atmospheric absorption and/or scattering of radiation emitted by a remote source (solar, lunar, stellar, or, more recently a GPS satellite). Limb sounding has significant advantages over nadir sounding (i.e., viewing straight down) or near-nadir sounding. Firstly, scanning the instrument field of view vertically across the atmospheric limb can give atmospheric profile information with greater vertical resolution than is typically possible from nadir sounders. In addition, complexities associated with emission or reflection of radiation by the planetary surface can be avoided. Finally, by viewing a significantly longer atmospheric path than nadir sounders, limb viewing instruments can achieve a stronger signal to noise for observations of tenuous atmospheric trace gases. However, this same long path length (typically a few 100 km) results in a poorer horizontal resolution than is possible with nadir sounding instruments. With the exception of the infrared Mars Climate Sounder instrument (MCS, McCleese et al., 2007) on the Mars Climate Orbiter, limb sounding observations have been confined to those of Earth’s atmosphere and are the focus of the discussion in this entry. Principles and techniques Limb radiances and line broadening Each limb view is associated with a particular “tangent height” – the closest distance from the limb ray to the Earth’s surface. High tangent height views typically give small signals, due to the tenuous atmosphere at these altitudes. As tangent altitudes decrease, atmospheric emission or scattering strengthens, increasing the observed signals. Eventually, the atmosphere becomes sufficiently opaque that signals from lower regions in the atmosphere are absorbed by the layers above and not seen by the instrument. At this point, radiances tend to remain fairly constant with decreasing tangent altitude (or to change only slightly, due to second-order geometrical effects) and are said to be “saturated” or “blacked out,” as the signal continues to derive largely from the lowermost nonopaque layers. Refraction is significant for limb rays in the lower atmosphere but is generally negligible above
20 km.

LIMB SOUNDING, ATMOSPHERIC

Molecular spectral lines are broadened in the atmosphere by a combination of the ensemble of Doppler shifts from the thermal motion of the molecules (“Doppler” broadening) and by collisions with other molecules (“collision” or “pressure” broadening). The latter generally dominates line widths of infrared and microwave signals up to
60 km, while for visible and ultraviolet wavelengths, Doppler broadening dominates throughout the bulk of the atmosphere. Pressure broadening of spectral lines can provide valuable information on the vertical distribution of trace gases (in addition to the information gained by vertically scanning the instrument field of view) with frequencies further from line centers conveying information on lower regions of the atmosphere, where lines are broad enough to contribute to the observed signals. For wavelengths where pressure broadening is insignificant, vertical distribution information can still be obtained by observing in multiple spectral regions having different atmospheric absorptions (and thus penetration depths).

Solar occultation and related observations Observation of the atmospheric absorption of solar radiation (“direct sun” measurements) has a long heritage in atmospheric science (e.g., the observations of ozone pioneered in the 1920s by Dobson). Solar occultation is a natural extension of these ground-based techniques (and similar observations from balloon and aircraft vantage points). An instrument on a low Earth-orbiting spacecraft can perform an occultation observation during sunrise and sunset on each of
14 orbits per 24 h period. Typically, occultation instruments observe a narrow portion of the solar disk and track this as it rises or sets through the atmosphere. The strong solar signal provides excellent signal to noise and obviates the need to cool the instrument or its detectors. The observations of the sun above the atmosphere, before sunset or after sunrise, can be used to ensure a stable instrument calibration. A succession of solar occultation instruments have provided a long record of atmospheric composition observations including the Stratospheric Aerosol and Gas Experiment (SAGE) I, II, and III series of instruments (McCormick et al., 1989), the Polar Ozone and Aerosol Measurement (POAM) instruments (Lucke et al., 1999), and the Halogen Occultation Experiment (HALOE) on the Upper Atmosphere Research Satellite (UARS) (Russell et al., 1993). Occultation observations were also made by the Atmospheric Trace Molecule Spectroscopy (ATMOS) instrument flown on the Space Shuttle ATLAS program (Gunson et al., 1996), and the Improved Limb Atmospheric Spectrometer instruments (ILAS I and II) on the Japanese Advanced Earth Observing Satellites (ADEOS I and II). The Scanning Imaging Absorption Spectrometer for Atmospheric CHartograpy instrument (SCIAMACHY, Bovensman et al., 1999) on the European Envisat performs solar occultation measurements in addition to limb and nadir imaging. Most recently, the

345

Atmospheric Chemistry Experiment’s Fourier Transform Spectrometer and Measurement of Aerosol Extinction in the Stratosphere and Troposphere Retrieved by Occultation (ACE/FTS and ACE/MAESTRO) instruments have been continuing and augmenting this record (Bernath et al., 2005). Currently, ACE is the only operating mission employing solar occultation, though other concepts are in formulation including plans to fly a SAGE III instrument on the International Space Station. While offering good vertical resolution and outstanding signal to noise and calibration stability, solar occultation instruments are fundamentally limited by orbital geometry to making only
30 observations per 24 h period. While some instruments augment this coverage with observations of lunar or stellar occultations, these, by definition, have poorer signal to noise than the solar occultation observations. As described below, limb sounding observations of atmospheric emission or of scattered solar radiation offer comparable vertical resolution to occultation but have the advantage that observations can be made on a nearglobal basis daily.

Radio occultation Observations of the atmospheric occultation of signals broadcast by GPS are a more recent development. In this technique, observations of refractive phase shift, as opposed to atmospheric absorption in different spectral regions, form the basis for the measurement. GPS occultation yields information on atmospheric refraction, and in turn temperature and/or water vapor profiles. More information on this technique is given elsewhere in this volume. Microwave limb sounding Atmospheric microwave emissions are generally associated with molecular rotational transitions, theoretically enabling observation of any atmospheric species with a significant dipole moment. Microwave limb sounding instruments have made observations of a wealth of species in the frequency range from
60 GHz (5 mm wavelength) to
2.5 THz (120 mm). Microwave signals are unaffected by aerosols and all but the thickest clouds, as the wavelengths used are longer than the typical particle sizes. This enables microwave observations of atmospheric composition in a limb sounding geometry in regions that are too cloudy for observations at other wavelengths. To date, five spaceborne instruments employing limb sounding at microwave frequencies have flown: The Microwave Atmospheric Sounder (MAS) as part of the ATLAS payload on the Space Shuttle (Croskey et al., 1992), the Microwave Limb Sounder (MLS) instruments on the NASA UARS and Aura satellites (Barath et al., 1993; Waters et al., 2006), the Submillimeter Radiometer (SMR) on the Swedish Odin satellite (Murtagh et al., 2002), and, most recently, the Submillimeter-Wave Limb

346

LIMB SOUNDING, ATMOSPHERIC

Emission Sounder (SMILES) on the International Space Station (Ozeki et al., 2001). Microwave instruments can achieve arbitrarily fine frequency resolutions, enabling individual transition lines to be resolved in great detail. The observations of line shape enable simultaneous observation of both atmospheric pressure (largely affecting the width of a given transition line in the pressure-broadening regime) and species abundance (largely affecting the line strength). By combining inferred atmospheric pressure information with limb tangent altitude information, and assuming hydrostatic balance, atmospheric temperature profile information can be obtained. The field-of-view width for a microwave instrument is determined by the antenna size and wavelength employed, with narrower fields of view achieved for larger antennae and/or shorter wavelengths and somewhat large antennae dictated for many observations. For example, the Aura MLS instrument’s 1.6 m antenna has a field of view that is
3.5 km wide at the limb (full width, half maximum) at 200 GHz from a 700 km orbit. The lower vertical range of microwave limb sounding instruments is limited by continuum absorption from oxygen, nitrogen, and water vapor, with
8 km altitude typically being the deepest penetration.

Infrared limb sounding As with microwave limb sounding, infrared instruments observe thermal emission from the atmosphere, in this case mostly arising from molecular vibrational transitions. Again, collisional broadening enables determination of atmospheric pressure at the tangent point. Although not all infrared limb sounding instruments have the spectral resolution to resolve individual line shapes, pressure information can generally still be obtained from broader-band measurements. Scattering and emission from clouds pose a more significant limitation to infrared limb sounders than microwave instruments, particularly in the tropics where clouds are prevalent in the upper troposphere. In clearsky regions, infrared limb sounders can typically penetrate a few kilometers deeper than microwave sounders, but continuum absorption is, again, the ultimate limitation to this penetration. Infrared instruments can more easily achieve narrower fields of view than those in the microwave, and this can translate into a finer vertical resolution for the geophysical observations. However, the detectors typically need to be cooled (e.g., to
70 K) in order to achieve a scientifically useful signal to noise. Infrared limb sounding instruments have a long history in atmospheric science, starting with the Limb Radiance Inversion Radiometer (LRIR) on Nimbus 6, followed by the Limb Infrared Monitor of the Stratosphere (LIMS, Gille and Russell, 1984) and Stratospheric and Mesospheric Sounder (SAMS, Drummond et al., 1980) instruments on Nimbus 7. UARS included two infrared limb sounding instruments – the Cryogenic Limb Array

Etalon Spectrometer (CLAES, Roche et al., 1993) and the Improved Stratospheric and Mesospheric Sounder (ISAMS, Taylor et al., 1993). More recent infrared limb sounders include the Michelson Interferometer for Passive Atmospheric Sounding instrument (MIPAS, Fischer et al., 2008) on ESA’s Envisat spacecraft and the HighResolution Dynamics Limb Sounder (HIRDLS, Gille et al., 2008) on NASA’s Aura satellite. The Sounding of the Atmosphere using Broadband Emission Radiometry (SABER) instrument on the Thermosphere, Ionosphere, Mesosphere Energetics and Dynamics (TIMED) mission (Russell et al., 1999) is another recently launched infrared limb sounder mainly focusing on the chemistry and structure of the upper atmosphere.

Visible and ultraviolet limb sounding Limb viewing instruments at visible and ultraviolet wavelengths generally observe sunlight scattered by the atmospheric limb. These include the Optical Spectrograph and Infrared Imaging System (OSIRIS, Llewellyn et al., 2003) instrument on Odin and the planned limb sensor for the Ozone Mapping and Profiling Suite (OMPS) on the NPOESS (National Polar-orbiting Operational Environmental Satellite System) Preparatory Project (NPP). The SCIAMACY instrument on Envisat also includes a limb scattering capability. As collisional broadening does not significantly affect the (mainly electronic or vibronic) molecular transitions at these wavelengths, tangent pressure cannot be deduced from the observations, and the height registration of the resulting geophysical products is more critically reliant on independent knowledge of spacecraft pointing than is the case for longer wavelength observations. In addition to limb scattering sounders, past instruments have observed visible atmospheric airglow emissions in the upper atmosphere. These include the Wind Imaging Interferometer (WINDII, Shepherd et al., 1993) and the High-Resolution Doppler Imager (HRDI, Hays et al., 1993) instruments, both on UARS, which used these observations to deduce upper atmospheric dynamics. Inversion approaches for limb sounding instruments Although scanning the field of view of an instrument up and down the atmospheric limb enables high-resolution observations of vertical profiles, the observed signals are (as with nadir sounding) affected by emission, absorption, and scattering throughout the ray path. Disentangling the impact of each atmospheric layer on the observed signal and deducing vertical profiles of temperature and composition is a nontrivial task, commonly known as a “retrieval” or “inverse” calculation. A variety of techniques have been employed for limb sounding retrievals. The so-called onion peeling approach uses observations at the top of the limb scan to characterize the uppermost atmospheric region. This is then

LIMB SOUNDING, ATMOSPHERIC

accounted for when characterizing the next layer down, using lower-altitude limb views, and so on. A drawback of this technique is that the resulting profile depends strongly on the knowledge of the uppermost regions, where signal to noise is typically poor. The most commonly adopted approach for limb sounding retrievals is the well-established “optimal estimation” method (Rodgers, 2000), which seeks the atmospheric state that matches all the observed signals simultaneously (taking into account potential noise on each signal). Although more computationally intensive than simpler approaches, this need not be a barrier with modern computing capabilities. Indeed, the most computationally demanding part of the calculation is typically the “forward model” (the computation that estimates the signal that would be observed by the instrument for a given atmospheric state), which is a central part of all but the simplest retrieval approaches, and upon which the accuracy of the resulting geophysical profiles ultimately depends. Inhomogeneity along the limb line of sight can introduce biases in limb sounding retrievals, particularly in regions of strong atmospheric gradients. Some retrieval methods employ an iterative approach, whereby horizontal gradient information from a first pass is considered in a later retrieval step. In cases where the instrument line of sight is aligned with the spacecraft velocity vector, successive limb scans take multiple views through the same region of atmosphere, enabling a “tomographic” approach to the retrieval calculation to be taken (e.g., Livesey and Read (2000)).

Notable findings from limb sounding observations The near-global daily coverage and good vertical resolution of limb sounding instruments has provided a wealth of information on atmospheric structure and composition from the upper troposphere through to the thermosphere. The early observations from LIMS and SAMS set the stage, with zonal-mean information on the abundance of key stratospheric and mesospheric trace gases. The three atmospheric composition limb sounders (CLAES, ISAMS, and MLS) on UARS, along with the HALOE solar occultation instrument, provided valuable insights into the dynamics and chemistry of Earth’s stratosphere, most notably processes associated with chemical stratospheric ozone loss and the transport of air into and throughout the stratosphere. UARS observations also provided valuable information on the impact of volcanic gases and the resulting aerosols on the stratosphere, following the dramatic June 1991 eruption of Mt. Pinatubo in the Philippines (4 months before the UARS launch). In addition to stratospheric and mesospheric composition observations, UARS MLS provided unprecedented information on water vapor and ice clouds in the upper troposphere. The Aura MLS and HIRDLS instruments have enhanced this record providing the first daily global observations of upper tropospheric ozone, carbon

347

monoxide (MLS only), and nitric acid. The upper troposphere is an important region of the atmosphere for climate, as it is where water vapor (the strongest greenhouse gas) and ozone have their largest radiative impact.

Outlook At the time of writing, the only limb sounding instruments in operation are Aura MLS, Odin SMR and OSIRS, and Envisat MIPAS. The SMILES instrument experienced a critical failure after
6 months of operation, although plans are in formulation for a possible fix. While several new limb sounding instrument concepts are under formulation, to date none have been confirmed for launch. The most mature is the ESA Process Exploration through Measurements of Infrared and millimeter-wave Emitted Radiation (PREMIER) mission, which includes infrared and microwave limb sounding instruments observing the upper troposphere and lower stratosphere. Conclusion Limb sounding instruments provide a wealth of information on the composition, structure, and dynamics of Earth’s atmosphere, through observations of emitted or scattered radiation in an “edge on” viewing geometry. Limb sounding offers a valuable combination of good vertical resolution and near-global daily coverage, using wavelengths ranging from the microwave to the ultraviolet, and can provide observations from the upper troposphere to the middle thermosphere. Limb sounding observations have led to important discoveries concerning key dynamical and chemical processes in Earth’s stratosphere (including those processes associated with the “ozone hole”), and in the upper troposphere where water vapor and ozone have their strongest greenhouse forcing. Acknowledgment This research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the NASA. Bibliography Barath, F. T., et al., 1993. The upper atmosphere research satellite microwave limb sounder experiment. Journal of Geophysical Research, 98, 10, 751–10, 762. Bernath, P. D., et al., 2005. Atmospheric chemistry experiment (ACE): mission overview. Geophysical Research Letters, 32, L15S01, doi:10.1029/2005GL022386. Bovensman, H., Burrows, J. P., Buckwitz, B., Frerick, J., Noël, S., Rozanov, V. V., Chance, K. V., and Goede, A. P. H., 1999. SCIAMACHY: mission objectives and measurement modes. Journal of Atmospheric Science, 56, 127–150. Croskey, C. L., et al., 1992. The millimeter wave atmospheric sounder (MAS): a shuttle-based remote sensing experiment. IEEE Transactions on Microwave Theory and Techniques, 40(6), 1090. Drummond, J. R., Houghton, J. T., Peskett, G. D., Rodgers, C. D., Wale, M. J., Whitney, J. G., and Williamson, E. J., 1980. The stratospheric and mesospheric sounder on nimbus 7.

348

LIMB SOUNDING, ATMOSPHERIC

Philosophical Transactions of the Royal Society of London, 296(1418), 219–241. Fischer, H., et al., 2008. MIPAS: an instrument for atmospheric and climate research. Atmospheric Chemistry and Physics, 8, 2151–2188. Gille, J. C., and Russell, J. M., III, 1984. The limb infrared monitor of the stratosphere: experiment description, performance and results. Journal of Geophysical Research, 89, 5125–5140. Gille, J., et al., 2008. High resolution dynamics limb sounder (HIRDLS): experiment overview, recovery and validation of initial temperature data. Journal of Geophysical Research, 113, D16S43, doi:10.1029/2007JD008824. Gunson, M. R., et al., 1996. The atmospheric trace molecule spectroscopy (ATMOS) experiment: deployment on the ATLAS space shuttle missions. Geophysical Research Letters, 23, 2333–2336. Hays, P. B., Abreu, V. J., Dobbs, M. E., Gell, D. A., Grassi, H. J., and Skinner, W. R., 1993. The high-resolution doppler imager on the upper atmosphere research satellite. Journal of Geophysical Research, 98, 10, 713–10, 723. Livesey, N. J., and Read, W. G., 2000. Direct retrieval of line-of-sight atmospheric structure from limb sounding observations. Geophysical Research Letters, 27, 891–894. Llewellyn, E. J., et al., 2003. First results from the OSIRIS instrument on-board Odin. Sodankyla Geophysical Observatory Publications, 92, 1–47. Lucke, R. L., et al., 1999. The polar ozone and aerosol measurement (POAM) III instrument and early validation report. Journal of Geophysical Research, 104, 18785–18799. McCleese, D. J., et al., 2007. Mars climate sounder: an investigation of the thermal and water vapor structure, dust and condensate distributions in the atmosphere, and energy balance of the polar regions. Journal of Geophysical Research, 112, E05S06, doi:10.1029/1006JE002790. McCormick, M. P., Zawodny, J. M., Velga, R. E., Larsen, J. C., and Wang, P. H., 1989. An overview of SAGE I and II ozone measurements. Planetary and Space Science, 37, 1567–1586.

Murtagh, D., et al., 2002. An overview of the Odin atmospheric mission. Canadian Journal of Physics, 80, 309–319. Ozeki, H., et al., 2001. Development of superconducting submillimeter-wave limb emission sounder (JEM/SMILES) aboard the International Space Station. In Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series. Bellingham: SPIE, Vol. 4540, pp. 209–220. Roche, A. E., Kumer, J. B., Mergenthaler, J. L., Ely, H. A., Uplinger, W. H., Potter, J. F., James, T. C., and Sterritt, L. W., 1993. The cryogenic limb array etalon spectrometer (CLAES) on UARS: experiment description and performance. Journal of Geophysical Research, 98, 10, 763–10, 775. Rodgers, C. D., 2000. Inverse Methods for Atmospheric Science, Theory and Practice. Singapore: World Scientific. Russell, J. M., III, et al., 1993. The halogen occultation experiment. Journal of Geophysical Research, 93, 10, 777–10, 798. Russell, J. M., III, Mlynczak, M. G., Gordley, L. L., Tansock, J., and Esplin, R. 1999. An overview of the SABER experiment and preliminary calibration results. Proc. SPIE, 3756, doi:10.1117/ 12.366382, 277–288. Shepherd, G. G., et al., 1993. WINDII, the wind imaging interferometer on the upper atmosphere research satellite. Journal of Geophysical Research, 98, 10, 725–10, 750. Taylor, F. W., et al., 1993. Remote sensing of atmospheric structure and composition by pressure modulation radiometry from space: the ISAMS experiment on UARS. Journal of Geophysical Research, 98, 10, 799–10, 814. Waters, J. W., et al., 2006. The earth observing system microwave limb sounder (EOS MLS) on the Aura satellite. IEEE Transactions on Geoscience and Remote Sensing, 44, 1075–1092.

Cross-references GPS, Occultation Systems Stratospheric Ozone Trace Gases, Stratosphere, and Mesosphere

M

MADDEN-JULIAN OSCILLATION (MJO) Baijun Tian and Duane Waliser Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA

Synonyms 30–60 day oscillation; 40–50 day oscillation; Intraseasonal oscillation (ISO); Intraseasonal variability (ISV) Definition The MJO is a planetary-scale quasiperiodic oscillation of atmospheric wind and convective cloudiness anomalies that moves slowly eastward along the equator mainly over the tropical Indian and Pacific Oceans with a timescale on the order of 30–60 days. Introduction In 1971, Roland Madden and Paul Julian stumbled across a 40–50 day oscillation when analyzing the zonal (east– west) wind data from rawinsondes at Kanton Island (3
S, 172
W) over the equatorial western Pacific. Until the early 1980s, little attention was paid to this oscillation, which later became known as the MJO. Since the 1982– 1983 El Niño event, low-frequency variations in the tropics, both on intra-annual (less than a year) and interannual (more than a year) timescales, have received much more attention, and the number of MJO-related publications grew rapidly. The MJO turned out to be the dominant form of the intraseasonal (30–90 day) variability in the tropical atmosphere and has many important influences on the global weather and climate system. The MJO is a naturally occurring mode of variability of the tropical ocean–atmosphere system. It is characterized by an eastward propagation of large regions of both

enhanced and suppressed tropical convection, cloudiness and rainfall near the equator mainly over the tropical Indian and Pacific Oceans, and associated large-scale atmospheric circulation (wind) anomalies over the whole globe. The anomalous cloudiness or rainfall usually first emerges over the equatorial western Indian Ocean and intensifies and remains evident as it propagates eastward over the warm ocean waters of the equatorial eastern Indian Ocean and western Pacific, the so-called IndoPacific warm pool. This pattern of anomalous cloudiness and rainfall then generally weakens and disappears as it moves over the cooler ocean waters of the equatorial eastern Pacific, the so-called equatorial cold tongue. Along with this eastward-propagating pattern of equatorial cloudiness and rainfall anomalies, there also exist eastward moving distinct baroclinic patterns of lower- and upper-level atmospheric circulation anomalies in the tropics and subtropics. The circulation anomalies extend around the globe and are not confined to the eastern hemisphere as opposed to the cloudiness and rainfall anomalies. When the MJO moves eastward, it modulates the background cloud, rainfall, and circulation in the tropical Indian and Pacific Oceans on timescales shorter than a season but longer than a couple of weeks. The length of a typical MJO cycle is approximately 30–60 days but normally 40–50 days. Thus, the MJO is also known as the 30–60 day oscillation, 40–50 day oscillation, intraseasonal oscillation (ISO), or intraseasonal variability (ISV) after its typical timescale. A complete MJO cycle can be divided into two distinct phases according to the intensity of its convective activity and rainfall: convectively active (enhanced) phase or wet phase and convectively inactive (suppressed) phase or dry phase. The wet phase of the MJO cycle is characterized by enhanced tropical convection and large moist convective storms with higher cloud-top heights, more cloud cover, and heavier rainfall (thus more atmospheric latent heating)

E.G. Njoku (ed.), Encyclopedia of Remote Sensing, DOI 10.1007/978-0-387-36699-9, © Springer Science+Business Media New York 2014

350

MADDEN-JULIAN OSCILLATION (MJO)

than average. In contrast, the dry phase of the MJO cycle is typified by dry and clear conditions with lower cloud-top heights, less cloud cover and rainfall (thus less atmospheric latent heating) than normal. The MJO appears to be predictable with lead times of 2–3 weeks. This can help bridge the gap in environmental forecast skill between that of weather (lead times up to a few days) and that of seasonal-to-interannual climate predictions (lead times from a few months to a few years). Satellite remote sensing data have played an important role in the MJO studies during the last three decades because of the high spatial (a few kilometers) and temporal (3 h or daily) resolutions and global coverage of the satellite data especially over the tropical oceans where the rawinsondes are sparse and the global reanalyses have had large uncertainties. These satellite-based studies have significantly advanced our knowledge in the MJO description and mechanism as well as its global impacts. They have also led to considerable improvement in our numerical modeling capability and theoretical understanding of the MJO. This entry briefly reviews the central role of satellite remote sensing data in studying the description, mechanisms, and global impacts of the MJO.

Description The MJO and its eastward-propagating convective feature are most active during the Northern Hemisphere (boreal) winter season (November–April) when the Indo-Pacific warm pool is centered near the equator. During the Northern Hemisphere (boreal) summer season (May–October), the change in the large-scale wind patterns associated with the Asian summer monsoon results in the large-scale convective disturbances propagating northeastward, from the equatorial Indian Ocean into Southeast Asia. The discussions in this entry mainly focus on the boreal winter MJO events although many aspects of these discussions can be equally applied to the boreal summer MJO events. The MJO can be detectable in several important atmospheric and oceanic parameters, such as atmospheric cloudiness, atmospheric wind speed and direction, atmospheric temperature, atmospheric moisture, surface pressure, surface rainfall, sea surface temperature (SST), and surface heat and freshwater fluxes. However, the fundamental quantities related to the MJO are large-scale organized convection, cloudiness, rainfall, and tropospheric winds. Thus, our discussion of the MJO description mainly centers on how the remote sensing data help us understand the characteristics of large-scale convective cloudiness, rainfall, and tropospheric winds associated with the MJO. Outgoing longwave radiation (OLR) and infrared window radiance (or brightness temperature) provide broadband and narrowband measures of the total flux of longwave radiation lost to space at the top of the atmosphere. Deep convective clouds in the tropics have cold cloud tops and therefore have low values of OLR and brightness temperature. Typically, an OLR value of less

than 200 W m2 or a brightness temperature value of less than 220 K indicates the presence of deep convection in the tropics. Because of this simple property of OLR and brightness temperature, they have been widely used as a proxy for deep convection over the tropics, where the background longwave radiation from low clouds or surface is much higher. During the wet phase of the MJO, both OLR and window radiance can be significantly reduced due to higher and colder convective cloud tops. On the other hand, during the dry phase of the MJO, both OLR and window radiance can be significantly enhanced because of clear skies and lower and warmer cloud tops. Thus, OLR and infrared window radiance are good indicators of the convective intensity of the MJO. Over the tropical warm ocean waters where MJO convection is active, satellites are the only way to observe convective cloudiness with large spatial coverage. As a result, the satellite remote sensing data have played a central role in studying the general spatial and temporal structure and eastwardpropagating features of the MJO convection and cloudiness during the past three decades. Based on limited rawinsonde and surface station data, Madden and Julian speculated that the MJO is characterized by slowly eastward-propagating, large-scale oscillations in the tropical convective cloudiness over the equatorial Indian Ocean and western Pacific as the result of an eastward movement of large-scale atmospheric circulation cells oriented in the equatorial zonal plane. Evidence of such eastward-propagating clouds in satellite data was first presented by Arnold Gruber in 1974 and Abraham Zangvil in 1975 who both found large-scale eastward-propagating features near 40–50 days at the equator in the cloud brightness data obtained from the Environmental Science Services Administration (ESSA) satellites (ESSA 3 and 5). However, no further evidence was found until the early 1980s when NOAA OLR data and wind analyses from US National Meteorological Center (NMC) became available. These Advanced Very High Resolution Radiometer (AVHRR) OLR data started from the mid-1970s and were mainly from a series of polar-orbiting satellites, such as the scanning radiometer (SR) series and the Television Infrared Observation Satellite–Next Generation (TIROS-N) series. These OLR data have a twice-daily resolution and a good global coverage each day. By the early 1980s, almost 10 years of daily AVHRR OLR data were archived and available to the research community. In the early to mid-1980s, a series of observational papers on the MJO (at the time still referred to as the 40–50 day or 30–60 day oscillation) using the AVHRR OLR and NMC wind analyses appeared. These studies clearly demonstrated the existence of the slowly eastward propagation of the tropical cloudiness at the intraseasonal timescale and documented many detailed and important convective cloudiness and circulation features of the MJO. These papers also helped to bring the MJO to the attention of the scientific community. The NOAA polar orbiter satellites have been operating almost continuously over the

MADDEN-JULIAN OSCILLATION (MJO)

past 30 years. As a result, the NOAA OLR data have a relatively long record (over 30 years) and have been and are still extensively employed for studying the MJO. Figure 1 shows the infrared brightness for a MJO event from December 7 to 26, 1987, with each panel separated by 5 days. Each map covers the all longitudes (0–360
) between 20
S and 20
N. Bright white areas indicate cold high clouds, and dark regions indicate cloud-free or warm low cloud conditions. The slow eastward propagation of cold high clouds associated with the MJO is evident. The cold high clouds first form over the western equatorial Indian Ocean on December 7, 1987 and, over the course of the following 20 days, then intensify and propagate eastward across the equatorial Indian Ocean and the Maritime Continent to the equatorial western Pacific Ocean. In addition to the large-scale eastward-propagating pattern of MJO convective cloudiness, there exist many finescale structures within the convective cloudiness. The high spatial (100 km) and temporal (daily) resolution AVHRR OLR data from NOAA polar orbiters and much higher spatial (a few kilometers) and temporal (3 h) resolution window-channel infrared data from the geostationary satellites, such as Geostationary Meteorological Satellite (GMS) from Japan, are particularly useful in investigating the fine structure of the MJO convective cloudiness due to their high spatial and temporal resolutions. For example, the OLR data indicated many shortperiod, synoptic-scale convective systems within the planetary-scale 30–60 day fluctuations. Along the equator, these active convective systems move eastward with a phase speed of 10–15 m s1 and have a horizontal spatial scale of several thousand kilometers and a timescale of less than 10 days. These synoptic-scale, eastward-propagating convective systems within the MJO envelope are referred to as super cloud clusters or more recently convectively coupled Kelvin waves. The GMS infrared window radiance data have revealed that a super cloud cluster consists of many fine-scale cloud clusters. These fine-scale cloud clusters typically propagate westward along the equator with a lifetime of about 1–2 days. Although each cloud cluster moves westward, a super cloud cluster moves eastward due to the successive formation of a new cloud cluster east of the mature-stage cloud cluster. This suggested the existence of a hierarchy of convective activity within the MJO that is still an outstanding avenue of research of today. In the Tropics, surface rainfall is closely related to convective cloudiness and thus is another key quantity of interest to characterize the MJO. The tropical rainfall can be estimated from the satellite-observed infrared and microwave radiances. The surface rainfall can first be indirectly derived from infrared window radiance and OLR which are very sensitive to cloud-top temperatures that are indirectly tied to surface rainfall. The microwave radiances are very sensitive to the hydrometeors that directly result in surface precipitation and thus can be used more directly to retrieve surface precipitation. The microwavebased rainfall retrievals can be divided into passive and

351

Madden-Julian Oscillation (MJO), Figure 1 Infrared satellite observations for the global tropics (20
N–20
S) for (from the top) 03 GMT December 7, 12, 17, 22, and 26, 1987. Bright white areas indicate high clouds and deep convection, and dark regions indicate cloud-free conditions (Based on Global Cloud Imagery (GCI) data courtesy of M. Salby, University of Colorado).

active microwave (radar) retrievals. The passive retrievals can further be divided into the microwave emission-based (sensitive to cloud liquid water) and the microwave scattering-based (sensitive to ice particles and large water drops) rainfall retrievals. Starting from the 1990s, several global rainfall data have been generated from satelliteobserved infrared and microwave data and were instrumental in studying the MJO during the last two decades. For example, daily, global oceanic rainfall data retrieved based on microwave emission from the Microwave Sounding Unit (MSU) on the NOAA TIROS-N satellites was first used to study the MJO convective feature in the 1990s. During the late 1990s, the NOAA Climate Prediction Center (CPC) generated a global rainfall data set, referred to CPC Merged Analysis of Precipitation (CMAP), through the merged analysis of precipitation from several sources, such as gauges, satellites, and numerical model outputs. The satellite rainfall estimates for the CMAP includes the window infrared-based rainfall estimate from NOAA geostationary satellites (e.g., Geostationary Operational Environmental Satellites, GOES), the OLR-based rainfall estimate from the NOAA polarorbiting satellites, the microwave emission-based rainfall estimate from the MSU on the NOAA polar-orbiting

352

MADDEN-JULIAN OSCILLATION (MJO)

satellites and the Special Sensor Microwave/Imager (SSM/I) on the Defense Meteorological Satellite Program (DMSP) satellites, and the microwave scattering-based rainfall estimate from SSM/I. Similar merged global rainfall data were also produced by the Global Precipitation Climatology Project (GPCP) based on similar inputs. However, some subtle differences exist between the CMAP and GPCP precipitation data sets, such as diurnal cycle adjustment and atoll precipitation adjustments. Both CMAP and GPCP rainfall data with pentad (5 day) resolution have been extensively used to study the convective features of the MJO and identify MJO events. Launched in 1997, the Tropical Rainfall Measurement Mission (TRMM) satellite provided the first spaceborne precipitation radar (PR) (active microwave) to monitor global rainfall from space in addition to the passive TRMM Microwave Imager (TMI) instrument and the visible and infrared scanner (VIRS) instrument. The TRMM PR and TMI data have also been used to study the MJO. However, their spatial and temporal sampling is rather coarse. To alleviate these sampling deficiencies of the TRMM PR and TMI, the TRMM Multisatellite Precipitation Analysis (TMPA) project provides a calibration-based sequential scheme for combining precipitation estimates from multiple satellites, as well as gauge analyses where feasible, at fine scales (0.25
 0.25
and 3 hourly). The input satellite data for the TMPA are mainly from two sources: (1) passive microwave-based precipitation from the TMI on TRMM, the SSM/I on DMSP satellites, the Advanced Microwave Scanning Radiometer–Earth Observing System (AMSR-E) on Aqua, and the Advanced Microwave Sounding Unit-B (AMSU-B) on the NOAA polar-orbiting satellites; (2) window infrared-based precipitation data collected by the international constellation of geostationary satellites. This TMPA data set, also known as the TRMM 3B42, has relatively better retrieval accuracy and sampling at fine spatial and temporal scales. It has been used extensively for the recent MJO studies. In association with the eastward-propagating equatorial convective cloud and rainfall system are strong variations in lower- and upper-level large-scale atmospheric wind fields along the equator and in the subtropics. Unlike the convective cloudiness that is mostly confined over the equatorial Indian and western Pacific Oceans, the largescale wind anomalies of the MJO extend globally along the equator and into the subtropics. For example, along the equator, low-level zonal winds converge into the convective center, while upper-level zonal winds diverge away from the convective center. These lower- and upper-level zonal winds are interconnected through ascending (upward vertical movement) moist air within the convective center and descending dry air outside the convective center. These large-scale zonal winds propagate eastward together with the convective cloudiness along the equator and can reach into the western hemisphere (eastern Pacific, Atlantic, and Africa). In addition to these zonal winds along the equator are large-scale gyre circulations extending into the subtropics in both the lower

and upper troposphere that are tied to the eastwardpropagating convective cloudiness and zonal winds along the equator. For example, in the lower troposphere, a subtropical cyclonic couplet (counterclockwise in the Northern Hemisphere and clockwise in the Southern Hemisphere) flanks or lies to the west of the MJO convective region, while a subtropical anticyclonic couplet (clockwise in the Northern Hemisphere and counterclockwise in the Southern Hemisphere) lies to the east of the MJO convective region. On the other hand, in the upper troposphere, a subtropical anticyclonic couplet flanks or lies to the west of the MJO convective region, while a subtropical cyclonic couplet lies to the east of the MJO convective region due to the baroclinic nature of the tropical large-scale wind fields. The near equatorial largescale zonal wind anomalies are a Kelvin wave response to the MJO convective heating, while the off equatorial meridional wind anomalies are a Rossby wave response to the MJO convective heating. The Kelvin wave is named after Lord Kelvin, who studied water waves along a vertical side boundary under rotation conditions. In this case, the equator, where the vertical component of the Earth’s rotation vector changes sign, serves the vertical side boundary. The Rossby wave is due to the latitudinal variation of the vertical component of the Earth’s rotation and is named after C. G. Rossby, who was the first to clearly isolate the so-called Rossby wave dynamics (a balance between inertia and rotation). These large-scale patterns of convective cloudiness and wind fields are components of what are collectively referred to as equatorial waves or convectively coupled equatorial waves. It is the upper-level circulation features of these waves that allow the convective signatures of the MJO over the Indo-Pacific warm pool to influence weather “downstream” over the eastern Pacific and Atlantic (e.g., hurricanes and tropical cyclones) as well as the midlatitudes (e.g., precipitation extremes along the US west coasts). In terms of satellite observations, tropospheric winds are difficult to observe directly except via cloud tracking, such as the cloud-drift winds derived from the Multi-angle Imaging SpectroRadiometer (MISR) and NOAA geostationary satellites. However, the surface winds can be measured directly using the spaceborne radar scatterometers, such as the SeaWinds instrument on NASA’s Quick Scatterometer (QuikSCAT) satellite. The importance of the cloud-drift winds and QuikSCAT surface winds for the MJO study has been recognized but still in the early stages of exploration. The overall large-scale dynamic structure of the MJO, especially in the upper levels, is still mainly derived from the global reanalyses or radiosondes at the moment.

Mechanisms To understand the mechanisms responsible for the initiation and maintenance of the MJO, it is important to quantify the evolution of the thermodynamic environment and surface conditions associated with the MJO. In particular,

MADDEN-JULIAN OSCILLATION (MJO)

documenting the three-dimensional temperature and moisture structure of the MJO is crucial in advancing our theoretical understanding of the MJO. The availability of satellite-based temperature and moisture soundings makes these studies possible. For example, tropical mean tropospheric temperature, lower-troposphere (surface–300 mb) temperature, and upper-troposphere (500–100 mb) temperature were derived from the MSU channels in the early 1990s. These data were used to study the relationship between MJO convection and temperature anomalies in the 1990s. It was found that when the MJO is amplifying (e.g., over eastern Indian Ocean), convective heating anomalies are positively correlated to temperature anomalies. This implies production of eddy available potential energy (EAPE), which can in turn be used to drive atmospheric motion and sustain the MJO. When the MJO is decaying (e.g., over the eastern Pacific or east of the Date Line), temperature anomalies are nearly in quadrature with convective heating anomalies. As a result, their correlation and production of EAPE are small which is no longer an energy source for the MJO. For water vapor, the TIROS Operational Vertical Sounder (TOVS) provided the water vapor fields at five different levels in the troposphere in the 1990s. These data were used to study the three-dimensional structure and evolution of water vapor over the life cycle of the MJO in the early 2000s. The composite evolution of moisture shows markedly different vertical structures as a function of longitude. There is a clear westward tilt with the height of the moisture maximum associated with the MJO propagating eastward across the Indian Ocean. These disturbances evolve into nearly vertically uniform moist anomalies as they reach the western Pacific. Near-surface (below 850 mb) positive water vapor anomalies were observed to lead the convection anomaly by 5 days over the Indian Ocean and western Pacific. Upper-level positive water vapor anomalies were observed to lag the peak in the convection anomaly by 5–10 days, as the upper troposphere is moistened following intense convection. In the eastern Pacific, the moisture variations then become confined to the lower levels (below 700 mb), with upper-level water vapor nearly out of phase. While the MSU and TOVS provided useful initial insights into the three-dimensional temperature and moisture structure of the MJO, their vertical resolution was too low to describe the detailed vertical structure, especially near the tropopause and boundary layer. Recently, global atmospheric moisture and temperature profiles with a much higher vertical resolution were produced by the Atmospheric Infrared Sounder (AIRS)/Advanced Microwave Sounding Unit on the NASA Aqua satellite. The AIRS data provide an unprecedented opportunity to document the vertical moist thermodynamic structure and spatial–temporal evolution of the MJO. Figure 2 presents the pressure-longitude cross sections of the temperature anomaly and its relationship to the convective rainfall anomaly for a composite MJO cycle (from20 day to +20 day, separated by 10 days each). In the Indo-Pacific

353

warm pool, the temperature anomaly exhibits a trimodal vertical structure: a warm anomaly in the free troposphere (800–250 hPa) and a cold anomaly near the tropopause (above 250 hPa) and in the lower troposphere (below 800 hPa) for the wet phase and vice versa for the dry phase. The moisture anomaly also shows markedly different vertical structures as a function of longitude and the strength of the convection anomaly. Most significantly, the AIRS data demonstrate that, over the Indian Ocean and western Pacific, enhanced convection and precipitation is generally preceded in both time and space by a low-level warm and moist anomaly and followed by a low-level cold and dry anomaly. This zonal asymmetry in the low-level moisture and temperature anomaly provides a favorable moist thermodynamic condition for the eastward propagation of the MJO. Furthermore, the comparison between the AIRS observations and the National Center for Environmental Prediction (NCEP)/National Center for Atmospheric Research (NCAR) and NCEP/ Department of Energy (DOE) reanalyses revealed the poor representation of the low-level moisture and temperature structure associated with the MJO in these reanalyses particularly over the equatorial Indian and Pacific Oceans, where there are very few conventional data to constrain the reanalyses. Despite their reliance on (imperfect) numerical models, these reanalyses have been widely used as “observations” to validate MJO theories and model simulations. The ocean surface fluxes of heat (solar and infrared radiation, latent and sensible heat), mass (rainfall and water vapor flux), and momentum (wind) are considered important for the initiation and eastward propagation of the MJO by some researchers. Understanding the manner and degree the atmospheric components of the MJO are coupled to ocean surface fluxes is vital to understand the MJO dynamics and establish MJO theories. This coupling includes how the atmospheric components of the MJO influence the ocean surface heat, mass, and momentum fluxes and how the ocean surface provides the needed heat and moisture sources for the MJO convection. Surface heat fluxes have typically been very difficult to measure remotely from space although satellites offer the only viable way to estimate these quantities with regular temporal and spatial samplings over the tropical oceans. Apart from precipitation discussed above, satellite measurements of clouds and water vapor have been used with atmospheric radiative transfer models to provide estimates of solar and infrared radiation fluxes at the ocean surface. Moreover, satellite estimates of ocean surface winds (e.g., QuikSCAT and SSM/I discussed above) have been used in conjunction with water vapor measurements from satellites to construct estimates of latent heat (or evaporative) flux from the ocean. These types of observations, in conjunction with satellite SST retrievals (e.g., Advanced Very High Resolution Radiometer (AVHRR) or TMI), have been used to study how the MJO convection interacts with the ocean surface and explore the degree the ocean and atmosphere are coupled at intraseasonal and other

354

MADDEN-JULIAN OSCILLATION (MJO)

4 2

400

0

600

–2

800 1000

–4

Rainfall (mm/day)

Pressure (mb)

lag -4 0 200

0

600

–2

800 1000

–4

lag 0

0 200

4 2

400

0

600

–2

800 1000

–4

lag +2

0 200

4 2

400

0

600

–2

800 1000

–4

lag +4

0 200

4 2

400

0

600

–2

800 1000 0

30

60

90

120

150

180

210

240

270

300

330

–4 360

Rainfall (mm/day)

2

400

Rainfall (mm/day)

4

Rainfall (mm/day)

0 200

Rainfall (mm/day)

Pressure (mb)

Pressure (mb)

Pressure (mb)

Pressure (mb)

lag -2

Longtitude (E) 8S-8N Temperature MJO Anomaly –0.4

–0.3

–0.2

–0.1

0.0

0.1

0.2

0.3

0.4

Kelvin

Madden-Julian Oscillation (MJO), Figure 2 Pressure-longitude cross sections of equatorial mean (averaged from 8
S to 8
N) temperature anomaly for a composite MJO cycle based on AIRS data from 2002 to 2005. The color red denotes warm anomalies, while the color blue indicates cold anomalies. The superimposed solid black line denotes rainfall anomaly from TRMM (Reproduced from Tian et al., 2006, their Figure 3, by permission of American Meteorological Society).

timescales. Satellite-based ocean surface wind speed observations (e.g., QuikSCAT and SSM/I discussed above) have been used to derive momentum fluxes (e.g., wind stress) at the ocean surface. These observations have been critical in documenting and understanding the role of the MJO in influencing the development and evolution of El Niño and Southern Oscillation (ENSO) events. ENSO is the most important interannual variability in the coupled tropical atmosphere ocean system with a dominant

timescale of 2–7 years and has significant impacts on the global weather and climate. During the dry phase of the MJO, suppressed convection is associated with decreased cloud cover and increased surface insolation and anomalous surface easterlies. These anomalous surface easterlies act to decrease the surface wind speed because the background surface winds are weak westerlies in the equatorial Indian and western Pacific Oceans, hence decreasing the surface latent heat flux

MADDEN-JULIAN OSCILLATION (MJO)

355

Boreal Winter MJO Impacts 60N 40N 20N EQ 20S 40S 60S 0

30E

60E

90E

120E

150E

180

150W

120W

90W

60W

30W

0

Madden-Julian Oscillation (MJO), Figure 3 Geographical regions of a number of impacts of the boreal winter MJO on the global climate system. (1) Influencing tropical weather, alternative periods of wetter/drier conditions in the tropical Indian Ocean and western Pacific. (2) Modulating the diurnal cycle of tropical deep convection and rainfall in the tropical Indian Ocean and western Pacific. (3) Modulating the onsets and breaks of the Australian and South American monsoon systems. (4) Influencing ENSO cycle over the equatorial central and eastern Pacific. (5) Impacting tropical cyclone genesis over the tropical Indian Ocean and western Pacific. (6) Influencing the development of heavy rainfall events over US west coast. (7) Changing the subtropical total-column ozone over the eastern Hemisphere and Pacific Ocean. (8) Affecting the aerosol and air pollution over the equatorial Indian and western Pacific Oceans as well as the tropical Africa and Atlantic Ocean. (9) Influencing the ocean surface Chl across the tropical ocean coasts.

(or evaporation). Increased surface shortwave radiation and reduced surface evaporation contribute to the warming of SST for the dry phase. During the subsequent wet phase of the MJO, enhanced convection is associated with increased cloud cover and decreased surface insolation. As a result, the SST warming trend is arrested and a cooling trend is initiated. Subsequently, the continued cooling of the upper ocean is accelerated by increased westerly surface winds leading to enhanced surface evaporation and increased entrainment of cold water from below the thermocline. Then the wet phase is followed by another dry phase when SST warming occurs. Therefore, over the Indian Ocean and western Pacific, the enhanced convection is usually led by a warm SST anomaly to the east due to enhanced insolation and decreased evaporation and followed by a cold SST anomaly to the west due to decreased insolation and enhanced evaporation. When the convective anomaly approaches the Date Line, the surface evaporation anomaly and surface solar radiation anomaly tend to cancel each other. Thus, the SST anomaly is rather small over the eastern Pacific, so does the convective anomaly. This convection–SST phase relationship leads many scientists to believe that the MJO is a coupled mode of the tropical ocean–atmosphere system.

Impacts During the past three decades, the MJO has been shown to have important influences on various weather and climate phenomena over the globe at many timescales, such as the diurnal cycle, tropical weather, monsoon onsets and breaks, ENSO, tropical hurricanes and cyclones, extreme precipitation events, extratropical and high-latitude circulation, and weather patterns. Given evidence that the MJO is predictable with lead times of 2–3 weeks, the strong modulation of the global climate system by the MJO implies that many

other components of the global climate system may be predictable with similar lead times. Some examples of the MJO impacts are listed and described below, and the regions of impacts are illustrated in Figure 3. First, the MJO significantly impacts the tropical synoptic weather, such as alternative periods of wet and dry conditions, especially over the tropical Indian Ocean and west Pacific, through its influences of tropical rainfall and cloudiness (Figure 3). During the wet phase of the MJO, the tropical atmosphere is very moist and cloudy with heavy rainfall. In contrast, during the dry phase of the MJO, the tropical atmosphere experiences dry and clear conditions with plenty of sunshine. Second, the MJO can influence the diurnal cycle of tropical deep convection and rainfall through its effect on the background state over the equatorial Indian and western Pacific oceans (Figure 3). The diurnal cycle is enhanced over both land and water during the convectively active phase of the MJO, while it is reduced during the convectively suppressed phase of the MJO. However, the diurnal phase is not significantly affected by the MJO. Third, the MJO can substantially modulate the intensity of monsoon systems around the globe, such as the Australian and South American monsoons for boreal winter and the Asian and North American monsoons for boreal summer (Figure 3). The wet phase of the MJO can affect both the onset timing and intensity of the monsoon, while the dry phase of the MJO can prematurely end a monsoon and also initiate breaks during already existing monsoons. Fourth, there is evidence that the MJO influences the ENSO cycle (Figure 3). It was argued that the westerly wind bursts associated with the MJO over the equatorial western Pacific are an important trigger for an El Niño event. The MJO may not cause an El Niño event, but

356

MADDEN-JULIAN OSCILLATION (MJO)

can contribute to the speed of development, and perhaps the overall intensity of an ENSO cycle. Fifth, the MJO is known to modulate tropical cyclone activity in the Indian Ocean, Pacific Ocean, Gulf of Mexico, and Atlantic Ocean by providing a large-scale environment that is favorable (unfavorable) for storm development (Figure 3). For example, westerly wind anomalies at the surface in and just behind the area of enhanced convection of the MJO may generate cyclonic (anticyclonic) rotation north (south) of the equator, respectively. At the same time, in the upper levels, anticyclonic (cyclonic) rotation develops along and just behind the area of enhanced convection resulting in a means to reduce vertical wind shear and increase upper-level divergence – both of which are favorable for tropical cyclone development and intensification. The strongest tropical cyclones tend to develop during the wet phase of the MJO. As the MJO progresses eastward, the favored region for tropical cyclone activity also shifts eastward from the Indian Ocean to the Pacific Ocean and eventually to the Atlantic Ocean. Sixth, boreal winter extreme precipitation events along the US west coast are often connected with the pattern of tropical rainfall and circulation anomalies associated with the MJO (Figure 3). When the heavy tropical rainfall associated with the MJO is concentrated at the Maritime Continent, a strong blocking anticyclone is located in the Gulf of Alaska with a strong polar jet stream around its northern flank. During this time, the US west coast typically experiences a dry spell. When the enhanced tropical rainfall associated with the MJO shifts to the central Pacific and weakens, a deep low pressure system typically forms near the Pacific Northwest coast and can bring up to several days of heavy rain and possible flooding to the Pacific Northwest coast. These events are often referred to as “Pineapple Express” events, so named because a significant amount of deep tropical moisture traverses the Hawaiian Islands on its way toward western North America. Although most quantities/processes/phenomena of interest, such as hurricanes, monsoons, and extratropical circulation and weather patterns, are not wholly described by satellite data, the satellite-based OLR or rainfall data were usually used to identify the MJO events and often to characterize aspects of the impacts (e.g., rainfall). Some studies totally depend on the available satellite data. For example, in the study of the MJO impact on the diurnal cycle of tropical deep convection, the International Satellite Cloud Climatology Project cloud product was employed to characterize the diurnal cycle, and the TRMM 3B42 precipitation product was used to identify the MJO events. Recently, a number of studies have documented the MJO impacts on atmospheric composition, air quality, and biogeochemical cycle. This discovery has critically depended on the availability of satellite data. For example, the MJO

impact on tropical total-column ozone has been recently characterized using the satellite-observed tropical totalcolumn ozone from the AIRS and Total Ozone Mapping Spectrometer (TOMS). It was found that tropical total ozone intraseasonal variations are large ( 10 Dobson unit) and comparable to those in the annual and interannual timescales. These intraseasonal total ozone anomalies are mainly evident in the subtropics over the Pacific and eastern hemisphere, with a systematic relationship to the MJO convection and wave dynamics discussed earlier (Figures 3 and 4a). The subtropical negative ozone anomalies typically flank or lie to the west of the equatorial anomalous convection and are collocated with the subtropical uppertroposphere anticyclones generated by the equatorial anomalous convective heating. On the other hand, the subtropical positive ozone anomalies generally lie to the east of the equatorial anomalous convection and are collocated with the subtropical upper-troposphere cyclones generated by the equatorial anomalous convective heating. The subtropical ozone anomalies are anticorrelated with geopotential height anomalies near the tropopause and thus mainly associated with the ozone variability in the stratosphere rather the troposphere. Another example, the recent availability of multiple, global satellite aerosol products from TOMS, Moderate Resolution Imaging Spectroradiometer (MODIS), and Advanced Very High Resolution Radiometer (AVHRR) has made the investigation of the MJO modulation of aerosols possible. Large aerosol variations are found over the equatorial Indian and western Pacific Oceans where MJO convection is active, as well as the tropical Africa and Atlantic Ocean where MJO convection is weak, but the background aerosol level is high (Figures 3 and 4b). Although significant uncertainties still exist in the satellite aerosol retrievals, the satellite data indicate that the MJO and its associated cloudiness, rainfall, and circulation variability may systematically influence the aerosol variability. The impacts of the MJO on the carbon monoxide (CO) abundances in the tropical tropopause layer (TTL) were also recently reported based on the Aura Microwave Limb Sounder (MLS) CO data. The effects of the eastward propagation of MJO on CO abundances in the TTL are evident. This indicates that the anomalous deep convection associated with the MJO can inject CO from the lower troposphere up to the TTL. The availability of satellite-derived ocean surface chlorophyll (Chl) from Sea-viewing Wide Field-of-view Sensor (SeaWiFS) provided a means for the discovery of the MJO impacts on oceanic biology. It was found that the MJO produces systematic and significant variations in ocean surface Chl in a number of regions across the tropical Oceans, including the northern Indian Ocean, a broad expanse of the northwestern tropical Pacific Ocean, and a number of near-coastal areas in the far eastern Pacific Ocean (Figures 3 and 4c, d). This indicates a need to further investigate the MJO modulation of the biogeochemical cycle properties and higher levels of food chains.

MADDEN-JULIAN OSCILLATION (MJO)

357

Madden-Julian Oscillation (MJO), Figure 4 (a) Map of total-column ozone anomaly for a composite MJO cycle based on TOMS/ SBUV data from 1980 to 2006. The color red denotes high ozone anomalies, while the color blue indicates low ozone anomalies. The superimposed solid black line denotes rainfall anomaly from CMAP (Reproduced from Tian et al., 2007, by permission of American Geophysical Union). (b) As (a) but for TOMS aerosol index anomaly (Reproduced from Tian et al., 2008, by permission of American Geophysical Union). (c) As (a) but for CMAP rainfall anomalies. (d) As (a) but for SeaWiFS ocean surface Chl anomaly (Reproduced from Waliser et al., 2005, by permission of American Geophysical Union).

Summary The MJO is a large-scale quasiperiodic oscillation of tropical atmospheric circulation and convection anomalies that moves slowly eastward along the equator mainly over the tropical Indian and Pacific Oceans with a timescale on the order of 30–60 days. The MJO is the dominant form of the intraseasonal variability in the tropical atmosphere and has many important influences on the global weather and climate system. Since the 1970s, the satellite remote sensing data have played a fundamental role in advancing our knowledge in the MJO, particularly in terms of its description, theoretical mechanisms, and global impacts. First, the satellite data provided us the fundamental knowledge of the convective and dynamic features of the MJO. Second, the satellite data presented us the three-dimensional thermodynamic structure and the surface condition (e.g., SST and surface heat flux) evolution associated with the MJO that helped us to better understand the MJO and propose theoretical description. Third, the satellite data offered us the opportunity to discover

the global impacts of the MJO that have relevance to societal concerns such as extreme precipitation events, atmospheric composition, air quality, and biological markers in the ocean.

Acknowledgment This research was carried out at Jet Propulsion Laboratory, California Institute of Technology, under a contract with NASA. Bibliography Gottschalck, J., Kousky, V., Higgins, W., and L’Heureux, M., 2008. Madden-Julian oscillation. http://www.cpc.noaa.gov/products/ precip/CWlink/MJO/mjo.shtml Hendon, H. H., and Salby, M. L., 1994. The life-cycle of the Madden-Julian oscillation. Journal of the Atmospheric Sciences, 51(15), 2225–2237. Lau, W. K. M., and Waliser, D. E. (eds.), 2005. Intraseasonal Variability of the Atmosphere–Ocean Climate System. Heidelberg: Springer. 474 pp.

358

MAGNETIC FIELD

Madden, R. A., 2003. Tropical meteorology: intraseasonal oscillation (Madden-Julian oscillation). In Holton, J. R., et al. (eds.), Encyclopedia of Atmospheric Sciences. London: Academic, pp. 2334–2338. Madden, R. A., and Julian, P. R., 1971. Detection of a 40–50 day oscillation in the zonal wind in the tropical pacific. Journal of the Atmospheric Sciences, 28(7), 702–708. Madden, R. A., and Julian, P. R., 1972. Description of global-scale circulation cells in tropics with a 40–50 day period. Journal of the Atmospheric Sciences, 29(6), 1109–1123. Madden, R. A., and Julian, P. R., 1994. Observations of the 40–50day tropical oscillation: a review. Monthly Weather Review, 122(5), 814–837. Tian, B., Waliser, D. E., Fetzer, E. J., Lambrigtsen, B. H., Yung, Y. L., and Wang, B., 2006. Vertical moist thermodynamic structure and spatial-temporal evolution of the MJO in AIRS observations. Journal of the Atmospheric Sciences, 63(10), 2462–2485, doi:10.1175/JAS3782.1. Tian, B., Yung, Y. L., Waliser, D. E., Tyranowski, T., Kuai, L., Fetzer, E. J., and Irion, F. W., 2007. Intraseasonal variations of the tropical total ozone and their connection to the MaddenJulian oscillation. Geophysical Research Letters, 34(8), L08704, doi:10.1029/2007GL029451. Tian, B., Waliser, D. E., Kahn, R. A., Li, Q. B., Yung, Y. L., Tyranowski, T., Geogdzhayev, I. V., Mishchenko, M. I., Torres, O., and Smirnov, A., 2008. Does the Madden-Julian oscillation influence aerosol variability? Journal of Geophysical Research, 113(D12), D12215, doi:10.1029/2007JD009372. Waliser, D. E., 2006. Intraseasonal variability. In Wang, B. (ed.), The Asian Monsoon. New York: Springer/Praxis, pp. 203–257. Waliser, D. E., Murtugudde, R., Strutton, P., and Li, J. L., 2005. Subseasonal organization of ocean chlorophyll: prospects for prediction based on the Madden-Julian oscillation. Geophysical Research Letters, 32(23), L23602, doi:10.1029/2005GL024300. Zhang, C. D., 2005. Madden-Julian oscillation. Reviews of Geophysics, 43(2), RG2003, doi:10.1029/2004RG000158.

Cross-references Aerosols Radars

MAGNETIC FIELD Nils Olsen DTU Space, Technical University of Denmark, Lyngby, Denmark

Definition Core field. The main part of the geomagnetic field, caused by dynamo action in the Earth’s fluid outer core. Crustal field. The magnetic field contribution caused by (permanent or remnant) magnetized material in the Earth’s crust. Internal sources. Magnetic field contributions caused by electrical currents or magnetized material in the Earth’s interior. External sources. Magnetic field contributions produced by electric currents in the ionosphere or magnetosphere.

Introduction The Earth has a large and complicated magnetic field, the major part of which is produced by a self-sustaining dynamo operating in the fluid outer core. Magnetic field observations provide one of the few tools for remote sensing the Earth’s deep interior, especially regarding the dynamics of the fluid flow at the top of the core. However, what is measured at or near the surface of the Earth is the superposition of the core field and fields caused by magnetized rocks in the Earth’s crust, by electric currents flowing in the ionosphere, magnetosphere, and oceans, and by currents induced in the Earth by time-varying external fields. These sources have their specific characteristics in terms of spatial and temporal variations, and their proper separation, based on magnetic measurements, is a major challenge. Such a separation is a prerequisite for remote sensing by means of magnetic field observations. Data sources Prior to the satellite era, only near-surface (ground-based, marine, and airborne) magnetic observations were available for magnetic probing the Earth’s interior. Presently about 150 geomagnetic observatories monitor the time changes of the field; their data are available through the World Data Center system (e.g., www.ngdc.noaa.gov/ wdc, www.wdc.kugi.kyoto-u.ac.jp, www.wdc.bgs.ac.uk) and INTERMAGNET (www.intermagnet.org). However, the global distribution of the observatory network is very uneven, with large uncovered areas especially over the oceans. While observatories monitor the field change at a given fixed location, regional surveys map the magnetic field at a given epoch (short period fluctuations of external origin have to be removed from the raw data; for this purpose, measurements of the field change at a nearby observatory or a temporary station are used). However, only parts of the globe are presently covered by airborne and marine surveys. The obtained data have been used for instance in the preparation of the World Digital Magnetic Anomaly Map (WDMAM, Korhonen et al. (2007), see also http:// projects.gtk.fi/WDMAM/) and dedicated anomaly maps like EMAG2 (Maus et al., 2009). True global coverage with magnetic field observations is only possible with satellites. Although spaceborne probing began more than 60 years ago with the launch of the Sputnik 3 satellite in 1958, the data coverage that is necessary for global field modeling was first obtained by the POGO satellite series (Cain, 2007) which measured the magnetic field intensity between 1965 and 1972. The first high-precision vector measurements from space were taken by the Magsat satellite (Purucker, 2007) in 1979– 1980. More recently, the launch of the satellites Ørsted in February 1999 (Olsen, 2007), CHAMP in July 2000 (Maus, 2007), and SAC-C in November 2000 opened revolutionary new possibilities for probing the Earth from space. In the near future, the Swarm satellite constellation mission (Friis-Christensen et al., 2006, 2009), comprising

MAGNETIC FIELD

359

three satellites expected to be launched in late 2013, will provide even better possibilities for magnetic remote sensing the Earth’s interior.

Characteristics of the geomagnetic field The strength of the magnetic induction B, in following for simplicity denoted as “magnetic field,” varies at Earth’s surface between about 25,000 nT near the equator and about 65,000 nT near the poles (1 nT ¼ 109 T, with 1 T ¼ 1 tesla ¼ 1 V/s1 m2). By far the largest part (95 % or more at Earth’s surface) is due to dynamo action in the core (e.g., Roberts, 2007); magnetized material in the crust (e.g., Purucker and Whaler, 2007) accounts on average for only a few % of the total field but can locally reach magnitudes of several hundreds or even thousands of nT. Crustal magnetization consists of two parts: Induced magnetization is proportional, both in strength and direction, to the ambient field within which the rock is embedded. Were the core field to disappear, induced magnetization would vanish, too. Then, only the second type of magnetization, remnant magnetization, remains. As a general rule, remnant magnetization is weak in continental regions (where induced magnetization dominates), while both types of magnetizations are significant in oceanic areas. The top panel of Figure 1 shows the magnetic field intensity at Earth’s surface in 2010.0, and the middle panel shows the yearly change in field intensity (also for epoch 2010.0), based on the International Geomagnetic Reference Field (IGRF) (Finlay et al., 2010). Core and crust contribute to all spatial scales of the field, but the core field dominates for horizontal scales larger than 2,800 km (corresponding to spherical harmonic degrees n < 14) while the crustal field dominates for horizontal scales smaller than 2,800 km (n > 14). The bottom panel of the figure shows the crustal anomaly field (scales smaller than 2,400 km) close to the surface (4 km above the geoid), based on the EMAG2 model of Maus et al. (2009). The crustal field is static (at least on timescales of centuries or shorter), but the core field undergoes a significant time change, known as secular variation. The temporal change shown in the middle panel of Figure 1 is therefore caused by processes in the Earth’s core. In addition to these internal sources, there are contributions from electric currents in the ionosphere (90–1,000 km altitude) and the magnetosphere (at distance larger than several Earth radii); these are called external sources. They are very dynamic, ranging from a few nT during geomagnetic quiet conditions to several hundreds or even thousands of nT during disturbed times, with especially large amplitudes at polar latitudes. Figure 2 shows a typical spectrum of the magnetic North component for a site at mid-latitudes. Time changes with periods longer than 4 years are dominated by core processes, while those at shorter periods, and especially signals with seasonal and daily periodicity, are caused by ionospheric and magnetospheric sources. See Schmucker (1985) for details on the

20000

30000

40000

50000

60000

70000

[nT]

−200

−100

0 [nT/yr]

−200 −60 −48 −36 −20 −4

4 [nT]

200

100

20

36

48

60 200

Magnetic Field, Figure 1 Earth’s magnetic field intensity (top) and its time change (middle) in 2010, at Earth’s surface. Bottom: intensity of magnetic field anomalies at 4 km altitude above the geoid.

external field variations ELF, sferics, ULF, pulsations, DP, Dst, and Sq. These short period changes, which are caused by external currents, produce secondary, induced, currents in the Earth’s interior, the magnetic field of which adds to that of the primary, external, currents. The observed magnetic time variations are thus a superposition of primary (inducing) and secondary (induced) contributions. Analysis of these time changes

360

MAGNETIC FIELD

Period [days] 10–6

10–4

10–2

100

102

104

11 yrs

Validity of pre-Maxwell approximation 106 ELF - sferics

DP - substorms

Dst - storms 1010

Sq

ULF - pulsations

10

60 Hz and harmonics (anthropogenic origin)

105

Annual and semi- annual variation

Daily variation and harmonics

100 Schumann resonance oscillations

10–2

100

external I internal variations dominate

Spectral power [nT2/Hz]

27 day solar rotation

2

4 yrs

Spectral amplitude [nT/Hz1/2]

104

10–5 seconds 1

10–4

minutes

10 1

hours

10

1

10

days 1

months

10

1

years

10 1

10

Period [secs] 10–2

100

102

104

100 10–10

106

108

10–6

Magnetic Field, Figure 2 Amplitude of magnetic field variations (north component) at middle latitudes in dependence on period, determined using data from the sites Socorro/USA (1 ms to 1 s), Kakioka/Japan (2 s to 2 h), and Niemegk/Germany (longer than 2 h). Note that the pre-Maxwell approximation Equation 1 is only valid for periods 1 s.

provides information on the electrical conductivity of the Earth’s interior (e.g., Constable, 2007). Although field changes of internal as well as external origin occur at all timescales, a common practice in separating them relies on their different temporal variations. Over the last 150 years, the axial dipole component of the Earth’s magnetic field has decayed by nearly 10 %. This is ten times faster than the natural decay, in case the dynamo was switched off. The current decay rate is characteristic of magnetic reversals, which – as paleomagnetic data have shown – occur on average about once every half million years. Geographically, the recent dipole decay is largely due to changes in the field beneath the South Atlantic Ocean, connected to the growth of the South Atlantic anomaly. The core field and its temporal change (secular variation) directly reflect the fluid flow in the outermost core and provide a unique observational constraint on geodynamo theory. However, only the part of the core field that varies on timescales longer than, say, 1 year is

observable at the Earth’s surface; shorter fluctuations are heavily attenuated due to the electrical conductivity of the mantle. Hence, variations with timescales longer than a few years are usually attributed to processes in the core, whereas those with periods shorter than 1 year are attributed to external field contributions. Yet interesting features occur at intermediate timescales. A serious limitation regarding the investigation of core processes at timescales of months to years is the effect of geomagnetic variations of external origin, since they contribute significantly on timescales up to that of the 11 year solar cycle.

Basic equations Magnetic field investigation of core and crust is typically done in the quasi-static approximation, which requires that the timescales in consideration are longer (1 s) compared to the time required for light to pass the length scale of interest ( T 22:9
C T

(52c) Equations 49a, b–51a, b do not explicitly take into account the effect of ice density upon the real part of the permittivity or direction of wave propagation in the ice medium. The correlation coefficient for Equations 49a, b–51a, b is, in general, between 0.7 and 0.8.

371

In general, the absorption coefficient ka increases with increasing frequency and temperature. The loss of multiyear ice is considerably lower than that of first-year ice, mainly due to much lower salinity. The values for first-year lowsalinity sea ice show that the temperature dependence of the absorption coefficient is high at temperatures near 0
C. At 10 GHz, the differences between the results for columnar, frazil, and multiyear ice are substantial. In the 1–10 GHz range, the penetration depth is between 100 and 5 cm for first-year ice and 500 and 30 cm for multiyear ice. Sensors operating at X-band provide information on sea ice mainly from the topmost 5–80 cm, depending on ice type, salinity, and temperature, whereas the corresponding numbers for L-band sensors are 40–500 cm.

Dielectric properties of soils Soils consist of bulk soil, air, and water. A soil’s textural composition is usually given in terms of weight percentages of sand, silt, and clay. Sand includes particles with diameters in the range between 0.05 and 2.0 mm, silt includes particles with diameters in the 0.002–0.05 mm range, and clay includes particles with diameters smaller than 0.002 mm. The water contained in the soil is usually divided into bound water and free water, although the transition between these is not sharp. Molecules of bound water are contained in the first few molecular layers surrounding the soil particles; these are tightly held by the soil particles due to the matric and osmotic forces (Baver et al., 1977). The amount of bound water is proportional to the total surface area of the soil particles, which depends on the soil particle size distribution and mineralogy: The smaller the particles, the more there can be bound water. Hence, clay has more bound water than sand. Water molecules located further away from the soil particle surface can move within the soil medium and are referred to as free water. From the electromagnetic point of view, soil is a heterogeneous medium consisting of bulk soil, air, bound water, and free water. Its dielectric properties as a function of frequency depend on the soil bulk density (compaction), soil composition (particle size distribution and mineralogy), the volume fractions of bound and free water, the salinity of the soil solution, and temperature (Dobson et al., 1985). Wang and Schmugge (1980) showed that various dielectric mixing formulas describing soil as a two-component system (soil with free water) fail to describe the complex permittivity of wet soil realistically at 1.4 and 5 GHz. They were the first to explicitly include both free water and bound water in their empirical mixing formula by examining two moisture regions: (a) water contents less than the maximum bound water fraction and (b) water contents higher than the bound water fraction. They included two free parameters in the model and, by optimizing their values, were able to explain the behavior of the complex permittivity, especially its real part. Wang (1980) modeled the soil–water system with

372

MICROWAVE DIELECTRIC PROPERTIES OF MATERIALS

a Debye-like relaxation over a finite band of relaxation frequencies and, by adjusting two free parameters, adequately predicted the behavior of data over the 0.3–1.4 GHz range. The two free parameters were the width of the activation energy of the soil solution and the mean relaxation frequency of the soil–water mixture at a given frequency and for a given soil and water content. Hallikainen et al. (1985) performed dielectric measurements of five soil types at 12 frequencies between 1.4 and 18 GHz. Soil moisture values from 0 % to the highest moisture contents that can be supported by that soil type without drainage taking place were used; additionally, data for frozen soils were acquired as well. They developed for each frequency polynomial expressions, separately for real and imaginary parts of the permittivity, dependent on volumetric moisture content and the percentage of sand and clay contained in the soil. Dobson et al. (1985) developed two dielectric models for wet soil based on the data set of Hallikainen et al. (1985), a theoretical model and a semiempirical model. The theoretical model accounts explicitly for the presence of bound water adjacent to hydrophilic soil particle surfaces and employs a fourcomponent dielectric mixing model that describes the soil–water system consisting of dry soil solids as a host medium with randomly distributed and randomly oriented disc-shaped inclusions of bound water, bulk water, and air. The bulk water component characteristics in the model depend on frequency, temperature, and salinity. Based on comparisons with data, the theoretical model was determined to be an appropriate formulation, and it yields values that describe well the observed effects of frequency and soil type. However, its accuracy is limited by the uncertainty concerning the dielectric properties of bound water. The driving force in the development of the semiempirical model of Dobson et al. (1985) was to develop a user-friendly, frequency-dependent model that is based on readily measured soil characteristics including volumetric moisture and weight fractions of sand and clay. The final expressions for the real and imaginary parts of the permittivity are

1a 0 r  m (53a) e0ws ¼ 1 þ b ðeas  1Þ þ mbv e0a v fw rs h 00 i1a e00ws ¼ mbv e00a fw

(53b)

where a is a constant shape factor, rb is soil bulk density, rs ¼ 2.66 g/cm3 is soil specific density, es is the soil solid permittivity, mv is the volumetric moisture, b0 and b00 are the soil-texture-dependent coefficients for computing the real and imaginary part, respectively, and efw is the permittivity of free water. The optimum value for the shape factor was determined to be a ¼ 0.65. The relative permittivity of soil solids is es ¼ ð1:01 þ 0:44rs Þ2  0:062 and the soil-texture-dependent coefficients are

(54)

be0 ¼ 0:01ð127:48  0:519S  0:152CÞ

(55a)

be00 ¼ 0:01ð1:33797  0:603S  0:166CÞ

(55b)

where S and C are the percentage of sand and clay, respectively. The dielectric properties of free water are computed using an expression similar to that for saline water: seff rs  rb ew0  ew? j

(56) ew ¼ ew? þ 1 þ j2pf tw 2pf e0 rs mv where the effective conductivity is seff ¼ 1:645 þ 1:939rb  0:02013S þ 0:01594C (57a) The above expressions were determined using the whole soil dielectric data set between 1.4 and 18 GHz. For the frequency range of 0.3–1.3 GHz, another expression for the effective conductivity was developed (Peplinski et al., 1995): seff ¼ 0:0467 þ 0:2204rb  0:4111S þ 0:66144C (57b) For a reader interested in the dielectric properties of soils at a certain frequency between 1.4 GHz and 18 GHz, the polynomial expressions in Hallikainen et al. (1985) are useful. Mironov et al. (2004) developed a generalized refractive index mixing dielectric model for moist soils incorporating the dielectric characteristics of bound water. The above studies show that the complex permittivity of wet soil is a compressed version of that for slightly saline water. In the 1.4–18 GHz range, the real part of the permittivity increases with increasing volumetric moisture and decreasing frequency, whereas the imaginary part increases with increasing moisture and increasing frequency. At any given moisture content and at all frequencies, the real part is roughly proportional to sand content and inversely proportional to clay content; however, this effect decreases with increasing frequency. The effect of soil texture on the imaginary part varies with frequency. At 1.4 GHz, the real part increases with increasing clay content for moisture levels above 0.2 cm3/cm3, whereas at 4–6 GHz, it is nearly independent of soil texture at all moisture levels and at frequencies of 8 GHz and above it decreases with increasing clay content. This behavior is obviously due to the ionic conductivity (strongest at low frequencies) and the relation of volume fraction of bound water to soil specific surface. At moisture levels close to saturation, e may reach values up to 23-j3 for sand (mv ¼ 0.35) and 33-j9 for clay (mv ¼ 0.50) at 1.4 GHz. At 18 GHz, the corresponding values are 13-j8 and 18-j12, respectively. The penetration depth for wet soil decreases drastically with increasing moisture content and frequency;

MICROWAVE DIELECTRIC PROPERTIES OF MATERIALS

consequently, retrieval of soil moisture from microwave radiometer and radar data works best at low frequencies. Even at 1.4 GHz, information only on surface moisture is obtained. The permittivity of soils decreases drastically at temperatures below 0
C (Hallikainen et al., 1985). Between 11
C and 24
C, both the real and imaginary parts of permittivity still depend on temperature, demonstrating that not all water is frozen. Recently, Mironov et al., (2010) conducted dielectric measurements on organic rich permafrost soil from 1 to 16 GHz and from 30
C to þ25
C. They also extended the previously developed dielectric soil model of Mironov et al. (2004) to cover frozen soil.

Dielectric properties of vegetation The characteristics of plants and trees vary substantially, including the size, density, moisture content, and geometry. Hence, the physical characteristics of vegetation material are not discussed in this presentation. For a detailed discussion on timber characteristics, the reader is referred to Dinwoodie (2000). Dielectric properties of wood are discussed in detail in Torgovnikov (1993); they have been primarily examined for industrial purposes, especially for microwave drying of timber at 2.45 GHz. The dielectric properties of plants including trunks, stalks, and leaves for remote sensing have been measured and modeled by El-Rayes and Ulaby (1987), and Ulaby and El-Rayes (1987) in the 0.2–20 GHz range, and monitored in situ at L-band by McDonald et al. (1999). Mätzler (1994) measured and modeled various leaves at frequencies 1–94 GHz. Ulaby et al. (1987) measured the propagation constant for vegetation canopies with vertical stalks between 1.6 and 10.2 GHz. Dielectric spectroscopy of Scots pine at frequencies up to 1 GHz has been done by Tomppo et al. (2009). Ulaby and El-Rayes (1987) modeled the complex permittivity of vegetation as a simple additive mixture of plant material, free water, and bound water due to their observation that bound water has a substantially lower relaxation frequency than free water. At 22
C, their Debye–Cole dual-dispersion model for the permittivity of vegetation is

75:0 18s j ev ¼ er þ vfw 4:9 þ 1 þ j 0:556f f " # (58) 55:0 þ vbw 2:9 þ 1 þ ð j 0:556f Þ0:5 where er is the permittivity of bulk vegetation material, f is the frequency in GHz, s is the ionic conductivity of the free-water solution in S/m, and vfw and vbw are the relative fractions of free and bound water, respectively. The volumetric moisture content consists of free water and bound water: Mu ¼ vfw þ vbw ¼

Mg r 1  Mg ð1  rÞ

(59)

373

where Mg denotes the gravimetric moisture content and r is the bulk density. If Mg and r are known, the following expressions can be used for obtaining higher accuracy: er ¼ 1:7 þ 3:2Mv þ 6:5Mv2

(60)

vfw ¼ Mv ð0:82Mv þ 0:166Þ

(61)

vbw ¼

31:4Mv2

1 þ 59:5Mv2

(62)

The Debye–Cole dual-dispersion model has been determined to provide an estimate of e0 v for leaves, stalks, branches, and trunks with an accuracy of 5 %. If the salinity S is known in order to provide an estimate for the conductivity (S/m) via s ¼ 0:16S  0:0013S 2

(63)

similar accuracies apply to e00v.

Summary Dielectric properties of natural media at microwave frequencies have been reviewed, and the effects of various physical parameters to the behavior of their permittivity have been discussed. Since most natural media are heterogeneous, theoretical derivation of their dielectric properties is not straightforward. Measurements provide useful information for model development and evaluation, but only within the range of the covered physical characteristics and employed frequencies. The permittivity of pure and saline water is the key to modeling the behavior of most natural media; gaps still occur for its values at nearfreezing temperatures, especially at high microwave frequencies. More experimental data are needed to confirm the behavior of wet snow at frequencies above 18 GHz. Measurements of the permittivity of sea ice for various salinities and ice types, especially for high brine volumes and at temperatures above 5
C, would substantially benefit modeling of the behavior of sea ice. Permittivity measurements of various types of vegetation are needed to test presently available semiempirical models and further develop them. Bibliography Anderson, D. L., 1960. The physical constants of sea ice. Research, 13, 310–318. Assur, A., 1960. Composition of Sea Ice and Its Tensile Strength. Wilmette: U.S. Army Snow, Ice and Permafrost Research Establishment. Baver, L. D., Gardner, W. H., and Gardner, W. R., 1977. Soil Physics. New York: Wiley. Bogorodskii, V. V., and Khokhlov, G. P., 1975. Electrical properties of ice in the ice edge zone of the Bering Sea at 10 GHz frequency. In Proceedings of the Final Symposium on the Results of the Joint Soviet-American Expedition. Leningrad: Gidrometeoizdat, pp. 219–233 (in Russian). Böttcher, C. J. F., 1973. Theory of Electrical Polarisation, 2nd edn. Amsterdam: Elsevier, Vol. 1.

374

MICROWAVE DIELECTRIC PROPERTIES OF MATERIALS

Böttcher, C. J. F., Van Belle, O. C., Bordewijk, P., and Rip, A., 1978. Theory of Electrical Polarisation, 2nd edn. Amsterdam: Elsevier, Vol. 2. Colbeck, S. C., 1982. The geometry and permittivity of snow at high frequencies. Journal of Applied Physics, 53, 4495–4500. Cumming, W., 1952. The dielectric properties of ice and snow at 3.2 cm. Journal of Applied Physics, 23, 768–773. Debye, P., 1929. Polar Molecules. New York: Dover. Dinwoodie, J. M., 2000. Timber, Its Nature and Behaviour, 2nd edn. New York: Taylor & Francis. Dobson, M. C., Ulaby, F. T., Hallikainen, M. T., and El-Rayes, M., 1985. Microwave dielectric behavior of wet soil – part II: dielectric mixing models. IEEE Transactions on Geoscience and Remote Sensing, 23, 35–46. Ellison, W., 2006. Microwave dielectric properties of water. In Mätzler, C., Rosenkranz, P. W., Battaglia, A., and Wigneron, J. P. (eds.), Thermal Microwave Radiation: Applications for Remote Sensing. Stevenage: Institute of Engineering and Technology. IET Electromagnetic Waves Series, Vol. 52, pp. 431– 455. El-Rayes, M. A., and Ulaby, F. T., 1987. Microwave dielectric spectrum of vegetation – part I: experimental observations. IEEE Transactions on Geoscience and Remote Sensing, 25, 541–549. Evans, S., 1965. Dielectric properties of ice and snow – a review. Journal of Glaciology, 5, 773–792. Frankenstein, G., and Garner, R., 1967. Equation for determining the brine volume of sea ice from 0.5
C to 22.9
C. Journal of Glaciology, 6, 943–944. Hallikainen, M. T., 1983. A new low-salinity sea-ice model for UHF radiometry. International Journal of Remote Sensing, 4, 655–681. Hallikainen, M. T., and Winebrenner, D., 1992. The physical basis for sea ice remote sensing. In Carsey, F. (ed.), Microwave Remote Sensing of Sea Ice. American Geophysical Union, Geophysical Monograph 68, Chap. 3, Washington, DC, pp. 29–46. Hallikainen, M. T., Ulaby, F. T., and Abdel-Razik, M., 1982. Measurements of the dielectric properties of snow in the 4–18 GHz frequency range. In Proceedings of the 12th European Microwave Conference. Helsinki, pp. 151–156. Hallikainen, M. T., Ulaby, F. T., Dobson, M. C., and El-Rayes, M., 1985. Microwave dielectric behavior of wet soil – part I: empirical models and experimental observations. IEEE Transactions on Geoscience and Remote Sensing, 23, 25–34. Hallikainen, M. T., Ulaby, F. T., and Abdelrazik, M., 1986. Dielectric properties of snow in the 3 to 37 GHz range. IEEE Transactions on Antennas and Propagation, 34, 1329–1340. Hallikainen, M. T., Ulaby, F. T., and Van Deventer, T. E., 1987. Extinction behavior of dry snow in the 18 to 90 GHz range. IEEE Transactions on Geoscience and Remote Sensing, 25, 737–745. Hoekstra, P., and Cappillino, P., 1971. Dielectric properties of sea and sodium chloride ice at UHF and microwave frequencies. Journal of Geophysical Research, 76, 4922–4931. Klein, L. A., and Swift, C. T., 1977. An improved model for the dielectric constant of seawater at microwave frequencies. IEEE Transactions on Antennas and Propagation, 25, 104–11. Liebe, H. J., Hufford, G. A., and Manabe, T., 1991. A model for the complex permittivity of water at frequencies below 1 THz. International Journal of Infrared and Millimeter Waves, 12, 659–675. Mätzler, C., 1994. Microwave (1–100 GHz) dielectric model of leaves. IEEE Transactions on Geoscience and Remote Sensing, 32, 947–949. Mätzler, C., 1996. Microwave permittivity of snow. IEEE Transactions on Geoscience and Remote Sensing, 34, 573–581. Mätzler, C., and Wegmüller, U., 1987. Dielectric properties of freshwater ice at microwave frequencies. Journal of Physics D: Applied Physics, 20, 1623–1630. Errata, 1988. 21:1660.

Mätzler, C., Aebischer, H., and Schanda, E., 1984. Microwave dielectric properties of wet snow. IEEE Journal of Oceanic Engineering, 9, 366–371. McDonald, K. C., Zimmermann, R., Way, J. B., and Chun, W., 1999. Automated instrumentation for continuous monitoring of the dielectric properties of woody vegetation: system design, implementation, and selected in situ measurements. IEEE Transactions on Geoscience and Remote Sensing, 437, 1880–1894. Mironov, V. L., De Roo, R. D., and Savin, I. V., 2010. Temperaturedependable microwave dielectric model for an Arctic soil. IEEE Transactions on Geoscience and Remote Sensing, 48, 2544–2556. Mironov, V. L., Dobson, M. C., Kaupp, V. H., Komarov, S. A., and Kleshchenko, V. N., 2004. Generalized refractive mixing dielectric model for moist soils. IEEE Transactions on Geoscience and Remote Sensing, 42, 773–785. Peplinski, N. R., Ulaby, F. T., and Dobson, M. C., 1995. Dielectric properties of soils in the 0.3–1.3 GHz range. IEEE Transactions on Geoscience and Remote Sensing, 33, 803–807. Correction: 33: 1340. Polder, D., and Van Santen, J. H., 1946. The effective permeability of mixtures of solids. Physica, 12, 257–271. Sackinger, W. M., and Byrd, R. C., 1972. Reflection of Mm waves from snow and sea ice, IAEA Report 7203, Institute of Arctic Environmental Engineering, University of Alaska, Fairbanks. Sihvola, A., 1999. Electromagnetic Mixing Formulas and Applications. London: Institution of Electrical Engineers. IEE Electromagnetic Wave Series, Vol. 47. Stogryn, A., 1986. A study of the microwave brightness temperature of snow from the point of view of strong fluctuation theory. IEEE Transactions on Geoscience and Remote Sensing, 24, 220–231. Stogryn, A., 1987. An analysis of the tensor dielectric constant of sea ice at microwave frequencies. IEEE Transactions on Geoscience and Remote Sensing, 25, 147–158. Stogryn, A., and Desargeant, G. J., 1985. The dielectric properties of brine in sea ice at microwave frequencies. IEEE Transactions on Antennas and Propagation, 33, 523–532. Tiuri, M., Sihvola, A., Nyfors, E., and Hallikainen, M. T., 1984. The complex dielectric constant of snow at microwave frequencies. IEEE Journal of Oceanic Engineering, 9, 377–382. Tomppo, L., Tiitta, M., Laakso, T., Harju, A., Venäläinen, M., and Lappalainen, R., 2009. Dielectric spectroscopy of scots pine. Wood Science and Technology, 43, 653–667. Torgovnikov, G. I., 1993. Dielectric Properties of Wood and WoodBased Materials. Berlin: Springer. Ulaby, F. T., and El-Rayes, M. A., 1987. Microwave dielectric spectrum of vegetation – part II: dual-dispersion model. IEEE Transactions on Geoscience and Remote Sensing, 25, 550–557. Ulaby, F. T., Moore, R. K., and Fung, A. K., 1986. Microwave Remote Sensing – Active and Passive, III: From Theory to Applications. Dedham: Artech House. Ulaby, F. T., Tavakoli, A., and Senior, T. B. A., 1987. Microwave propagation constant for a vegetation canopy with vertical stalks. IEEE Transactions on Geoscience and Remote Sensing, 25, 714–725. Vant, M. R., 1976. A Combined Empirical and Theoretical Study of the Dielectric Properties of Sea Ice over the Frequency Range 100 MHz to 40 GHz. Technical Report, Carleton University, Ontario, Ottawa. Vant, M. R., Ramseier, R. O., and Makios, V., 1978. The complexdielectric constant of sea ice at frequencies in the range 0.1–40 GHz. Journal of Applied Physics, 49, 1264–1280. Von Hippel, A., 1954. Dielectrics and Waves. Boston: MIT Press. Wang, J. R., 1980. The dielectric properties of soil-water mixtures at microwave frequencies. Radio Science, 15, 977–985. Wang, J. R., and Schmugge, T. J., 1980. An empirical model for the complex dielectric permittivity of soils as a function of water content. IEEE Transactions on Geoscience and Remote Sensing, 18, 288–295.

MICROWAVE HORN ANTENNAS

MICROWAVE HORN ANTENNAS Yahya Rahmat-Samii Department of Electrical Engineering, University of California at Los Angeles, Los Angeles, CA, USA

Definition Horn Antenna. An antenna with metallic flare walls which provides a transition between waves propagating in a waveguide and electromagnetic waves directivity radiated into space. Microwave horn antennas Horns are one of the simplest and most widely used microwave antennas. Interests in horn antennas date back to the turn of the nineteenth century and then considerably revived during World War II. Microwave horn antennas are essentially a device to make a transition from waves propagating in a waveguide into electromagnetic signals transmitting in another open medium such as free space. They can have metallic walls, dielectric material, or the combination of both, and occur in various shapes and sizes to fulfill many practical applications, such as communication systems, remote sensing, radio frequency heating, reference sources for other antenna testing and evaluation. Other than being a stand-alone directive power transmitter/receiver, horns are also used as feeds for other antennas such as reflectors, lenses, and compound antennas. Their widespread application is due to their simple, solid geometry and excellent performance when beam directivity is required (Love, 1976; Balanis, 1988, 1996; Olver et al., 1994; Bird and Love, 2007). Figure 1 summarizes the primary categories of horns, and Figure 2 shows some representative examples of well-known horn designs. Rectangular pyramidal horns

375

are probably the most commonly used class of horns. One of the primary applications of a pyramidal horn is the standard gain antenna since its gain may be calculated very accurately by knowing their dimensions. The beamwidths in the two principal planes can also be independently controlled by varying the rectangularaperture dimensions. Some other forms of rectangular horns with nonlinear profiles are also used to increase aperture efficiency and lower side lobes compared with a linear flare. Circular geometries are also in widespread use. The axial symmetry of conical horns allows them to generate any polarization of the dominant mode which makes them well suited for circular polarization. The beamwidths, however, are usually unequal in the two principal planes. In order to overcome this problem, the dual-mode and corrugated (hybrid-mode) conical horns have been developed. Higher-order modes are excited so that the aperture field of the horn is modified in such a way as to produce radiation patterns with axial symmetry and very low cross-polarization.

Horn specifications Ample types of horns can provide various radiation performances, and for every specific application, it is necessary for an antenna designer to consider some critical specifications which dictate the choosing of an appropriate type of horn antennas. Bandwidth The frequency band over which the system is to operate is usually specified as a range of frequencies (for single-band applications) where the antenna must satisfy a required return loss (or VSWR) and provide a radiation pattern adequate for the application. Consider an antenna required to operate from fmin to fmax. The bandwidth, in general, can be calculated as

Horn Antennas Rectangular

Circular

Other Types

Smooth wall

Smooth-wall conical

Diagonal

Corrugated

Corrugated

Elliptical

Dielectric-lined

Scalar

Dielectric

Profiled

Dielectric-lined

Multimode

Coaxial

Ridged

Profiled Longitudinal corrugation Multimode Ridged

Microwave Horn Antennas, Figure 1 A summary of primary types of horn antennas.

376

MICROWAVE HORN ANTENNAS

Microwave Horn Antennas, Figure 2 Representative types of horn antennas: (a) pyramidal horn, (b) circular horn, and (c) corrugated horn.

Bandwidth ð%Þ ¼

2ðfmax  fmin Þ  100: ðfmax þ fmin Þ

z

Based on this definition, a bandwidth below 20 % is generally considered narrowband whereas a bandwidth over 20 % is considered wideband. Often, there is a specific frequency determined in specifications as the design frequency at which the design provides optimum performances. For single-band applications, affi practical choice pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi is a design frequency of fc ¼ fmin fmax for narrowband operations and fc ¼ 1.2 fmin for wideband operations.

Gain and aperture efficiency The antenna gain takes into account both the directional capability and efficiency of the antenna. Among many different types of gains, two definitions are usually used to describe horn antennas (Balanis, 1996). The most commonly accepted definition of gain only considers the accepted power and removes the power reflected from the input. In this definition, an ideal horn without losses is assumed and the gain equals directivity as the efficiency is 100 %. Another definition, called absolute gain, retains the reflected power from the input, as this is an intrinsic property of the antenna. For horn antennas, the results of these two definitions are usually very similar since the reflection coefficient is small. The difference is a factor of 1/(1  |Gin|2), where Gin is the reflection coefficient at the input. Another useful measure for comparing the performances of horn antennas is the aperture efficiency. It is defined as the ratio of the antenna gain in the direction (y, f) and the gain of the same antenna aperture with ideal uniform illumination given by aperture ¼

l2 Gðy; jÞ; 4pA

where A is the total area of the aperture, l is the wavelength at the operating frequency, and G(y, f) is the gain function. In many applications it is assumed that the aperture efficiency is in the boresight direction of the antenna.

2a y

2b

x

Microwave Horn Antennas, Figure 3 A TE 10-mode openended rectangular waveguide feed with dimensions 2a and 2b.

It has to be noted that what is usually defined is the gain of the whole antenna system, not just the gain of the aperture antenna in isolation. Therefore, all other losses such as mismatch and ohmic losses have to be taken into consideration. A well-designed aperture antenna fed by a poorly designed feed system results in poor performance.

Polarization The radiated far-field of a horn antenna can be decomposed into two components: The one in the direction of the desired polarization is called co-polarization, and the orthogonal component is called cross-polarization. Polarization performance measurement can determine, in general, how much power is radiated in the undesired polarization which cannot be received by receiving antennas. The most common measurement is the peak cross-polarization level for linear polarization and the axial ratio for circular polarization. The measurement of the cross-polarization can be achieved in different ways. For linearly polarized horns, the co-polar pattern is measured by aligning the test antenna with the

120

60

100

50 FIRST SIDE LOBE BEAMWIDTH

80

40 FIRST-NULL BEAMWIDTH HALF-POWER BEAMWIDTH

60

30

40

20 FIRST SIDE LOBE MAXIMUM MAGNITUDE

20

10

0 0

0.5

1.0

1.5

2.0

2.5

b

0 3.0

160

80

140

70

FIRST SIDE LOBE BEAMWIDTH 60

120 BEAMWIDTH-DEGREES

70 RELATIVE MAXIMUM MAGNITUDE OF FIRST SIDE LOBE-DECIBELS DOWN

140

BEAMWIDTH-DEGREES

a

377

FIRST-NULL BEAMWIDTH 100

50

FIRST SIDE LOBE MAXIMUM MAGNITUDE

80

40

30

60 HALF-POWER BEAMWIDTH

20

40

10 3.0

20 0

0.5

APERTURE DIMENSION(b)-WAVELENGTHS

1.0

1.5

2.0

RELATIVE MAXIMUM MAGNITUDE OF FIRST SIDE LOBE-DECIBELS DOWN

MICROWAVE HORN ANTENNAS

2.5

APERTURE DIMENSION(a)-WAVELENGTHS

Microwave Horn Antennas, Figure 4 Different beamwidths and first side lobe levels for TE 10-mode open-ended rectangular waveguide on ground plane: (a) E-plane and (b) H-plane.

a

b

c 2b1

2a

2b1

2b 2a1

2a1

Microwave Horn Antennas, Figure 5 Typical rectangular horn antennas: (a) E-plane sectoral horn, (b) H-plane sectoral horn, and (c) pyramidal horn.

polarization of the distant source antenna, and the cross-polar pattern is obtained by rotating the source antenna by 90
and repeating the radiation pattern measurement.

Functioning environment Consideration of the working environment in which the antenna is likely to operate is important in design process as well. In many cases this too will dictate the choice of materials and the mechanical aspects of the system. One consideration is the potential effects of surrounding objects on the horn’s performance. A horn antenna may work perfectly well in isolation but give poor performance when surrounded by nearby scatterers. A typical effect is

caused by conductive surrounding objects which are fitted afterward in forms of blockage or radome effect. Other considerations are any special needs for horn antennas such as pressurization, high-power handling, low passive intermodulation level, robustness, and low-loss requirements.

Conventional horn antennas Rectangular horn Rectangular horns are constructed by gradually flaring a rectangular waveguide to provide a smooth transition from the input to free space. If the flare angle is small enough so that the higher-order modes can be

378

MICROWAVE HORN ANTENNAS

30

b/λ

E /λ

a/λ

GAIN-DECIBELS

H /λ

20

H

E

10 b a 0 0

10

20

30

a/λ, b/λ,

40

50

E /λ, H /λ

Microwave Horn Antennas, Figure 6 Gain characteristics of pyramidal horns of optimum-gain design.

ABSOLUTE VALUE OF RELATIVE FIELD STRENGTH

1.0

0.8

q a

0.6 /λ−(a/λ)2/2 H -PLANE 0.4 E -PLANE

0.2

0

0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

(a/λ) sin q

Microwave Horn Antennas, Figure 7 Universal patterns of pyramidal horns of optimum-gain design.

neglected, the transverse electric field of the dominant mode of the waveguide is a good first approximation to the aperture field at the horn mouth. For reference, some important characteristics of the open-ended rectangular waveguide (Figure 3) are shown in Figure 4.

Depending on to what plane its opening tapers rectangular horns may be classified as E-plane sectoral, H-plane sectoral, or pyramidal horns (Figure 5). The type, direction, and amount of taper can have a significant effect on the overall performance of the horn. The radiation characteristics can be determined using the aperture field

MICROWAVE HORN ANTENNAS

379

100

50

/λ)

20

TH

L/λ ≅ 0.3(D/λ)2 GdB ≅ 20 log [xD/λ] −2.82

10

(L

NG

E LL

IA

AX

D/λ)

ER (

ET IAM

D

D/λ, L/λ

5

2 CIRCULAR WAVEGUIDE

1 d

0.5

D

L

0.2 0.1

8

10

12

14

16

18 20 22 GAIN (G)-DECIBELS

24

26

28

30

Microwave Horn Antennas, Figure 8 Gain characteristics of conical horns of optimum-gain design.

b

0

FREQUENCY –9600 MHz FLARE HALFANGLE–6.25⬚

−10

CIRCULAR GUIDE

MODE MATCHING SUPPRESSOR IRIS

HORN FLARE

PHASING SECTION

HORN APERTURE

TAPER

2a0

2a1

2a

g

CALCULATED

−20 Eq E -DECIBELS

a

−30

CROSS POLARIZATION MEASURED

−40

−50

TE11 TE11

TE11 TE11,TM11

−60 TE11,TM11

90

60

30

0

30

q -DEGREES

Microwave Horn Antennas, Figure 9 Dual-mode conical Potter horn: (a) horn structure and (b) E-plane pattern.

60

90

380

MICROWAVE HORN ANTENNAS

L

r rap=a r(z) rth

l

HP

Microwave Horn Antennas, Figure 10 Profiled conical corrugated horn.

method or GTD construction (Jull, 1973; Huang et al., 1983). The latter is also capable of predicting the near field and back-radiated field. Considerable amounts of design data are available on rectangular horns, and the reader is referred to the just-mentioned references (Balanis, 1988, 1996; Bird and Love, 2007). For pyramidal horns, the geometrical parameters may be chosen to achieve the so-called optimum-gain design (Balanis, 1988). Figure 6 shows the relationship between the parameters to design an optimum-gain horn for a specified gain (the overall efficiency of these horns is typically about 50 %). Figure 7 is the plot of the E- and H-plane patterns of such horns.

Conical horns The geometry of a circular horn is shown in Figure 8. In contrast to pyramidal horns which are typically fed by rectangular waveguides, the circular horn is usually fed by circular waveguides. The aperture field of these horns can be constructed in a fashion similar to that of pyramidal horns by simply multiplying the aperture field of the circular waveguide by a quadratic phase term (Balanis, 1988). The resulting integral for the computation of the radiated field can then be evaluated numerically. Figure 8 gives the proper horn dimensions for constructing an optimum-gain horn for a specified gain. These horns typically possess more symmetric E- and H-plane patterns than do their pyramidal horn counterparts. They can also be used more effectively to create circularly polarized fields. Multimode, corrugated, matched horns, and others The requirements for designing more advanced remote sensing systems have resulted in the generation of a new class of feeds. These feeds in particular are used in large reflectors for radio astronomy and remote sensing reflectors. Among them one may refer to multimode (Potter, 1963; Thomas, 1970;

Bhattacharyya and Goyette, 2004), corrugated (hybrid mode) (Clarricoats and Olver, 1984; James, 1984; Thomas et al., 1986; Olver et al., 1994), and matched horns (Rudge and Adatia, 1975; Bahadori and Rahmat-Samii, 2006). Multimode horns were invented to equalize the pattern asymmetry of single-mode horns. For instance, in the Potter horn (Potter, 1963) the TM11 mode is generated along with the dominant TE11 mode of a circular horn. Although this new TM11 mode does not have any appreciable effect on the H-plane radiation pattern, when it is properly phased and combined with the TE11 mode, it can effectively alter the E-plane aperture distribution, which results in a symmetric radiation pattern. All these favorable features, however, cannot be properly realized until the feed aperture diameter exceeds about 1.3 l. For example, Figure 9 shows the radiation pattern of a dual-mode Potter horn as reported in Potter (1963). A partial conversion of the TE11 mode energy to a TM11 mode happens in the flared section of the horn, while the straight section enforces the condition that both modes have the proper phase relation at the aperture which can be maintained over a bandwidth of less than 10 %. There are also available other types of multimode horns, which result from a combination of modes such as TE10, TE12, and TM12 in a square-aperture pyramidal horn. Corrugated horns have become the main choice of antenna feeds in recent advanced communications, radar, and remote sensing applications where demanding performance is required. First introduced in the 1960s (Minnett and Thomas, 1966; Simmons and Kay, 1966), they can provide superior radiation performances such as symmetric patterns and low cross-polarization levels compared with conical horns. Corrugated horns are capable of creating similar boundary conditions at all polarizations which result in similar tapers in the aperture field distribution in all planes. Due to these boundary conditions, symmetric radiation

MICROWAVE HORN ANTENNAS

381

100

qO

90

Da/λ

CORRUGATED HORN HALF FLARE ANGLE

80

qO−70⬚

20 dB HALF-BEAMWIDTH-DEGREES

70

65⬚ 60⬚

60

55⬚ 50⬚

50

45⬚ 40⬚

40

35⬚ 30⬚

30

25⬚

20

15⬚

20⬚

10⬚ 5⬚

10

0 1

2

3

4

5

6

7

8

9

10

NORMALIZED APERTURE DIAMETER (DO/λ)

Microwave Horn Antennas, Figure 11 20 dB half beamwidths of corrugated horns.

patterns can be obtained at levels as low as 25 dB in both the E- and H-planes. A corrugated horn can be realized by grooving the E-plane of a pyramidal horn or the entire wall of a circular horn with, typically, six or more slots (corrugations) per wavelength (Figure 10). For circular corrugated horns, Figure 11 shows the plots of the pattern widths at 20 dB level as a function of the opening angle. The existence of the corrugations, especially near the waveguide-horn junction, affects the VSWR of the horn. The usual practice is to begin the corrugations at a small distance from the junction. These horns are also classified as hybrid-mode horns because they support modes in which both longitudinal E- and H-field components are present. In a circular corrugated horn, a natural mixture of TE11 and TM11 results in the generation of a hybrid-mode HE11. In contrast to dual-mode horns there is no need of a mode converter, and therefore, the

hybrid-mode horns have typically wider bandwidths. In particular, one version of these hybrid horns (scalar horns), which has a large flare angle with a relatively short horn, has radiation characteristics with little dependence on frequency. There now exists advanced computational and optimization methods (Sinton et al., 2002) to tailor corrugated horn antennas for various applications including profiled corrugated horns with shorter lengths than the standard straight corrugated horns. There are other horn feeds which are properly tailored to specifically overcome some undesirable characteristics of reflectors. For example, matched horn feeds (Rudge and Adatia, 1975; Bahadori and Rahmat-Samii, 2006) are used to significantly reduce the generation of unwanted cross-polarized fields in an offset parabolic reflector illuminated with a tilted horn feed. The basic concept behind these designs is to match the horn

382

MICROWAVE RADIOMETERS

aperture field distribution with the receiving focal plane distribution of the reflector. There also exists a class of dielectric horns with some unique properties, as described in Oh et al. (1970), Baldwin and McInnes (1973), and Jha et al. (1975), and other configurations (Love, 1962; Vokurka, 1979).

Summary Microwave horn antennas are widely used as standard gain antennas for antenna testing or evaluation or feeds for other antennas due to their simple and solid geometries and excellent performances. Among many different types of horns, rectangular pyramidal horns and conical horns, multimode and corrugated horns are the most popular and practical configurations, and a selection of a specific type is made based on the design specifications such as bandwidth, gain and aperture efficiency, polarization, and some other unique radiation characteristics. With the advances in computational and optimization methods, more sophisticated designs such as profiled corrugated horns and matched horns are developed for demanding applications in radars, satellite communications, remote sensing, deep-space telemetry, and radio astronomy. Bibliography Bahadori, K., and Rahmat-Samii, Y., 2006. “A Tri-Mode Horn Feed for Gravitationally Balanced Back-to-Back Reflector Antennas”, IEEE Antennas and Propagation Society Symposium, pp. 4397– 4400, Albuquerque, NM. Bahadori, K., and Rahmat-Samii, Y., 2009. “Tri-Mode Horn Feeds Revisited: Cross-pol Reduction in Compact Offset Reflector Antennas,” IEEE Trans. Antennas Propag, vol. 57, no. 9, pp. 2771–2775. Balanis, C. A., 1988. Horn antennas. In Lo, Y. T., and Lee, S. W. (eds.), Antenna Handbook. New York: Van Nostrand Reinhold. Balanis, C. A., 1996. Antenna Theory: Analysis and Design, 2nd edn. New York: Wiley. Baldwin, R., and McInnes, P. A., 1973. Radiation patterns of dielectric loaded rectangular horns. IEEE Transactions on Antennas and Propagation, 21, 375–376. Bhattacharyya, A., and Goyette, G., 2004. A novel horn radiator with high aperture efficiency and low cross-polarization and applications in arrays and multibeam reflector antennas. IEEE Transactions on Antennas and Propagation, 52, 2850–2859. Bird, T. S., and Love, A. W., 2007. Horn antennas. In Volakis, J. (ed.), Antenna Engineering Handbook, 4th edn. New York: McGraw-Hill. Clarricoats, P. J. B., and Olver, A. D., 1984. Corrugated Horns and Microwave Antennas. London: Peregrinus. Huang, J., Rahmat-Samii, Y., and Woo, K., 1983. A GTD study of pyramidal horns for offset reflector antenna applications. IEEE Transactions on Antennas and Propagation, 31, 305–309. James, G. L., 1984. Design of wide-band compact corrugated horns. IEEE Transactions on Antennas and Propagation, 32, 1134–1138. Jha, R. K., Misra, D. K., and Jha, L., 1975. Comparative study of dielectric and metallic horn antennas. In IEEE Antennas and Propagation Society International Symposium, pp. 16–18. Jull, E. V., 1973. Errors in the predicted gain of pyramidal horns. IEEE Transactions on Antennas and Propagation, 21, 25–31.

Love, A. W., 1962. The diagonal horn antenna. Microwave Journal, 5, 117–122. Love, A. W. (ed.), 1976. Electromagnetic Horn Antennas. New York: IEEE Press. Minnett, H. C., and Thomas, B. M., 1966. A method of synthesizing radiation patterns with axial symmetry. IEEE Transactions on Antennas and Propagation, 14, 654–656. Oh, L. L., Peng, S. Y., and Lunden, C. D., 1970. Effects of dielectrics on the radiation patterns of an electromagnetic horn. IEEE Transactions on Antennas and Propagation, 18, 553–556. Olver, A. D., Clarricoats, P. J. B., Kishk, A. A., and Shafai, L., 1994. Microwave Horns and Feeds. New York: IEEE Press. Potter, P. D., 1963. A new horn antenna with suppressed sidelobes and equal beamwidths. Microwave Journal, 6, 71–78. Rudge, A. W., and Adatia, N. A., 1975. New class of primary-feed antennas for use with offset-parabolic-reflector antennas. Electronics Letters, 11, 597–599. Simmons, A. J., and Kay, A. F., 1966. The scalar feed – A high performance feed for large paraboloid reflectors. In IEE Conference Publications, Vol. 21, pp. 213–217. Sinton, S., Robinson, J., and Rahmat-Samii, Y., 2002. Standard and micro genetic algorithm optimization of profiled corrugated horn antennas. Microwave and Optical Technology Letters, 35, 449–453. Thomas, B. M. A., 1970. Prime-focus one- and two-hybrid-mode feeds. Electronics Letters, 6, 460–461. Thomas, B. M. A., James, G. L., and Greene, K. J., 1986. Design of wide-band corrugated conical horns for Cassegrain antennas. IEEE Transactions on Antennas and Propagation, 34, 750–757. Vokurka, V. J., 1979. Elliptical corrugated horn for broadcasting satellite antennas. Electronics Letters, 15, 652–654.

Cross-references Electromagnetic Theory and Wave Propagation Emerging Technologies, Radar Microwave Radiometers Microwave Radiometers, Conventional Radars

MICROWAVE RADIOMETERS Niels Skou National Space Institute, Technical University of Denmark, Lyngby, Denmark

Synonyms Passive microwave radiometer (PMR) Definition The microwave radiometer is a calibrated receiver that measures properties of the natural emission from the environment as picked up by an antenna system. Introduction: what is radiometry about? All bodies at a temperature above the absolute zero (0 K ¼ 273
C) radiate power, according to Planck’s law. At microwave frequencies the Rayleigh-Jeans approximation holds, and the radiation is proportional to physical temperature. Actually, this is only true for the

MICROWAVE RADIOMETERS

so-called blackbodies, which are perfect emitters. Natural bodies radiate less, and we introduce the term emissivity (e) which is a number between 0 and 1 describing how well the body radiates relative to a blackbody. Within radiometry the radiated power is expressed as the so-called brightness temperature, TB, so that TB ¼ e · Tphys. The brightness temperature of a blackbody is thus equal to its physical temperature, while all natural bodies will have brightness temperatures lower than that. The brightness temperature of a natural body can also be understood as that physical temperature a blackbody would have to have in order to emit the power in question. Typical values related to planet Earth range from almost 0 K (looking up toward free space) to about 300 K. The radiated energy can be picked up by an antenna in order to be measured by a radiometer. Unfortunately antennas are not perfect: a perfect antenna would have a sharply defined main lobe so that when this is pointed toward the target, only this contributes to the received power. But the main lobe is not with sharp cutoff, so some power is received from the surroundings of the target. Furthermore, antennas have side lobes far from the main lobe meaning that some power is also received from other directions. The process can be expressed by the following equation: Z Z 1

TB ðy; fÞ Gðy; fÞdO: (1) TA ¼ 4p 4p

This convolution integral, where G(y, f) is the antenna radiation pattern, shows how the received power, denoted as the so-called antenna temperature TA, is a gain-weighted summation of the brightness temperatures TB(y, f) from each direction (see Ulaby et al., 1981).

Sensitivity As stated above, the task of the radiometer is to measure the power picked up by the antenna. So in fact the radiometer is basically a calibrated microwave receiver. The basic and simplest radiometer, the total power radiometer, is illustrated in Figure 1a. The received power is amplified with a gain G, a certain bandwidth B around the center frequency is selected by the filter, the microwave signal is detected by a square law detector, and since we are handling noise-like signals, a certain integration time t is required. Finally, it is indicated that we cannot make a receiver without introducing an additional noise TN. The output will represent a power measure expressed as P ¼ k · B · G · (TA + TN) where k is Boltzmann’s constant: 1.38 · 1023 J/K. As already stated the signals are noiselike, so the output fluctuates, but the fluctuations are smoothed by the integration. The level of fluctuations, the standard deviation of the output to be specific, is expressed as the radiometer sensitivity, DT, and it is calculated from Eq. 2:

383

TA þ TN DT ¼ pffiffiffiffiffiffiffi : B:t

(2)

Stability and accuracy A fundamental problem with this simple and direct implementation of the radiometer is that the output, P ¼ k · B · G · (TA + TN), is dependent on stability of the gain G and the receiver noise temperature TN. Hence, stability may be problematic, and frequent calibration is required, or a more developed radiometer principle must be employed (Skou and LeVine cpoo). If k, B, G, and TN are not only constants but known constants, we have no absolute accuracy problems: a given TA results in a given P that can be calculated. Such knowledge of the constants is rarely available, leading to the necessity for calibration. This illustrates the fundamental difference between stability and accuracy: stability is a highly appreciated virtue of an instrument, but a stable instrument need not be accurate. The steps toward accuracy include the calibration process. In the following we will describe a slightly different aspect of absolute accuracy, which stresses the care that must be exercised when designing or working with radiometers. Consider losses in a signal path – it could be the waveguide connecting the antenna with the radiometer input or a passive component in the radiometer front end. Let ‘ denote the fractional loss (or the absorption coefficient) and T0 the physical temperature. T1 is the input temperature and the output temperature is then given by T2 ¼ T1 (1  ‘) + ‘ · T0. The difference between output and input is TD ¼ T2 T1 ¼ ‘ (T0  T1). If T1 is 100 K and T0 is 300 K, a loss as small as 0.01 dB (‘ ¼ 0.0023) results in a difference, TD, of 0.5 K. Bearing in mind that the losses of a real signal path are much greater than 0.01 dB, the physical temperature of the path must be measured and used for correction of the measured brightness temperature. The corresponding losses must be known to an accuracy of better than 0.01 dB and must remain stable within the same limits. Consider a mismatch, for example, at the input of a radiometer; with a reflection coefficient r, we similarly find TD ¼ r (TRAD  T1) where TRAD is the microwave temperature as seen from the point of reflection into the radiometer. TRAD is typically 300 K and if T1 again is assumed to be 100 K, a reflection coefficient of 26 dB will give an error (TD) of 0.5 K. Care must be exercised to obtain reflection coefficients better than 26 dB. Radiometer types The conventional radiometers are illustrated to the left in Figure 1, while the more advanced types are illustrated by the correlation radiometer to the right. The total power radiometer has already been discussed. It is the simplest and basic radiometer that has the optimum sensitivity but may have problems with stability.

384

MICROWAVE RADIOMETERS

a

c G

x2

B

TA

τ

∫

G

TA1

x2

B

τ

∫

VOUT1

VOUT TN1

TN

Total power radiometer

b

FS G

TA

B

x2

τ ±1

TR

TN

∫

Dicke radiometer

real

complex corr. G

TA2

imag x2

B

τ

∫

VOUT

TN2

VOUT2

Correlation radiometer

Microwave Radiometers, Figure 1 Two conventional radiometer types: (a) total power and (b) Dicke and (c) the correlation radiometer.

The classical way of alleviating stability problems is to employ switching as done in the Dicke radiometer (Figure 1b). The idea here is that the radiometer in fact measures the difference between the unknown antenna temperature and an internal, known reference. The Dicke radiometer has proven very good behavior and has been widely used. An enhanced version of the Dicke radiometer is the noise injection radiometer (NIR) in which a nullbalancing technique is employed resulting in total independence of gains and receiver noise temperatures. A price has to be paid for the enhanced stability of Dicke type switching radiometers: the sensitivity is degraded by a factor of 2 compared with that of the total power radiometer. More advanced radiometers are very often based on the correlation radiometer shown in Figure 1c. It basically consists of two total power radiometers, but in addition to the usual detected outputs, the signals before detection in the two channels are cross correlated to produce two additional outputs: the real and imaginary part of the correlation. The correlation radiometer is the building block of interferometric, synthetic aperture radiometer systems and often also of polarimetric radiometer systems.

Radiometer antennas An important part of any radiometer system is its antenna. The purpose of the antenna is to collect the emitted energy from a target and present it to the radiometer input. Thus, the antenna radiation pattern determines the spatial resolution of the radiometer system. Important antenna parameters are main beam efficiency and side lobe level. A good beam efficiency and low side lobes ensure that the radiometer system primarily measures what the antenna points toward and not so much the surroundings. Of paramount importance for a radiometer antenna is low loss following the previous accuracy discussion.

Hence, the most popular antenna types are the microwave horn and reflector antennas. Horns are bulky, low-gain devices and as such normally not used as primary spaceborne antennas, but they play an important role in ground-based and airborne systems and as feeds for reflector antennas. Especially Potter horns and dual-mode horns are favored due to excellent radiation patterns. Reflector antennas – typically based on offset parabolic reflectors – have been used extensively in spaceborne systems. They feature low loss and good radiation patterns, and, very importantly, they can serve a wide range of frequencies (e.g., 5–90 GHz) through the use of clusters of feed horns. For the antennas discussed here, where the electrical aperture is roughly equal to the physical aperture, it is very simple to estimate the width of the main beam. The 3 dB beamwidth is slightly larger than the reciprocal of the aperture measured in wavelengths of the operating frequency: y ¼ 1:4 :

l : D

(3)

The factor 1.4 reflects the need for high beam efficiency. A 1 m reflector antenna used at 30 GHz (1 cm wavelength) will thus have a 0.014 (¼ 0.8
) beamwidth. Looking straight down from a satellite in an 800 km orbit, the footprint on Earth will be 11.2 km. Spaceborne radiometer systems are not high-resolution devices like optical scanners or synthetic aperture radars, but they do supply unique data for geophysical interpretation with frequent coverage of the globe.

Calibration The purpose of calibration is to establish the relation between the input brightness temperature and the

MICROWAVE RADIOMETERS

a

385

b RADIOMETER TO

TB

ANTENNA HORN

TA RADIOMETER

AL FOIL

ANTENNA FOAM INSULATION MICROWAVE ABSORBER MICROWAVE ABSORBER LIQUID NITROGEN

Microwave Radiometers, Figure 2 Antenna target calibrator, principle, and practical layout.

radiometer output – usually digital numbers following analog to digital conversion of the radiometer output voltage. Assuming that the radiometer transfer is linear, two calibration points are required, typically one hot around 300 K and one cold in the 0–77 K range. The most obvious solution is a cooled microwave termination. A well-matched load will generate a noise temperature equal to its physical temperature – and the physical temperature is easily measured. If the load is left at room temperature, it provides a hot calibration point, and if it is submerged into liquid nitrogen, it will be a cold load at around 77 K. However, the latter is not quite as easy as it sounds: heat flow into the load must be prevented, and the loss in the input transmission line must be corrected for. An alternative solution is a cooled target viewed by a suitable antenna connected to the radiometer (see Figure 2). A microwave absorber (normally used to cover the inside walls of radio anechoic chambers) will emit a brightness temperature equal to its physical temperature T0. Under ideal conditions, the antenna will sense nothing but the brightness temperature from the absorber and TA ¼ TB ¼ T0. Figure 2 also shows a practical layout of this concept. The radiometer is connected to an antenna horn through a short waveguide (very low losses!). The horn views a microwave absorber soaked with liquid nitrogen. The absorber and the liquid nitrogen are contained in an insulated metal bucket, and the excess opening of the bucket is covered by aluminum foil. In this way the antenna is only able to pick up energy from the absorber, which is cooled to 77 K by the nitrogen. There is generally no problem with losses and heat flow in the antenna and the waveguide, liquid nitrogen is readily available, and the setup is cheap and simple. Overall this is a very useful radiometer calibrator – and before liquid nitrogen is applied it provides also the hot calibration point. Finally, it shall be noted that pointing an antenna toward free space will provide a cold calibration point of only a few K. The problem is that the atmosphere will give

a significant contribution to the temperature, and it must be carefully accounted for. However, this is not the case for satellite-borne systems, and here such a sky view is often used for the cold calibration point. Going to the more advanced correlation radiometers, also shown in Figure 1, calibration is somewhat more complicated. First, a basic calibration of the two channels is carried out as just described, but in addition to this, the calibration of the correlation channels requires generation of a pair of known signals with a known amount of correlation between them. This issue is outside the scope of the present text.

Conclusion Microwave radiometers are sensitive receivers requiring special attention to stability and accuracy. Several types have been developed over the years each with its special virtue. The antenna is an important part of the radiometer system determining what is actually measured by the radiometer. Any radiometer must be carefully calibrated – not only after its development but also a frequent check of calibration must be carried out at proper intervals when the radiometer is in use. More information can be found in (Skou and LeVine, 2006). Bibliography Skou, N., and LeVine, D., 2006. Microwave Radiometer Systems, Design and Analysis. Boston: Artech House. Ulaby, F. T., Moore, R. K., and Fung, A. K., 1981. Microwave Remote Sensing. Norwood: Artech House, Vol. 1.

Cross-references Calibration, Microwave Radiometers Microwave Horn Antennas Microwave Radiometers Microwave Radiometers, Conventional Microwave Radiometers, Correlation Microwave Radiometers, Interferometers Microwave Radiometers, Polarimeters

386

MICROWAVE RADIOMETERS, CONVENTIONAL

MICROWAVE RADIOMETERS, CONVENTIONAL Niels Skou National Space Institute, Technical University of Denmark, Lyngby, Denmark

Synonyms Passive microwave radiometer (PMR) Definition The conventional microwave radiometer is a calibrated receiver that measures the power from the environment as picked up by an antenna system. Introduction The emission from the environment is received by the antenna. This can typically be horizontally or vertically polarized. The conventional radiometer measures the power corresponding to one of these polarizations, and it is basically a single channel receiver. If both polarizations must be measured, two independent radiometer receivers are employed. In-depth treatment of conventional radiometers is found in Ulaby et al. (1981) and in Skou and LeVine (2006). Total power radiometer The simplest and most direct implementation of a radiometer is found in the total power radiometer (TPR) as illustrated in Figure 1a. The gain in the radiometer has been symbolized by an amplifier with a gain G and the frequency selectivity by a filter with a bandwidth B (centered around some given frequency). We cannot make a receiver without introducing an additional noise TN. The microwave power has to be detected to find some measure of its mean. In the present case, it is very attractive to use a square-law detector: then the output voltage will be proportional to the input power and hence the input temperature. Finally, we indicate where the integration takes place: the signal from the detector is smoothed by the integrator to reduce fluctuations in the output, and the longer is the integration time, the more smoothing. The output can be expressed as (1) VOUT ¼ c ðT A þ T N Þ G where c is a constant. VOUT is totally dependent on TN and G. These can for some applications not be regarded as stable enough to satisfy reasonable requirements to stability and accuracy. In other cases, however, the total power radiometer is very useful, namely, where frequent calibration, for example, once every few seconds, is possible. The sensitivity (standard deviation of the output signal) of the total power radiometer is given by TA þ TN (2) DTTPR ¼ pffiffiffiffiffiffiffiffiffi B t

Dicke radiometer In 1946, R. H. Dicke found a way of alleviating the stability problems in radiometers (Dicke, 1946). By using the radiometer not to measure directly the antenna temperature TA but rather the difference between this and some known reference temperature TR, the sensitivity of the measurement to gain and noise temperature instabilities is greatly reduced (see Figure 1b). The input of the radiometer is rapidly switched between the antenna temperature and the reference temperature. The switch frequency FS could be 1,000 Hz. The output of the square-law detector is multiplied by +1 or 1, depending on the position of the Dicke switch, before integration. The input to the integrator is then V1 ¼ c (TA + TN) G in one half period of FS and V2 ¼ c (TR + TN)

G in the second half period. Provided that the switch frequency FS is so rapid that TA, TN, and G can be regarded as constants over the period and that the period is much shorter than the integration time, the output of the radiometer is found as VOUT ¼ c ðT A  T R Þ G

(3)

It is seen that TN has been eliminated, while G is still present, although with less weight. Now G multiplies the difference between TA and TR, where TR is reasonably chosen to be in the same range as TA, while in the total power case, G multiplied the sum of TA and the rather large TN. The Dicke principle has proven to be very useful, and Dicke radiometers (DR) have been used extensively over the years. A price has to be paid, however, for the better immunity to instabilities. Since only half of the measurement time is spent on the antenna signal (the other half is spent on the reference temperature), the sensitivity is degraded by a factor of 2 compared with the total power radiometer: TA þ TN DTDR ¼ 2 pffiffiffiffiffiffiffiffiffi B t

(4)

Noise injection radiometer The noise injection radiometer (NIR) represents the final step toward stability; that is, the output is independent of gain and noise temperature fluctuations (Goggins, 1967; Hardy et al., 1974). From Equation 3 it is seen that the output from a Dicke radiometer is zero (independent of G and TN) if the reference temperature and the antenna temperature are equal. The noise injection radiometer is a specialization of a Dicke radiometer in which this condition is continuously fulfilled by a servo loop. In almost any case encountered in Earth remote sensing, the antenna temperature is below some 300 K (emissivities between 0 and 1 are multiplied by the physical temperature). The reference temperature in a Dicke radiometer is conveniently equal to the physical

MICROWAVE RADIOMETERS, CONVENTIONAL

a

387

b G

B

x2

TA

TA

τ

Ú TN

TA′

τ

x2

TN

Ú

VOUT

Dicke radiometer

d

Dicke radiometer

TI

B

±1 TR

TA′ = TA + TI

TA

G

VOUT

Total power radiometer

c

FS

VOUT ≈ 0

TA

G

B

x2

LPF A/D

Loop gain

Noise injection radiometer

TR

TN

Processor

OUT

Hybrid radiometer

Microwave Radiometers, Conventional, Figure 1 (a) Total power radiometer, (b) Dicke radiometer, (c) noise injection radiometer, and (d) hybrid radiometer.

temperature in the microwave front end, that is, 300–320 K. In Figure 1c we show how the output TI of a variable noise generator is added to the antenna signal TA, so that the resultant input (TA0 ) to the Dicke radiometer is equal to the reference temperature (TR) and a zero output results. A servo loop adjusts TI to maintain the zero output condition or rather the near-zero output condition: the loop gain can be made large but not infinite. From Equation 3 we have VOUT ¼ c ðTA 0  T R Þ G ¼ 0 and as TA0 ¼ TA + TI, we find TA ¼ TR  TI. TR is a known constant, and knowledge of TI is required to find TA. The accuracies of the Dicke radiometer part of the NIR and of the loop gain are (given large loop gain) completely insignificant for the accuracy with which we determine TA. This is solely dependent on the accuracy of TI. Accurate and stable noise sources with variable output can be made and are used for “injecting” the required signal TI into the input line, so that TI and TA are added. The sensitivity of the noise injection radiometer is very close to that of the Dicke radiometer: TR þ TN DTNIR ¼ 2 pffiffiffiffiffiffiffiffiffi B t

(5)

Hybrid radiometer The radiometers as described hitherto in this entry are the classical receiver types, and their implementation is indicated in the classical way using, for example, analog integration after detection and analog subtraction of antenna and reference signals (in the Dicke radiometer). And indeed many radiometers are still implemented this way. But with the advent of analog to digital converters and digital processing, other implementation forms are possible and often used. This is illustrated in Figure 1d.

As soon as possible following detection, the signal is A to D converted – only a low-pass filter is indicated to condition the signal bandwidth to the sampling frequency of the converter. The signal from the converter is led to some kind of digital processor, typically a PC or an FPGA (field-programmable gate array), where suitable data handling takes place. This can typically be digital integration to the required integration time t, as well as subtraction of the antenna signal and the reference signal. Since these processes are under computer control, flexibility becomes a key word, and the distinction between total power and Dicke radiometer vanishes to some extent: if the processor operates the input switch rapidly and regularly, we can regard it as a Dicke case, while if the measurement situation is such that the antenna signal can be measured (with interruptions when the switch points to the reference signal) without loss of data, then we have a total power case with frequent calibration. A classical Dicke radiometer spends half of its time measuring the well-known reference temperature and thus only half of its time doing its real job, namely, measure the unknown antenna temperature. Thus over the years researchers have considered a better duty cycle for the antenna measurements (in turn potentially leading to improved sensitivity). However, having analog subtraction of the signals, the 50 % duty cycle is instrumental to having simple and stable circuitry. But with the subtraction done digitally this is no longer a limitation, and optimized duty cycles can be found on a case-by-case basis. A word of warning should, however, be stated: the naive notion that spending more time on the unknown antenna signal will lead to improved sensitivity is not necessarily true! When more time is spent on the antenna, less time is left for the reference and this in turn leads to an increased standard deviation for that measurement. Thus, the final standard deviation after subtraction might not decrease.

388

MICROWAVE RADIOMETERS, CONVENTIONAL

However, there are possible improvement schemes: since the reference temperature and the receiver noise temperature are assumed relatively stable while the antenna temperature may change rapidly, averaging over several reference temperature measurements will reduce the standard deviation of this measurement and thus allow a non-50 % duty cycle to be employed. Initial instrument fluctuations and averaging times must of course be considered carefully. See further discussion in Tanner et al. (2003) and Racette and Lang (2005).

Implementation The radiometer is merely a very sensitive microwave receiver, and like any receiver, it can be implemented as a direct receiver or, by use of a mixer and a local oscillator, as a superheterodyne receiver. In the direct receiver, all amplification takes place at the input RF frequency, and all selectivity is determined by filters in the same frequency range. In the superheterodyne receiver most of the amplification takes place at the much lower intermediate frequency (IF), and selectivity is determined by a combination of filters at RF and IF levels. Microwave FET amplifiers with excellent noise figures and square-law detectors covering the frequencies well beyond 40 GHz are available. Hence, the general trend is that conventional radiometers be implemented as direct receivers at frequencies below some 40 GHz, while higher-frequency implementations seem to favor the superheterodyne receiver – due to unavailability of amplifiers or very expensive amplifiers. Another very important issue when implementing microwave radiometers is the concept of thermal stabilization. The best and simplest way to achieve a stable

a

radiometer is to carefully temperature-stabilize all important components of the receiver – first and foremost the microwave front end.

Imaging considerations A microwave radiometer is often mounted on an aircraft or a satellite in order to cover large areas. If a radiometer with its antenna is mounted on a satellite and the antenna points straight down, the forward movement of the satellite will facilitate measurements on the ground along a straight line (the nadir path of the satellite). Coverage of the entire Earth by such “profiles” will require an enormous number of orbits! A dramatic increase in mapping efficiency results from scanning the antenna (see Figure 2a). The antenna rotates about a vertical axis, and the footprint will cover a wide swath on the Earth dependent on satellite altitude and incidence angle. Other scanning methods are possible, but the rotating scan is attractive due to constant incidence angle on the ground and the lack of accelerations associated with reciprocating scans. In a scanning system it is obvious that there is only a limited time for the radiometer to carry out its measurement before the footprint moves to another position within the swath. The so-called dwell time per footprint is short. The users in general want small footprints (or high spatial resolution, to put it differently). As technology evolves, high-resolution systems become possible, but a small footprint results in rapid rotation (mechanical problems) and in a very small dwell time per footprint, hence in a short integration time, which, through our radiometer sensitivity formula, directly translate into poor sensitivity! The solution to this fundamental problem is offered by the so-called push-broom concept illustrated in Figure 2b.

b ANTENNA SCAN SATELLITE VELOCITY VECTOR

SATELLITE VELOCITY VECTOR NADIR PATH

NADIR PATH

SWATH USEFUL SWATH

ANTENNA FOOT – PRINT

Microwave Radiometers, Conventional, Figure 2 (a) Conical scan and (b) push broom.

ANTENNA FOOT –PRINT

MICROWAVE RADIOMETERS, CORRELATION

In the push-broom radiometer system a multiple-beam antenna covers the swath while the satellite moves forward. A host of radiometer receivers are connected to an equal number of antenna feeds, producing individual beams to sense the Earth simultaneously. The obvious advantages of the push-broom system compared to a scanning system are no moving antenna to cause problems in the satellite design and much larger dwell time per footprint, hence better sensitivity. For the scanning radiometer systems, the requirement for receiver sensitivity is severe. At the same time, frequent calibration is easily achieved: once per scan, while the antenna is anyway looking away from the swath, the receiver is calibrated. Hence, the total power radiometer is an obvious candidate for such systems, due to its optimal sensitivity and since potential instabilities are taken care of by frequent calibration. For the push-broom system, requirements for receiver sensitivity are greatly relaxed due to the much larger dwell time per footprint as compared to the scanner situation. At the same time, frequent calibration is not attractive, as all receivers are always busy sensing the Earth. Hence, the push-broom situation favors a trading of sensitivity for stability, and in conclusion, a Dicke type of switching radiometer is preferred – maybe the NIR.

Conclusion Microwave radiometers are sensitive receivers requiring special attention to stability and accuracy. Several types have been developed over the years, and in general stability comes at a price: degraded sensitivity. So, when designing a satellite-borne imaging system, an important trade-off between sensitivity, stability, and imaging properties must be carried out. Bibliography Dicke, R. H., 1946. The measurement of thermal radiation at microwave frequencies. The Review of Scientific Instruments, 17, 268–279. Goggins, W. B., 1967. A microwave feedback radiometer. IEEEAES, 3, 83–90. Hardy, W. N., Gray, K. W., and Love, A. W., 1974. An S-band radiometer design with high absolute precision. IEEE-MTT, 22, 382–390. Racette, P. E., and Lang, R. H., 2005. Radiometer design analysis based upon measurement uncertainty. Radio Science, 40, 1–22. Skou, N., and LeVine, D., 2006. Microwave Radiometer Systems, Design and Analysis. Norwood: Artech House. Tanner, A. B., Wilson, W. J., and Pellerano, F. A., 2003. Development of a high-stability L-band radiometer for ocean salinity measurements. In IEEE, Proceedings of IGARSS’03, pp. 1238–1240. Ulaby, F. T., Moore, R. K., and Fung, A. K., 1981. Microwave Remote Sensing. Norwood: Artech House, Vol. 1.

Cross-references Calibration, Microwave Radiometers Microwave Horn Antennas Microwave Radiometers

389

Microwave Radiometers, Correlation Microwave Radiometers, Interferometers Microwave Radiometers, Polarimeters Reflector Antennas

MICROWAVE RADIOMETERS, CORRELATION Christopher Ruf Department of Atmospheric, Oceanic and Space Sciences, University of Michigan, Ann Arbor, MI, USA

Definition and overview A microwave radiometer measures statistical properties of the electrical field incident on its antenna. A total power radiometer, for example, measures the variance of the field strength, which is proportional to the brightness temperature of the source of the electric field. A correlation radiometer measures the statistical covariance between two incident electric fields. The most common types of correlation radiometers are autocorrelation spectrometers (Ruf and Swift, 1988), spatial interferometers (Kerr et al., 2000), and coherent detection polarimeters (Piepmeier and Gasiewski, 2001). Autocorrelation spectrometers measure the correlation between one electric field and a time-delayed version of itself. Spatial interferometers measure the correlation between the electric field at two locations. Polarimeters measure the correlation between two polarization components of an electric field. Since variance is defined as the statistical correlation of a signal with itself, a total power radiometer can be thought of as a special case of all three correlation radiometers – a spectrometer with zero time delay, an interferometer with no spatial separation, or a polarimeter with a common polarization component. Autocorrelation spectrometer The autocorrelation of a time-varying electric field is given by * *  t; rÞi RðtÞ ¼ hE ðt; rÞEðt *

(1)

where Eðt; rÞ is the electric field at time t and position r, * denotes complex conjugation, t is the time delay between the two versions of the electric field being correlated, and hi denotes a statistical expectation operator. In practice, the expectation operation is approximated by a time average because, in most cases, second-order statistics of the electric field associated with microwave thermal emission can be considered to be stationary over short time intervals. The autocorrelation is typically sampled over a suitable range of time delays, t. The Fourier transform of the autocorrelation with respect to the time delay is its power spectrum. For this reason, the autocorrelation of a radiometric signal is often measured in order to determine the spectral dependence of its brightness temperature. The Fourier transform that recovers the power *

390

MICROWAVE RADIOMETERS, INTERFEROMETERS

spectrum from the autocorrelation is performed in software as part of data post-processing. Autocorrelators are used to determine the Tb spectrum when the spectrum is to be resolved with sufficiently fine resolution that a bank of narrow band filters is impractical.

Spatial interferometer The correlation between the electric field at two positions is given by V ðsÞ ¼ hA1 fEðr; OÞg A2 fEðr  s; OÞgi *

*

*

*

*

(2) *

where Eðr; OÞ is the electric field at position r arriving from angular direction O, An{} denotes the angular reception sensitivity of antenna n (n ¼ 1, 2) to the incident electric field, and s* is the separation between the two versions of the electric field being correlated. This correlation statistic is referred to as the visibility of the brightness temperature distribution. The visibility is sampled over a suitable range of separations between antenna pairs. If the two antennas have overlapping angular reception sensitivities, then the Fourier transform of the visibility with respect to the separation, *s, is related to the angular dependence of the power density of the incident electric field received by the antennas. The power density is, in turn, proportional to the brightness temperature of the source of the electric field. Spatial interferometers (also called Fourier synthesis imagers) are used to determine the angular dependence of the brightness temperature, Tb (O). The Fourier transform required to convert measured visibilities to a Tb image is performed in software as part of the data postprocessing. Interferometers are used to image the Tb when its angular variation is to be resolved with sufficiently fine resolution that a large antenna, capable of comparable spatial resolution, is impractical.

Coherent detection polarimeter The correlation between two linearly polarized components of the electric field is given by   (3) Cpg ¼ Ep ðt; *rÞ Eq ðt; *rÞ where the subscripts p and q denote the two polarization components being correlated. A common polarization component was assumed in Equations 1 and 2 for the two versions of the electric field being correlated. If p and q are the same, then Cpq is proportional to that polarization component of the brightness temperature of the source of the electric field. If p and q are orthogonal, then the real and imaginary components of Cpq are proportional to the third and fourth Stokes parameters in brightness temperature. This type of correlation radiometer differs from the other two in that the correlation measurement itself, and not a Fourier transform of it, is typically the measurement of fundamental interest.

Conclusions Correlation microwave radiometers measure the statistical covariance between different components (in time, space, or polarization) of the electric field radiated by a source of thermal emission. Certain properties of the brightness temperature associated with that source (its spectrum, its angular distribution, or its polarization state) can be determined from the correlation measurements. Correlation radiometers are typically used when determination of those properties would be otherwise difficult or impossible. Bibliography Kerr, Y., Font, J., Waldteufel, P., and Berger, M., 2000. The soil moisture ocean salinity mission (SMOS). Earth Observation Quarterly, 66, 18. Piepmeier, J. R., and Gasiewski, A. J., 2001. Digital correlation microwave polarimetry: analysis and demonstration. IEEE Transactions on Geoscience and Remote Sensing, 39(11), 2392. Ruf, C. S., and Swift, C. T., 1988. Atmospheric profiling of water vapor density with a 205–235 GHz autocorrelation radiometer. Journal of Atmospheric and Oceanic Technology, 5(4), 539–546.

Cross-references Microwave Radiometers, Interferometers Microwave Radiometers, Polarimeters

MICROWAVE RADIOMETERS, INTERFEROMETERS Manuel Martin-Neira European Space Agency (ESA-ESTEC), Noordwijk, The Netherlands

Synonyms Aperture synthesis radiometers; Synthetic, interferometric radiometers Definition Interferometry. Technique of using the mean product of two random signals to infer some characteristics of the source that generated them. Aperture synthesis. Application of interferometry to an array of antennas to form an image of the signal source equivalent to one that would be observed by an antenna the size of the whole array. Introduction Microwave interferometric radiometers are a particular class of microwave radiometers, which work on the same principles as radio telescopes (Krauss, 1996; Thompson et al., 1988): interferometry and aperture synthesis. In turn, the development of the first aperture synthesis radiometers for space applications has led to a fundamental review of the very principles of radio astronomy

MICROWAVE RADIOMETERS, INTERFEROMETERS

391

exemplified in the Corbella equation. Instead of being pointed to the sky, these radiometers observed the Earth surface from ground-based, airborne, and spaceborne platforms. However, interferometric radiometers present very different features by comparison with radio telescopes, namely, the small size of the antenna elements and the short spacing at which they are clustered together, both of the order of the wavelength. Aperture synthesis radiometers have been developed to achieve fine spatial resolution in those applications where a single scanning antenna is mechanically too complex to realize. Since an interferometric radiometer does not need to scan to make an image, this has been the preferred solution in those cases. The original driving scientific applications have been the mapping of soil moisture and ocean salinity from space (Swift, 1993) and, more recently, atmospheric observations from geostationary orbit.

Historical development of aperture synthesis in remote sensing While the origin of radio astronomy dates back to 1950 (Ryle et al., 1950), its application to Earth observation was only considered in the late 1970s by the University of Berne (Schanda, 1979) and the beginning of the 1980s by engineers at NASA Goddard Space Flight Center in collaboration with the University of Massachusetts at Amherst (LeVine and Good 1983; Ruf et al., 1988; Tanner, 1990; Swift et al., 1991). The objective behind was to map the Earth’s soil moisture and ocean salinity, two important geophysical parameters never measured before at global scale. The first interferometric radiometer that was built had a synthetic beam in only one dimension, using the real aperture antenna pattern in the other. This was NASA’s ESTAR (electronically steered thinned array radiometer), an aircraft demonstrator of such a hybrid instrument; see Figure 1 (LeVine et al., 1992; Swift, 1993). Subsequent developments followed elsewhere with different variations as that using the motion of the platform to save in the required number of receivers (Komiyama, 1990), this being equivalent to the use of Earth rotation in radio astronomy. Aperture synthesis in two dimensions was developed in Europe during the 1990s. The Technical University of Denmark constructed a laboratory demonstrator (Laursen et al., 1994; Skou, 2004), and the European Space Agency (ESA) started the research of an L-band spaceborne MIRAS (Microwave Imaging Radiometer with Aperture Synthesis) (Martín-Neira, 1993; Goutoule et al., 1994; Bayle et al., 2002). ESA’s study involved French scientists at the Centre d’Etudes Spatiales de la Biosphère (CESBIO) (Kerr et al., 2000) and benefited with the participation of radio astronomers from the Observatoire du Midi Pyrenees (Lannes and Anterrieu 1994; Anterrieu et al., 2002). The Polytechnic University of Catalonia (UPC) in Barcelona played an important role in defining the requirements and calibration strategy for MIRAS

Microwave Radiometers, Interferometers, Figure 1 NASA’s electronically steered thinned array radiometer: the first onedimensional airborne radiometer (Courtesy David LeVine, NASA).

Microwave Radiometers, Interferometers, Figure 2 Helsinki University of Technology’s two-dimensional airborne interferometer (HUT-2D) (Courtesy Martti Hallikainen, TKK).

(Camps, 1996; Torre et al., 1996). Even more crucial was UPC’s research on the completed ESA’s MIRAS demonstrator, which led to the Corbella equation (Corbella et al., 2004), a fundamental correction to the formulation used by radio astronomers. The Helsinki University of Technology embarked in the manufacturing of an airborne two-dimensional microwave interferometer, the HUT-2D (see Figure 2) (Rautiainen et al., 1999), the first airborne two-dimensional aperture synthesis radiometer that provided good quality images of the Earth surface (Kainulainen et al., 2007). The calibration strategy of SMOS was first tested on HUT-2D (Colliander et al., 2007).

392

MICROWAVE RADIOMETERS, INTERFEROMETERS

Microwave Radiometers, Interferometers, Figure 3 ESA’s soil moisture and ocean salinity (SMOS) mission (Courtesy Yann Kerr, CESBIO).

In 1999, the SMOS (soil moisture and ocean salinity) mission was selected by ESA as second Earth Explorer Opportunity Mission, carrying MIRAS as only payload (McMullan et al., 2008). SMOS was launched 2 November 2009 being the first aperture synthesis radiometer flown in space; see Figure 3. It has successfully demonstrated the technique paving the way for its application in other areas. In fact, aperture synthesis has been proposed from geostationary orbit (Ruf, 1990), and higher frequency interferometers are now considered viable for Earth observation satellites flying in low Earth orbit. In the 2000s, several other ground-based and airborne microwave interferometric radiometers have been developed by different groups around the world, as NASA Goddard’s ESTAR-2D, JPL’s and ESA’s geostationary sounder demonstrators (Christensen et al., 2007), the C- and X-band one-dimensional interferometers of the Chinese Center of Space Science and Application Research’s (Wu et al., 2005; Yan et al., 2005), or ESA’s Airborne MIRAS in Europe.

Basic principles in remote sensing aperture synthesis A microwave aperture synthesis radiometer consists of a collection of N antennas arranged in an adequate geometry for optimum imaging. Every antenna is connected to a microwave receiver, which filters out all frequencies of the incoming radiation except for a narrow band, in such a way that the output signal is quasi-monochromatic, with slowly varying but random amplitude and phase. The receiver output signal is then a random process with Gaussian statistics and is best represented by its complex analytic signal.

Each of the N  (N  1)/2 possible pairs that can be formed with all the antennas of the collection is called a baseline. The two elements of a baseline receive the radiation from a distant target with some delay with respect to each other depending on the angle of arrival of the signal relative to the baseline. The average product of the two receiver output signals of every baseline, named correlation or visibility, is then computed. The visibility presents maxima and minima (null value), depending on whether the corresponding spatial delay is or not an integer number of the central wavelength. This means that some directions yield peak visibility values, while others do not contribute to it at all, what is equivalent to a spatial filtering. Every baseline acts then as a spatial filter: the longer the baseline, the shorter the spatial wavelength of the filter. Baselines oriented in different directions provide spatial filtering along those same directions. Thus, the set of all visibilities provides the spatial frequency content of the scene inside a frequency domain limited by the physical extent and location of the antenna elements. The image can then be recovered by a Fourier Transform of the visibilities. This is the formulation of the Van Cittert–Zernike theorem. Nonetheless, as explained later, this theorem would not suffice to describe the operation of an aperture synthesis radiometer, which obeys the Corbella equation instead.

Spatial resolution and windowing By the properties of the Fourier Transform, the angular resolution of the interferometric radiometer is determined by the extension of the spatial frequency domain. Since a given baseline can be taken in one direction as well as in its opposite direction, the spatial frequency support has a size twice the maximum baseline length, this fixing the angular resolution. The angular resolution of an aperture synthesis radiometer can now be compared with that of a real aperture instrument. An antenna of the same physical size as the synthetic radiometer would produce a far-field distribution equal to the Fourier Transform of the aperture illumination, following standard antenna theory. The square of the far field gives the radiated power that, when inverse Fourier Transformed, yields the spatial convolution of the original field illumination with itself. For a uniformly illuminated square aperture, the self-spatial convolution becomes a square pyramid with a side twice the physical length of the aperture. This is equivalent to applying a pyramidal weighting function to the visibilities collected by the synthetic radiometer. In summary, the same spatial resolution of a real aperture with a given illumination can be achieved by an interferometric radiometer of the same physical size and a proper weighting of its visibility function. This is the principle of aperture synthesis or aperture thinning. Real aperture radiometers never use uniform illumination as the side lobes become too high. Instead strong tapering is applied to increase the amount of energy collected through the main beam. Similarly, the visibility

MICROWAVE RADIOMETERS, INTERFEROMETERS

samples of an aperture synthesis radiometer are multiplied by weighting functions, or windows, that achieve a better beam efficiency than when using a pyramidal weighting. As an example, SMOS weighs the visibility function with a Blackman window.

Field of view The signal received by each element of an interferometer is affected by its antenna pattern. As the gain of the element decreases substantially outside its main beam, the radiometer becomes quite insensitive away from boresight. Imaging is therefore only feasible within the main beam of the element, which defines the field of view of the interferometric radiometer. Small antenna elements are preferred because of their wide beam, which avoids the need for mechanical scanning, a major advantage of interferometric radiometers. This is an important difference with respect to radio astronomy, where mechanically scanned large dish antennas are used. The field of view of an aperture synthesis radiometer is hence very wide and constrained by the element pattern, but not only, as explained below. Spatial frequency sampling and aliases The element pattern of a microwave interferometer is chosen to fit the size of the area to be imaged, or swath, in each particular application. For instance, an interferometer flying in a low Earth orbit at 800 km altitude will map the complete Earth in less than 3 days if its field of view is 60
wide (900 km on ground), while from geostationary orbit 18
suffices to cover the Earth disk. In matching the element pattern to the area of interest, the physical size of the element becomes defined, which in turn determines the pitch or minimum spacing possible between a pair of adjacent antenna elements. The resulting element spacing is usually larger than the maximum sampling period of the spatial frequency domain that guarantees the absence of grating lobes, and aliases appear in the image surrounding the alias-free area. The alias borders, like the element pattern, limit the field of view in aperture synthesis radiometry. The extension of the alias zones depends on the physical arrangement of the antenna elements. For any given spacing, the alias is minimized when the antennas are placed on a hexagonal grid with a pitch the same as the spacing. Examples of such hexagonal geometries are the triangle, the hexagon, the Y (three lines at 120
), and, in general, any snowflake configuration, which lead to visibility domains with the same shapes (triangle, hexagon, a six-point star). An alternative geometry is the rectangular one, like a square, a rectangle, a cross, a U, or a T, all leading to a square or rectangular coverage in the spatial frequency domain. The maximum distances between elements (normalized to the wavelength) before aliases appear are 0.577 and 0.5 for the hexagonal and rectangular geometries, respectively, which shows the benefit of the former.

393

Radiometric sensitivity The sensitivity of an aperture synthesis radiometer is comparable to that of a real aperture radiometer. As an example, consider a U-shape interferometer with N elements per arm. On the one hand, the sensitivity is proportional to the collecting area, given by the area of each antenna element a times the number of them 3N, that is, 3Na. On the other, the sensitivity is proportional to the square root of the integration time, or, what is equivalent, proportional to the square root of the number of resolution cells inside the field of view. In alias-free conditions, this number equals 2N. Finally, the sensitivity is also proportional to the volume of the weighting window, normalized by the area of its base, which for a pyramidal window gives 1/3. The final sensitivity of the interferometer is thus determined by 2N2a. A uniformly illuminated real aperture radiometer of collecting area A has 4 resolution cells in its field of view, and its sensitivity is thus proportional to 2A. If both instruments have the same physical size, A ¼ N2a, the sensitivity of the real aperture radiometer becomes 2N2a, the same as that of the aperture synthesis radiometer. The Corbella equation The Van Cittert–Zernike theorem, basis of radio astronomy, is not compatible with the Bosma theorem (Wedge and Rutledge 1991). This became apparent for the first time within ESA’s MIRAS Demonstrator Pilot Project in 2002. A two-dimensional aperture synthesis radiometer was placed inside an anechoic chamber, and all measured visibilities were zero, as expected from the Bosma theorem. This theorem states that for any passive network (as an array of antennas inside an anechoic chamber) in thermal equilibrium with its terminations (as the isolators following the antennas), the cross-correlation of the outcoming noise waves must be null. But according to the Van Cittert–Zernike theorem, the measured correlations should have been equal to the Fourier Transform of the element antenna pattern. The theoretical foundations of aperture synthesis were revised and a new equation established, compatible with the Bosma theorem, which correctly explains aperture synthesis, and by extension, radio astronomy: the Corbella equation. According to this equation, the visibility function is proportional to the difference or contrast in physical temperature between the target and the instrument. High correlations are expected when looking to the cold sky, and very small correlations when inside an anechoic chamber at similar physical temperature as the instrument. The Corbella equation has been verified from ground, airborne and space. The flat target transformation According to the Corbella equation, any uncertainty in the knowledge of the antenna pattern is amplified by the instrument–target temperature contrast. Thus, it is desirable to minimize such contrast. This is possible if the flat target response (FTR) of the interferometer is acquired

394

MICROWAVE RADIOMETERS, INTERFEROMETERS

first. The FTR is the visibility function measured when imaging an unpolarized uniform brightness temperature distribution, which is stable over time. In practice, the cold sky near the galactic poles of the Milky Way is a good approximation of such a target, and the FTR can be measured by pointing the aperture synthesis radiometer to the galactic poles. The FTR is like a mask formed by ripples due to system imperfections, mainly antenna errors. When this mask is scaled down to the temperature contrast between the instrument and an Earth scene, and is subtracted from the measured visibilities, the instrument errors are cancelled to a large extent. Mathematically, the Corbella equation is transformed by this linear combination with the result that the instrument physical temperature is replaced by the average temperature of the scene. The final temperature contrast is then reduced to that between each pixel of the scene and the average of all the pixels. The transformation of the Corbella equation using the FTR is known as the flat target transformation (FTT) and is an essential step in aperture synthesis imaging (Martín-Neira et al., 2008).

Polarimetry in aperture synthesis An aperture synthesis radiometer is inherently a polarimetric type of instrument because of its very wide field of view. Horizontal H and vertical V polarizations on the target are transformed into v and h polarizations on the element antenna according to a different rotation matrix depending on the direction. Since the H and V fields appear mixed together in the v and h fields in the instrument polarization frame, a nonzero vh cross brightness temperature is measured even from a target with no correlation between H and V. Reversely, to recover the horizontal and vertical brightness temperatures of the target, it is necessary to measure the vh cross brightness temperatures. When these are not available, a singularity along the image diagonals appears, and pixels in those areas have to be discarded. They can be recovered though when, due to the platform motion, they move to directions outside the singularities. Calibration An aperture synthesis radiometer is a complex instrument and requires several calibration steps. The first one is the flat target transformation, explained above, for which a cold sky view to acquire the flat target response must have been performed before hand. The second step is the correction of any comparator offset in the digital circuits and quadrature error of the demodulator. Then comes the phase correction, achieved by injecting correlated noise to all receivers internally, and the amplitude calibration, based on two standard temperatures, as the cold sky and an internal matched load at a well-monitored physical temperature. Finally, an uncorrelated load is used to remove any correlation offsets due to residual internal interference. In addition, to form an image, the antenna patterns must have been characterized a priori, and the

fringe-washing function has to be determined during operation. The fringe-washing function takes into account the decorrelation of the signal with time delay across the array and is estimated from early, punctual, and late correlations.

Summary and conclusions Aperture synthesis radiometry has developed starting in the early 1980s mainly driven by the need to map soil moisture and ocean salinity from space. During this development, the theoretical fundamentals had to be revised, and a new formulation, the Corbella equation, was found, which describes correctly the operation of this type of instruments, and, in fact, of radio telescopes as well. A consequence of the new formulation is the flat target transformation, a powerful method to calibrate out antenna errors in aperture synthesis. The spatial resolution and sensitivity of an aperture synthesis radiometer are comparable to those of a real aperture radiometer. Field of view, collecting area, and integration time take different values in each case, but balance each other, yielding the same net result at the end. ESA’s SMOS mission carrying the first aperture synthesis radiometer into space was launched 2 November 2009. Scientists have been able to assess the ultimate benefits of this new remote sensing technique. Following SMOS success, a next generation of interferometric radiometers might follow not only from low Earth orbit but also from geostationary orbit, already under study. Bibliography Anterrieu, E., Waldteufel, P., and Lannes, A., 2002. Apodization functions for 2-D hexagonally sampled synthetic aperture imaging radiometers. IEEE Transaction in Geoscience and Remote Sensing, 40(12), 2531–2542. Bayle, F., Wigneron, J.-P., Kerr, Y. H., Waldteufel, P., Anterrieu, E., Orlhac, J.-C., Chanzy, A., Marloie, O., Bernardini, M., Sobjaerg, S., Calvet, J.-C., Goutoule, J.-M., and Skou, N., 2002. Twodimensional synthetic aperture images over a land surface scene. IEEE Transaction in Geoscience and Remote Sensing, 40(3), 710–714. Camps, A., 1996. Application of Interferometric Radiometry to Earth Observation. PhD thesis, Barcelona, Polytechnic University of Catalonia. Christensen, J., Carlstrom, A., Ekstrom, H., Emrich, A., Embretsen, J., De Maagt, P., and Colliander, A., 2007. GAS: the geostationary atmospheric sounder. In IGARSS-2007 Proceedings, Barcelona, pp. 223–226. Colliander, A., Lemmetyinen, J., Uusitalo, J., Suomela, J., Veijola, K., Kontu, A., Kemppainen, S., Pihlflyckt, J., Rautiainen, K., Hallikainen, M., and Lahtinen, J., 2007. Ground calibration of SMOS: NIR and CAS. In IGARSS Proceedings, pp. 3631–3634. Corbella, I., Duffo, N., Vall-llossera, M., Camps, A., and Torres, F., 2004. The visibility function in interferometric aperture synthesis radiometry. IEEE Transactions on Geoscience and Remote Sensing, 42(8), 1677–1682. Goutoule, J. M., Kraft, U., and Martin-Neira, M., 1994. MIRAS: preliminary concept of a two-dimensional L-band aperture synthesis radiometer. In MicroRad’94 Proceedings, Rome. Kainulainen, J., Rautiainen, K., Tauriainen, S., Auer, T., Kettunen, J., and Hallikainen, M., 2007. First 2D interferometric

MICROWAVE RADIOMETERS, POLARIMETERS

radiometer imaging of the Earth from an aircraft. IEEE Transactions on Geoscience and Remote Sensing Letters, 4(2), 241–245. Kerr, Y. H., Wigneron, J. P., Ferrazzoli, P., and Waldteufel, P., 2000. Soil moisture and vegetation biomass retrievals using L-band, dual polarised and multi angular radiometric data in preparation of the SMOS mission. In IGARSS-2000 Proceedings, Hawaii, Vol. 3, pp. 1244–1246. Komiyama, K., 1990. Super-synthesis radiometer (SSR) for the remote sensing of the Earth. Technical Report. Ibaraki: Electrotechnical Laboratory. Krauss, J. D., 1996. Radio Astronomy. New York: McGraw-Hill. Lannes, A., and Anterrieu, E., 1994. Image reconstruction methods for remote sensing by aperture synthesis. In IGARSS-94 Proceedings, Pasadena, Vol. 4, pp. 2228–2230. Laursen, B., and Skou, N., 1994. A spaceborne synthetic aperture radiometer simulated by the TUD demonstrator model. In IGARSS-94 Proceedings, Pasadena, Vol. 3, pp. 1314–1316. LeVine, D. M., and Good, J. C., 1983. Aperture synthesis for microwave radiometers in space. NASA Technical Memorandum 85033. Greenbelt, MD: Goddard Space Flight Center. LeVine, D. M., Griffis, A., Swift, C. T., and Jackson, T. J., 1992. ESTAR: a synthetic aperture microwave radiometer for measuring soil moisture. In IGARSS-92 Proceedings, Houston, TX, Vol. II, pp. 1755–1757. Martín-Neira, M., 1993. MIRAS: a two-dimensional passive aperture synthesis radiometer. In Presentation at PIERS’93. Pasadena, CA: JPL. Martín-Neira, M., Suess, M., Kainulainen, J., and MartínPorqueras, F., 2008. The flat target transformation. IEEE Transactions on Geoscience and Remote Sensing, 46(3), 613–620. McMullan, K., Brown, M., Martin-Neira, M., Rits, W., Ekholm, S., Marti, J., and Lemanczyk, J. 2008. SMOS: the payload. IEEE TGARS, 46, 594–605. Rautiainen, K., Valmu, H., Jukkala, P., Moren, G., and Hallikainen, M., 1999. Four-element prototype of the HUT interferometric radiometer. In IGARSS-99 Proceedings, Hamburg, Vol. 1, pp. 234–236. Ruf, C. S., 1990. Antenna performance for a synthetic aperture microwave radiometer in geosynchronous Earth orbit. In IGARSS-90 Proceedings, pp. 1589–1592. Ruf, C. S., Swift, C. T., Tanner, A. B., and LeVine, D. M., 1988. Interferometric synthetic aperture microwave radiometry for the remote sensing of the Earth. IEEE Transactions on Geoscience and Remote Sensing, 26(5), 597–611. Ryle, M., Smith, F. G., and Elsmore, B., 1950. A Preliminary Survey of the Radio Stars in the Northern Hemisphere. London: Royal Astronomical Society. Schanda, E., 1979. Multiple wavelength aperture synthesis for passive sensing of the Earth’s surface. In International Symposium Digest. Seattle, WA: IEEE Antennas and Propagation Society, p. 762. Skou, N., 2004. Spaceborne L-band radiometers: pushbroom or synthetic aperture? In IGARSS-04 Proceedings, Vol. 2, pp. 1264–1267. Swift, C. T., 1993. ESTAR – The Electronically Scanned Thinned Array Radiometer for Remote Sensing Measurement of Soil Moisture and Ocean Salinity. Goddard Space Flight Center, Washington DC: NASA, p. 4523. Swift, C. T., Le Vine, D. M., and Ruf, C. S., 1991. Aperture synthesis concept in microwave remote sensing of the Earth. IEEE Transactions on Microwave Theory and Techniques, 39, 1931– 1935. Tanner, A. B., 1990. Aperture Synthesis for Passive Microwave Remote Sensing: The Electronically Scanned Thinned Array Radiometer. PhD thesis, Amherst, University of Massachusetts. Thompson, A. R., Moran, J. M., and Swenson, G. W., 1988. Interferometry and Synthesis in Radio Astronomy. New York: Wiley.

395

Torre, F., Camps, A., Bara, J., Corbella, I., and Ferrero, R., 1996. On-board phase and modulus calibration of large aperture synthesis radiometers: study applied to MIRAS. IEEE Transactions on Geoscience and Remote Sensing, 34(4), 1000–1009. Wedge, S. W., and Rutledge, D. B., 1991. Noise waves and passive linear multiports. IEEE Microwave and Guided Wave Letters, 1(5), 117–119. Wu, J., Liu, H., Yan, J., Ban, S., Dong, X., and Jiang, J., 2005. Research activity on synthetic aperture radiometry in CSSAR/ CAS. In PIERS 2005 Proceedings, Hangzhou. Yan, J., Wu, J., Liu, H., Dong, X., and Jiang, J., 2005. Design and implementation of digital correlator for CAS synthetic aperture radiometer. In PIERS 2005 Proceedings, Hangzhou.

Cross-references Microwave Radiometers Microwave Radiometers, Conventional Microwave Radiometers, Correlation Microwave Radiometers, Polarimeters Observational Systems, Satellite Sea Surface Salinity Soil Moisture

MICROWAVE RADIOMETERS, POLARIMETERS David Kunkee The Aerospace Corporation, Los Angeles, CA, USA

Synonyms Microwave polarimeter; Polarimetric radiometer Definition Microwave Polarimetric Radiometer. Microwave radiometer capable of measuring at least three of the four modified Stokes’ parameters of the incident field. Polarimetry. Measurement of the polarization characteristics of incident electromagnetic radiation. Introduction Polarimetric microwave radiometers are capable of measuring at least three Stokes’ parameters of the incident microwave radiation or radiometric scene. Traditional dual-polarized radiometers measure vertically and horizontally polarized brightness temperatures or, equivalently, the first two Stokes’ parameters of the input scene. In order to completely characterize the spatially averaged input scene, measurements of third and fourth Stokes’ parameters are required (Jackson, 1975), and these measurements can be performed utilizing coherent (direct) or incoherent techniques. Direct measurements are performed by polarization-correlating radiometers to measure the third and/or fourth Stokes’ parameters. Indirect methods include a combination of linearly polarized brightness temperatures typically oriented at 45
and 45
with respect to the principle directions in order to derive the third Stokes’ parameter and by combing left-and right-hand circular polarized brightness temperatures to derive the fourth

396

MICROWAVE RADIOMETERS, POLARIMETERS

Stokes’ parameter. Several airborne microwave polarimeters have been part of airborne experiments conducted between 1993 and 2003 in order to investigate application of microwave polarimetry to retrieval of sea surface wind direction (Yueh et al., 1995). The satellite-based wind direction system (WindSat), developed by the US Naval Research Laboratory (NRL) and launched in January 2003, performed the first space-based polarimetric measurements (Gaiser et al., 2004).

Background Airborne and space-based polarimetric radiometers measure the intensity of the Earth’s naturally occurring upwelling thermal emission. This arbitrarily polarized, non-or partially coherent thermal radiation is completely characterized by the four parameter-modified Stokes’ vector: 0 1 Iv B Ih C W C I ¼B @ U A ðm2 sr HzÞ V where Iv and Ih are the spectral intensities of the vertically and horizontally polarized field components and U and V are the in-phase and quadrature correlations, respectively, between the orthogonally polarized field components ^n and ^h. The spectral intensity of this emission can be described by the brightness temperature vector, T B :

D E 1 2 jE j n Tn B D E C B Th C B C 2 l2 jE j C B B C h ¼ T B¼ I ¼ @ C     TU A B k @ 2Re En E A h   TV 2Im En Eh 0

1

0

where Ev and Eh represent the vertically and horizontally polarized electric fields, respectively, h i represents a time-averaged quantity. Microwave polarimeters measure at least three of the four parameters. There are two general methods in passive microwave remote sensing in order to measure the third and fourth Stokes’ parameter: (1) cross-correlation between two orthogonally polarized measurements (polarization-correlating radiometer) and (2) measurements of polarized brightness temperature measurements using multiple polarizations. For example, using linearly polarized brightness temperatures at orientations of 45
with respect to the orientation of the vertical polarization vector, the third Stokes’ parameter can be found simply by differencing the two measurements: TU ¼ T45  T45. Similar approaches can be used to measure TV. One of the first reports on radiometer designs for sea surface microwave emission polarimetry was provided by Dzura (1992).

Polarization-correlating radiometer (direct method) An example of an analog polarization-correlating radiometer is given below. Figure 1 shows the block diagram of a three-channel polarization-correlating radiometer. f 1 u 2 A

LA

B

aA r

Úo (·)dt

GA

uoA

uA

udA

uAin A f

V

Δj

Δl

1 u 2 U

uoU

aU

−

r

Úo (·)dt 1 (v + v ) B 2 A

H B

udU

aB

uBin LB

B

r

GB 1 u 2 B

+

Úo (·)dt

uoB

uU

∑ −

uB

udB

Microwave Radiometers, Polarimeters, Figure 1 Block diagram of a three-channel polarization-correlating radiometer (Gasiewski and Kunkee, 1993).

MICROWAVE RADIOMETERS, POLARIMETERS

The measurement sensitivity for vU (and therefore TU) is proportional to the geometric mean of the characteristics of vA and vB; hence, if channels A and B are identical, channel U will have the same sensitivity (DTRMS) as channels A and B. Phase adjustments are made by adjusting the relative oscillator phase (Df) and/or relative line length (Dl) before the combiner. The phase can be adjusted to provide response to in-phase (U) or phase quadrature (V) components; however, in order to measure V (fourth Stokes’ parameter), a single-side band configuration must be used in order to avoid cancelation of the signal from the contributions of the high and low side bands. The post-detection analog signal summation in the scheme shown in Figure 1 is typically performed by using digital electronic circuitry (Piepmeier and Gasiewski, 2001).

Polarization combining radiometer (indirect method) An alternate method is to combine brightness temperature measurements at multiple polarizations to derive U or V indirectly. This method is also called the incoherent method and is typically applied through measurements of linearly polarized brightness temperatures offset 45
from the original basis to derive the third Stokes’ parameter TU and using left- and right-hand-polarized brightness temperatures to derive the fourth Stokes’ parameter, TV. Using a unitary rotational transform, the relationship between linearly polarized brightness temperatures rotated j degrees is expressed by 1 0 Tv0 B Th0 C C T0 ¼ B @ TU 0 A TV 0 30 2 1 Tv cos2 j sin2 j 0:5sin2 j 0 6 sin2 j C B cos2 j 0:5sin2 j 0 7 7B Th C ¼6 2 4 sin 2j sin2 2j cos 2j 0 5@ TU A TV 0 0 0 1 ¼ U ðjÞT where T 0 is the measurement made with the rotated polarization basis. Therefore, by using T45 ¼ Tu0 ðj ¼ 45
Þ, the third Stokes’ parameter, TU ¼ T45  T45. A popular way of deriving brightness temperature measurements at  45
is through the use of polarization combining networks (Yueh et al., 1995). Note that in the above equation, the fourth Stokes’ parameter, TV, is rotationally invariant; however, it can be shown that TV ¼ TR  TL where TR and TL are the right- and left-hand circularly polarized brightness temperatures of the incident wave, respectively (Tsang et al., 1985).

Ocean surface wind measurements One of the primary applications of microwave polarimetry is to improve ocean surface wind (OSW) retrievals

397

performed with dual-polarized, Tv and Th, brightness temperatures at 18 and 37 GHz. Dual-polarized radiometric measurements from space-based radiometers are currently and routinely used to accurately observe ocean surface wind speed to better than 1 m/s accuracy, Wentz (1997). The relationship between wind-roughened ocean surfaces and upwelling brightness temperature is well understood and was first explored by Hollinger (1971). Ocean surface wind speed measurements have been performed from operational microwave radiometers since the first special sensor microwave/imager (SSM/I) radiometer was launched and began operating in 1987 (Goodberlet et al., 1990: see also “Ocean, Measurements and Applications”). The first measurements showing a variation of brightness temperature as a function of the relative angle between the surface wind direction and the polarization plane were reported by Bespalova et al. (1979) as part of an airborne radiometry experiment conducted by the Space Research Institute at Moscow. A study by Wentz (1992) showed that there was a systematic bias in wind speed retrievals from SSM/I compared to ocean buoys that was dependent on the angle of observation with respect to the surface wind direction. The directional dependence was used to demonstrate that wind direction could also be retrieved using microwave brightness temperatures. However, the presence of multiple solutions, or directional ambiguities, limits the utility of wind direction derived using dual-polarized microwave brightness temperatures. Building upon theoretical modeling and laboratory measurements of polarized microwave emission from wave-covered surfaces (Yueh et al., 1994; Gasiewski and Kunkee, 1994), the utility of the third Stokes’ parameter for ocean surface wind direction retrieval was demonstrated through airborne experiments that measured polarimetric brightness temperatures over the ocean (Yueh et al., 1995). In order to demonstrate wind direction retrievals from space, the WindSat microwave polarimetric radiometer was developed by the US Naval Research Laboratory and launched in 2003 (Gaiser et al., 2004). Wind vector (speed and direction) retrievals using WindSat data (Bettenhausen et al., 2006) are now assimilated into numerical weather forecasting (NWP) models and used to improve forecasts (Candy et al., 2009).

Other applications Measurements of the third and fourth Stokes’ parameters by WindSat over the Greenland ice sheet have shown responses related to the asymmetrical features of polar ice sheets and potentially the microphysical structure of snow crystals (Li et al., 2008). Over desert, polarimetric measurements are also sensitive to the asymmetrical features of the surface such as the structure of sand dunes (Narvekar et al., 2007). Investigations have also shown the WindSat polarimetric measurements to be highly sensitive to RFI signals, and in some cases, showing responses before contamination is apparent in the vertically or horizontally polarized channels (Ellingson and Johnson, 2006; see also “RFI”).

398

MICROWAVE SUBSURFACE PROPAGATION AND SCATTERING

Summary Microwave polarimetry involves the measurement of the third and fourth Stokes’ parameter of the input radiometric “scene.” Traditional dual-polarized radiometers measure only the first two parameters of the Stokes’ vector and miss some characteristics of the brightness temperature scene. Microwave polarimeters can be designed for direct (coherent) polarimetric measurements or indirect (incoherent) measurements that are derived through post-detection combination of polarized measurements. Motivation to develop airborne and space-based microwave polarimeters is provided by improvements to passive ocean surface wind vector measurements. The WindSat radiometer was the first space-based microwave polarimeter, providing global measurements of all four Stokes’ parameters. These measurements are used to retrieve ocean surface wind vectors and improve weather forecast quality of NWP models. Other potential applications of microwave polarimetry are under development and include snow and ice characterization, land monitoring, and RFI detection and characterization. Bibliography Bespalova, E. A., Veselov, V. M., Glotov, A. A., et al., 1979. Sea ripple anisotropy estimates from variations in polarized thermal emission of the sea. Doklady Akademii Nauk SSSR, 246(6), 1482–1485. Bettenhausen, M. H., Smith, C. K., Bevilacqua, R. M., Wang, N., and Gaiser, P. W., 2006. A nonlinear optimization algorithm for WindSat wind vector retrievals. IEEE Transactions on Geoscience and Remote Sensing, 44(3), 597–609. Candy, B., English, S. J., and Keogh, S., 2009. A comparison of the impact of quickscat and windsat wind vector products on met office analyses and forecasts. IEEE Transactions on Geoscience and Remote Sensing, 47(6), 1632–1640. Dzura, M. S., Etkin, V. S., Khrupin, A. S., Pospelov M. N., and Raev M. D., 1995. Radiometers-polarimeters: principles of design and applications for sea surface microwave emission anisotropy. In IEEE International Geoscience and Remote Sensing Symposium, 1995. IGARSS’95, Vol. 2, pp. 1432–1434. Ellingson, S. W., and Johnson, J. T., 2006. A polarimetric survey of radio frequency interference in C-and X-bands in the continental United States using windsat radiometry. IEEE Transaction on Microwave Theory and Techniques, 44(3), 540–548. Gaiser, P. W., et al., 2004. WindSat spaceborne polarimetric radiometer: sensor description and early orbit performance. IEEE Transaction on Microwave Theory and Techniques, 42(11), 2347–2361. Gasiewski, A. J., and Kunkee, D. B., 1993. Calibration and applications of polarization correlating radiometers. IEEE Transaction on Microwave Theory and Techniques, 41(5), 767–773. Gasiewski, A. J., and Kunkee, D. B., 1994. Polarized microwave emission from water waves. Radio Science, 29(6), 1449–1466. Goodberlet, M. A., Swift, C. T., and Wilkerson, J. C., 1990. Ocean surface wind speed measurements of the special sensor microwave/imager (SSM/I). IEEE Transactions on Geoscience and Remote Sensing, 28(5), 823–828. Hollinger, J. P., 1971. Passive microwave measurements of sea surface roughness. IEEE Transactions on Geoscience Electronics, GE-9(3), 165–169. Jackson, J. D., 1975. Classical Electrodynamics. New York: Wiley, p. 848.

Li, L., Gaiser, P. W., Albert, M. R., Long, D. G., and Twarog, E. M., 2008. WindSat passive polarimetric signatures of the greenland ice sheet. IEEE Transactions on Geoscience and Remote Sensing, 46(9), 2622–2631. Narvekar, P. S., Jackson, T. J., Bindlish, R., Li, L., Heygster, G., and Gaiser, P. W., 2007. Observations of land surface passive polarimetry with the windsat instrument. IEEE Transactions on Geoscience and Remote Sensing, 45(7), 2019–2028. Piepmeier, J. R., and Gasiewski, A. J., 2001. Digital correlation microwave polarimetry: analysis and demonstration. IEEE Transactions on Geoscience and Remote Sensing, 39(11), 2392–2410. Tsang, L., Kong, J. A., and Shin, R. T., 1985. Theory of Microwave Remote Sensing. New York: Wiley, p. 613. Wentz, F. J., 1992. Measurement of oceanic wind vector using satellite microwave radiometers. IEEE Transactions on Geoscience and Remote Sensing, 30(5), 960–972. Wentz, F. J., 1997. A well-calibrated ocean algorithm for special sensor microwave/imager. Journal of Geophysical Research, 102(C4), 8703–8718. Yueh, S. H., Nghiem, S. V., Kwok, R., Wilson, W. J., Li, F. K., Johnson, J. T., and Kong, J. A., 1994. Polarimetric thermal emission from periodic water surfaces. Radio Science, 29, 87–96. Yueh, S. H., Wilson, W. J., Li, F. K., Nghiem, S. V., and Ricketts, W. B., 1995. Polarimetric measurements of sea surface brightness temperatures using an aircraft K-band radiometer. IEEE Transactions on Geoscience and Remote Sensing, 33(1), 85–92.

MICROWAVE SUBSURFACE PROPAGATION AND SCATTERING Alexander Yarovoy Delft University of Technology, Delft, The Netherlands

Definitions Subsurface. Natural materials (soils, rock, snow, ice) below air–ground interface. Ground-penetrating radar. Radar for subsurface sensing. It locates, images, and characterizes changes in electrical and magnetic properties of subsurface materials. Introduction The first description of microwaves use for subsurface sensing is attributed to a German patent by Leimbach and Löwy from 1910 (Daniels, 2004). In this patent, propagation of microwaves between pairs of vertically buried dipole antennas has been used to detect any subsurface objects with higher conductivity than the surrounding medium. Only monochromatic electromagnetic waves have been considered in this patent. The first use of electromagnetic pulses with a broad spectrum to determine the structure of buried objects is attributed to Hülsenbeck (Hülsenbeck et al., 1926). It was noted that any dielectric variation, not necessarily involving conductivity, would also produce reflections. The first ever experiments with subsurface microwave propagation to sound the depth of a glacier were performed in Austria in 1929 (Stern, 1929, 1930). These experiments were largely forgotten

MICROWAVE SUBSURFACE PROPAGATION AND SCATTERING

until the late 1950s when US Air Force radars were seeing through ice as planes tried to land in Greenland but misread the altitude and crashed into the ice. This accident together with lunar soil investigations within Apollo program (Simmons et al., 1972) triggered research on microwave penetration into subsurface, including not only ice sounding but also mapping soil properties and the water table. This fundamental research and simultaneous development of ultra-wideband microwave technology resulted, in the early 1970s, in wide use of microwave subsurface propagation and scattering for characterization and mapping of subsurface. Since then, the range of applications has been expanding steadily.

Dielectric properties of soil, rock, ice, and snow Propagation velocity and attenuation of microwaves is controlled by the electrical and magnetic properties of subsurface. Regarding the subsurface, three different types of natural earth materials should be distinguished: rocks, soils, and ice/packed snow. At microwave frequencies, electrical properties (which in most cases are more important than the magnetic properties) are dominantly controlled by density and by the chemistry, fine structure (composition of liquid/gas/solid components), and content of water. Below we concentrate mainly on rock and soil properties. Following Heimovaara et al. (1994), soils can be considered as a four-component system: irregularly shaped solid particles, air, free water, and bound water. The bound water refers to the first few (up to 10) molecular layers of water near solid surfaces that are rotationally hindered by surface forces. The frequencydependent complex dielectric permittivity of soils can be described then with a four-component complex dielectric mixing model based on the volumetric mixing of the refractive indices of the soil components. It should be noted that free water and bound water have different dielectric permittivities and thus different impact of the electrical properties of the material. Similarly to this approach, Wobschall (1977) has considered rocks as a three-phase system: irregularly shaped particles, air-or water-filled voids (pores), and crevices. In both cases, presence of moisture strongly influences electrical properties of material. Solid particles of soils and rocks have different physical properties depending on their material. In the most simple and frequently used approach (Wang and Schmugge, 1980; Dobson, et al., 1985), the soil solid particles are considered to be a mixture of sand particles of diameter d > 0.005 cm; silt, 0.0002 cm < d < 0.005 cm; and clay, d < 0.0002 cm, weight content of which is expressed in percentage of total weight of soil. Dielectric permittivities of dry sand, clay, and silt are slightly different (see Table 1). More important is that clay, silt, and sand have different abilities to bind the water. It is shown in Wang and Schmugge (1980) that the quantity of bound water in soil depends on the volume of clay fraction in it and

399

Microwave Subsurface Propagation and Scattering, Table 1 Typical values of relative dielectric permittivity of some dry soils and rocks at microwave frequencies (Daniels, 2004) Material

Relative dielectric permittivity

Sand Loam Clay Granite Limestone Sandstone Silt

2.4–6 4–10 2–6 5 7 2–5 3–10

the quantity of bound water increases with the volume of clay. This is explained by a large specific area of clay surface compared to other soil fractions. Sand and silt particles are also covered with bound water films; however, the amount of bound water on sand and silt particles is less than 0.1 % (Bojarskii et al., 2002). Further discussions on dielectric properties of soils and rocks can be found in Hoekstra and Delaney (1974), Hipp (1974), De Loor (1983), and Hallikainen et al. (1985). Providing a deep insight in the interaction of microwaves with rocks and soils, the abovementioned approaches require detailed knowledge of many physical parameters of soils, which are often not known during the field work. Thus, a large number of approximate equations have been proposed to compute dielectric permittivity of soil as a mixture of different components (Shutko and Reutov, 1982; Dobson et al., 1985; Sen et al., 1981). Among them, the so-called Bruggeman–Hanai–Sen (BHS) mixing model (Sen et al., 1981) is widely used nowadays:  C ðem  er Þ eewr p¼ em  ew where er is bulk soil relative dielectric permittivity, em is the dielectric permittivity of dry soil, ew is the dielectric permittivity of water, p is fractional porosity (volume of voids/total volume), and C is a shape factor (1/3 for spherical grains). When the water content is large, wet rock or soil can be considered simply as dry solid/water mixture and simply empirically stated Topp equation (Topp et al., 1980) is frequently used: er ¼ 3:03 þ 9:3 y þ 146:0 y2  76:3 y3 where er is bulk soil relative dielectric permittivity and y is volumetric water content. Increase of the water content in general is responsible for increase of relative dielectric permittivity of the material. In order to use different dielectric mixture formulas, knowledge of dielectric permittivity of water and dry soils is required. The complex-valued dielectric permittivity of bulk water at microwave frequencies can be described by

400

MICROWAVE SUBSURFACE PROPAGATION AND SCATTERING

the Debye theory and the Cole–Cole model (Cole and Cole, 1941; Heimovaara, 1994) is frequently used: eðoÞ ¼ e0  je00 ¼ e? þ

es  e? 1b

1 þ jðwtÞ

j

sdc we0

where es is the low-frequency (static) dielectric permittivity of water (which is of about 88 at a temperature of 0
C), e? is the high (infinite)-frequency dielectric permittivity and equals 4.25 for the 0
C, t is the relaxation time of water and b ¼ 0.0125 and is a factor that accounts for the possible spread in relaxation frequencies, sdc is the DC conductivity at a given temperature, and e0 is the permittivity of free space, which is 8.854 1012 F/m. The relaxation time of water molecules, which equals 7.7 1012 s at 27
C, drastically increases due to surface forces, and in monomolecular layers of water covering, the soil particles can reach the value of 5.0 1010 s at 27
C (Bojarskii et al., 2002). As a result, the bound water exhibits relaxation at frequencies below 300 MHz and has, at microwave frequencies, considerably lower values for the dielectric permittivity than free water. The permittivity of bound water is temperature dependent, resulting in a substantial increase in bulk dielectric permittivity with temperature at microwave frequencies for soils with high surface area (Or and Wraith, 1999). In an idealized representation, the water relative permittivity remains constant at high (roughly above 10 GHz) and low (below 100 MHz) frequencies. In a transition region over a frequency band, the real part of complex dielectric permittivity considerably decreases, while imaginary part exhibits a maximum. Solid parts of most soils and rocks have, when dry, a relative permittivity in the range 2–9 (see Table 1) (Daniels, 2004). Very often, it is assumed that the relative permittivity of dry soils and rock is frequency independent at microwave frequencies. Total losses in natural rocks and soils are mainly due to conductivity and polarization losses. According to Olhoeft (see, e.g., Olhoeft, 1987), water plays the dominant role in rock and soil conductivity through ionic charge transport through water-filled pore spaces in rocks and soils. Furthermore, conductivity might be due to presence of metals and other good conductors. Conductivity results in dissipation of microwave energy into heat. Similarly to soils, dielectric properties of snow and seawater ice depend on texture and volumetric amount of water and ice crystal (Bogorodskii et al., 1983; Boyarskii and Tikhonov, 1994; Kovacs et al., 1995). Relaxation frequencies of water molecule in ice crystals lie in kHz region. Thus at microwave frequencies, pure ice has almost constant with frequency dielectric permittivity of about 3.17 and very low losses. Dry snow can be considered as a mixture of ice crystals and air. Depending on density of snow, its relative dielectric permittivity varies from 1.2 to 2. Dielectric permittivity of wet snow depends largely on amount of water in it.

Seawater ice has different properties than freshwater ice, because seawater ice consists of pure ice, air, brine, and possibly solid salts. Furthermore, the sea ice has a multilayer structure, which depends on periodic processes of growth, deformation, and melting. Depending on its age and internal structure, several types of sea ice are distinguished. Relative dielectric permittivity of ice may vary from 3 till 18–30. Losses of coherent energy of microwaves in ice are determined by ice conductivity and noncoherent scattering on spatial heterogeneities.

Attenuation and dispersion of microwaves in subsurface To simplify the mathematical description of microwave propagation in natural subsurface, a complex-valued apparent dielectric permittivity is used: e ¼ e0  je00 where e0 is a bulk dielectric permittivity of the subsurface and e00 stands for losses associated with both conductivity and polarization (dipolar) losses: s e00 ¼ þ e00 relaxation jo The polarization losses are caused mainly by dielectric relaxation of water and are due to the polarity of the two hydrogen atoms of the water molecule. The plane electromagnetic wave propagating in such medium can be presented as a function of distance z and time t: Eðz; tÞ ¼ E0 eaz ejðotbzÞ where E stands for the electric field component of the field, a is the attenuation factor, and b is the phase constant. Both parameters are determined by the complex-valued apparent dielectric permittivity of the medium: qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 12 e00 2 me0 1 þ e0  1 a¼o 2 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 12  00 2 0 1 þ ee0 þ 1 b ¼ o me2 The distance d at which the electric field is attenuated in e times is called the skin depth and is equaled to 1/a. The skin depth is widely used in practice to estimate penetration depth of microwaves into subsurface or thickness of the top subsurface layer which determines microwave emission from the subsurface. It can be seen that the attenuation factor a by ignoring frequency dependence of complex-valued apparent dielectric permittivity is linearly related to frequency. Thus, the skin depth in the very first approximation also linearly decreases with frequency.

MICROWAVE SUBSURFACE PROPAGATION AND SCATTERING

Attenuation of microwaves in dry subsurface at 100 MHz varies from typically less than 1 dB/m for snow and sand up to 10 dB/m for some clay, limestone, and sandstone (Daniels, 2004). The attenuation factor slowly increases with frequency in the range from 100 MHz till 1 GHz, while at the frequencies above 1 GHz, it sharply increases due to dielectric relaxation of water. At, for example, 3 GHz, microwave attenuation in subsurface varies typically between several dozens and several hundreds dB/m. The phase constant b determines the phase velocity of microwave propagation in subsurface: 2 0sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1312  00 2 o 4me0 @ e 1 þ 0 þ 1 A5 v¼ ¼ b 2 e In the case of small losses and far away from relaxation frequencies, the microwave propagates with the phase velocity: c 1 u ¼ pffiffiffiffi ¼ pffiffiffiffiffiffiffiffiffiffiffiffi m0 e0 er er which is frequency independent. In the general case, finite conductivity of subsurface and the dielectric relaxation of water are responsible for dependence of phase velocity on frequency. This dependence is called dispersion. For microwaves, the dispersion is typically observed at the frequencies above 1 GHz. By propagation of wideband microwave pulses in subsurface, the dispersion and attenuation cause not only decrease of the pulse magnitude but also distortion of the pulse waveform.

Subsurface scattering Subsurface is heterogeneous not only on microscopic but also on macroscopic scale. On top of regular geological structure created by multiple strata of different soils/ rocks, there is a finer irregular structure created by local spatial inhomogeneities. These are created by spatial variations of moisture, all kind of inclusions (e.g., stones or man-made objects), and animal burrows. While propagating in subsurface, microwaves scatter on interfaces between different strata as well as on local inhomogeneities. Fields scattered from large-scale heterogeneities are used in subsurface remote sensing for geological structure reconstruction and minerals prospecting (Fung, 1994; Daniels, 2004). It has been shown by Yarovoy et al. (2000) that the presence of a low-loss (e.g., sand) layer above rock at certain angles and polarizations of the incident field will enhance microwave backscattering from either internal interface (rock surface) or from air– ground interface. Field scattered on small-scale (local) inhomogeneities is called subsurface clutter. At medium and/or high microwave frequencies, typical size of many abovementioned local inhomogeneities becomes

401

comparable with the wavelength of electromagnetic field. As a consequence, strength of clutter at these frequencies might become comparable to useful (searched) reflections, resulting in masking these reflections. It is widely believed that at frequencies above 1 GHz, the subsurface clutter becomes the major limiting factor for subsurface sensing.

Subsurface sensing Subsurface propagation and scattering of microwaves makes a basis for subsurface sensing. Nowadays, three major approaches are used: ground-penetrating radar (GPR) and airborne and spaceborne remote sensing. GPR GPR is a surface-based active radar system (Daniels, 2004). In order to achieve reasonable penetration depth, frequencies below 1 GHz are typically used in GPR. At a frequency of 1 GHz, penetration depth varies from about 1–2 m (for dry sandy soil or dry rock) to about 10 cm in water-saturated clay. At a frequency of 100 MHz, penetration depth might reach several 100 m (Cook, 1975). Due to its short-range operation, GPR has to radiate very short (of an order of a few ns) microwave pulses into subsurface, which requires ultra-wide operational bandwidth of the system. Commercial GPR systems have typically octave bandwidth (ratio between the highest and lowest operational frequency equals 2), while some experimental systems have this ratio above 10. While ultra-wide operational bandwidth is one key feature of GPR, another one is ground-coupled antennas. The latter is needed to avoid, as much as possible, reflection of microwaves from air–ground interface and maximize transmitted into the ground microwave power. GPR nowadays are widely used for a wide scope of applications starting from geological prospecting and ending with landmine detection and classification. Airborne sensing (CARABAS, radiometers) Airborne subsurface sensing has been performed both by active (radars) and passive (radiometers) systems. First active systems have been used for polar ice profiling and mapping of central areas of Kalimantan. Furthermore, CARABAS (airborne radar developed by FOA in Sweden and operating at the frequency band 25–85 MHz) has demonstrated capabilities to image buried pipelines in desert conditions (Hellsten et al., 1996), while FOLPEN (radar developed by SRI International and operating at the frequency band 200–400 MHz) has produced images of buried metal-cased antitank landmines in the Yuma Desert (Grosch et al., 1995). While performance of active radar systems is limited by maximal allowed transmitted power and field breakdown in the antennas, passive systems are free from these limitations. Airborne microwave radiometers have been successfully used in arid regions for search of subsurface

402

MICROWAVE SUBSURFACE PROPAGATION AND SCATTERING

water, water leakage from irrigation channels, and other purposes. Achieving probably similar to active systems penetration depth, passive systems however cannot provide 3D subsurface images as well as comparable to active systems cross-range resolution.

Spaceborne sensing Similar to airborne systems, satellite-based systems use both radar and radiometer for subsurface sensing. In particular, SIR-A and SIR-C results demonstrated that the L-band radar has penetration capabilities up to 2 m in rigid areas, revealing details of buried rock structures (Elachi et al., 1984). Conclusion Having almost 100 years of history, subsurface remote sensing became recently a fast-growing area of science and engineering with steadily expanding range of applications. This subject lies in a border area of three disciplines: microwave interaction with subsurface materials at microscopic level is studied by geophysics; microwave propagation and scattering in lossy, dispersive, and heterogeneous media is studied by electromagnetics; and extraction of information from microwaves radiated or emitted from the subsurface is studied by remote sensing (including GPR). The main research focus in this area remains on interaction of microwaves with heterogeneities of soils and rocks at microscale in order to determine effective dielectric permittivity of matter and development of statistical models for microwave field scattered from subsurface heterogeneities at macroscale. Bibliography Bogorodskii, V., Bentli, C., and Gudmandsen, P., 1983. Russian Radio Glaciology. Leningrad: Gidrometeoizdat. In Russian. Bojarskii, D. A., Tikhonov, V. V., and Komarova, N. Y., 2002. Model of dielectric constant of bound water in soil for applications of microwave remote sensing. Progress in electromagnetic Research, 35, 251–269. Boyarskii, D. A., and Tikhonov, V. V., 1994. Microwave effective permittivity model of media of dielectric particles and applications to dry and wet snow. In Proceedings of Geoscience and Remote Sensing Symposium. Vol. 4, p. 2065. Cole, K. S., and Cole, H. R., 1941. Dispersion and absorption in dielectrics – I. Alternative current characteristics. Journal of Chemical Physics, 9, 341. Cook, J., 1975. Radar transparencies of mine and tunnel rocks. Geophysics, 40, 865. Daniels, D. J. (ed.), 2004. Ground-Penetrating Radar, 2nd edn. London: The Institution of Electrical Engineers. De Loor, G. P., 1983. The dielectric properties of wet materials. IEEE Transactions on Geoscience and Remote Sensing, 21, 364. Dobson, M. C., Ulaby, F. T., Hallikainen, M. T., and El-Rayes, M. A., 1985. Microwave dielectric behavior of soil – part II: dielectric mixing models. IEEE Transactions on Geoscience and Remote Sensing, 23, 35. Elachi, C. H., Roth, L. E., and Schaber, G. G., 1984. Spaceborne radar subsurface imaging in hyperarid regions. IEEE Transactions on Geoscience and Remote Sensing, 22, 383.

Fung, A. K., 1994. Microwave Scattering and Emission Models and Their Applications. Norwood: Artech House. Grosch, T. O., Lee, C. F., Adams, E. M., Tran, C., Koening, F., Tom, K., and Vickers, R. S., 1995. Detection of surface and buried mines with an UHF airborne SAR. Proceedings of SPIE, 2496, 110. Hallikainen, M. T., Ulaby, F. T., Dobson, M. C., El-Rayes, M. A., and Wu, L. K., 1985. Microwave dielectric behaviour of wet soil – part I. Empirical models and experimental observations. IEEE Transactions on Geoscience and Remote Sensing, 15, 25. Heimovaara, T. J., 1994. Frequency-domain analysis of time-domain reflectometry waveforms 1. Measurement of the complex dielectric permittivity. Water Resources Research, 30, 189. Heimovaara, T. J., Bouten, W., and Verstraten, J. M., 1994. Frequency domain analysis of time domain reflectometry waveforms 2. A four-component complex dielectric mixing model for soils. Water Resources Research, 30, 201. Hellsten, H., Ulander, L. M., Gustavsson, A., and Larsson, B., 1996. Development of VHF CARABAS II SAR. Proceedings of SPIE, 2747, 48. Hipp, J. E., 1974. Soil electromagnetic parameters as functions of frequency, soil density and soil moisture. Proceedings of the IEEE, 62, 98. Hoekstra, P., and Delaney, A., 1974. Dielectric properties of soils at UHF and microwave frequencies. Journal of Geophysical Research, 79, 1699. Hülsenbeck, R., et al., 1926. German patent No. 489434. Kovacs, A., Gow, A. J., and Morey, R. M., 1995. The in-situ dielectric constant of polar firn revisited. Cold Regions Science and Technology, 23, 245. Olhoeft, G. R., 1987. Electrical properties from 10–3 to 10 + 9 Hz – physics and chemistry. In Banavar, J. R., Koplik, J., and Winkler, K. W. (eds.), Physics and Chemistry of Porous Media II. New York: American Institute of Physics. Or, D., and Wraith, J. M., 1999. Temperature effects on soil bulk dielectric permittivity measured by time-domain reflectometry: a physical model. Water Resources Research, 35, 371. Sen, P. N., Scala, C., and Cohen, M. H., 1981. A self-similar model for sedimentary rocks with application to the dielectric constant of fused glass beads. Geophysics, 46, 781. Shutko, A. M., and Reutov, E. M., 1982. Mixture formulas applied in estimation of dielectric and radiative characteristics of soil and grounds at microwave frequencies. IEEE Transactions on Geoscience and Remote Sensing, 20, 29. Simmons, G., Strangway, D. W., Bannister, L., Baker, R., Cubley, D., La Torraca, G., and Watts, R., 1972. The surface electrical properties experiment. In Kopal, Z., and Strangway, D. W. (eds.), Lunar Geophysics. Dordrecht: Reidel, p. 258. Stern, W., 1929. Versuch einer elektrodynamischen Dickenmessung von Gletschereis. Gerlands Beitrage zur Geophysik, 23, 292. Stern, W., 1930. Uber Grundlagen, Methodik und bisherige Ergebnisse elektrodynamischer Dickenmessung von Gletschereis. Zeitschrift Gletscherkunde, 15, 24. Topp, G. C., Davis, J. L., and Annan, A. P., 1980. Electromagnetic determination of soil water content: measurement in coaxial transmission lines. Water Resources Research, 16, 574. Wang, J. R., and Schmugge, T. J., 1980. An empirical model for the complex dielectric permittivity of soils as a function of water content. IEEE Transactions on Geoscience and Remote Sensing, 18, 288. Wobschall, D., 1977. A theory of the complex dielectric permittivity of soil containing water: the semidisperse model. IEEE Transactions on Geoscience and Remote Sensing, 15, 49. Yarovoy, A. G., de Jongh, R. V., and Ligthart, L. P., 2000. Scattering properties of a statistically rough interface inside a multilayered medium. Radio Science, 35, 455.

MICROWAVE SURFACE SCATTERING AND EMISSION

MICROWAVE SURFACE SCATTERING AND EMISSION David R. Lyzenga College of Engineering, Naval Architecture and Marine Engineering, University of Michigan, Ann Arbor, MI, USA

Definition Surface scattering. The process by which microwave radiation incident upon a solid or liquid surface is wholly or partially redirected away from that surface. Microwave emission. The process by which microwave radiation originates from a solid or liquid surface due to the rotational or vibrational motions of molecules near the surface. Introduction This topic deals with the processes through which a solid or liquid surface influences the electromagnetic field (at microwave frequencies) in the vicinity of that surface and the manner in which these processes are determined by the physical and chemical properties of the surface. The processes of surface scattering and emission, though distinctly different, are related by Kirchhoff’s law and can both be described by means of a single function, the bidirectional reflectance. The bidirectional reflectance function is defined in the following section, and the dependence of this function on the physical and chemical properties of the surface is discussed in the subsequent section. For active microwave (radar and scatterometry) purposes, the scattering properties of surfaces are more commonly described in terms of the normalized radar cross section, or cross section per unit area, which is related to the bidirectional reflectance as discussed near the end of this section. Bidirectional reflectance The bidirectional reflectance of a given surface, denoted here by the symbol r(m,f;m0 ,f0 ), may be loosely described as the proportion of the radiance incident upon the surface from the direction (m0 ,f0 ) which is scattered or reflected from the surface into the direction (m,f). Here m refers to the cosine of the angle between the direction of the reflected radiation and the surface normal, and f refers to the azimuthal angle of the reflected radiation about the surface normal. The symbols m0 and f0 similarly refer to the direction from which the incident radiation arrives at the surface. If we denote the surface normal direction by the unit vector n and the direction of propagation of the reflected and incident radiation by the unit vectors k and k0 , respectively, then m ¼ k n and m0 ¼  k0 n. For a specular surface, all of the incident radiation is reflected into the direction m ¼ m0 and f ¼ f0 + p, whereas for a diffusely reflecting surface, the radiation incident from a given direction is scattered into a range of angles.

403

The bidirectional reflectance may be defined more rigorously, albeit implicitly, by the equation Z 1 Z 2p rðm; f; m0 ; f0 ÞLðm0 ; f0 Þm0 dm0 df0 Lðm; fÞ ¼ 0

0

(1)

where L(m,f) is the reflected radiance and L(m0 ,f0 ) is the incident radiance (Nicodemus, 1965). For a system in thermal equilibrium, the incident radiance is isotropic and is equal to Bv ðTs Þ ¼ l22 ehv=kThvS 1 where n ¼ c/l is the frequency, h is Planck’s constant, k is Boltzmann’s constant, and TS is the surface temperature. In this case, the sum of the reflected and emitted radiation is also equal to Bn(TS). The emitted radiance in the direction (m,f) is then equal to e(m,f) Bn(TS), where Z1 Z2p eðm; fÞ ¼ 1  0

rðm; f; m0 ; f0 Þm0 dm0 df0

(2)

0

is the surface emissivity. This relationship is known as Kirchhoff’s law for unpolarized radiation. At microwave frequencies, hn < < kTS for temperatures typical of the Earth’s surface, and Bn(TS)  2kTS/l2 (Rayleigh-Jeans law). Thus, the emitted radiance is proportional to the product of the surface temperature and the emissivity and is often converted into units of temperature (K) and referred to as the brightness temperature, this being defined as the temperature of a blackbody that emits the same radiance as the actual surface, i.e., TB ¼ e TS. Thus far, we have only considered the case of unpolarized radiation. However, the quantities defined above can be extended to the case of polarized radiation as well. There are various definitions of the state of polarization, but for microwave radiation, the most common definition is in terms of the modified Stokes parameters Th, Tv, U, and V (see entry on “Radiation, Polarization, and Coherence”). A complete description of the reflectivity would require 16 components, corresponding to each possible combination of incident and reflected polarization states. However, for most remotesensing applications, we are primarily interested in the reflection of atmospheric (and cosmic) radiation which is virtually unpolarized, and we can combine these into four components corresponding to the reflection of unpolarized radiation into each of the four polarization states. For notational purposes, we indicate that the bidirectional reflectance is polarization dependent by using the bold-faced symbol r(m,f;m0 ,f0 ), to be understood as a vector with components corresponding to the four modified Stokes parameters. Since the blackbody radiation is unpolarized, the blackbody emissivity vector can be written as «b ¼ [1,1,0,0], and Kirchhoff’s law for polarized radiation becomes Z1 Z2p «ðm; fÞ ¼ «b  0

0

rðm; f; m0 ; f0 Þdm0 df0 :

(3)

404

MICROWAVE SURFACE SCATTERING AND EMISSION

Just above the surface of the Earth, the upwelling radiation field includes contributions from both the downwelling (atmospheric and cosmic) radiation that is reflected from the surface and the radiation that is emitted from the surface. The reflected contribution can be written in terms of the brightness temperature as Z 1 Z 2p rðm; f; m0 ; f0 ÞTd ðm0 ; f0 Þdm0 df0 T r ðm; fÞ ¼ 0

0

(4)

where Td(m0 ,f0 ) represents the (unpolarized) downwelling radiation at the surface, and the surface-emitted component can be written as T e ðm; fÞ ¼ TS «ðm; fÞ

Z

¼ TS «b  TS

1 0

Z

2p 0

rðm; f; m0 ; f0 Þdm0 df0 : (5)

Combining these expressions, the total upwelling brightness temperature just above the surface can be written as Tðm; fÞ ¼T e þ T r Z1 Z2p ¼TS «b þ 0 0

rðm; f; m0 ; f0 Þ

(6)

0

½Td ðm0 ; f Þ  TS dm0 df0 : Measured at spacecraft altitudes, the observed brightness temperature would of course be reduced by attenuation in the atmosphere and augmented by the upwelling radiation emitted from the atmosphere.

Dependence of bidirectional reflectance on surface properties The interest in making measurements of the microwave radiation field above the surface of the Earth is to obtain information on geophysical or biological processes occurring at the surface or in the atmosphere or on the physical or chemical state of the surface and/or atmosphere. The types of information and the methods used to extract this information are discussed in other entries (see, e.g., Land-Atmosphere Interactions, Evapotranspiration; Ocean-Atmosphere Water Flux and Evaporation; Ocean, Measurements and Applications; Cryosphere, Measurements and Applications). In this section, we discuss some of the relationships that underlie these applications. As discussed in the previous section, the effects of the surface on the radiation field above the surface can be described in principle by the surface temperature and the bidirectional reflectance vector r(m,f; m0 ,f0 ). This quantity can be defined for a specific surface but more often refers to a statistical ensemble of surfaces with similar properties. It could in principle be measured, but as can

be imagined, this is a difficult and time-consuming process since it involves many different combinations of incidence and reflection angles and must be repeated for many surfaces to obtain a stable statistical mean. Thus it is desirable to model the bidirectional reflectivity when possible, although measurements are necessary for model validation and can sometimes be used directly if some assumptions are made about the directional dependence. Models have been developed, even for complex surfaces such as vegetation canopies, as discussed in the Measurements and Applications entries referenced above. A review of these models is beyond the scope of this entry, but in general the reflectivity depends on the surface roughness, or the geometrical shapes of the elements comprising the surface or canopy, as well as the dielectric constant of the component materials. Since the dielectric constant is influenced by such factors as the water content of soils and vegetation, the salinity and phase (liquid or ice) of the ocean surface, and possibly other chemical constituents, the radiation field potentially contains information about these factors. In addition, the roughness of the ocean surface is a function of the wind speed and direction (as well as, possibly, other factors such as the presence of surface slicks, converging or diverging surface currents, and atmospheric stability conditions). As a result, radiometric measurements from aircraft and satellite platforms have been shown to be very useful for measuring surface winds over the ocean on regional and global scales and for observing some of the other features mentioned above on smaller scales. Polarimetric measurements such as those provided by WINDRAD (Yueh et al., 1999) and WindSat (Gaiser et al., 2004) are particularly useful for separating atmospheric from surface effects and for inferring the wind direction as well as wind speed. For the ocean surface, the bidirectional reflectance is most commonly modeled using a two-scale model (Wentz, 1975; Yueh, 1997) although other models such as the small-slope approximation (Irisov, 1997; Johnson and Zhang, 1999) have also been used. Such modeling also requires a representation of the ocean wave spectrum and its dependence on the wind speed. The spectral model of Durden and Vesecky (1985) has been widely used for this purpose, but some modifications have also been suggested (Lyzenga, 2006) in order to improve comparisons with satellite data. The shortwave (centimeter-scale) portion of the ocean wave spectrum has a strong influence on the microwave surface reflectivity, and there are large uncertainties in this portion of the spectrum due to the difficulty of measuring the surface on these scales. The effects of wave breaking and surface foam on the reflectivity are also incompletely understood, although considerable work has been done in this area. In the radar literature, the scattering properties of objects are most often described in terms of the radar cross section s, which is defined as 4p times the reflected power per unit solid angle divided by the power density (power per unit area) incident on the object. It can be defined for any scattering angle but is most commonly used for

MISSION COSTS OF EARTH-OBSERVING SATELLITES

backscatter, i.e., for the case in which the source of radiation and the receiver are collocated. This definition applies to discrete objects but was later extended to the case of scattering surfaces by dividing the reflected power by the surface area illuminated or resolved by the radar. This dimensionless quantity is denoted by the symbol so and referred to as the normalized radar cross section, or the cross section per unit area. The relationship between the radar cross section and the bidirectional reflectance can be found by considering the radiation scattered from a unit surface area. For the case of a distant source, the incident radiance can be written as Lðm0 ; f0 Þ ¼ Eo dðm0  mo Þdðf0  fo Þ

(7)

where Eo is the incident power per unit area normal to the direction of the incident radiation (as in the definition of the radar cross section). The scattered radiance for this case is given by Equation 1 as Lðm; fÞ ¼ mo Eo rðm; f; mo ; fo Þ:

(8)

Following from the definition of the radiance, the reflected power per unit solid angle is equal to m L(m,f). Applying the definition of the radar cross section, we then have s
¼ 4p

mLðm; fÞ ¼ 4pmmo rðm; f; mo ; fo Þ Eo

(9)

in agreement with Tomiyasu (1988). For the backscatter case, of course, m ¼ mo .

Conclusion The effect of the Earth’s surface on the naturally occurring microwave radiation field above the surface can be described in terms of the surface temperature and the bidirectional reflectance of the surface. The bidirectional reflectance is a function of the geometry and composition of the surface. For land surfaces, the reflectance is strongly dependent on the water content of the soil and/or vegetation. For water surfaces, the reflectance depends on the salinity, the phase (liquid or solid), and the surface roughness. For the liquid ocean, the surface roughness is primarily dependent on the wind speed but may also be influenced by surface slicks, currents, and atmospheric stability effects. Bibliography Durden, S. P., and Vesecky, J. F., 1985. A physical radar crosssection model for a wind-driven sea with swell. IEEE Journal of Oceanic Engineering, 10, 445–451. Gaiser, P. W., St Germaine, K. M., Twarog, E. M., Poe, G. A., Purdy, W., Richardson, D., Grossman, W., Jones, W. L., Spencer, D., Golba, G., Cleveland, J., Choy, L., Bevilacqua, R. M., and Chang, P. S., 2004. The WindSat spaceborne polarimetric microwave radiometer: sensor description and early orbit performance. IEEE Transactions on Geoscience and Remote Sensing, 42, 2347–2361.

405

Irisov, V. G., 1997. Small-slope expansion for thermal and reflected radiation from a rough surface. Waves in Random Media, 7, 1–10. Johnson, J. T., and Zhang, M., 1999. Theoretical study of the small slope approximation for ocean polarimetric thermal emission. IEEE Transactions on Geoscience and Remote Sensing, 37, 2305–2316. Lyzenga, D. R., 2006. Comparison of WindSat brightness temperatures with two-scale model predictions. IEEE Transactions on Geoscience and Remote Sensing, 44, 549–559. Nicodemus, F. E., 1965. Directional reflectance and emissivity of an opaque surface. Applied Optics, 4, 767–773. Tomiyasu, K., 1988. Relationship between and measurement of differential scattering coefficient (so) and bidirectional reflectance distribution function (BDRF). IEEE Transactions on Geoscience and Remote Sensing, 26, 660–665. Wentz, F. J., 1975. A two-scale model for foam-free sea microwave brightness temperatures. Journal of Geophysical Research, 80, 3441–3446. Yueh, S. H., 1997. Modeling of wind direction signals in polarimetric sea surface brightness temperatures. IEEE Transactions on Geoscience and Remote Sensing, 35, 1400–1418. Yueh, S. H., Wilson, W. J., Dinardo, S. J., and Li, F. K., 1999. Polarimetric microwave brightness signatures of ocean wind directions. IEEE Transactions on Geoscience and Remote Sensing, 37, 949–959.

Cross-references Cryosphere, Measurements and Applications Land-Atmosphere Interactions, Evapotranspiration Ocean-Atmosphere Water Flux and Evaporation Ocean, Measurements and Applications Ocean Measurements and Applications, Ocean Color Radiation, Polarization, and Coherence

MISSION COSTS OF EARTH-OBSERVING SATELLITES Randall Friedl and Stacey Boland Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA

Synonyms Earth-observing mission costs; Spaceborne mission costs Definition Earth-observing satellite mission. An instrumented spacecraft deployed in space that is designed to observe select geophysical parameters of the Earth for purposes of environmental science or monitoring. A satellite mission requires a launch vehicle, observing instruments, and a spacecraft bus which houses the instruments as well as systems for guidance, navigation, control, communications, command and data, power, thermal control, propulsion, and structures. Satellite mission cost. The total cost of implementing a satellite mission. The cost includes all of the component

406

MISSION COSTS OF EARTH-OBSERVING SATELLITES

costs as well as costs for design, system engineering, operating and data analysis software, and mission operations. The cost does not typically include scientific analysis of the data collected by the mission.

Introduction Observing the Earth from the vantage point of space has transformed our understanding of the Earth as an integrated system of physical and biological components. The number of Earth-observing satellites has steadily increased with time, with more than 100 satellite missions currently operating. The desire for new satellite missions remains high, especially since many of the geophysical parameters of interest to scientists (e.g., 44 essential climate variables as defined by the Global Climate Observing System Program) are still not observed from space and other key parameters need to be sustained over many satellite lifetimes to enable investigations of long-term climate trends. The need for new missions is tempered by the relatively high cost, currently ranging from $100 M to $3 B, to deploy and operate a satellite mission. Understanding the factors that determine the costs of satellite missions is critical to developing and sustaining an effective system of Earth-observing satellites. Cost factors Transporting instrumented science payloads to space requires powerful launch vehicles. The relatively small number (i.e., ~30 total, ~12 USA) of available launch vehicles is divided into four classes, as defined by the US Federal Aviation Administration Office of Commercial Space Transportation, that reflect lift capability as follows: small (5,000 lb to Low Earth Orbit or LEO), medium (5,001–12,000 lb to LEO), intermediate (12,001–25,000 lb to LEO), and heavy ( 25,000 lb to LEO). Examples of launch vehicles across the four classes are Pegasus (small), Delta II (medium), Atlas IIAS (intermediate), and Atlas V (heavy). While the cost of a launch vehicle increases with each step-up in class, the per pound payload cost is roughly the same across classes when the launch vehicle lift capability is fully utilized. There is thus a substantial cost incentive to either design a payload to fit into the smallest possible launch vehicle or to combine payloads to take advantage of the full capability of a larger-class vehicle. Most of the past and current Earth science missions range between 1,000 and 10,000 lb (instruments plus spacecraft) and were launched on smalland medium-class launch vehicles. While launch vehicle costs can be directly correlated to performance requirements in terms of a “price per pound” metric, this is not the case for estimating the cost of the satellite instruments. Science instrument payloads typically represent the most unique and challenging components of Earth-observing missions. Payload designs are generally mission unique, resulting from a careful optimization of science requirements against physical and programmatic constraints (e.g., cost, schedule, risk). Instrument

costs can thus be driven by a variety of factors. Increasingly stringent requirements on temporal and spatial resolution, for example, might necessitate larger signal collection optics and/or antennas as well as highly sensitive detectors and precision signal processing electronics, driving payload mass, volume, and component costs higher. Attempts to reduce payload mass and volume might necessitate incorporation of new, yet expensive, lightweight technologies. Spacecraft costs, like launch vehicle costs, are generally performance driven. Instrument acodation requirements, particularly mass, power, data rate, and configuration constraints, greatly dictate requirements for the platform, or spacecraft bus, that houses the instrument. While many standardized platforms are now available, most missions have the need for some degree of spacecraft customization (e.g., tighter thermal control, improved pointing accuracy or knowledge, larger onboard data storage capacity, additional power allocations). The cost of customization is to a large extent dependent on whether new technology development is required. Mission operations costs are driven, to a large extent, by mission data volume and latency requirements which can drive ground network, data processing, and archival costs. The spacecraft’s onboard data storage capacity (i.e., solid-state recorder size) is also a determining factor in the required frequency of data downlinks and/or number of ground stations.

Cost estimation methods Three methodologies are generally used to estimate mission costs, namely, analogy, parametric, and “grass roots.” An analogy-based estimate represents the simplest approach to cost estimation. It relies on direct comparisons to analogous ongoing or previous missions. Provided there is little deviation from the chosen analog, rapid and accurate mission cost estimates can be obtained using this method. However, since very few missions share identical subsystems or components, analogy-based estimates are not widely applicable without including cost adjustments based on analysis of differences between the target mission and the analog. Such adjustments typically employ parametric analyses in order to enable quantitative extrapolations between systems with varying degrees of differences. Analogy-based estimates best apply to mission lines which involve launching a series of similar (or near-identical) satellites (e.g., the NOAA POES series of weather satellites) and those which incorporate incremental performance improvements between generations (e.g., the NASA TOPEX-Poseidon, Jason, OSTM-sustained research satellite series). Parametric cost estimates are based on cost driver relationships derived from analyses of historical data. The primary cost drivers used for parametric studies typically include mass, power, and subsystem complexity. For example, experience shows that mission costs increase roughly linearly with mass within a given launch vehicle

MISSION OPERATIONS, SCIENCE APPLICATIONS/REQUIREMENTS

class, allowing derivation of the “price per pound” metric cited above. Once developed, parametric relationships are particularly useful for performing rapid trade studies during the initial scoping of mission concepts, particularly for concepts without direct historical analogy. Grassroots cost estimates involve detailed consideration of each element of a mission or component of an instrument. For each element or component, a cost estimate is generated and basis of estimate is documented, describing the technique used to derive the estimate. For a true grassroots estimate to be performed, the mission or instrument design must be specified to a fairly high degree to enable rigorous quantitative cost estimates based on material and labor costs allocated over a defined implementation schedule. Consequently, grassroots estimates find their greatest utility during later stages of mission development. During concept development, a “quasigrassroots” mechanism can be employed in which the grassroots cost estimating structure is preserved, including documentation of a basis of estimate, while leveraging analogy or parametric cost estimates for the individual components for which detailed requirements, designs, and/or schedules are not well known.

Cost estimation accuracy The ability to conceive, plan, and implement an effective and affordable multi-satellite observing system rests in large part on understanding the costs of individual missions. In particular, decisions regarding mission priorities and sequencing are intimately tied to considerations of cost to benefit. Because there are relatively small numbers of satellite missions and most of them are partially, or wholly, unique in design, the ability to accurately estimate cost is limited in the beginning phases of a mission. Analysis of historical mission cost data reveals a tendency for substantial increases (10–50 %) in estimates for lower-cost Earth missions when going from initial to later mission phases. These increases have been traced primarily to initial underestimations of instrument and spacecraft components. In many cases, the cost increases reflect more stringent measurement requirements stemming from refinement of the mission science goals, rather than inherent inaccuracies in the estimation methodology. Acknowledgment This research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the NASA. Bibliography Bearden, D., 2005. Perspectives on NASA mission cost and schedule performance trends. In NASA Goddard Flight Center Symposium. Committee on Earth Observation Satellites mission database, available at http://www.ceos.org/ Emmons, D. L., Bitten, R. E., and Freaner, C. W., 2006. Using historical NASA cost and Schedule growth to set future

407

program and project reserve guidelines, IEEEAC paper #1545, Version 3. FAA Commercial Launch Vehicle website. http://www.faa.gov/ about/office_org/headquarters_offices/ast/launch_data/. Futron Corporation, 2002. Space transportation costs: trends in price per pound to orbit 1990–2000. www.futron.com. Global climate observing system program’s essential climate variables. Available from http://www.wmo.int/pages/prog/gcos/ index.php?name¼essentialvariables. Larson, W. J., and Wertz, J. R., 1999. Space Mission Analysis and Design. El Segundo: Microcosm, Vol. 3. NASA, 2008. Cost estimation handbook. Available from http://ceh. nasa.gov. Wertz, J. R., and Larson, W. J., 2007. Reducing Space Mission Cost. New York: Springer.

Cross-references Global Climate Observing System Mission Operations, Science Applications/Requirements Observational Systems, Satellite

MISSION OPERATIONS, SCIENCE APPLICATIONS/ REQUIREMENTS David L. Glackin Los Angeles, CA, USA

Synonyms Environmental data record (EDR) requirements Definition Science applications. Environmental phenomena to be measured in the Earth’s atmosphere, oceans, land surface, solid Earth, ice cover, and near-Earth space environment. Science requirements. A numerical specification of the phenomena that must be measured and a corresponding set of attributes including spatial resolution, spatial reporting interval, spatial coverage, measurement uncertainty (accuracy/precision), long-term measurement stability, mapping uncertainty, reporting frequency, and timeliness of data delivery. Introduction The science applications are described thoroughly elsewhere in this volume (see Atmospheric General Circulation Models; Ocean, Measurements and Applications; Cryosphere, Measurements and Applications; LandAtmosphere Interactions, Evapotranspiration; Trace Gases, Troposphere - Detection from Space). The thrust of this entry is the science requirements and how they are motivated, established, codified, analyzed, and iterated. David L. Glackin: deceased.

408

MISSION OPERATIONS, SCIENCE APPLICATIONS/REQUIREMENTS

Motivation for and establishment of requirements In an ideal world, science requirements should be motivated by a problem in need of a remote sensing instrument (or set of instruments), not vice versa. Reality is sometimes different, but environmental remote sensing missions, especially large ones, typically have a large user constituency behind them that provides the rationale for the mission. Mission requirements may be driven by the need for new types of data, improvements in the quality of old types of data, or continuity of data for studies of long-term trends. Ideally, the requirements flow should be from the top down. The science requirements should drive the instrument concepts. Those concepts in turn should drive the instrument designs, which should in turn drive the platform (satellite, aircraft, etc.) concept and design. Examples of important aspects of the platform design that the science requirements drive are pointing accuracy and stability and the choice of satellite orbit (see Observational Systems, Satellite). As an adjunct to the science requirements, the postlaunch data reduction, analysis, and research activities should be considered up front. Too often, the tendency is to focus on building the instruments and the platform to meet the requirements and to delay consideration of the analysis and research. The latter approach can have large systems impacts, and it is best to consider these issues up front while setting requirements. For example, the calibration/validation phase of a mission is a major postlaunch activity and is the one in which the instrument calibration is established and in which the quality of the science data products is validated (see Calibration and Validation). If the science requirements lead to data products that are exceedingly difficult to validate, unexpected amounts of postlaunch resources may be required (see Mission Costs of Earth-Observing Satellites). Science requirements are usually established by user working groups, which are organized by scientific discipline. It is important to involve the instrument systems engineers early in this process. They understand the hardware in an end-to-end systems sense and can help the user communities to understand the cost and risk implications of their requirements. Often, scientific “desires” must be tempered with technological reality. Sometimes, it may be possible to achieve half of what is desired at one-tenth the cost. Science requirements are typically iterated between these two groups of people until a realistic set of requirements is established. Science requirements are often couched in terms of “thresholds” and “objectives.” The thresholds are the minimum set of requirements that must be met. The objectives represent the performance that the users would like to meet, if meeting them does not inordinately drive the hardware design. Remote sensing missions may in the end deliver data products that fall between these two benchmarks. Codification of requirements The science requirements must be established carefully. Clarity is of the utmost importance. If the requirements

are not clear up front, much time will be wasted later in clarifying the requirements to the instrument and platform vendors and then in rewriting the requirements. Definitions must be clear and comprehensive. Units must be consistent and understandable. Requirements must be consistent with basic physics. There should be no conflicting requirements. Very significant is the establishment of requirements for measurement quality. It is possible to generate major confusion over items as seemingly simple as measurement accuracy, precision, and uncertainty. The latter can be a major stumbling block in the process of requirements establishment and should be treated thoughtfully.

Requirements analysis and iteration Science requirements should be thoroughly analyzed after they are initially established, for clarity, consistency, comprehensiveness, lack of conflict, and impact on the end-to-end mission. This is the point at which the design engineers, the hardware vendors, and the operators should engage with the system engineers to ensure that the set of requirements is sound. At this stage, iterations of the requirements are usually made, in consultation with the user community. It is paramount that experienced people think through these issues up front. Otherwise, problems can crop up during the course of the program that can be very wasteful of resources. It is particularly important to define the “driving requirements” (Wertz and Larson, 1991). These are the ones that dictate some of the major hardware characteristics. “Requirements creep” is almost always a problem in any large remote sensing mission. Once the user communities start to see what the system can do, they usually think of more things that it might do. One solution for that is to define these items as “Pre-Planned Product Improvements” (P3I) that might be addressed later, given sufficient resources. Conclusion An intelligent process of establishing the science requirements for a mission can make the difference between mission success and failure or between delivering the hardware on time and under budget and delivering it late and far over budget. Getting the science requirements right at the beginning and thinking things through thoroughly up front is key to mission success. Bibliography Wertz, J. R., and Larson, W. J., 1991. Space Mission Analysis and Design. Dordrecht: Kluwer.

Cross-references Calibration and Validation Cryosphere, Measurements and Applications Ocean, Measurements and Applications

O

OBSERVATIONAL PLATFORMS, AIRCRAFT, AND UAVS Jeffrey Myers NASA/Ames Research Center, Airborne Science and Technology Laboratory, University of California, Santa Cruz, Moffett Field, CA, USA

Synonyms Airborne platforms; Remote sensing aircraft; Suborbital platforms Definition Airborne observational platforms are any aircraft, manned or unmanned, which is equipped to obtain remotely sensed data or in situ atmospheric samples above the surface of the Earth. Introduction Airborne platforms are typically used to obtain observations of high spatial or temporal resolution that cannot be practically achieved with satellite- or ground-based systems. These include remote sensing observations of dynamically evolving events (e.g., volcanic eruptions or convective weather systems) which could not be made by satellites due to their fixed revisit schedules or in situ air sampling of atmospheric constituents at varying altitudes. Because they can selectively operate during clearsky conditions, they are also often used for remote sensing in areas which have persistent cloud cover. Imagery in some tropical, mountainous, and arctic regions, for example, cannot be obtained from polar-orbiting satellites on any regular basis. The spatial resolution of imagery obtained from an airborne platform can be precisely controlled by varying the altitude of the aircraft above the ground. Pixel sizes

produced from a sensor with a 2.5 mrad instantaneous field of view (IFOV), for example, can thus be varied between 3 and 50 m with the same instrument, when operated at altitudes of 1,200 and 19,800 m, respectively. This variability allows the imagery to be tailored to the requirements of the specific phenomenon being studied. Most satellite systems, especially those with the infrared spectral bands of interest to earth scientists, are designed for broadscale global observations at fixed spatial resolutions, which may be too coarse for many detailed geophysical process studies. Spatial resolutions of airborne infrared imaging systems can exceed 3 m and can be as high as 1 cm in the visible and near-infrared wavelengths, depending on platform altitude. Airborne platforms are also used as test beds for new instrument development. They provide an economical method of testing prototype satellite sensors prior to launch and also for collecting the empirical data needed to develop the associated retrieval algorithms for new orbital systems. Once a satellite system is placed in orbit, airborne sensors are often used to conduct simultaneous, high-resolution observations for data validation and instrument calibration purposes.

Overview of platform types Manned aircraft A wide variety of airborne platforms are used for Earth observations, ranging from small single-engine propeller planes with minimal modifications to large specially adapted multiengine commercial and military reconnaissance aircraft. Unmanned Aerial Vehicles (UAVs), airships, and balloons also fill unique observational niches. These platforms are modified to varying degrees to accommodate either remote sensing or air sampling payloads. The particular measurement requirements, together with budget constraints, indicate the most appropriate platform for a given application. Fundamental aircraft

E.G. Njoku (ed.), Encyclopedia of Remote Sensing, DOI 10.1007/978-0-387-36699-9, © Springer Science+Business Media New York 2014

410

OBSERVATIONAL PLATFORMS, AIRCRAFT, AND UAVS

variables to be considered include operating range and altitude, payload capacity, flight endurance, and cost of operation. There is typically a trade-off with most aircraft between the weight of the payload and the amount of fuel to be carried, which in turn affects the operating range. If the instruments to be flown are relatively lightweight and the projected operating area is small and close to an airfield, a light one- or two-engine private plane is by far the most economical. Typically unpressurized, they usually operate at altitudes below 7,500 m with ranges under 600 km. As the size of the instrument payload and/or the extent of the study area increases, larger and more capable platforms are required. Turbine or turbo-charged twin engine propeller aircraft offers the next level of performance with a significant increase in payload and range, pressurized cabins, and operating altitudes up to 10,000 m. These often offer the best compromise between performance and cost-effectiveness for many applications. Corporate jets or larger commercial transport aircraft, which can fly above 12,000 m and at much higher speeds, can carry heavy payloads and cover even larger areas in a single flight but are relatively expensive to operate. Special-purpose high-altitude platforms, such as the NASA ER-2 (a Lockheed U-2 reconnaissance aircraft derivative) or the Russian M-55 Geophysica, can fly into the lower stratosphere and operate over very large geographical areas. These aircraft are often used for atmospheric research but are still more expensive and require specialized ground support equipment. For remote sensing missions, the very long atmospheric path length below these aircraft makes the atmospheric correction task more problematic. Imaging through a nearly full atmospheric column can be an advantage for the validation of satellite data, however. Flying at 20,000 m altitude, these aircraft are above 95 % of the Earth’s atmosphere, allowing for very accurate at-sensor response comparisons between airborne and orbital systems.

UAVs: unmanned aerial vehicles Also known as Unmanned Aerial Systems (UAS), these include a full range of size classes from very small handlaunched aircraft resembling model airplanes to the large high-altitude observational systems used by military services and some science agencies. The complexities of operation, together with payload and range capabilities, also vary conversely. Most UAVs are remotely piloted by a person on the ground, using a radio link to send commands to the flight control system on the aircraft. However, some are configured to operate autonomously, using an onboard computer with a preprogrammed flight plan, together with a satellite Global Positioning System (GPS) and an inertial guidance system. UAVs offer several significant advantages over manned platforms. For example, they can operate in hazardous environments (e.g., near volcanic activity or in toxic plumes) without endangering a human pilot. Some UAVs also have endurances in excess of 20 h, which enables sustained observations of

an evolving phenomenon. Airships, both rigid dirigibles and nonrigid types, also have this capability. Apart from vehicles designed for completely autonomous operation, some form of communication link is generally required to operate a UAV. This can consist of a simple line-of-sight radio link as used on small model airplanes, which limits the range of operation, or a larger satellite communication (“satcom”) system, which extends the operating range over the horizon. The most elementary satellite communications systems use mobile satellite telephone technology with small omnidirectional antennas and conventional low-speed data modems. The higher data-rate satcom systems found on larger UAVs require large directional antennas and are significantly heavier and more expensive to operate. The smaller UAVs can be very economical overall to operate, while the larger military-type platforms, which use high-speed satellite communications links and complex ground stations with professional pilots, can be considerably more expensive. Regardless of size, however, almost all UAVs are subject to some form of government air traffic control regulations, which can make flight operations – especially over urban areas – highly problematic. In the USA, even the smallest model airplane types are subject to federal airspace restrictions when used for any purpose other than recreation. Access to the local airspace for a UAV should never be assumed and indeed has been a major obstacle to their wider proliferation in the private and scientific sectors.

Payload accommodations Airborne observation platforms require varying degrees of mechanical modification to suit their intended purpose. These can range from simple external instrument mounting fixtures on an otherwise unmodified light aircraft to permanent airframe modifications that may include optical windows, air sampling probes, or external attachment points for sensor pods. Remote sensing aircraft These aircraft are usually fitted with optical camera windows at various angles for nadir, side, or zenith viewing or external pods with similar accommodation. If the windows are fitted in an aircraft with a pressurized cabin, they must be specially engineered with adequate safety margins. Windows in an unpressurized aircraft (typically operating at altitudes less than 3,000 m) are less problematic to install. Commercially available aircraft optical windows typically have wavelength transmissions between 180 nm and 2.5 mm. Infrared windows are also available, although generally in smaller sizes, and are much more expensive. Many infrared instruments instead are designed to view through an open aperture in the aircraft fuselage to avoid window cost and/or spectral transmission issues. In the case of a pressurized aircraft, this open observation port, together with the instrument itself, is enclosed beneath a sealed housing which preserves the

OBSERVATIONAL PLATFORMS, AIRCRAFT, AND UAVS

aircraft cabin pressurization. These installations often include an external sliding door that covers the open aperture to protect the instrument when not in use.

Airborne radar platforms Airborne radar or passive microwave systems require various forms of antennas, which are either integral to the airframe or carried in external pods. Flat-panel phased-array antennas which have no moving parts can be affixed directly to the outside of the aircraft. Mechanically pointed antennas are typically mounted under domes or in external pods made of a material that is transparent to the relevant frequencies. Depending on the particular system, these antennas can be quite large and may require extensive modifications to the platform. They may also have large electrical power requirements, which can require special modifications to the aircraft power system. Air sampling platforms These platforms are equipped with inlets to bring the external ambient air into measurement devices mounted either inside the cabin or in externally mounted pods. These usually consist of external probes or booms on the aircraft nose, fuselage, or wings, which extend the inlets into the free airstream. The inlet probes must extend beyond the boundary layer of the platform and be otherwise free of any turbulence or aerodynamic effects from the aircraft itself, to ensure unbiased sampling of free air at ambient pressure. The location of the probes on the aircraft is usually determined by modeling the airflow around the particular fuselage, under the airspeed and attitude parameters (pitch, roll, yaw) expected during sampling operations. Data systems Observation platforms are also often fitted with ancillary data systems to augment the primary instrument measurements. These may include the basic recording of aircraft state parameters, such as time, navigational position and attitude, altitude, and heading, as well as external air temperatures and pressures. These data are usually provided by some combination of satellite GPS and the aircraft’s own Inertial Navigation System (INS) or Air Data Computer. For applications requiring very precise knowledge of platform attitude and location, such as the precision geo-location of image data, or for quantitative air sampling, stand-alone Inertial Measurement Unit (IMU) and differential GPS units may be used. In the air sampling case, this may be coupled with external pressure transducers, which allows for the derivation of external winds around the platform. Other considerations The operating environment for instruments onboard an aircraft is typically demanding and may impose stresses on the equipment that are not always conducive to good

411

scientific measurement. Outside air temperatures vary radically with altitude, often causing instruments to cool considerably unless mounted inside a pressurized cabin. Those with thermal infrared channels viewing through an open port are particularly subject to severe environmental stresses, including temperature cycling, water condensation, and optical contamination. The use of heaters to maintain the temperature of critical electrical and optical components above the local dew point, and hermetic sealing of optics and spectrometers wherever possible, has been proven to increase data quality as well as the longevity of the systems themselves. The deposition of contaminants onto glass surfaces (lenses, windows, etc.) can also compromise measurement quality, so regular cleaning of these surfaces is necessary. The electrical power found on many commercial aircraft, usually 28 V Direct Current, together with the associated electrical grounding circuitry, can be highly variable in quality. Instruments should incorporate internal electrical power conditioning whenever possible to mitigate adverse effects. Mechanical vibration is another potential problem source for delicate instrumentation and each platform has its own vibration frequency spectrum which changes with engine power and control surface settings. Excessive vibration can affect the optical alignment of spectrometers, induce microphonic noise in infrared detectors, and cause the failure of electrical connections. The G-forces associated with landings and in-flight turbulence can also be substantial. Some form of vibration isolation between instrument and airframe, tailored to the weight of the instrument and the specific vibration regime of the platform, is often indicated. In general, instrumentation should be ruggedized to obtain consistent results in the airborne environment.

Summary Airborne observational platforms consist of a variety of types, including conventional and unmanned aircraft, and have a broad range of technical capabilities and operating costs. Aircraft used for Earth observing require varying degrees of modification, depending on their intended applications. This entry provides a brief overview of platform types, mission-specific modifications, and their relevant operating envelopes. Bibliography Cox, T., Nagy, C., Skoog, M., and Somers, I., 2004. Civil UAV capability assessment. Suborbital Science Program Internal Document. NASA. Fladeland, M., and Schoenung, S., 2007. NASA Earth science requirements for suborbital observations. Airborne Science Program Internal Document. NASA. Henderson, F. M., and Lewis, A. J. (eds.), 1998. Manual of Remote Sensing. New York: Wiley. Principles & Applications of Imaging Radar, Vol. 2. Jackson, M. (ed.), 2009. Manual of Remote Sensing. New York: Wiley. Earth Observing Platforms & Sensors, Vol. 1.1.

412

OBSERVATIONAL SYSTEMS, SATELLITE

King, M., Menzel, W. P., Grant, P. S., Myers, J. S., Arnold, G. T., Platnick, S. E., Gumley, L. E., Tsay, S. C., Moeller, C. C., Fitzgerald, M., Brown, K. S., and Osterwisch, F. G., 1996. Airborne scanning spectrometer for remote sensing of cloud, aerosol, water vapor, and surface properties. Journal of Atmospheric and Oceanic Technology, 13(4), 777–794. Kramer, H. J., 1996. Observation of the Earth and Its Environment, 3rd edn. New York: Springer. McDonnell Aircraft Company, Marketing Division, 1982. Reconnaissance Handy Book. St. Louis: McDonnell Aircraft. Miller, R., Del Castillo, C., and McKee, B. (eds.), 2005. Remote Sensing of Coastal Aquatic Environments. Dordrecht: Springer. Pepi, J. W., 1994. Fail-safe design of an all BK-7 glass aircraft window. In Window and Dome Technologies and Materials IV, SPIE, Vol. 2286, September 28, 1994. Rencz, A. N., and Ryerson, R. A. (eds.), 1999. Manual of Remote Sensing. New York: Wiley. Remote Sensing for the Earth Sciences, Vol. 3. Stefanutti, L., and Sokolov, L., 1999. The M-55 geophysica as a platform for the airborne polar experiment. Journal of Atmospheric and Oceanic Technology, 16(10), 1303–1312. Ustin, S., 1999. Manual of Remote Sensing. New York: Wiley. Remote Sensing for Natural Resource Management and Environmental Monitoring, Vol. 4.

Cross-references Remote Sensing, Physics and Techniques

OBSERVATIONAL SYSTEMS, SATELLITE David L. Glackin Los Angeles, CA, USA

Acronyms ADEOS ADM AIM ALISSA ALOS AMI ASCAT ATSB CALIPSO CAST CBERS CERES CHAMP CHRIS

Advanced Earth Observing Satellite Atmospheric Dynamics Mission Aeronomy of Ice in the Mesosphere l’Atmosphere par Lidar Sur SAliout Advanced Land Observing Satellite Active Microwave Instrument Advanced Scatterometer Astronautic Technology Sdn Bhd (Malaysia) Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation Chinese Association for Science and Technology China-Brazil Earth Resources Satellite Clouds and the Earth’s Radiant Energy System CHAllenging Minisatellite Payload Compact High-Resolution Imaging Spectrometer

David L. Glackin: deceased.

CNES COMS

Centre National d’Etudes Spatiales Communications, Oceans and Meteorology Satellite CONAE National Commission for Space Activities (Argentina) COSMIC Constellation Observing System for Meteorology, Ionosphere, and Climate COSMO Constellation of small Satellites for Mediterranean basin Observation CSA Canadian Space Agency CSIRO Commonwealth Scientific and Industrial Research Organisation DARA German Space Agency DLR German Aerospace Center DMC Disaster Monitoring Constellation DMSP Defense Meteorological Satellite Program DoD Department of Defense EADS European Aeronautic and Space Company EarthCARE Earth Clouds, Aerosols and Radiation Explorer EO-1 Earth Observer-1 EOS Earth Observing System ERBS Earth Radiation Budget Satellite EROS Earth Resources Observation Satellite ERS European Remote Sensing ESA European Space Agency ETM+ Enhanced Thematic Mapper Plus EUMETSAT European Organisation for the Exploitation of Meteorological Satellites FEDSAT Federation Satellite FY Feng Yun GCOM Global Change Observation Mission GEO Geostationary Earth Orbit GER Geophysical and Environmental Research Corporation GERB Geostationary Earth Radiation Budget GFZ German Research Centre for Geosciences GISTDA Geo-Informatics and Space Technology Development Agency (Thailand) GLM GEO Lightning Mapper GMES Global Monitoring for Environment and Security GOCE Gravity field and steady-state Ocean Circulation Explorer GOES Geostationary Operational Environmental Satellite GOMS Geostationary Orbit Meteorological Satellite GPM Global Precipitation Measurement GRACE Gravity Recovery and Climate Experiment HJ Huan Jing HY Hai Yang ICESat Ice, Cloud, and land Elevation Satellite IJPS Initial Joint Polar System INPE National Institute of Space Research (Brazil) IRS Indian Remote Sensing ISRO Indian Space Research Organisation

OBSERVATIONAL SYSTEMS, SATELLITE

JAXA JPL KARI KOMPSAT LDCM LEO LIS LITE MADRAS MDA MEO MetOp MIS MOMS MSG MTG NASA NMP NOAA NPOESS NPP NRL NSPO NTU OCO OLI OMI OSTM PARASOL POES PROBA RDL RISAT ROCSAT RSSS SAC SAOCOM SAR ScaRab SeaWiFS SIR SMAP SMOS SNSB SPIN-2 SPOT SRTM SSTL

Japanese Aerospace Exploration Agency Jet Propulsion Laboratory Korea Aerospace Research Institute Korea Multi-Purpose Satellite Landsat Data Continuity Mission Low Earth Orbit Lightning Imaging Sensor Lidar In-space Technology Experiment Microwave Analysis and Detection of Rain and Atmospheric Structures MacDonald, Dettwiler and Associates, Ltd. Medium Earth Orbit Meteorological Operational Microwave Imager/Sounder Modular Optoelectronic Multispectral Scanner Meteosat Second Generation Meteosat Third Generation National Aeronautics and Space Administration New Millennium Program National Oceanic and Atmospheric Administration National Polar-orbiting Operational Environmental Satellite System NPOESS Preparatory Project Naval Research Laboratory National Space Organization (Taiwan) Nanyang Technological University (Singapore) Orbiting Carbon Observatory Operational Land Imager Ozone Monitoring Instrument Ocean Surface Topography Mission Polarization and Anisotropy of Reflectances for Atmospheric Sciences coupled with Observations from a Lidar Polar-orbiting Operational Environmental Satellite Project for On-Board Autonomy Research and Development Laboratory Radar Imaging Satellite Republic of China Satellite Remote Sensing Satellite System Satelite de Aplicaciones Cientificas Satellites for Observation and Communication Synthetic Aperture Radar Scanner for Radiation Budget Sea-viewing Wide-Field Sensor Shuttle Imaging Radar Soil Moisture Active Passive Soil Moisture Ocean Salinity Swedish National Space Board Space Information-2 Systeme Pour l’Observation de la Terre Shuttle Radar Topography Mission Surrey Satellite Technology Ltd.

THEOS TOMS TOPEX TPFO TRMM TUBITAK UAV UCAR USGS ZY

413

Thailand Earth Observation System Total Ozone Mapping Spectrometer TOPography EXperiment for Ocean Circulation TOPEX/Poseidon Follow-On Tropical Rainfall Measuring Mission The Scientific and Technological Research Council of Turkey Uninhabited Aerial Vehicle University Corporation for Atmospheric Research United States Geological Survey Zi Yuan

Synonyms Earth observation satellites; Earth observation systems; Remote sensing satellites; Remote sensing systems Definition Satellite Observational System. A system for the study of the Earth and its environment (atmosphere, oceans, ice, land surface, solid Earth, and near-Earth space environment) from space. The end-to-end system includes (1) the spacecraft (the platform that hosts the remote sensing instruments and provides them with pointing, power, commands, data storage, etc.), (2) the remote sensing instruments (often called “sensors”) that observe and measure the characteristics of the Earth’s environment, and (3) the ground-based mission operations. The term “satellite” is usually used for the combination of the spacecraft and instruments, but in some communities, the term “spacecraft” is used for that combination, with the term “satellite” indicating the platform. Introduction to satellite observational systems Although the first weather satellite, TIROS I, was launched in 1960, the field of satellite-based remote sensing of Earth really began to take form in the 1970s (see Remote Sensing, Historical Perspective). The launches of Landsat-1 in 1972, Skylab in 1973, Nimbus-7 in 1978, and Seasat in 1978 set the stage for modern environmental remote sensing (Glackin, 2004). Since the pioneering work of the 1970s, the field of satellite environmental remote sensing has steadily evolved. Before 1990, only six nations owned environmental satellites (China, France, India, Japan, the USA, and the former USSR), but since then, the number has quintupled. The new nations (Glackin and Peltzer, 1999; Kramer, 2002; eoPortal, 2008) include the ones in Africa (Algeria, Morocco, and Nigeria), Asia (Indonesia, Malaysia, South Korea, Taiwan, Thailand, and Ukraine), Europe (Belgium, Germany, Italy, Portugal, and the UK), the Middle East (Egypt, Iran, Israel, Pakistan, and Turkey), North America (Canada), Scandinavia (Sweden), and South America

414

OBSERVATIONAL SYSTEMS, SATELLITE

(Argentina, Brazil, and Chile), as well as the country of Australia. Many of these nations have procured their satellites from other nations, as described below. Pre-1990 programs primarily involved systems of high cost and complexity, but post-1990 the focus shifted to include missions involving smaller satellites of lower complexity and cost and greater commercial involvement, including hybrid government/commercial systems (see Public-Private Partnerships) and fully commercial systems (see Commercial Remote Sensing). Examples of hybrid government/commercial systems are the US Landsat, the French SPOT (Systeme Pour l’Observation de la Terre), and the Canadian RADARSAT satellites. In such hybrid programs, industry is expected to share a portion of the development costs and to recoup their investment by developing a commercial market for the data products (see Cost Benefit Assessment). Fully commercial systems have met with some success, although many efforts have been canceled. These include an Australian hyperspectral (see Remote Sensing, Physics and Techniques) mining satellite called ARIES, satellites for precision agriculture called GEROS from Geophysical and Environmental Research Corp. (GER) and Resource21 from Boeing/GDE Systems, the “One-Meter System” from GDE Systems, the Radar-1 SAR (Synthetic Aperture Radar) system from Research and Development Laboratory (RDL), and AVStar, a system from AstroVision to provide color imagery of the Earth from GEO (Geostationary Earth Orbit) (see Glackin and Peltzer, 1999.) The five latter systems were all from the USA. UoSAT-5, launched by the British company Surrey Satellite Technology Ltd (SSTL; SSTL, 2008) in 1991, was the first commercial microsat. OrbView-1, launched by the US company ORBIMAGE (now part of GeoEye) in 1995, was the first commercial weather satellite (Glackin and Peltzer, 1999). In the 1990s, governments also pioneered the concept of buying end-user data products rather than satellites. The first example involved a NASA (the National Aeronautics and Space Administration) data buy of data from OrbView-2/SeaStar, the first satellite launched to provide commercial ocean color imagery in 1997 (GeoEye, 2008). Commercial high-resolution remote sensing imagery (see Commercial Remote Sensing) became a reality when the former USSR began selling 5 m resolution panchromatic (see Remote Sensing, Physics and Techniques) imagery in 1987, followed by Russian sales of 2 m resolution panchromatic imagery in 1992 under the Kosmos/SPIN-2 (Space Information-2) program. Both were digitized from returned film. The US industry felt that SPIN-2 could be the harbinger of a lucrative market that might expand internationally, from which they would be excluded because of existing restrictions on imagery resolution (NASA, 2007). As a result of intense lobbying by industry, in 1994, the first licenses were granted allowing commercial US systems with a resolution of 1 m to be built and flown. Still, in 1998, the highest-resolution satellite imagery commercially

available was from SPIN-2 and from India’s IRS-1C and -1D (IRS ¼ Indian Remote Sensing) satellites with 5.8 m panchromatic resolution (Jacobsen, 2008). The first commercial high-resolution remote sensing satellite with 1 m resolution, IKONOS, was launched by the US company Space Imaging (now part of GeoEye; GeoEye, 2008) in 1999, followed by others of similar resolution from several countries (India, Israel, Russia, and South Korea). Those in turn were followed in 2007 by highresolution SAR systems from Canada (RADARSAT-2 with 3 m resolution) and Germany (TerraSAR-X with 1 m resolution). The commercial high-resolution satellites will be discussed further below. Much of their success has relied on the existence of governments as “anchor-tenants,” i.e., principal customers. Sales to the commercial sector have not proven to be as viable as many practitioners envisioned in the 1990s. Beginning in 1991, remote sensing satellites in the microsatellite (microsat) category (10–100 kg) were launched, followed in 1997 by satellites in the minisatellite (minisat) category (100–500 kg). The first such minisat was NASA’s TOMS/Earth Probe carrying the Total Ozone Mapping Spectrometer (TOMS/EP, 2008). These microsats and minisats have enabled much of the international proliferation in this field. Many of the newer countries involved in this field have used the technological know-how of other countries to build much of the space-based hardware. They have used foreign-partnership programs for technology transfer to build up an indigenous capability for constructing the hardware, usually to complement their existing capability in analyzing remote sensing data. A leader in the technology transfer field for nearly 20 years has been the very successful British company SSTL (SSTL, 2008). They have provided microsats to countries including Algeria (Alsat-1/DMC), Chile (FASat-Alfa and Bravo), China (Beijing-1/DMC), Nigeria (NigeriaSat-1/DMC), Portugal (PoSat-1), South Korea (KITSAT-1 and -2), Thailand (TMSat), and Turkey (BilSat-1/DMC). DMC stands for Disaster Monitoring Constellation, an effort in international collaboration established by SSTL for the monitoring of disasters from space using simple multispectral cameras. The former TRW in the USA provided minisats to South Korea (for KOMPSAT-1; Korea Multi-Purpose Satellite-1; KOMPSAT, 2008) and Taiwan (for ROCSAT-1/Formosat-1; Republic of China Satellite-1), used principally for ocean color monitoring (ROCSAT, 2008). Coupled with the trend toward more small satellites, especially in the minisat class, has been a desire for downsized but highly capable remote sensing instrumentation. Smaller, lighter instruments mean smaller, lighter spacecraft to support them, which means smaller launch vehicles, all equating to lower cost (see Mission Costs of Earth-Observing Satellites). The international proliferation of small satellites has certainly not meant that large satellites have gone out of fashion. In 2002 alone, three high-mass satellites

OBSERVATIONAL SYSTEMS, SATELLITE

(eoPortal, 2008) were launched [SPOT-5 from France at 3,000 kg, ADEOS-2 (Advanced Earth Observing Satellite-2) from Japan at 3,700 kg, and Envisat from ESA at 8,000 kg]. Many more medium-to-large (>500 kg) satellites have been launched and are planned. While many systems hosting multispectral instruments have been launched and are operating successfully, relatively few hyperspectral instruments have been launched. Following some failures and cancelations, the first success was NASA’s NMP (New Millennium Program) EO-1 (Earth Observer-1) satellite in 2000, carrying the Hyperion instrument with 220 spectral bands (EO-1, 2008). This was followed by the PROBA (Project for On-Board Autonomy) technology demonstration satellite from the UK and Belgium in 2001, carrying the CHRIS (Compact High-Resolution Imaging Spectrometer) hyperspectral imager (CHRIS, 2008). CHRIS data have received broad application. In 2004, NASA’s Earth Observing System (EOS) Aura satellite hosted the hyperspectral Ozone Monitoring Instrument (OMI) from the Netherlands and Finland (Aura, 2008). It observes ozone in the ultraviolet (UV) from 0.27 to 0.38 mm (see Remote Sensing, Physics and Techniques). Future hyperspectral instruments have been announced by Argentina/Brazil (HSI, 2007), Germany (EnMAP, 2008), India (IMS-1, 2008), Italy (Hypseo, 2007), and South Africa (ZASat-2, 2008). Hyperspectral imagers can quickly generate gigabytes of data and provide a particular challenge for satellite data handling and transmission systems (see Mission Operations, Science Applications/Requirements and Mission Costs of Earth-Observing Satellites). The application of lidar (see Remote Sensing, Physics and Techniques) to spaceborne remote sensing remains a fledgling field. The first lidar to fly in space designed to perform remote sensing of the Earth’s atmosphere was NASA’s LITE (Lidar In-space Technology Experiment) that flew on the shuttle in 1994 (Kramer, 2002). The Russian Balkan-1 lidar was launched on the Mir-Spektr module in 1995 (Kramer, 2002), while the French ALISSA lidar was launched on the Mir-Priroda module in 1996 (Kramer, 2002) to study clouds, aerosols, and the Earth’s surface (ALISSA ¼ l’Atmosphere par Lidar Sur SAliout). Currently operating lidars are on the US/French CALIPSO (Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation) satellite, launched in 2006, with 100 m laser spot size at the Earth’s surface and NASA’s EOS ICESat (Ice, Cloud, and land Elevation Satellite), launched in 2003, with 70 m spot size. Planned spaceborne lidars are ESA’s in 2008, to measure the tropospheric wind profile for the first time; the European/ Japanese EarthCARE (Earth Clouds, Aerosols and Radiation Explorer) satellite, to fly in 2013, with N1) for the radiation in water, yielding 2(N2–N1) terms of the total reflecting region. Thus, the solution of the discrete ordinate method applied to the atmosphere (superscripted with a) can be written for the pth layers as:

2Zrsur ðZ; z; jÞ ¼ AðzÞyðZ; z; jÞ

yðZ; z; jÞ ¼ R 2p 0

4pZrsur ðZ; z; jÞ R1 dj 0 rsur ðZ; z; jÞZdZ

(103)

It should be noted that replacing surface phase function wsur ðZ; z; jÞ with the appropriate probabilistic functions yðZ; z; jÞ is not quite correct because the transformation, although having some advantages, simultaneously violates the fundamental physical properties of reciprocal angular symmetry wsur ðZ; z; jÞ ¼ wsur ðz; Z; jÞ. Thus, the input optical models for the atmosphere we can enter into the analysis of appropriate informational levels are connected with the parameterization and structuring of surface optical quantities. The parameterization of the bidirectional surface reflectance is given the minimization’s condition: XXX jyðZi ; zj ; jk Þ  ~yðZi ; zj ; jk Þj2 ¼ s2 mini;j;k i

j

k

(104) Afterward, the previous procedures of structuring, due to the use of Legendre’s polynomials series expansion and Fourier series expansion, can be repeated for the discussed surface case without any great difficulty. The same remarks hold for the procedure of calibration and the sum of structured elements. The main point is the necessity of connecting the total number of the coefficients and functions calculated separately for the atmosphere and underlying surfaces. Thus, as results we have input a unified optical model for an atmosphere-surface system with four levels of informational content, which is needed to accurately define successive radiation fields modeling.

Discrete ordinate method The discrete ordinate method isolates the azimuth dependence of Eqs. 5 and 6, expanding the phase function into the Legendre polynomial. Using the addition theorem for spherical harmonics and taking into account the orthogonality properties of the Legendre polynomials, we obtain that the phase function can be written as: wðt; gÞ ¼

2N 1 X m¼0

ð2  d0m Þpm ðt; Z; zÞcos mðj  j0 Þ (105)

Ipm
ðt;
Zai Þ ¼

N1 h X j¼1

a Cjp gjp ð
Zai Þexpðkjpa tÞ

i a ð
Zai Þexpðkjpa tÞ þ Up ðt;
Zai Þ þCjp gjp

(107) where i ¼ 1, . . . N1 and p  N1. The same is true for the pth layers in water (superscripted with w): Ipm
ðt;
Zwi Þ ¼

N1 h X j¼1

w Cjp gjp ð
Zwi Þexp ðkjpw tÞ

i w þ Cjp gjp ð
Zwi Þ exp ðkjpw tÞ þ Up ðt;
Zwi Þ

(108) where i ¼ 1, . . . N2 and N1 < p  N1 + N2. The eigenvalues kjpa, kjpw and the eigenvectors gjpa, gjpw are determined solving the homogeneous associate of Eq. 19 with the source terms put equal to zero. The term Upm(t,
Zi) is the particular solution. The coefficients C
jp are determined by: 1. The boundary conditions at the top of the atmosphere and at the bottom of the ocean

RADIATIVE TRANSFER, SOLUTION TECHNIQUES

2. Radiation continuity conditions at each interface between the layers 3. Fresnel’s equations at the atmosphere water interface This leads to a system of linear algebraic equations: AX ¼ B

(109)

whose solution is X ¼ A1 B. Here, A contains information about the properties of the optical system and the bidirectional reflectance distribution function (BRDF). The order of A is defined by the unknown coefficient C
jp. The column vector X contains information on C
jp, the matrix B contains information about the particular solutions as well as the lower boundary emissivity. Since the phase function is the sum of Rayleigh (molecular) scattering plus the Mie scattering due to the effect of aerosol in atmosphere, an improved solution was proposed by T. Nakajima and Tanaka (1981), which computes the Rayleigh scattering exactly. The theoretical approach can be found in Levoni et al. (2001). This approach improves the accuracy of the solution and reduces the computational time, which largely depends on the number of momenta used for Lagrange polynomials. Spurr (2001) has developed a solution based on the recognition that if the integral properties of one of the layers in a system are perturbed, there is no need to repeat the entire computation from scratch. By its analysis, he deduced that the solution of the problem can be defined by a matrix form of the type: AX0 ¼ B0

(110)

In this relation, which is similar to the previous one Eq. 109, the vector column B depends on the layer parameters that have been changed. Since the matrix A, unperturbed, is solved by LU decomposition, the solution of Eq. 110 can be found by back-substitution using the LU factorized form of A. The Spurr model, named LIDORT, also contains perturbations in layer thermal emission as well as surface albedo. The first radiative transfer model, LOWTRAN (Kneizys et al., 1988), and the successive version, MODTRAN (Berk et al., 1989), use the discrete ordinate methods, even with a limited number of momenta of Legendre polynomials. The success of the discrete ordinate methods is due to their flexibility, at least in the frame of remote sensing, so various other authors have adopted this approach, dressing the solutions with the optical features of the atmosphere and the surface to obtain a suitable model. These include the SBDART (Ricchiazzi et al., 2007), libRadtran (Mayer et al., 2005), and, in the IR, HARTCODE (Miskolczi et al., 1998).

Statistical modeling or Monte Carlo methods Monte Carlo methods provide approximate solutions to the radiative transfer equation by sampling the trajectories

619

of a large number of photons. There are at least three Monte Carlo methods: 1. Forward with bin averaging, where the photons are averaged over various chosen ranges of solid angles. 2. Forward with discrete angles, where the radiation is estimated at each collision for particular predetermined angles. 3. Backward or adjoint with discrete angles, where the photons start at particular angles at the sensor and retrace back the path. Such a method is the most useful for spherical atmospheres. In solar radiation problems, the photons are introduced with random initial position at the boundary of the domain. Their directions are chosen according to the illumination conditions. The photons then begin their flight. In an efficient Monte Carlo program, a statistical weight is associated with each photon. In practice, a sample transmission is chosen uniformly between 0 and 1. The photon travels along its path until the cumulative extinction along the photon’s path reaches the value that corresponds to this transmission. The photon experiences a scattering event in which the scattering angle is determined by choosing a random number and comparing this to the cumulative phase function. After a photon’s new direction is determined with a new transmission, this process repeats until the photon leaves the domain or is absorbed. During scattering the absorption is normally modeled by multiplying the photon weight (initially 1) by the single scattering albedo of the medium doing the scattering. At each collision this weight is multiplied by the ratio of the total scattering cross section to the total cross section for all processes, corrected for the absorption probability. The photon trajectory is terminated when the statistical weight falls below a preassigned value (usually 105). In highly variable media, a technique called maximum cross section is often used to reduce the per-photon variance of flux. The cumulative extinction at each step is scaled by the maximum extinction within the domain, and the likelihood of scattering at the end of each step depends on the ratio of the local to the maximum extinction. Radiative quantities are computed by these weighted photons. The atmosphere may contain more than one optically active component (e.g., aerosols, gases, and clouds often coexist). The extinction accumulated along the photon trajectory is the extinction due to all components. Scattering may be treated either by averaging the single scattering phase functions (weighted by the extinction and single scattering albedo) or by choosing the phase function for a single component based on the relative amounts of extinction. A technique called Russian roulette can help to speed up calculations in absorbing media. At each scattering event, the photon weight is compared with a random number. If the weight exceeds the random number, the photon trajectory is terminated, otherwise the photon weight is reset to 1. In Cornish-Gilliflower release, for instance (http://code.google.com/p/i3rc-monte-carlo-model/), the

620

RADIATIVE TRANSFER, SOLUTION TECHNIQUES

implementation Russian roulette is performed only when the photon weight is less than one-half. Intensity is computed using a technique called local estimation. At each scattering event, the total extinction to the boundary from the scattering location is computed in each direction at which intensity is desired. The intensity at the exiting location is then incremented by the product of the transmission along that direction, the phase function of the scattering particle evaluated at the angle between the direction the photon is traveling and the intensity direction, and the photon weight. The phase functions of large particles (i.e., those much larger than the wavelength of light being simulated) have large forward peaks due to diffraction. These peaks may be orders of magnitude larger than the phase function, only a few degrees off-axis. If the forward peak happens to align with one of the directions at which intensity is being computed, the contribution from an individual scattering event can be very large, leading to large variance between calculations (and hence large error estimates). In order to reduce this variance, the phase function used for local estimation can be replaced by a hybrid that replaces the original phase function at small angles by a Gaussian of user-specified width. The hybrid phase function is constructed to be continuous and properly normalized.

From 1D to 3D Monte Carlo Let us now apply the previous concepts to a homogeneous atmosphere with extinction coefficient b(r) and phase function P(r, o · o0 ). One can calculate the 1D Monte Carlo radiative transfer using a random number generator (numerical algorithm) to produce the random numbers, rn, between 0 and 1 with a probability distribution function pdf ¼ 1. From rn, one can generate another set of random numbers as rx ¼ ln(rn) with pdf ¼ exp(rx) and rx between 0 and ?. Then, in order to determine the trajectory of the photons, one can: 1. Determinate the starting position of a photon given by x0 with direction Z, j. 2. Generate a photon path length using random numbers x ¼ rx using the random numbers rx. 3. Calculate a new photon position x0 + x called the event point (or collision point). 4. Analyze what can happen with the photon at this event point by generating the random number rn and comparing it with the single scattering albedo L. If rn > L, then the photon is absorbed and the cycle returns to point 1 for a new photon. If rn < L, the photon is scattered and the process continue to the next step. 5. Find a new direction for the scattered photon using the phase function to calculate the cumulative probability function to relate the scattering angle to a random number. 6. Repeat, starting from the point 3, until all photons are analyzed.

Let us consider now the inhomogeneous atmosphere. We can split it into a homogeneous P grid. The point of interaction (or event point) rx ¼ Dj Dj bðrÞ where Dj ¼ step 1 size in each grid and the free path length is l ¼ bðrÞ . In order to estimate the accuracy of the method with respect to the number of photons, one needs to estimate the Monte Carlo error given by the expectation and variance of random variables. They can be computed by the obtained number N of independent values xi(i ¼ 1, 2, 3 . . . N) of a random variable rn, each weighted by the probability with which the value is taken. A detailed analysis is described in Marshak and Davis (2005). In order to obtain a 3D radiative transfer equation, one needs to reconvert the plane-parallel radiative transfer Eqs. 5 and 6 into one equation, taking into account three spatial dimension (x, y, and z) and two angular dimensions (y,j). This procedure is different than the one we have used in the plane-parallel system previously described, where the azimuth angles and zenith were, respectively, represented by the Fourier series and discrete ordinates. Since in 3D radiative transfer the spatial dimensions cannot be solved analytically, as the optical depth is in plane-parallel transfer, they must be solved in other way. One way is to use the so-called spherical harmonics discrete ordinate method (SHDOM) developed by Pincus and Evans (2009). We do not go into such a method because our purpose here is to describe the solution for the Monte Carlo method. Let us now assume a cylinder to be infinitesimally small and located somewhere in the atmosphere. It is oriented along an axis o that is not pointing directly at the target. Rewrite now the integrodifferential radiative transfer radiation Eq. 5 for an unpolarized beam that describes the change of radiation I(r, o) into direction o as b ðr; lÞ oHIðr; oÞ ¼  bðrÞIðr; lÞ þ s 4p Z 0 Iðr; o ÞPðr; o o0 Þdo0

(111)

4p

þ Sðr; oÞ where o is the direction vector with component (sinycosj, sinysenj, cosy) and y and j are zenith and azimuth angles; b(r) is the extinction coefficients; and bs(r) is the scattering coefficient. Let us rearrange this equation into an integral equation:   1 b ðrÞ 1 oH Iðr;oÞ ¼  s 1þ bðrÞ bðrÞ 4p Z Iðr;o0 Þ Pðr;o o0 Þdo0 4p

þ Sðr;oÞ

1 bðrÞ (112)

RADIATIVE TRANSFER, SOLUTION TECHNIQUES s ðrÞ where the ratio bbðrÞ is the single scattering albedo. Such integral equation can be solved with the aid of Green’s function. In fact, let us transform it into a simple form:

Lx f ðxÞ ¼ uðxÞ

(113)

where

  1 d Lx ¼ 1 þ aðxÞ dx

is the operator of the left part of the equation, f (x) is the radiance, and u(x) is the right part of Eq. 112, continuous function in the D domain selected. The extinction coefficient b(r) has been replaced by a(x). In order to solve the Eq. 113, we look for a function g: ~ Cn(D) ! C(D) such that L(g(h)) = h where y(t) N ¼ g(h(t)). This is a convolution equation of the form y ¼ g h so that the solution is Z b yðtÞ ¼ gðt  xÞ hðxÞdx

with the optical density t(r, r0 ) between r and r0. We obtain the convolution for by integrating x 2 [a,b] where b is the distance from r0 to the boundary of the medium along the direction o. Then,   rr0 d o  jrr 0j Go ðr; r0 Þ ¼ bðr0 Þ exp ðtðr; r0 ÞÞ jr  r0 j2 (118) The d function selects r that fits to the direction o and the reference point r0. Then, Eq. 112 can be written in terms of the collision density f(r, o) ¼ b(r)I(r, o) by the convolution with Go(r, r0 ) and multiplication with b(r): Z Z Lðr0 Þ Pðr0 ; o o0 Þ f ðr; oÞ ¼ 4p A

a

~g ðtÞ ¼ dðtÞ L

0

0

 expðtðr; r ÞÞ

Let us assume now that Z b f ðxÞ ¼ kðx; x0 Þuðx0 Þdx0

(119)

where C(r, o) ¼ Cthermal(r,o) + Csolar(r,o) Z Cthermal ðr; oÞ ¼bðrÞ ba ðr0 ÞBl ðT ðrÞÞ A

so the Green’s function g(t) can be defined as

rr dðo  jrr Þ 0 j2

jr 

r 0 j2

dr0 (120)

(114)

a

where x ¼ a is the position at the boundary along –o. Applying the previous assumptions to Eq. 113, we obtain aðxÞkðx; x0 Þ þ

4p

 f ðr0 ; o0 ÞbðrÞexpðtðr; r0 ÞÞ   rr0 d o  jrr 0 j2 do0 dr0 þ Cðr; oÞ  2 0 jr  r j

a

~ on D. Note that where g(t) is the Green’s function for L if we take h(t) ¼ d(t) Z b gðt  xÞdðxÞdx ¼ gðtÞ yðtÞ ¼

621

d kðx; x0 Þ ¼ aðxÞdðx  x0 Þ dx

for x 6¼ x0 the right side is zero and, therefore,  Z x  0 0 aðtÞjdtj kðx 6¼ x ; x Þ ¼ c: exp  x0

(115)

(116)

with c is a constant in the regions x < x0 and x > x0 where the solution is expected to be continuous. The d function implies a step function in c for which the Green’s function of the operator Lx is 8 > > > > <

0 if x < x0  Z x  0 0 Gw ðx;r0 Þ bðr0 Þ exp  bðr þ ðt  x ÞoÞdt otherwise > > x0 > |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl ffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} > : tðr;r0 Þ

(117)

and Z

Ysun ðr0 Þ   rr0 d o  jrr 0j  expðtðr; r0 ÞÞ dr0 2 0 jr  r j (121)

Csolar ðr; oÞ ¼ F0 ðlÞbðrÞ

A

Then the transformation of the radiative transfer equation (RTE) into an integral equation (IRTE) yields Z Z f ðr; oÞ k½ðr0 ; o0 Þ !ðr; oÞ (122) A 4p 0 0 0 0 f ðr ; o Þdo dr þ Cðr; oÞ where the transport kernel k is k½ðr0 ; o0 Þ ! ðr;oÞ ¼ kp ½ðr0 ;o0 Þ ! ðr; oÞ

  rr0 d o  jrr 0j jr  r0 j2 (123)

622

RADIATIVE TRANSFER, SOLUTION TECHNIQUES

with the transport transition density kp ½ðr0 ; o0 Þ ! ðr; oÞ given by Lðr0 Þ Pðr0 ; o o0 Þf 4p ðr0 ; o0 ÞbðrÞ exp ðtðr; r0 ÞÞ

kp ½ðr0 ; w0 Þ ! ðr; wÞ ¼

(124) which can be interpreted as a probability density of photons to perform the transition ðr0 ; o0 Þ ! ðr; oÞ. The factor k and kp can be found by integration over the medium. Let us rewrite Eq. 122 in operator form as f ¼Kf þC

(125)

whose solution can be represented as a Neumann series (i.e., by the order of scattering): f ¼ C þ KC þ K 2 C þ K 3 C þ ¼

? X

K nC

n¼0

(126) which converges for all bounded media with 0  L  1. If the forward Monte Carlo for radiative transfer is based on IRTE, the backward Monte Carlo is based on the principle of reciprocity that describes the radiative process in the reverse direction. Applying the time reversal operator, that is, introducing the minus sign before the direction o in Eq. 112, we obtain the so-called adjoint RTE containing the adjoint transport operator K+. Thus, the solution for an adjoint integrodifferential equation with delta function source and homogeneous boundary condition is f þ ¼ K þf þ þ C

(127)

where C is the detector function. The solution of Eq. 127 can be found as for forward time direction: fþ ¼

? X

K þZ C

(128)

Z¼0

showing the solution of the RTE in the form of Newman series can be expressed by nested integrals. Monte Carlo methods are ideal for computing numerical solutions of previous integrals. Considering a natural event, the set of possible results can be continuous or discrete. The probability of an event can be measured using the so-called probability density function pdf, while the cumulative distribution function cdf is the integral of pdf. All the terms we have introduced, that is, the phase function and the free path length, can be computed using the pdf and cdf from which it is also possible to derive the expectation value, higher momentum, variance, and standard deviation of a random variable. The physics of radiation transport is merged with stochastic methods to create

numerical algorithms that estimate the collision density of the radiance. The methods described are based on the principle of reciprocity, either to simulate the natural process in forward direction, the photon source in the sun, or to describe the backward or adjoint process, the trajectory originating at the detector’s position. In this case, the initial distribution of propagation corresponds to the field of view (FOV) of the detector. One key for the implementation of Monte Carlo simulation is the ray tracer simulating the propagation of light. A ray tracer uses random numbers to decide over the occurrence of events like extinction of light, scattering, and absorption, according to the local physical conditions. The trajectory is completed when the photon is absorbed or leaves the atmosphere. Monte Carlo methods are exhaustively described in books by Marchuck et al. (1980) and Marshak and Davis (2005). A series of different papers deals with different aspects of Monte Carlo methods; among the others we cite Iwabuchi (2006) and Baker et al. (2003), which describe numerical techniques to reduce the computational demands of Monte Carlo simulations; O’Hirok and Gautier (1998), which describes techniques for broadband Monte Carlo simulations; and Cornet et al. (2010) and Battaglia and Mantovani (2005), which describes techniques to include polarization into Monte Carlo simulations. The latter also focuses on microwave radiation and anisotropic media. Tanaka and Ellison (1996; 2000) describe thermal infrared Monte Carlo simulations.

Summary The radiative transfer equations in an atmosphere-surface coupled system are a complex integrodifferential equations system requiring proper solution techniques. Despite advances in computer science, solutions remain time consuming and, to ensure the required accuracy, they require proper selection. This entry deals with some possible solution techniques. In the Introduction, we cite the classical texts and have selected the most advanced one in order to give the most effective overview. Further, we describe the selected methods by a cascade process by which the different techniques can be used singly or all together. The methods we have selected are the Fourier expansion, the iterative method, primary scattering, Gauss-Seidel, Discrete Ordinate, and Monte Carlo. All these methods are interconnected; the solution of one method can be also used for another method, as will be explained in the text. Even the statistical approach or Monte Carlo uses some solutions described by the other numerical selected methods. We introduce the rigorous analytical approach by the solutions embedded into the iterative and Fourier approach. In order to give a complete picture, we also introduce the input models of atmosphere and surface and how to use them. The bibliography provides sources for some of the software, which is usually in the public domain.

RADIATIVE TRANSFER, SOLUTION TECHNIQUES

Bibliography Barker, H. W., Goldstein, R. K., and Stevens, D. E., 2003. Monte Carlo simulation of solar reflectances for cloudy atmospheres. Journal of the Atmospheric Sciences, 60(16), 1881–1894. Battaglia, A., and Mantovani, S., 2005. Forward Monte Carlo computations of fully polarized microwave radiation in non isotropic media. Journal of Quantitative Spectroscopy and Radiative Transfer, 95, 285–308. Berk, A., Bernstein, L. S., and Robertson, D. C., 1989. Modtran MODTRAN: A Moderate Resolution Model for LOWTRAN 7. Burlington: Spectral Sciences. Bohren, C. F., and Clotiaux, E. E., 2006. Fundamentals of Atmospheric Radiation. Weinheim: Wiley. Chandrasekhar, S., 1960. Radiative Transfer. New York: Dover. Cornet, C., C-Labonnote, L., and Szczap, F., 2010. Threedimensional polarized Monte Carlo atmospheric radiative transfer model (3DMCPOL): 3D effects on polarized visible reflectances of a cirrus cloud. Journal of Quantitative Spectroscopy and Radiative Transfer, 111, 174–186, doi:10.1016/j. jqsrt.2009.06.013. De Haan, J. F., Bosma, P. B., and Hovenier, J. W., 1987. The adding method for multiple scattering calculations of polarized light. Astronomy and Astrophysics, 183, 371–391. Goody, R. M., and Yung, Y. L., 1989. Atmospheric Radiation Theoretical Basis. New York: Oxford University Press. http://code.google.com/p/i3rc-monte-carlo-model/ IAMAP Radiation Commission, 1975. Standard Procedures to Compute Atmospheric Radiative Transfer in a Scattering Atmosphere. Boulder: NCAR, Vol. I. IAMAP Radiation Commission, 1980. Standard Procedures to Compute Atmospheric Radiative Transfer in a Scattering Atmosphere. Boulder: NCAR, Vol. II. Iwabuchi, H., 2006. Efficient Monte Carlo methods for radiative transfer modeling. Journal of the Atmospheric Sciences, 63(9), 2324–2339. Kneizys, F. X., Shettle, E. P., Abreu, L. W., Chetwynd, J. H., and Anderson Lowtran, G. P., 1988. Users Guide to LOWTRAN 7. Hanscom AFB: Air Force Geophysics Lab. Kondratiev, K. Y., Kozoderov, V. V., and Smokty, O. I., 1992. Remote Sensing of the Earth from Space: Atmospheric Correction. Heidelberg: Springer. Lenoble, J., 1985. Radiative Transfer in Scattering and Absorbing Atmospheres: Standard Computational Procedures. Hampton: Deepak Publishing, p. 300. Lenoble, J., 1993. Atmospheric Radiative Transfer. Hampton: Deepak Publishing, p. 532. Levoni, C., Cervino, M., Guzzi, R., and Torricella, F., 1997. Atmospheric aerosol optical properties: a data base of radiative characteristics for different component and classes. Applied Optics, 36, 8031–8041. Levoni, C., Cattani, E., Cervino, M., Guzzi, R., and Di Nicolantonio, W., 2001. Effectiveness of the MS-method for computation of the intensity field reflected by a multilayer plane-parallel atmosphere: results from an accelerated yet accurate radiative transfer code. Journal of Quantitative Spectroscopy and Radiative Transfer, 4, 636–649. Liou, K. N., 1980. An Introduction to Atmospheric Radiation. New York: Academic. Liou, K. N., 1992. Radiation and Cloud Processes in the Atmosphere. New York: Oxford University Press. Marchuk, G., Mikhailov, G., Nazaraliev, M., Darbinjan, R., Kargin, B., and Elepov, B., 1980. The Monte Carlo Methods in Atmospheric Optics. Berlin: Springer, p. 208. Marshak, A., and Davis, A. B. (eds.), 2005. 3D Radiative Transfer in Cloudy Atmospheres. Springer: Berlin, p. 686.

623

Mayer, B., Emde, C., Buras, R., Hamann, U., and Kylling, A., 2005. LibRadTran: library for radiative transfer. http://www.libradtran. org/doku.php. Miskolczi, F., Rizzi, R., Guzzi, R., and Bonzagni, M., 1998. A new high resolution transmittance code and its application in the field of remote sensing. In Lenoble, J., and Geleyn, J. F., (eds.), International Radiation Symposium. Deepak Publishing: Hampton VA, USA. Nakajima, T., and Tanaka, M., 1981. Algorithms for radiative intensity calculations in moderate thick atmospheres using a truncation approximation. Journal of Quantitative Spectroscopy and Radiative Transfer, 40, 51–69. O’Hirok, W., and Gautier, C., 1998. A three-dimensional radiative transfer model to investigate the solar radiation within a cloudy atmosphere. Part I: spatial effects. Journal of the Atmospheric Sciences, 55(12), 2162–2179. Petty, G. W., 2006. A First Course in Atmospheric Radiation, 2nd edn. Madison: Sundog. Pincus, R., and Evans, K. F., 2009. Computational cost and accuracy in calculating three-dimensional radiative transfer: results for new implementations of Monte Carlo and SHDOM. Journal of the Atmospheric Sciences, 66(10), 3131–3314. Preisendorfer, R., 1965. Radiative Transfer on Discrete Spaces. New York: Pergamon. Ricchiazzi, P., Shiren, Y., and Gautier, C., 2007. SBDART: a practical tool for plane-parallel radiative transfer in the earth’s atmosphere. Earth Space Research Group, Institute for Computational Earth System Science University of California, Santa Barbara, http://www.paulschou.com/tools/sbdart/. Sobolev, V., 1975. Light Scattering in Planetary Atmospheres. New York: Pergamon. 442p. Spurr, R. J., 2001. A General Discrete Ordinate Approach to the Calculation of Radiances and Analytical Weighting Functions with Applications to Atmospheric Remote Sensing. PhD thesis, Eindoven Tech. University. Stamnes, K., 1982. On the computation of angular distributions of radiation in planetary atmosphere. Journal of Quantitative Spectroscopy and Radiative Transfer, 28, 47–51. Stamnes, K., and Swanson, R., 1981. A new look at the discrete ordinate method for radiative transfer calculation in anisotropically scattering atmosphere. Journal of the Atmospheric Sciences, 38, 387–399. Stamnes, K., Tsay, S. C., Wiscombe, W. J., and Jayaweera, K., 1988. Numerical stable algorithm for discrete ordinate radiative transfer in multiple scattering and emitting layered media. Applied Optics, 27, 2502–2509. Takara, E. E., and Ellingson, R. G., 1996. Scattering effects on longwave fluxes in broken cloud fields. Journal of the Atmospheric Sciences, 53(10), 1464–1476. Takara, E. E., and Ellingson, R. G., 2000. Broken cloud field longwave-scattering effects. Journal of the Atmospheric Sciences, 57(9), 1298–1310. Thomas, G. E., and Stamnes, K., 1999. Radiative Transfer in Atmosphere and Ocean. Cambridge: Cambridge University Press, p. 517. Van De Hulst, H. C., 1980. Multiple Light Scattering. Tables, Formulas and Applications. New York: Academic. Verstraete, M. M., Pinty, B., and Dickinson, R. E., 1990. A physical model of the bidirectional reflectance of vegetation canopies. I – Theory II – inversion and validation. Journal of Geophysical Research, 95, 11755–11775. (ISSN 0148–0227), Research supported by ESA, CNRS, and NCAR. World Climate Programme. 1982. WCP-43 Tropospheric aerosols: review and current data on physical and optical properties (computed by Harris, F. S., and Gerber, H. G.), WMO Geneve. Yanovithskij, E. G., 1997. Light Scattering in Inhomogeneous Atmosphere. Heidelberg: Springer.

624

RADIATIVE TRANSFER, THEORY

Definition The radiative transfer theory, also called transport theory, is the theory describing the wave propagation through a medium characterized by a random distribution of scatterers. It usually applies to electromagnetic radiation, but it can be generalized to acoustic radiation. The radiative transfer theory is one of the two theories developed to deal with radiation absorption and multiple-scattering problems: the other approach is the so-called wave analytical theory. It is generally formulated in terms of specific intensity and can be extended to polarized radiation. The radiative transfer theory constitutes the physical basis of several remote sensing techniques due to its relative simplicity and capability to deal with multiple-scattering effects.

so that a formulation which includes all multiple-scattering terms is practically impossible to derive. The RTT, conversely, does not start with the wave equation, but simplifies the problem dealing with the transport of energy through a medium containing scatterers (Chandrasekhar, 1960; Ishimaru, 1978; Ulaby et al., 1986). The development of theory is almost heuristic as it is not rigorously derived from the wave equation as the wave analytical theory. Although diffraction and interference effects are included in the description of the absorption and scattering properties of a single scatterer, RTT does not actually include the interferential effects of the ensemble of the scatterers. As a matter of fact, RTT assumes that there is no correlation between the fields scattered by each particle (socalled independent-scattering assumption), so that the addition of the scattered power, rather than the addition of the scattered fields, holds. Indeed, RTT carries information on the mutual coherence of the diffuse field, as it will be introduced later on dealing with the assumptions and limitations of RTT at the end of this section (Ishimaru, 1975; Furutsu, 1975). The RTT can be expressed in terms of specific intensity which expresses the radiant power per unit area, solid angle, and frequency. If the specific intensity is scalar, as for the case of totally unpolarized light, the radiative transfer equation (RTE) becomes a single differential equation which is usually called the scalar RTE (Ulaby and Elachi, 1990). For vector fields, the scalar specific intensity can be generalized using the four Stokes parameters which lead to four coupled differential equations. The latter is sometimes called the vector RTE (Tsang et al., 1985). In the following sections, we will briefly introduce both scalar and vector RTEs.

Introduction The subject of radiative transfer is very transversal and, indeed, covers several research fields, including astrophysics, applied physics, optics, planetary sciences, atmospheric sciences, meteorology, and engineering. Generally speaking, we can define the radiative transfer theory (RTT), also called transport theory, as the theory describing the wave propagation through a medium characterized by a random distribution of scatterers (Chandrasekhar, 1960). It usually applies to electromagnetic radiation, but it can be generalized to acoustic radiation (Ishimaru, 1978; Tsang et al., 2000). The RTT is one of the two distinct theories developed to deal with radiation absorption and multiple-scattering problems: the other approach is the so-called wave analytical theory (WAT). The latter starts, in case of electromagnetic waves, from the basic Maxwell differential equations which are usually transformed into an equivalent wave equation (Ishimaru, 1978; Tsang et al., 1985). The wave analytical theory is, in principle, able to treat absorption and scattering to any order together with diffraction and interference effects. However, as intuitive, the detailed approach of the WAT is very cumbersome and mathematically complicated,

Brief history of radiative transfer theory The RTT was pioneered by astrophysicists who opened the field at the beginning of the twentieth century. The work of Schuster (1905) on the investigation of the light radiation transfer through a foggy atmosphere appears to be the first paper discussing the importance of multiple scattering. Schuster basically formulated the problem in terms of upward and downward light beams, the basic assumption for the so-called two-stream approximation of radiative transfer. The latter was also employed by Schwarzschild (1906) to explain the limb darkening of the sun. In his effort to understand the physical interior of a star, Eddington (1916) developed a first-order expansion of intensity in terms of Legendre polynomials, leading to the so-called Eddington approximation of radiative transfer. In the same years, Schwarzschild introduced the concept of the medium emission as well as absorption within the context of thermodynamic equilibrium, a subject widely used to deal with thermal infrared radiation in molecular atmosphere with negligible scattering. These pioneering works were collected in a volume by Menzel in the middle of the 1960s (1966).

Cross-references Aerosols Cloud Properties Fields and Radiation Land-Atmosphere Interactions, Evapotranspiration Land Surface Roughness Ocean-Atmosphere Water Flux and Evaporation Optical/Infrared, Scattering by Aerosols and Hydrometeors Radiative Transfer, Theory

RADIATIVE TRANSFER, THEORY Frank S. Marzano Department of Information Engineering, Sapienza University of Rome, Rome, Italy Centre of Excellence CETEMPS, University of L’Aquila, L’Aquila, Italy

RADIATIVE TRANSFER, THEORY

In the 1950s, in his landmark book, Chandrasekhar presented the subject of radiative transfer in plane-parallel atmospheres as a branch of mathematical physics and developed numerous solution methods and techniques, including the consideration of polarized radiation (Chandrasekhar, 1960). An equally fundamental treatise is the one of Sobolev (1975). In the 1960s and 1970s, many efforts were carried out to provide a robust physical basis of the RTT and its relation with the wave analytical theory (Ishimaru, 1978; Tsang et al., 1985). The basic differential equation of RTT is equivalent to Boltzmann’s equation (also known as the BoltzmannMaxwell collision equation), used in the kinetic theory of gases (Sommerfeld, 1956). There is a noteworthy equivalence between RTT and the neutron transport theory which was well established by Davison (1958). The formulation of RTT is, indeed, flexible and capable of treating many physical problems. It has been successfully applied to the problems of atmospheric radiation, underwater visibility, marine biology, photographic optics, and wave propagation through planets, stars, and galaxies.

Specific intensity The fundamental quantity in the RTT is the scalar specific intensity I, sometimes also called radiance (indicated by L) or brightness (indicated by B). The concept of the specific intensity can be introduced considering a flow of wave energy in a random medium where the frequency, phase, and amplitude of the wave undergo random variations in time and space. This implies that the magnitude and direction of its power flow density vector (also called Poynting vector in electromagnetics) vary continuously in time and space (Ishimaru, 1978; Tsang et al., 1985). At a given point r and a direction defined by the unit vector ^s, the scalar specific intensity I(r, ŝ) is measured in W m2 sr1 Hz1 (where sr stands for steradian or solid unit angle). Referring to Figure 1, the amount of elementary power dP (in Watts) flowing within a solid angle dO through an elementary area dA oriented along

625

^ in a frequency interval the direction of the unit vector n dn is given by dP ¼ Iðr; ^sÞ cos ydAdOdn

(1)

Indeed, Equation 1 describes the radiation flux emitted from a surface dA along ŝ and it is often referred as the forward (or outward) specific intensity I+ ¼ I(r, ŝ) (also called surface intensity). Conversely, if the point r is taken on the surface dA, we can express the backward (or inward) specific intensity I ¼ I(r,-ŝ) incident upon the surface along ŝ (also called field intensity). In space, the inward power upon dA should be identical to the outward power emitted from the same surface, so that: I+ ¼ I. In electromagnetics, the specific intensity I(r, ŝ) can be statistically related to the Poynting vector S. If the latter at a given point r is a random function of time and each component of S varies in time, the tip vector of S moves randomly in time. If this random variability is expressed by a probability density function pS, we can then define the specific intensity as the ensemble average value of the random vector S. This means that the I(r, ŝ) is basically the sum of all random Poynting vectors S whose tips are located within a solid unit angle in the direction ŝ. The same interpretation can be easily extended to acoustic waves (Ishimaru, 1978).

Flux and energy density The forward flux density F+ (measured in W m2 Hz1) passing through an elementary area dA on a finite surface A is obtained by integrating Equation 1 over a semisolid angle 2p+ in the forward interval (0  y  p/2) along the ^: direction n ð ^Þ ¼ ^dO Iðr; ^sÞ^s n (2) F þ ðr; n 2pþ

Similarly, for the backward flux density F through dA in the backward direction ð^ nÞ and for a semisolid angle 2p in the backward interval (p/2  y  p), we can define: ð  ^Þ ¼ Iðr; ^sÞ^s ð^ nÞdO (3) F ðr; n 2pþ

sˆ

nˆ θ dΩ dA(r)

Radiative Transfer, Theory, Figure 1 Geometry for a specific intensity emitted by an elementary surface dA.

For a radiating surface, the forward flux density F+ is often called radiant emittance or radiant excitance (Elachi, 1987). For receiving surface, the backward flux density F is also called irradiance. The net flux density can be expressed as the component of a flux density vec^ equal to the difference between the fortor F(r) along n ward flux density and the backward flux density. For the flux density vector F(r), it holds over the whole solid angle: ð (4) FðrÞ ¼ Iðr; ^sÞ^sdO 4p

626

RADIATIVE TRANSFER, THEORY

Other useful definitions related to the specific intensity are the energy density u(r) and the average intensity U(r), defined by ð 1 c (5) Iðr; ^sÞdO ¼ U ðrÞ uðrÞ ¼ c 4p 4p

where c is the wave velocity with u measured in J m3 Hz1 and u measured in W m2 Hz1. Note that Equation 4 is derived considering that the energy density du(r) can be expressed as the ratio of the energy I dA dO dn dt and the volume dA c dt it should occupy. As an important example, we can consider the case when the specific intensity I(r, ŝ) is independent of the direction ŝ so that the radiation is said to be isotropic. In ^ this circumstance, the total net flux density F along n can be expressed by ð ^ ¼ Iðr; ^sÞ cos ydO ¼ I0 ðrÞp (6) FðrÞ ¼ FðrÞ n 4p

where I0 is average specific intensity value. The previous relationship is called the Lambert law which can be also expressed in terms of power from Equation 1 as a cosine law: P ¼ P0 cosy with P0 the average power.

Invariance and boundary conditions It can be proved that the specific intensity is invariant along the ray path in free space (Ishimaru, 1978; Tsang et al., 2000). Even though this property might seem to contradict the intuitive spreading of power flux in free space, it should be reminded that the specific intensity is defined per unit solid angle from Equation 1. The proof of the specific intensity invariance can be approached by considering two points r1 and r2 in free space separated by a distance r along the direction ŝ and two small areas dA1 and dA2 perpendicular to ŝ. Let I(r1, ŝ) and I(r2, ŝ) be the specific intensity at r1 and r2, respectively. The power received by dA2 can be expressed in two ways: (1) as I(r1, ŝ) dA1 dO1 dn from Equation 1 and (2) as I(r2, ŝ) dA2 dO2 dn. The previous expression must be equal, so that I(r1, ŝ) dA1 dO1 ¼ I(r2, ŝ) dA2 dO2. But it holds dA1 ¼ r2 dO2 and dA2 ¼ r2 dO1, so that it results I(r1, ŝ) ¼ I(r2, ŝ). Once dealing with applications, the random medium is very likely inhomogeneous. This means that when looking for the specific intensity within the considered volume, we need to know the boundary surface conditions in terms of specific intensity (Tsang et al., 1985; Tsang et al., 2000; Ulaby et al., 1986). It is customary in RTT to divide the boundary conditions into three categories: (1) isotropic surface, (2) planar surface, and (3) rough surface. In the first case, the surface is also called Lambertian as the behavior of the specific intensity I(r, ŝ) is independent of the direction ŝ and follows that of the Lambert law. This can be expressed in terms

I þ ðrs Þ ¼ I  ðrs Þ

(7)

where rs is the position vector belonging to the boundary surface S. In the second case, the surface is called Fresnelian as the behavior of the specific intensity I(r, ŝ) is regulated by the Snell law (i.e., reflection angle equal to incident angle) and the relation between the reflected and transmitted wave amplitude, with respect to the incident one, follows the Fresnel coefficient (see Figure 2). If G|| and G⊥ are the reflection coefficients of the electric field polarized parallel and perpendicular to the plane of incidence, respectively, for unpolarized (scalar) incident specific intensity I and reflected specific intensity I, it holds ^ ^sr ¼ n ^ ^si I þ ðrs ; ^sr Þ ¼ jGð^si Þj2 I  ðrs ; ^si Þ with n

(8)

where ^sr and ŝi are the unit vector for the reflection and incident directions, whereas |G|2 ¼ 0.5 (|G|||2 + |G⊥|2) is the unpolarized reflectance (or reflectivity) of the Fresnel boundary surface (Tsang et al., 1985). For the surface transmittance, a similar argumentation can be carried out considering the transmission Fresnel coefficients and the power conservation at the surface so that the incident power should be equal to the sum of the reflected and transmitted ones (Ishimaru, 1978). In the third general case, the surface may be called natural and there is not necessarily an analytical law relating the incident and the reflected specific intensity (Ulaby et al., 1986). If gd ð^sr ; ^si Þ is the power differential reflectivity of the boundary surface, it can be written in a general way: ð I þ ð^rs ; ^sr Þ ¼ gd ð^sr ; ^si ÞI  ðrs ; ^si ÞdOi (9) 2p

where dOi is the solid angle including all incident directions in 2p (p/2  y  p) and ^sr spans the forward interval (0  y  p/2). The characterization of gd will depend on the surface geometry and composition.

Radiative transfer equation After examining the invariance property and the boundary conditions, we are ready to analyze the propagation of the fundamental quantity in RTT, the specific intensity I(r, ŝr), through a medium containing a random distribution of scatterers. The latter scatter, absorb, and emit the wave energy, and all these characteristics should be included into a differential equation expressed in terms of the specific intensity. The scalar radiative transfer equation (RTE) can be derived by considering an unpolarized specific intensity I(r, ŝ) incident upon a cylindrical elementary volume dV with a cross section dA and a length ds, as in Figure 3. The volume dV ¼ ds dA contains a random distribution of scatterers with different properties described by the parameters qi (i ¼ 1  Q). The latter can represent the size, shape, and orientation variability of the scatterers and are

RADIATIVE TRANSFER, THEORY

627

Radiative Transfer, Theory, Figure 2 Boundary condition for specific intensity at the planar surface S separating two media with different refractive indexes.

dA

I(r, sˆ) dΩ dΩ′

I(r, sˆ′)

Radiative Transfer, Theory, Figure 3 Propagation and interaction of the specific intensity upon an elementary volume containing a random distribution of scatterers.

usually described by a joint probability density function pq(q1,. . .,qQ), usually normalized to one. In case we refer to spherical particles with different diameters, then the Q scatterer parameters reduce to q1 ¼ D and pD(D). In order to derive the RTE, we need (1) first, to characterize the volumetric interaction in terms of absorption and scattering and (2) then to consider the balance of the power along the ray, taking into account thermal emission, incident radiation, and multiple scattering.

Volumetric interaction parameterization The volumetric interaction between the incident radiation and an ensemble of scatterers needs to take into account both absorption and scattering. In this respect, we can exploit the assumption of independent scattering, which is a negligible correlation among the scattered fields. This means that the absorbed and scattered power of each scatterer can be simply summed up to provide the total absorbed and scattered power (Ishimaru, 1978; Tsang et al., 2000). For a single scatterer at a point r and characterized by some properties qi, we can define the absorption cross section sa ðr; ^si ; q1 ; . . . ; qQ Þ (measured in m2) with respect to

an incident wave along the unit direction ŝi. The cross section sa is the ratio between the power dPa, absorbed by the scatterer itself, and incident power flux density Si and will depend on the scatterer geometry, orientation, and composition. Considering an ensemble of the previous scatterers, randomly distributed after pq(q1,. . .,qQ), and summing over all the absorbed powers dPa, the volumetric absorption coefficient ka(r, ŝi) is defined as ð ka ðr;^si Þ ¼ Nq sa ðr;^si ; q1 ; ... ;qQ Þ (10) Dq pq ðq1 ;.. .; qQ Þdq1 ;. ..; dqQ where Nq is the number of scatterers per unit volume (measured in m3) and Dq is the domain of integration. This means that ka is measured in m1. A similar argument can be used to introduce the differential scattering cross section sd ðr; ^sr ; ^si ; q1 ; . . . ; qQ Þ (measured in m2) which, however, depends not only on the incident direction ŝi but also on the scattering direction ŝr. The cross section sd is given by the ratio of the power flux intensity r2dSr, scattered at a distance r in the far field by the scatterer itself, and the incident power flux density Si

628

RADIATIVE TRANSFER, THEORY

along ŝi. Considering a scatterer ensemble and summing over all the scattered power flux intensities r2dSr, the volumetric differential scattering coefficient kd(r, ŝr, ŝi) may be defined as ð kd ðr; ^sr ; ^si Þ ¼ Nq sd ðr; ^si ; q1 ; . . . ; qQ Þ (11) Dq pq ðq1 ; . . . ; qQ Þdq1 ::dqQ From Equation 11, it is straightforward to derive the volumetric scattering coefficient ks(r, ŝi), as it is the integral of kd over all scattering direction in 4p: ð ks ðr;^si Þ ¼ kd ðr;^sr ;^si ÞdOr ¼Nq 4p

ð

ss ðr;^si ; q1 ; . . . ; qQ Þpq ðq1 ; . . . ; qQ Þdq1 ::dqQ q

(12) where Or is the scattering solid angle including all directions ŝr and ss is the scattering cross section of each scattered, measured in m2. From Equations 11 and 12, it is also derived that the scattering cross section ss is the integral of differential cross section sd over the whole solid angle. Note that the coefficient ks(r, ŝi), measured in m1, is consistent with the coefficient ka(r, ŝi), whereas ss is the ratio between the scattered power dPr and the incident power flux density Si. Finally, for an ensemble of scatterers at a point r, where both volumetric absorption and scattering take place, it is customary to define the volumetric extinction coefficient ke(r, ŝi) as ke ðr; ^si Þ ¼ ka ðr; ^si Þ þ ks ðr; ^si Þ

(13)

where the extinction is supposed to be sum of the absorption and scattering effects. A further important quantity is the volumetric scattering phase function pðr; ^sr ; ^si Þ, defined as kd ðr; ^sr ; ^si Þ (14) pðr; ^sr ; ^si Þ ¼ 4p ks ðr; ^si Þ where its angular normalization over the whole solid angle is 4p. Sometimes, the phase function is defined with respect to ke, instead of ks, so that its angular normalization yields the product between 4p and the volumetric albedo (e.g., Ishimaru, 1978). The latter is defined as wðr; ^si Þ ¼

ks ðr; ^si Þ ke ðr; ^si Þ

(15)

As an example, we can consider a random distribution of spherical particles of diameter D0. In this case, the joint probability distribution reduces to a Dirac impulse function: pD(D) ¼ d(DD0). From Equations 10, 11, 12, 13, 14, and 15, the volumetric interaction parameters are then simply equal to

ka ðr; ^si Þ ¼ND sa ðr; ^si Þ; ss ðr; ^si Þ wðr; ^si Þ ¼ se ðr; ^si Þ

ks ðr; ^si Þ ¼ ND ss ðr; ^si Þ;

(16)

Differential equation formulation Considering the geometry of Figure 3, we can then suppose the random scatterers characterized by the volumetric interaction parameters previously defined. From Equation 13, the reduction of the specific intensity I(r, ŝ) due to its propagation through the volume dV ¼ dsdA is equal to dIðr; ^sÞ ¼ ds½ka ðr; ^sÞ þ ks ðr; ^sÞ Iðr; ^sÞ ¼ dske ðr; ^sÞIðr; ^sÞ

(17)

On the other hand, the specific intensity increases 0 ), incident on dV from another because a portion of I(r, ŝ 0 direction ŝ , is scattered into the direction ŝ and is added to I(r, ŝ). This contribution along ŝ at a point r can be derived from the differential scattering coefficient0 kd(r, 0 by kd (r, ŝ, ŝ ), over ŝr, ŝi) by integrating I(r, ŝ ), weighted 0 all the incident directions ŝi ¼ ŝ within the entire solid angle: ð (18) dIðr; ^sÞ ¼ þds kd ðr; ^s; ^s0 ÞIðr; ^s0 ÞdO0 4p

where O’ is the solid angle including all directions incident on the volume dV from directions different from ^s. The term in Equation 18 is often called the multiplescattering pseudo-source. Within the volume dV, if a local thermodynamic equilibrium can be supposed, the Kirchhoff law holds. The latter states that a medium that absorbs radiation at a given frequency, at the same time emits radiation at the same frequency. The rate at which emission takes place is a function of temperature and frequency and is related to the Planck law which for a black body (BB) gives the following isotropic specific intensity IBB: IBB ðrÞ ¼

hn3 1 2 hn=KT c e 1

(19)

where h and K are the Planck and Boltzmann constants, respectively, c is wave velocity, n the frequency, and T the temperature. For hn < < KT (valid at microwaves), IBB  KT/l2 with l the wavelength. It is easy to show that for the thermal emission specific intensity, it holds: dIðr; ^sÞ ¼ þdsXðr; ^sÞ ¼ þdska ðr; ^sÞIBB ðr; ^sÞ

(20)

which shows that the emission power X per unit volume, solid angle, and frequency of a medium is equal to the product ka times IBB.

RADIATIVE TRANSFER, THEORY

Adding all the contributions in Equations 17, 18, and 20, we get the integrodifferential equation of radiative transfer: dIðr; ^sÞ ¼ ½Iðr; ^sÞ^s ¼ ke ðr; ^sÞIðr; ^sÞ ds ð 0

0

0

kd ðr; ^s; ^s ÞIðr; ^s ÞdO þ ka

þ

(28)

(21)

where the second term expresses the divergence of the specific intensity vector. Equation 21 can be also rearranged by introducing the so-called optical distance t, defined by (Chandrasekhar, 1960) ð (22) tðr; ^sÞ ¼ ke ðr; ^sÞds 4p

By using the definition of the scattering phase function in Equation 14, the RTE in Equation 21 is also given by ð dIðr; ^sÞ wðr; ^sÞ ¼  Iðr; ^sÞ þ pðr; ^s; ^s0 ÞIðr; ^s0 Þ dt 4p 4p

dO0 þ ½1  wðr; ^sÞ IBB ðrÞ (23) At microwaves, it is useful to introduce the brightness temperature TB(r, ŝ), defined as K TB ðr; ^sÞ l2

(24)

where the brightness temperature is also defined as the product of the body emissivity times physical temperature. Through Equation 24, the RTE becomes ð dTB ðr; ^sÞ wðr; ^sÞ ¼  TB ðr; ^sÞ þ pðr; ^s; ^s0 Þ dt 4p (25) 4p TB ðr; ^s0 ÞdO0 þ ½1  wðr; ^sÞ T ðrÞ

It is very often convenient to separate the total specific intensity into two terms, the reduced incident specific intensity Ic(r, ŝ) and diffuse specific intensity Id(r, ŝ): Iðr; ^sÞ ¼ Ir ðr; ^sÞ þ Id ðr; ^sÞ

(26)

Substituting the previous equation into Equation 24, we obtain 8 dI ðr;^sÞ < rdt ¼ Ir ðr;^sÞ :

dId ðr;^sÞ dt

sÞ ¼ Id ðr;^sÞ þ wðr;^ 4p

Ð 4p

J ðr; ^sÞ ¼½1  wðr; ^sÞ IBB ðrÞ ð wðr; ^sÞ pðr; ^s; ^s0 ÞIc ðr; ^s0 ÞdO0 þ Jt ðr; ^sÞ þ 4p 4p

4p

Iðr; ^sÞ ¼

629

pðr;^s;^s0 ÞId ðr;^s0 ÞdO0 þ J ðr;^sÞ

:

(27) where the source function J(r, ŝ) is given by the thermal term Jt and reduced incident term Jr:

Note that from Equation 27, the reduced incident specific intensity is equal to the incident specific intensity reduced by the medium extinction coefficient ke. The boundary conditions for the specific intensity at the surface S of medium containing random distribution scatterers may be various: (1) if no diffuse intensity is entering the medium, it holds Id(r, ŝ) ¼ 0 on S; (2) if the incident intensity Ii is a collimated beam directed along ŝ0, then it holds Ii(r, ŝ0) ¼ I0 d(ŝ–ŝ0) on S so that Ir(r, ŝ0) ¼ I0 d(ŝ–ŝ0) exp(t); and (3) if the incident intensity Ii is a diffuse one itself, then it holds Ii(r, ŝ0) ¼ I0 on S so that Ir(r, ŝ0) ¼ I0 exp(t).

Integral formulation and received power The RTE for Id(r, ŝ) in Equation 27 is first-order differential equation with respect to s. It is sometimes convenient to deal with the integral formulation especially for complex geometries. Supposing that at the boundary surface S in r ¼ r0 it holds Id(r0, ŝ) ¼ 0 (see Figure 4), the general integral solution for Ir(r, ŝ) and Id(r, ŝ) is 8 t < Ir ðr;^sÞ ¼ Ii ðr0 ;^sÞe Ðs ðtt0 Þ wðr0 ;^sÞ Ð : Id ðr;^sÞ ¼ e pðr0 ;^s;^s0 ÞId ðr0 ;^s0 ÞdO0 þ J ðr0 ;^sÞ ds0 4p 0

4p

(29) where Ii(r0, ŝ) is the incident specific intensity and t0 is optical distance between r0 and r0 with s the distance between r0 and r. In actual measurements of the specific intensity, it is necessary to take into account the characteristics of the receiver. The latter can be conveniently described by receiving equivalent area Ae(ŝ, ŝ0), measured in m2, whose maximum is supposed to be along ŝ0. If I(r, ŝ) is incident upon a receiver, placed at r and characterized by Ae(ŝ, ŝ0), the receiver power per unit frequency PR(ŝ0) (measured in W/Hz) is given by ð PR ðr; ^s0 Þ ¼ Ae ðr; ^s0 ; ^sÞI ðr; ^sÞ (30) DO dO ¼ PRr ðr; ^s0 Þ þ PRd ðr; ^s0 Þ where DO is the angular domain integration and the last term is due to the decomposition of I into Ir and Id. The receiver equivalent area can be, for example, either a pencil beam (e.g., highly directive microwave antennas or optical detector) or a near-isotropic radiator (e.g., short dipole).

630

RADIATIVE TRANSFER, THEORY

Radiative Transfer, Theory, Figure 4 Incident, coherent, and diffuse specific intensity in a volume V of random scatterers.

Radiative transfer of partially polarized radiation The inclusion of polarization, while unnecessary for acoustic waves, may be important for electromagnetic waves. All electromagnetic waves in random media are, in general, partially polarized because, even though the incident wave is linearly polarized, the scattered wave is generally elliptically polarized and its polarization should randomly vary in time and space. The scalar RTE can be generalized to partially polarized wave by replacing the scalar specific intensity with the vector specific intensity I(r, ŝ), measured in W m2 sr1 Hz1, whose components are the Stokes parameters Ih(r, ŝ), Iv(r, ŝ), U(r, ŝ), and V(r, ŝ). The vector specific intensity can be defined starting from an elliptically polarized plane wave, representing an electrical field E(r, ŝ):   ^ þ Ev ^v ejk r (31) Eðr; ^sÞ ¼ Eh h where ^h and ^v are the unit horizontal and vertical polarization vectors with Eh and Ev their complex amplitudes, j is the imaginary unit, and k is the propagation constant along ŝ. The Stokes specific intensity vector I(r, ŝ) is then given by 3 2 3 2 Iv jE v j2 7 6 Ih 7 6 jE h j2 7 7 6 (32) Iðr; ^sÞ ¼ 6 4 U 5 ¼ 4 2 ReðEv Eh  Þ 5  V 2 ImðEv Eh Þ Note the first two Stokes parameters are also called modified parameters, as the original ones introduced by Stokes were I ¼ Iv + Ih and Q ¼ Iv–Ih. Moreover, sometimes the third term is divided by Z, the medium intrinsic impedance. An effective representation of the Stokes vector is through the so-called Poincarè sphere where in the Q, U, and V space, the I value is graphed with respect to the ellipticity and rotation angles (e.g., linear polarization shows a zero ellipticity angle and a 0 or 90 rotation angle if vertical or horizontal polarization is dealt with, respectively). For a completely polarized monochromatic wave,

it holds the following identity: I2 ¼ Q2 + U2 + V2, whereas for partially polarized wave, it holds the inequality I2 > Q2 + U2 + V2. The degree of polarization dp is defined as the ratio between the square root of (Q2 + U2 + V2) and I2: If dp ¼ 1, the wave is completely polarized, whereas if dp ¼ 0, the wave is completely unpolarized. If 0 < dp < 1, the wave is said to be partially polarized and it corresponds to a point inside the Poincarè sphere (Ulaby and Elachi, 1990). The vector analog of Equation 21 is the vector RTE: ð dIðr; ^sÞ ^ ^ ¼ ke ðr; sÞI ðr; sÞ þ kd ðr; ^s0 ÞI ðr; ^s0 ÞdO0 ds 4p

þ ka ðr; ^sÞIBB ðrÞ (33) where I is the 4x1 Stokes column vector, ke(r, ŝ) is the 4  4 extinction matrix, kd(r, ŝ, ŝ0 ) is the 4  4 differential scattering matrix, and ka(r, ŝ) is the 4  1 absorption column vector. Note that very often (and unfortunately), the matrix kd is also called phase matrix P even though it does not reflect the scalar definition of the phase function in Equation 14. The matrix parameters in the vector RTE of Equation 31 can be related to the single-scattering property of each scatterer contained in the random medium. In a linear medium, we can relate the scattered field Es(r, ŝ) by a scatterer to the incident field Ei(r, ŝ0 ) through the complex scattering 2  2 matrix S(ŝ, ŝ0 ):





Svv Svh Eih Esv ¼ Es ðr; ^sÞ ¼ Esh Shv Shh Eiv (34) 0 0 ¼ Sð^s; ^s ÞEi ðr; ^s Þ It is worth noting that the scattering matrix S can be expressed in any arbitrary orthonormal unit system, different from the one defined by the horizontal and the vertical polarizations as in Equation 34 (where usually the reference plane contains a preferential vertical axis and either the incident or scattered direction). An example is the

RADIATIVE TRANSFER, THEORY

so-called scattering-plane orthonormal system which contains both the incidence direction and scattered direction where the polarization states are distinguished as either parallel or orthogonal to the scattering plane itself (Tsang et al., 2000). The scattering-plane reference system is convenient if the scatterer has symmetries. By using the previous relation and Equation 32, we can easily show that, for a single scatterer, the scattered Stokes vector Is(r, ŝ) is related to the incident Stokes vector Ii(r, ŝ) by means of the so-called Stokes or Muller 4  4 matrix L(ŝ, ŝ0 ): 0

0

Is ðr; ^sÞ ¼ Lð^s; ^s ÞIi ðr; ^s Þ

(35a)

or explicitly:

(35b) The differential scattering matrix is then given, in analogy to the scalar case, as the weighted average of the Muller matrix L over the probability density function pq: ð     0 ^ ^ kd ðr; s; s Þ ¼Nq L r; ^s; ^s0 ; q1 ::; qQ pq q1 ; ::; qQ q 0

dq1 ::d ð qQ Pðr; ^s; ^s Þ     kd ðr; ^s; ^s Þ ¼Nq L r; ^s; ^s0 ; q1 ::; qQ pq q1 ; ::; qQ 0

q

dq1 ::dqQ Pðr; ^s; ^s0 Þ (36) ð

    kd ðr; ^s; ^s Þ ¼Nq L r; ^s; ^s0 ; q1 ::; qQ pq q1 ; ::; qQ q

Dq

(38) In the general case, for nonspherical particles, 4  4 extinction matrix ke is a non-diagonal matrix having the following general structure: 3 2ReðMvv Þ 0  ReðMvh Þ  ImðM Þ 6 0 2ReðMhh Þ  ReðMhv Þ ImðM Þ 7 7 6 ke ðr;^sÞ ¼ 6 7 4 2ReðMhv Þ 2ReðMvh Þ  ReðMvv þ Mhh Þ ImðMvv Þ 5

The previous matrix is also called the phase matrix P, as indicated in the third term of Equation 36. The latter is essentially based on the assumption of incoherent addition of the diffuse Stokes parameters (Tsang et al., 2000). In the same manner, the extinction matrix ke can be related to the complex scattering matrix S. In particular, the optical theorem provides a useful relationship for computing the extinction cross section in terms of the field scattered in the forward direction, relative to the incident field. If the latter is t polarized, where m ¼ v or h, the ppolarized extinction cross section sem is 4p Im½Smm ðr; ^s ¼ ^s0 ; ^s0 Þ k

2ImðMhv Þ 2ImðMvh Þ  ImðMvv þ Mhh Þ  ReðMvv Þ

(39) Where Mmn ¼

ð j2pNq Nq Smn ðr;^s; q1 ; ::; qQ Þpq ðq1 ;::; qQ Þdq1 ::dqQ k q

(40) with m,n ¼ h,v. The latter expressions can be derived from the theory of attenuation of the coherent wave through a random medium, adopting the so-called Foldy’s approximation (Tsang et al., 1985). For the special case of a random medium with spherical particles, the 4  4 extinction matrix ke simplifies to a diagonal matrix with all identical elements kesph ¼ kehh ¼ kevv: 3 2 0 0 0 kesph 6 0 kesph 0 0 7 7 6 ke ðr; ^sÞ ¼ 6 7 (41) 4 0 0 kesph 0 5 0

dq1 ::dqQ Pðr; ^s; ^s0 Þ

sem ðr; ^sÞ ¼

evaluated in the forward-scattering direction ŝ ¼ ŝ0 . The m-polarized extinction coefficient kem is given by the weighted average over the probability density function pq: ð     ket ðr; ^sÞ ¼ Nq set r; ^s; q1 ::; qQ pq q1 ; ::; qQ dq1 ::dqQ

2

3    2 3 Re Svh Svv jSvv j2 jSvh j2 Isv 7    6I 7 6 7 jShv j2 jShh j2 Re Shh Shv 6 sh 7 6 7 6 7¼6       7     4 Us 5 6 4 2 Re Svv Shv 2Re Svh Shh 2 Re Svv Shh þ Svh Shv  Im 5           Vs Im Svv Shh þ Svh Shv Re 2 Im Svv Shv 2Im Svh Shh 2

0

631

(37)

where Im is the imaginary-part operator and the scattering term Smm is the mm scattering amplitude of the particle,

0

0

kesph

Note that the absorption coefficient ka, usually due to gases, is an unpolarized quantity within the expression of the total extinction matrix ke. The emission vector of the random distribution of particles within the vector RTE is expressed through the absorption vector ka in Equation 31. Its expression might be fairly complicated in the general case and can be derived from the fluctuation-dissipation theorem for a single scatterer and the incoherent addition of Stokes parameters (Tsang et al., 1985). It can be shown that for random media of nonspherical particles, it holds 3 2 ka1 ð^sÞ 6 k ð^sÞ 7 7 6 a2 ka ðr; ^sÞ ¼ 6 7 (42) 4 ka3 ð^sÞ 5 ka4 ð^sÞ

632

RADIATIVE TRANSFER, THEORY

Where ð ka1 ð^sÞ ¼ ke11 ð^sÞ 

½kd11 ð^s;^s0 Þ þ kd21 ð^s;^s0 Þ dO0

4p

ð

ka2 ð^sÞ ¼ ke22 ð^sÞ 

½kd12 ð^s;^s0 Þ þ kd22 ð^s;^s0 Þ dO0

4p

ð ka3 ð^sÞ ¼ 2ke13 ð^sÞ  2ke23 ð^sÞ þ 2 ½kd13 ð^s;^s0 Þ þ kd23 ð^s;^s0 Þ dO0 4p

ð

ka4 ð^sÞ ¼ 2ke14 ð^sÞ þ 2ke24 ð^sÞ  2 ½kd14 ð^s;^s0 Þ þ kd24 ð^s;^s0 Þ dO0 4p

(43) where the elements kemn and kdmn are the mn elements of ke and kd, respectively, with m,n ¼ 1,2,3,4. Equations 42 and 43 state that the emission of the particle ensemble in the forward direction is related to the absorption by the particle ensemble in the backward direction. Of course, the boundary conditions for the vector RTE need to be specified for the Stokes vectors as well. For example, in case of a planar dielectric interface, the Fresnel theory can be generalized so that the reflected Stokes vector Ir is given by 2

6 6 6 Ir ðr; ^sÞ ¼ Rðyi ÞIi ðr; ^sÞ ¼ 6 6 4

Rv ðyi Þ

0

0

Rh ðyi Þ

0

0

0

0

0

0

3

7 7 0 0 7     7   7 Re Gv Gh Im Gv Gh 5     Re Gv Gh Im Gv Gh

(44) where R is the surface reflectance matrix and Rv(yi) ¼ |Gv(yi)|2 and Rh(yi) ¼ |Gh(yi)|2 are the vertically and horizontally polarized, respectively, surface reflectance coefficients at the incident angle yi (satisfying the Snell law) with Gv and Gh, the Fresnel reflection field coefficients. For a Lambertian surface, the surface reflectance matrix is a matrix with all terms equal to zero except R11 and R22 both equal to 0.5.

Numerical and approximate solutions The radiative transfer equation, in its scalar and vector form, is an integrodifferential equation which does not have analytical solutions, except in some special cases. Approximations and numerical techniques are usually adopted for solving the RTE (Chandrasekhar, 1960; Sobolev, 1975; Ishimaru, 1978; Tsang et al., 1985; Ulaby et al., 1986). Among these techniques, we enumerate (1) the iterative solution technique, (2) the discrete-ordinate technique and the invariant imbedding method, (3) the invariant imbedding technique, (4) the two-flux approximate solution, (5) the Eddington approximate solution, and (6) the diffusion approximate solution. In the iterative solution, the radiative transfer equation is cast into an integral form and then solved iteratively to obtain the zero-, first-, and second-order solutions

(e.g., Tsang et al., 1985). This technique is applicable when the random medium is weakly scattering, that is, when the albedo w is small. The s solution is computed by ignoring scattering except for its contribution to extinction (the multiple-scattering term is disregarded). In principle, we can obtain an accurate solution by iterating many times. But, in practice, the iteration beyond the second order requires a substantial amount of computational efforts. The discrete-ordinate eigen-analysis technique provides a fairly accurate numerical solution for the RTE and has been extensively used in literature (Chandrasekhar, 1960; Ishimaru, 1978). In this technique, the specific intensity and the phase matrix are first discretized into a finite number of directions by means of a quadrature, typically using a Fourier series expansion in the azimuth angle. It can be shown that for the incident plane wave case, all the Fourier components are independent of each other. Moreover, for the vertically and horizontally polarized waves, the first two terms of the Stokes vector are even functions and the last two terms are odd functions of the azimuth angle, respectively. Then, by means of the Gaussian quadrature typically, the resulting matrix differential equation for the discrete zenith angles is solved by eigen analysis. The discrete-ordinate technique is especially useful for plane-parallel media, but other methods, equally effective, should be also mentioned. Among these, the invariant imbedding (also called adding-doubling or matrix doubling in the scalar and vector RTE, respectively) is based on the general relationships between the reflection and transmission for a finite elementary slab and the formulation of integral equations relating them for a layered planeparallel problem (Tsang et al., 1985). For a one-dimensional (1D) problem, instead of using a series expansion technique or a discrete-ordinate approach, an approximate solution can be used in order to reduce the computational burden. The two-flux theory is, in this respect, very effective if the incident radiation is diffuse and the medium is dull (Kubelka and Munk, 1931). In this case, the RTE is reduced to two coupled differential equations expressed in terms of the forward F+ and backward F fluxes. A drawback of the two-flux theory is that the equation coefficients are not constant and may depend on the angular characteristics of the intensity (they become constant only if the wave is completely diffused). In case of an incident collimated beam, the twoflux theory can be extended to the four-flux theory and successfully applied (Ishimaru, 1978). For layered media, the Eddington solution is an effective approximate solution, based on the angular expansion of the specific intensity and the phase function in a series of Legendre polynomials with unknown coefficients and truncated at the first order (Eddington, 1916). Forwardscattering corrections of the basic solution can be used to improve the accuracy of the results in high albedo media (Joseph et al., 1976; Marzano, 2006). In case of highly scattering media, the diffusion approximation can be used to reformulate the RTE. In this case,

RADIATIVE TRANSFER, THEORY

the diffuse specific intensity is expressed as the average intensity plus a term which accounts for an increase in magnitude in the forward direction of the net flux (Ishimaru, 1978). Coupled with proper boundary conditions, the diffusion approximation takes the form of the fundamental steady-state diffusion equation. In case of isotropic scattering, the phase function can be assumed constant and the formulation further simplifies and can provide useful solutions when particles are much smaller than the wavelength (Menzel, 1966).

Limits and relation with the wave analytical theory As already mentioned, the radiative transfer theory describes the power intensity balance of the transmission, absorption, and scattering processes and heuristically derived from considerations on energy conservation, as apparent in Equation 21. The RTE is based on the assumption that the correlation between the fields scattered by the different particles in the random medium is negligible and the addition of power, rather than the addition of fields, holds. However, the conventional RTE as in Equation 21 may be not valid for densely distributed media and fails to account for the backscattering enhancement effect which may be significant under certain circumstances (Kuga and Ishmaru, 1989). Backscattering enhancement is a phenomenon in which a strong increase in the diffuse intensity can be observed in the backward direction when the random medium is illuminated by a plane wave, caused by the constructive interference of the two waves propagating in the opposite directions in the random medium. The intensity enhancement may be twice as large as the unenhanced intensity and the angular width may be much smaller than 1 . The backscattering enhancement cannot be explained by the radiative transfer theory as it does not include the phase correlation between waves. In order to explain these effects, we need to resort to the field theory. In fact, the same problem of the RTT can be cast in terms of field quantities and their statistical characterization. When this approached is followed, we can refer to the analytical wave theory (or multiple-scattering theory), whose general formulation is well summarized by the integral equation set of Twersky-Foldy (Twersky, 1964; Ishimaru, 1978). The analytical wave theory generally starts from the Maxwell wave equation of a scalar field C(r), which can represent either a component of the electric or magnetic field or a pressure wave and describes the various processes of multiple scattering in a detailed way, usually resorting to the concise formalism of the diagram method. The general result of is that the total field C(r) is composed by (1) the incident wave, (2) multiply scattered waves involving chains of successive scattering through different scatterers, and (3) multiply scattered waves containing all the paths which go through a scatterer more than once. This general picture is consistent with that schematically shown in Figure 1. The Twersky theory adopts the so-called expanded representation which

633

retains only the previous groups (1) and (2) but disregards the effects due to group (3). When the number of scattering events becomes large, the errors associated to the Twersky approximation become very small. Starting from the Twersky expanded representation, it is then possible to provide a statistical description of the random scalar field C(r), distinguishing between the coherent (or average) field Cc(r) and the diffuse (also incoherent or fluctuating) field Cd(r): cðrÞ ¼ cc ðrÞ þ cd ðrÞ

(45)

where Cc(r) ¼ and < Cd(r) > ¼0 with the angle brackets indicating ensemble average. It can be shown that the total field intensity < |C(r)|2>, which can be decomposed into the sum of coherent intensity | < C(r) > |2 and diffusive intensity < |Cd(r)|2 > from Equation 45, satisfies the so-called Twersky-Foldy integral equations. The latter are, in their turn, consistent with the first-order smoothing approximation of the more rigorous Dyson and Bethe-Salpeter equations (e.g., Furutsu, 1975). Considering that both the RTE in Equation 21 and the Twersky-Foldy integral equations describe the same multiple-scattering phenomenon, even though from two different points of view, a link between the two theories is expected (Ishimaru, 1975). Indeed, if GC(r,Dr) is the correlation function (or mutual coherence function) of the random field C(r) in a given position r and at a distance Dr, it can be shown that it holds the following approximate equality (Furutsu, 1975): ð (46) Gc ðr; 0Þ ¼ffi I ðr; ^sÞdO 4p

The previous relation indicates that the angular integral of the total specific intensity I (i.e., the radiation intensity) is proportional to the total field intensity < |C(r)|2>. Equation 46 can be generalized showing that the field correlation function GC is related to the Fourier transform of the specific intensity I. It shows that the description of multiple-scattering effects, based on the radiative transfer theory, does include information concerning the field quantities in the form of mutual coherence function (Ishimaru, 1978). It is worth noting that if the particles are sparsely distributed, the extinction coefficient ke is linearly proportional to the particle number concentration Nq, as in Equation 13. This is a valid assumption if the fractional volume density is less than 1 %. Beyond 1 %, the extinction coefficient is no longer linearly proportional to the number concentration, but starts decreasing in a way dependent on the size parameter, dielectric constant, and particle volume fraction. This behavior in dense media has been demonstrated by an extensive experimental and theoretical analysis (Kuga and Ishimaru, 1989; Tsang et al., 1985). The previously mentioned Twersky model is relatively simple, but applicable only for small particles with respect to the wavelength. For volume fractions

634

RADIO-FREQUENCY INTERFERENCE (RFI) IN PASSIVE MICROWAVE SENSING

higher than 10 %, the quasi-crystalline approximation (QCA) with either the Percus-Yevick pair correlation function or with the coherent potential approximation has shown to be effective solutions (Tsang et al., 1985).

Bibliography Chandrasekhar, S., 1960. Radiative Transfer. New York: Dover. Davison, B., 1958. Neutron Transport Theory. London: Oxford University Press. Eddington, A. S., 1916. On the radiative equilibrium of stars. Monthly Notices of the Royal Astronomical Society, 77, 16–35. Elachi, C., 1987. Introduction to the Physics and Technology of Remote Sensing. New York: Wiley. Fung, A. K., 1994. Microwave Scattering and Emission Models and Their Applications. Norwood: Artech House. Furutsu, K., 1975. Multiple scattering of waves in a medium of randomly distributed particles and derivation of the transport equation. Radio Science, 10, 29–44. Ishimaru, A., 1975. Correlation function of a wave in a random distribution of stationary and moving scatterers. Radio Science, 10, 45–52. Ishimaru, A., 1978. Wave Propagation and Scattering in Random Media. New York: Academic, Vol. I and II. Ishimaru, A., and Kuga, Y., 1982. Attenuation constant of coherent field in a dense distribution of particles. Journal of the Optical Society of America, 72, 1317–11320. Joseph, J. H., Wiscombe, W. J., and Weinman, J. A., 1976. The delta Eddington approximation for radiative flux transfer. Journal of Atmospheric Sciences, 33, 2452–2459. Kubelka, P., and Munk, F., 1931. Ein beitrag zur optic der farbanstriche. Zeitschrift für Technische Physik, 12, 593–603. Kuga, Y., and Ishimaru, A., 1989. Backscattering enhancement by randomly distributed very large particle. Applied Optics, 28, 2165–21169. Marzano, F. S., 2006. Modeling antenna noise temperature due to rain clouds at microwave and millimeter-wave frequencies. IEEE Transactions of Antennas and Propagation, 54, 1305–1317. Menzel, D. H. (ed.), 1966. Selected Papers on the Transfer of Radiation. New York: Dover. Schuster, A., 1905. Radiation through a foggy atmosphere. Astrophysics Journal, 21, 1–22. Schwarzschild, K., 1906. Nachrichten königlichen gesellschaft wissensch. Göttingen Mathematisch-Physikalische Klasse, 195, 41–64. Sobolev, V. V., 1975. Light Scattering in Planetary Atmospheres. New York: Pergamon. Sommerfeld, A., 1956. Thermodynamics of Statistical Mechanics. New York: Academic. Tsang, L., and Ishimaru, A., 1987. Radiative wave equations for vector electromagnetic propagation in dense nonsparse media. Journal of Electromagnetic Waves and Applications, 1, 52–72. Tsang, L., Kong, J. A., and Shin, R. T., 1985. Theory of Microwave Remote Sensing. New York: Wiley. Tsang, L., Kong, J. A., and Ding, K.-H., 2000. Scattering of Electromagnetic Waves: Theories and Applications. New York: Wiley. Twersky, V. 1964. On propagation in random media of discrete scatterers. In Stochastic processes in mathematical physics and engineering: Proceedings of a Symposium in Applied Mathematics of the American Mathematical Society, New York, pp. 84–116. Ulaby, F. T., and Elachi, C. (eds.), 1990. Radar Polarimetry for Geoscience Applications. Norwood: Artech House. Ulaby, F. T., Moore, R. K., and Fung, A. K., 1986. Microwave Remote Sensing. Norwood: Artech House, Vol. III.

Wen, B., Tsang, L., Winebrenner, D. P., and Ishimaru, A., 1990. Dense medium radiative transfer theory: comparison with experiment and application to microwave remote sensing and polarimetry. IEEE Transactions on Geoscience and Remote Sensing, 28, 46–59.

Cross-references Electromagnetic Theory and Wave Propagation Fields and Radiation Media, Electromagnetic Characteristics Optical/Infrared, Radiative Transfer Radiation, Multiple Scattering Radiation, Polarization, and Coherence Radiation, Volume Scattering Radiative Transfer, Solution Techniques Remote Sensing, Physics and Techniques

RADIO-FREQUENCY INTERFERENCE (RFI) IN PASSIVE MICROWAVE SENSING David Kunkee The Aerospace Corporation, Los Angeles, CA, USA

Synonyms Electromagnetic interference (EMI) Definition Radio-frequency interference (RFI) in passive microwave remote sensing occurs when anthropogenic (man-made) signals contaminate calibrated radiometric brightness temperature measurements of naturally occurring background thermal radiation resulting in anomalous measurements. Introduction Radio-frequency interference (RFI) to space-based passive microwave remote sensing systems became an important consideration after brightness temperature measurements from the Advanced Microwave Scanning Radiometer on NASA’s Earth Observing System (AMSR-E) showed widespread contamination from anthropogenic emissions. The AMSR-E, part of NASA’s Aqua mission, was launched on May 4, 2002, and was the first space-based radiometer since the scanning multichannel microwave radiometer (SMMR) (the SeaSat 7 mission ended in October 1987 and carried the last SMMR) to measure upwelling brightness temperatures near 6.8 GHz. Contamination (RFI) of brightness temperature measurements by SMMR radiometers at 6.6 GHz (ending in October 1987) was not significant over North America. However, beginning in 2002, substantial contamination of radiometric brightness temperatures near 6.9 GHz, from the AMSR-E, was found over North America as well as other populated regions of the world (Li et al., 2004). Contaminated brightness temperatures near 6.8 GHz, a spectral region important for remote sensing

RADIO-FREQUENCY INTERFERENCE (RFI) IN PASSIVE MICROWAVE SENSING

of all-weather sea surface temperature and soil moisture retrievals, helped initiate the current research in RFI mitigation for passive microwave measurements and focus attention on protection of the RF spectrum for scientific uses.

Background Passive microwave radiometers respond to the naturally occurring upwelling background radiation from the Earth’s surface and atmosphere. The power contained in these emissions is a function of bandwidth and determined by the formula P ¼ kTB, where k is Planck’s constant, 1.038  1023 J s2, T ¼ T is the effective brightness temperature in Kelvin, and B is the effective RF bandwidth of the radiometer in Hz. When the radiometer responds to anthropogenic signals either within or outside of the radiometer’s primary passband, the total RF power received by the radiometer increases; PT ¼ kTBB + PRFI where PRFI is mean power of the anthropogenic signal(s) during the radiometer’s period of integration adjusted by the radiometer’s antenna beam pattern and receiver passband response characteristics. As a result, the effective brightness temperature measured by the radiometer is larger than RFI the scene brightness temperature dTB ¼ PkB . This value represents the direct level of contamination. In general, the level of contamination, dTB, is a function of radiometer bandwidth and strength of interfering signal. Larger bandwidth radiometers may tolerate higher levels of interference; however, they are sensitive to emissions over a wider range of frequencies. It is also important to

635

understand that out-of-band (OOB) signals also contribute to PRFI. The OOB signals can be a significant consideration especially when strong anthropogenic emissions occur in the same spectral region that the radiometer is operating. Another form of RFI that may impose errors on measured brightness temperatures occurs when strong OOB signals are present at the receiver (radiometer) input and impact the operating characteristics of the receiver. The OOB signals may otherwise be rejected by the receiver channelization filter; however, if the signals are of sufficient strength to impact operation of the receiver low noise amplifier (LNA), errors can occur in the measured data. These “indirect” effects from strong OOB signals may result in a net decrease or increase in measured brightness temperature depending on the nature of the response of the receiver front end and RF electronics.

Examples of RFI in several microwave bands One of the most extensive examples of contaminated brightness temperatures can be seen in measurements of over North America from AMSR-E near 6.9 GHz (Figure 1). The data indicate that widespread utilization of this spectral region by other radio services (the spectrum segment is allocated to the fixed service (FS) and mobile service (MS) according to the current NTIA allocation chart (http://www.ntia.doc.gov/osmhome/redbook/ redbook.html)) was causing substantial contamination to brightness temperatures measured by the AMSR-E vertically and horizontally polarized channels at 6.9 GHz.

AMSR–E RFI Index, 6.9 GHz V–Pol, 7/11/02, 7/12/02 50

Latitude (deg)

40

30

20 −130

0

−120

10

−110

−100 −90 Longitude (deg) 20

30

−80

40

−70

50

−60

60

Radio-Frequency Interference (RFI) in Passive Microwave Sensing, Figure 1 RFI is displayed as the perturbation from a zero-mean (natural emission) level determined by differencing 6- and 10 GHz brightness temperatures measured by AMSR-E. Perturbations of up to 50 K are common across the USA affecting more than 50 % of the total land area with RFI > 5 K. The pervasive nature of the interference resulted in reduced quality of soil moisture retrievals using AMSR-E because 6.9 GHz data could not be used (Li et al., 2003).

636

RADIO-FREQUENCY INTERFERENCE (RFI) IN PASSIVE MICROWAVE SENSING

k 330 310 290 270

40

250 230 210

30

190 −120

−70 −110

−90

−100

170

−80

150 k 330 310 290

40

270 250 230 210

30

190 170 2013

1656

−100

150

1834

Radio-Frequency Interference (RFI) in Passive Microwave Sensing, Figure 2 (Top) Passive microwave brightness temperatures at 6.6 GHz from SMMR in October 1987. Essentially no RFI is seen over the North America in this period. (Bottom) Brightness temperature data at 6.9 GHz from AMSR-E in May 2003. The black spots represent high levels of anthropogenic emission primarily over regions of California and Arizona that saturate the AMSR-E radiometer. The red spots over most of the remaining areas of the USA represent contaminated brightness temperature measurements.

Contamination of background brightness temperatures is shown by departure from the mean difference of 6.9 and 10.7 GHz brightness temperatures in the figure. Positive values indicate RFI at 6.9 GHz. Contamination of brightness temperature measurements in the 6 GHz region was not found during the period of SMMR measurements from 1978 to 1987 (Njoku et al., 1980); brightness temperatures over North America were essentially contamination free as shown by data from SeaSat Nimbus 7 in contrast to AMSR-E data from 2002 (Figure 2). The red and black

spots in the lower part of the figure represent anthropogenic sources causing localized errors in measurements. In this example there is a high correlation of contaminated brightness temperatures with population centers. RFI contamination to passive microwave brightness measurements has been observed in the 10 GHz region by the NRL WindSat radiometer, AMSR-E, and NASA’s Tropical Rainfall Measuring Mission (TRMM) Microwave Imager (TMI). Two common types of occurrences have been noted. First, land-based anthropogenic

RADIO-FREQUENCY INTERFERENCE (RFI) IN PASSIVE MICROWAVE SENSING

637

k 330

320 55 310

300

50

290

280

270 45 260 10

–5

5

0 1301

10 250

1122

Radio-Frequency Interference (RFI) in Passive Microwave Sensing, Figure 3 Passive microwave brightness temperatures at 6.9 GHz from AMSR-E in June 2003 over France and the UK. The black spots represent saturation of the AMSR-E response indicating severe RFI. In this example occurrences are correlated with national boundaries.

emissions have caused localized contamination of brightness temperature measurements (Figure 3). Second, reflections of geostationary broadcast transmissions to Earth from the ocean cause broad areas of contamination to measurements of brightness temperature (Figure 4). This contamination adversely impacts retrievals of sea surface temperature (SST) and sea surface wind (SSW) that rely directly on passive microwave brightness temperature data from the 10 GHz band. Anomalies in sea surface wind retrievals near North America have also been attributed to RFI from geostationary broadcasts in the 18.6–18.8 GHz band segment. Flagging of potential anomalies in the WindSat 18.7 GHz data from NRL’s retrieval algorithm is shown (Figure 5). Blue denotes fewer anomalies and red denotes a higher occurrence of anomalies. The first occurrence of these retrieval anomalies coincides with the activation of the DirecTV 10 geostationary broadcast satellite in early October 2007. The satellite operates within the band 18.3–18.8 GHz at 103 West longitude. In late July 2008, DirecTV 11 also went online at 99 West longitude with the same coverage area and downlink frequencies as DirecTV 10. The anomalies, found only during the descending phase of WindSat, are attributed to reflection of the satellite transmissions from the

ocean surface. The anomalies correlate precisely with the nationwide coverage maps of DirecTV 10 and DirecTV 11.

RFI mitigation techniques After the AMSR-E brought RFI considerations to the forefront, development of RFI mitigation techniques has been under way by many researchers in passive microwave remote sensing. Techniques include frequency subbanding (Gasiewski et al., 2002), detection of higher-order statistics (Ruf et al., 2006; Misra et al., 2008; De Roo et al., 2007; Piepmeier et al., 2008), and frequency-and timedomain methods (Johnson et al., 2006; Niamsuwan et al., 2005; Güner et al., 2007). Frequency sub-banding utilizes comparisons of measured brightness temperatures from each sub-band to determine if any single sub-band or multiple sub-bands are contaminated with RFI. Timedomain methods can be effective against pulsed emissions by utilizing short-duration blanking to eliminate the effect of the pulsed emissions. Higher-order statistics of the radiometric signal are generally sensitive to anthropogenic emissions (non-Gaussian) but are not sensitive to naturally occurring variations in scene brightness temperature. Therefore, measurements of higher-order statistics of the

638

RADIO-FREQUENCY INTERFERENCE (RFI) IN PASSIVE MICROWAVE SENSING

AMSR-E 10V RFI over Ocean (7/4/03) 65 60

Latitude (deg)

55 50 45 40 35 30 0

–10

10

20

30

Longitude (deg) 0

10

20

30

40

50

60

Radio-Frequency Interference (RFI) in Passive Microwave Sensing, Figure 4 Example of RFI incurred at 10 GHz from oceanic reflections of broadcasts in bands adjacent to those observed by AMSR (areas in green and yellow). In this example AMSR-E, operating in the EESS band 10.6–10.7 GHz, is experiencing > 40 K brightness temperature perturbations, a value that greatly exceeds minimum levels that degrade environmental models using SST data derived from AMSR-E (Courtesy Li Li US NRL).

signal such as the fourth moment (kurtosis) can be used to detect the presence of anthropogenic emissions and RFI and serve to enhance effective mitigation of its impact. Although increasingly effective RFI mitigation techniques are under development, 100 % RFI mitigation will not be possible. Limited information from the scene brightness temperatures is lost in the presence of detectable anthropogenic emissions. The impacts and observations of RFI are also continuing to increase. Concurrent growth in communications and wireless markets are driving new applications of RF transceivers along with new developments and applications in passive microwave remote sensing that require improved levels of precision or additional frequency bands have resulted in increasing reports and concerns of RFI. For example, in the post SMMR radiometer era, 10 GHz space-based brightness temperature measurements from the TMI radiometer began in 1997, 6 GHz measurements were available from AMSR-E in 2002, and currently, there are two space-based radiometers operating at 1.4 GHz and another planned for 2014. These missions require new levels of radiometric precision and stability in a region of the spectrum that has many active users. As a result, there is increasing attention on protection of the RF spectrum for scientific uses.

Frequency allocations and regulation In the USA, the National Telecommunications and Information Administration (NTIA) and Federal

Occurrence of 18.7 GHz RFI (Count, September - October 2007) 50

45

Latitude (deg)

40

35

30

25

20 –130

0

–120

–110

20

–100 –90 Longitude (deg) 40

60

–80

–70

80

–60

100

Radio-Frequency Interference (RFI) in Passive Microwave Sensing, Figure 5 Occurrence of anomalous ocean wind retrievals near the North American shore. The anomalies began with the commencement of transmissions from a new geostationary broadcast satellite near 18.7 GHz on 1 October 2007 (Courtesy Li Li US NRL).

RADIO-FREQUENCY INTERFERENCE (RFI) IN PASSIVE MICROWAVE SENSING

Communication Commission regulate the licenses for government and nongovernment uses, respectively, of the radio services. Licenses are issued in a manner consistent with the international allocation of frequencies as agreed by the International Telecommunication Union (ITU) and World Radio Conference (WRC) (ITU, 2001). WRCs are currently held every 2 years in order to address international frequency-coordination issues. One of the major tasks of the ITU is to provide recommendations for regulation of radio spectrum among the various radio services including the Earth Exploration Satellite Service (EESS) (passive microwave radiometry is classified as the Earth Exploration Satellite Service (EESS) – passive), radio astronomy service, satellite broadcast service, fixed service, and mobile service, for example. Recommendations are provided for EESS by committees at the US national and international levels called working party 7C (WP7C). The WP7Cs are composed of representatives that understand the regulatory and technical issues from agencies and organizations with an interest or mission to provide microwave remote sensing data for Earth-observation applications. One of the most important ITU recommendations relevant to EESS is ITU-R SA.1028 and SA.1029. These ITU recommendations suggest the maximum allowable anthropogenic emissions within sample bandwidths utilized for passive remote sensing and EESS to ensure interference-free operation. More information on the regulatory process related to scientific uses of the spectrum can be found in the Handbook of Frequency Allocations and Spectrum Protection for Scientific Uses (NRC, 2007).

Conclusion Interference to space-based passive microwave remote sensing systems has continued to increase after the discovery of pervasive contamination of AMSR-E 6.9 GHz brightness temperature data. Transmissions originating with broadcast satellite systems operating in geosynchronous orbit near 10.7 and 18.7 GHz as well as land-based emissions in the 6-, 10-, and 18 GHz bands have been shown to impact passive microwave measurements. Over land, weak signals originating from pointto-point communication systems (fixed service) are sensed by space-based passive microwave radiometers observing near 10.65 and 18.6 GHz. Over ocean, data from NRL’s WindSat show a negative impact to retrievals of sea surface temperature and ocean winds due to RFI at 10 GHz and recently due to RFI at 18.7 GHz near North America. In the past several years, the remote sensing and radio science communities have responded with research and improvements to RFI mitigation approaches; however, the effectiveness of the approaches is limited, and contaminated brightness temperature measurements result in quantifiable data loss even with effective RFI mitigation in place. From the regulatory perspective, allocations of radio spectrum to

639

specific radio services are determined by international agreement and documented by the ITU. Much of the spectrum is allocated to multiple radio services and sharing is required. In many cases, the requirements for noninterference to EESS systems by other radio services is challenging and not well understood. The ITU is also a source for recommendations concerning radio regulations and sharing criteria. The increasing pervasiveness of RFI to passive microwave measurements is a growing concern among remote sensing scientists. The Earth’s naturally occurring microwave emissions represent a unique resource for monitoring and understanding our planet with direct application to weather forecasting and climate.

Bibliography De Roo, R. D., Misra, S., and Ruf, C. S., 2007. Sensitivity of the kurtosis statistics as a detector of pulsed sinusoidal RFI. IEEE Transactions on Geoscience and Remote Sensing, 45(7), 1938–1946. Gasiewski, A. J., Klein, M., Yevgrafov, A., and Leuskiy, V., 2002. Interference mitigation in passive microwave radiometry. In IEEE International Geoscience and Remote Sensing Symposium, 2002. IGARSS’02, Toronto, Vol. 3, pp. 1682–1684. Güner, B., Johnson, J. T., and Niamsuwan, N., 2007. Time and frequency blanking for radio frequency interference mitigation in microwave radiometry. IEEE Transactions on Geoscience and Remote Sensing, 45(11), 3672–3679. ITU, 2001. Radio Regulations, International Telecommunications Union. Geneva. http://www.itu.int/net/home/index.aspx. Johnson, J. T., Gasiewski, A. J., Güner, B., Hampson, G. A., Ellingson, S. W., Krishnamachari, R., Niamsuwan, N., McIntyre, E., Klien, M., and Leuski, V. Y., 2006. Airborne radio-frequency interference studies at C-band using a digital receiver. IEEE Transactions on Geoscience and Remote Sensing, 44(7), 1974–1985. Li, L., Njoku, E. G., Im, E., Chang, P. S., and St. Germain, K., 2004. A preliminary survey of radio-frequency interference over the U.S. in aqua AMSR-E data. IEEE Transactions on Geoscience and Remote Sensing, 42(1), 380–390. Misra, S., Ruf, C., and Kroodsma, R., 2008. Detectability of radio frequency interference due to spread spectrum communication signals using the kurtosis algorithm. In IEEE International Geoscience and Remote Sensing Symposium, 2008. IGARSS’08, Boston, Vol. 2, pp. 335–338. Niamsuwan, N., Johnson, J. T., and Ellingson, S. W., 2005. Examination of a single pulse blanking technique for radio frequency interference detection and mitigation. Radio Science, 40(5), RS5–S03. Njoku, E., Stacey, J. M., and Barath, F. T., 1980. The seasat scanning multichannel microwave radiometer (SMMR): instrument description and performance. IEEE Journal of Oceanic Engineering, OE-5(2), 100–113. NRC, 2007. Handbook of Frequency Allocations and Spectrum Protection for Scientific Uses. Washington, DC: National Academies Press. Piepmeier, P. M., and Knuble, J., 2008. A double detector for RFI mitigation in microwave radiometers. IEEE Transactions on Geoscience and Remote Sensing, 46(2), 458–465. Ruf, C. S., Gross, S. M., and Misra, S., 2006. RFI detection and mitigation for microwave radiometry with an agile digital detector. IEEE Transactions on Geoscience and Remote Sensing, 44(3), 694–706.

640

RAINFALL

RAINFALL Ralph Ferraro NOAA/NESDIS, ESSIC/CICS, College Park, MD, USA

Synonyms Rain; Rainwater; Showers Definition Rainfall. A form of precipitation that reaches the Earth’s surface in a liquid phase. Introduction The remote sensing of rainfall is a vital component to the integrated observing of precipitation on the Earth. While weather radars and rain gauges are the primary source of rainfall estimation in many areas, they are typically restricted to populated areas on the Earth and can only extend out over water bodies 150 km or so. Thus, satellites serve to fill in these huge data voids, especially over unpopulated regions and oceans. In this section, only satellite-based methods are considered (see Water Resources; Global Earth Observation System of Systems (GEOSS)). Additionally, this section refers to precipitation that falls at the surface in the form of liquid (see Snowfall). A number of different methods are used to retrieve rainfall from satellites. The strength and weaknesses of the various methods that are described are summarized in Table 1, whereas details on the general classes of methods follow in the next three sections. Satellites (see Observational Systems, Satellite) that are used to estimate rainfall are generally categorized into LEO and GEO. Then, the various retrieval algorithms are typically classified on their observing spectrum (VIS, IR, PMW, AMW) or “multispectral” (i.e., use of one or more of these individual spectrums), whether they are objective or interactive (i.e., manual intervention) or whether they use multiple satellites or other information such as radar or gauges (e.g., “blended” algorithms). Visible and infrared methods Visible (VIS) and infrared (IR) techniques (see Optical/ Infrared, Radiative Transfer) were the first to be conceived and are rather simple to apply, while at the same

time they show a relatively low degree of accuracy. On the other hand, GEO weather satellite VIS and IR imagers uniquely provide the rapid temporal update cycle (e.g., 30 min or less) needed to capture the growth and decay of precipitating clouds. A complete overview of the early work and physical premises of VIS and thermal IR (10.5–12.5 mm) techniques is provided by Barrett and Martin (1981), while Kidder and Vonder Haar (1995) present some of the more recent results. The rainfall retrieval in these wavelengths is based on the fact that bright (optically thick) clouds are positively correlated with regions of convective rainfall (Woodley et al., 1971). On the other hand, clouds with cold tops in the IR imagery produce more rainfall than those with warmer tops (Scofield, 1987). However, the correspondence between cold tops and visible bright spots is far from perfect and is not always well correlated with surface rainfall (especially in stratiform rainfall regimes). Figure 1 presents an example of typical VIS and IR signatures associated with different types of rainfall. Various approaches have been developed to stress particular aspects of the sensing of cloud physics properties to settle differences between VIS and IR retrievals and measured rainfall. Following Barrett and Martin’s classification, the rainfall estimation methods can be divided in the following categories: cloud indexing, bi-spectral schemes, life history, and cloud model based. In order to describe some methodologies that appear after Barrett and Martin’s work, a new category that includes Numeric Weather Prediction (NWP)-adjusted schemes and multispectral algorithms has been added.

Cloud indexing methods Cloud indexing techniques assign a rain rate level to each cloud type identified in the satellite imagery. The simplest and perhaps most widely used is the one developed by Arkin (1979) during the GARP (Global Atmosphere Research Program) Atlantic Tropical Experiment (GATE) on the basis of a high correlation between radar-estimated precipitation and fraction of the area colder than 235 K in the IR. The scheme, named the GOES Precipitation Index (GPI) (Arkin and Meisner, 1987), assigns these areas a constant rain rate of 3 mm h1, which is appropriate for tropical precipitation over 2.5 latitude  2.5 longitude areas. The GPI is a standard for climatological

Rainfall, Table 1 Summary of satellite measurements and their attributes Observation spectrum

Satellite type

Sensor examples

Strength

Weakness

Visible (VIS)

Passive microwave (PMW)

GEO LEO GEO LEO LEO

Cloud type Cloud evolution Cloud temperature Cloud evolution Direct measure of rain, especially over ocean

Active microwave (AMW)

LEO

GOES imager AVHRR GOES imager AVHRR SSM/I AMSR-E TRMM PR CloudSat CPR

Cloud tops Indirect rain rate Cirrus contamination Indirect rain rate Temporal sampling Indirect rain rate (land) Narrow swath width Rain rate sensitivity

Infrared (IR)

Direct measure of vertical structure of rain

RAINFALL

641

Rainfall, Figure 1 Comparison of VIS (left image) and IR (right image) from May 20, 2008 – 19:45Z over Central America. The red circle shows a feature with high reflectivity but high temperature (presumably low warm clouds with no rain associated), while the purple square shows a region with low temperature and very high reflectivity (large convective clouds). The last feature is related with higher rain rates.

rainfall analysis (Arkin and Janowiak, 1991) and is regularly applied and archived for climatological studies (see Climate Monitoring and Prediction). A family of cloud indexing algorithms was developed at the University of Bristol. “Rain days” are identified from the occurrence of IR brightness temperatures (TB) below a threshold at a given location. The estimated rain days are combined with rain-per-rain day means that are spatially variable to produce rainfall estimations for extended periods (10 or more days). The most recent version of the Bristol technique uses variable IR rain/no-rain TB thresholds (Todd et al., 1995, 1999).

shrinking clouds. A major problem arises in the presence of cirrus anvils from neighboring clouds: they often screen the cloud life cycle underneath, leading to possible underestimates early in the day and overestimates toward the evening. Reasonable performances of this type of methods are obtained for deep convective storms, while contradictory results arise from their application to stratiform systems or weak convection (Amorati et al., 2000).

Bi-spectral methods Bi-spectral methods are based on the very simple, although not always accurate, relationship between cold and bright clouds and high probability of precipitation, which is characteristic of cumulonimbus. Lower probabilities are associated to cold but dull clouds (thin cirrus) or bright but warm clouds (stratus). The RAINSAT technique (Lovejoy and Austin, 1979; Bellon et al., 1980) screens out cold but not highly reflective clouds or those that are highly reflective but have a relatively warm top. The number of false alarms of the pure IR techniques is reduced. The algorithm is based on a supervised classification trained by radar to recognize precipitation from both VIS brightness and IR brightness temperature.

Cloud model-based techniques Cloud model techniques aim at introducing the cloud physics into the retrieval process for a quantitative improvement deriving from the overall better physical description of the rain formation processes. A onedimensional cloud model relates cloud-top temperature to rain rate and rain area in the Convective Stratiform Technique (CST) (Adler and Negri, 1988). Local minima in the IR brightness temperature are sought and screened to eliminate thin, nonprecipitating cirrus. To do so a slope parameter is calculated for each temperature minimum. Adler and Negri (1988) have established an empirical discrimination of thin cirrus in the temperature/slope plane using radar and visible imagery data. If the Tmin and its slope fall to the left of the discrimination line, the Tmin location is classified as thin cirrus (non-raining). A larger slope implies a more clearly defined minimum, corresponding to a thunderstorm.

Life-history methods A family of techniques that specifically requires GEO satellite imagery are the life-history methods that rely upon a detailed analysis of the cloud’s life cycle, which is particularly relevant for convective clouds. An example is the Griffith-Woodley technique (Griffith et al., 1978). In this case, the increasing cloud area is treated different from

NWP-adjusted and multispectral methods Numerical weather prediction (NWP) outputs are being incorporated widely in satellite rainfall estimations. These correction factors range from moisture correction (Vicente et al., 1998) to orographic-driven precipitation. In the latter case, Digital Elevation Model (DEM) and low tropospheric winds are used to adjust the final

RAINFALL

O2

Window

0.6

0.4

0.2

H2O Moisure Sounding

0.8

O2 Temperature

H2O

Window

Vertical Transmittance to Space

1.0

Window

642

0.0 0

50

100

150

200

250

300

Frequency (GHz)

Rainfall, Figure 2 Vertical transmittance from the surface to space in the 15  N annual atmosphere.

retrievals (Vicente et al., 2002). The Auto-Estimator technique (Vicente et al., 1998) proposed a moisture correction factor defined as the product of precipitable water (PW) in the layer from the surface to 500 mb times relative humidity (RH) (mean values between the surface and the 500 mb level) data. The PWRH factor decreases rainfall rates in very dry environments and increases them in very moist ones. The most recent version of this algorithm, called Hydro-Estimator, uses PW to adjust the power-law relation between cloud-top temperature and precipitation rates, while RH is applied to adjust the final rain rates (Scofield and Kuligowski, 2003). The GOES Multispectral Rainfall Algorithm, GMSRA, (Ba and Gruber, 2001) combines multispectral measurements of the GEO satellites to estimate rainfall. One of the principal innovations of GMSRA relative to previous IR/VIS algorithms is that it combines several cloud properties used in a variety of techniques in a single and comprehensive rainfall algorithm. To be specific, the technique uses cloud-top temperatures as a basis of rain estimation (e.g., Arkin and Meisner, 1987; Vicente et al., 1998), and it utilizes the effective radii of cloud particles (e.g., Rosenfeld and Gutman, 1994) and spatial and temporal temperature gradients (e.g., Adler and Negri, 1988; Vicente et al., 1998) to screen out non-raining clouds.

Passive microwave methods Unlike VIS and IR signals, microwave energy can penetrate clouds, in particular, cirrus clouds, and its signal has a strong interaction with precipitation-size drops and ice particles. This direct impact on microwave

measurements by hydrometeors allows for the quantitative detection of precipitation properties in the atmosphere as well as on the surface. It should be pointed out that passive MW (PMW) means naturally emitted radiation from the Earth’s surface and atmosphere that interacts with clouds and precipitation and is measured by a radiometer onboard a satellite. Most passive microwave radiometers launched to date operate in frequencies ranging from 6 to 190 GHz. At different frequencies, microwave radiometers observe different parts of the rain profile. Below 20 GHz, emission by precipitation-size drops dominates and ice particles above the rain layer are nearly transparent. Above 60 GHz, ice scattering dominates and the radiometers cannot sense the raindrops below the freezing layer. Both emission and scattering effects are important for frequencies between 20 and 60 GHz. In general, emission by liquid drops raises brightness temperature, while scattering by ice particles has the opposite effect. Besides the shift in dominating mechanism from emission to scattering with the increase of frequency, rain rate also plays a role by enhancing both the emission and the scattering signals. It is noted that the scattering by ice increases much more rapidly with frequency than scattering by liquid (Kidder and Vonder Haar, 1995). Other elements that have significant impacts on microwave radiation are shown in Figure 2. Window channels can measure down to the Earth’s surface and are strongly influenced by surface properties (i.e., vegetation, soil moisture) (see Land Surface Emissivity). Other frequencies are sensitive to oxygen or water vapor/cloud droplets absorption. These microwave properties set the foundation for the development of rainfall estimation schemes.

RAINFALL

There are two major categories in rainfall estimation using passive microwave radiometry: emission-based method and scattering-based method. They will be discussed in the following sections.

Emission-based methods Emission-based rainfall algorithms are mostly applicable over ocean because water surfaces are relatively homogeneous and provide a cold background due to low emissivity. Below certain threshold rain intensity, emission dominates rather than scattering over water especially at lower frequencies. Brightness temperature (TB) increases rapidly with rain rate in this range and provides the basis for many emission-based precipitation retrieval methods. The primary disadvantage of emission-based techniques is that the microwave signal can saturate with moderate to heavy rain rates (which is frequency dependent), thus providing a strong nonlinear relationship between rain rate and TB. The earliest efforts to estimate rain rate from satellite measurements used microwave data from the electrically scanning microwave radiometer (ESMR) on the Nimbus 5 and Nimbus 6 satellites and the scanning multichannel microwave radiometer (SMMR) on both Nimbus 7 and Seasat-A. Wilheit et al. (1977) employed a radiative transfer model to calculate the 19.35 GHz TB as a function of rain rate over ocean and used the 19.35 GHz measurements from ESMR-5 to retrieve rain rate. Weinman and Guetter (1977) utilized the weak polarization of 37 GHz radiances from ESMR-6 to discriminate convective rain over land from open water. Rain rate over water was expressed as a function of TB taking into consideration the polarization effects; a lower polarization was correlated to heavier rain rates. SMMR had five frequencies with both horizontal and vertical polarizations. Spencer et al. (1983) took advantage of the significant impact of raindrops on the thermal emission at lower frequencies (highest SMMR frequency is 37 GHz) and used multiple regression approach to relate SMMR TBs to rain rate. The first of the special sensor microwave/imager (SSM/ I) series was launched in 1987 and was followed by five more successful SSM/I instruments. The SSM/I has seven vertically and horizontally polarized channels and four frequencies at 19.35, 22.235, 37.0, and 85.5 GHz. This is a well-studied instrument for rainfall estimation. Ferraro and Marks (1995) developed an SSM/I algorithm linking oceanic rain rate with liquid water path (Weng and Grody, 1994) using ground-based radar measurements. The retrieval of liquid water path relies on the emission signatures at 19 and 37 GHz. This ocean scheme is part of the operational SSM/I rainfall algorithm as introduced in the next section. Mugnai and Smith (1984) were the first to combine cloud model with radiative transfer model for estimating TBs from convective clouds. Studies along this line of thinking followed (Olson, 1989; Smith et al., 1994). The Goddard profiling algorithm (GPROF) is one of the

643

inversion algorithms that computes radiative transfer based on output from cloud models (Kummerow et al., 1996, 2001). GPROF is applied to the Tropical Rainfall Measuring Mission (TRMM) Microwave Imager (TMI) for the retrieval of surface rain rate and rainfall profile. A large a priori database of atmospheric profiles and TB at the TMI frequencies is constructed using output of two cloudresolving models and a one-dimensional radiative transfer model. Then the algorithm utilizes a Bayesian inversion approach to derive instantaneous rain rate and rainfall profile by a weighted summation of all the profiles in the database. The weight is determined by how close each TB vector in the database resembles the observed vector. GPROF is also the primary rainfall algorithm being applied to the Advanced Microwave Sounding Radiometer-Earth Observing System (EOS) (AMSR-E) (Wilheit et al., 2003). Bauer et al. (2005) proposed a scheme for retrieving precipitation profiles over all surfaces by using microwave frequencies both in atmospheric windows between 18 and 150 GHz and in oxygen absorption complexes or sounding channels near 50–60 and 118 GHz. Hydrometeor profiles are created by applying cloud and convection schemes using information from European Centre for Medium-Range Weather Forecasts (ECMWF) short-range forecasts. A radiative transfer model and one-dimensional variational (1D-Var) retrieval framework are employed to achieve optimum hydrometeor profile retrievals.

Scattering-based methods Due to the higher but more varying emissivity of the land surface, the only reliable means of detecting rainfall over land is by isolating depressed TBs as a result of scattering by millimeter-sized ice particles that exist in most rain clouds (see Optical/Infrared, Scattering by Aerosols and Hydrometeors). Since the signal being captured is a result of ice particles instead of raindrops, the scattering-based rainfall estimation is an indirect measure of rainfall, as it relates the magnitude of the scattering near the freezing layer to surface rainfall. The launch of the SSM/I in 1987 provided the first opportunity to retrieve rain rate through scattering at higher frequency (85 GHz). Meteorological satellites launched since then have included channels at or above 85 GHz that are sensitive to scattering, including more recent sensors that contain measurement channels at or above 150 GHz. Thus, these measurements have been exploited to develop scattering-based rainfall algorithms. Spencer et al. (1989) first introduced the concept of polarization-corrected brightness temperature (PCT) which is a linear combination of the vertically and horizontally polarized brightness temperatures at 85.5 GHz. The coefficients in PCT can be adjusted so that PCT is only sensitive to the scattering in the upwelling microwave radiation but not to other atmospheric and land surface conditions including the contrast of land versus ocean.

644

RAINFALL

An operational SSM/I rain rate product generated at Fleet Numerical Meteorology and Oceanography Center (FNMOC) (Ferraro, 1997) combines a scattering-based global algorithm with the abovementioned SSM/I emission-based ocean algorithm (Ferraro and Marks, 1995). The global algorithm calculates a scattering index (SI) (Grody, 1991) which is the difference between the vertically polarized 85.5 GHz observation and an estimate of the non-scattering contribution from this channel. The latter is a linear combination of the observations at lower channels (19.35 and 22.235 GHz). The SI algorithms are calibrated with ground-based radar measurements to produce instantaneous rain rate for both ocean and land (Ferraro and Marks, 1995). The emission-based algorithm complements the SI algorithm by detecting rain systems over ocean that has little or no ice scattering. A decision tree is included in the algorithm to screen out surface scattering sources such as snow cover and desert. This widely used algorithm is also adapted into GPROF land rain rate algorithm as described by McCollum and Ferraro (2003). The relationships between rain rate and 85 GHz brightness temperature are recalibrated using TRMM Precipitation Radar (PR) data, and a new procedure is developed to estimate convective rainfall. Zhao and Weng (2002) took advantage of the highly scattering nature of 89 and in particular 150 GHz radiances from the Advanced Microwave Sounding Unit-B (AMSU-B) and retrieved ice water path (IWP) using scattering parameters measured at these two channels. The derived IWP is then converted into the surface rainfall rate (RR) through an IWP and rainfall rate relationship developed from cloud model results (Weng et al., 2003). This rain rate product is being operationally generated at National Environmental Satellite, Data, and Information Service (NESDIS) of National Oceanic and Atmospheric Administration (NOAA). This algorithm has also been applied to Microwave Humidity Sounder (MHS) with some modification since AMSU-B and MHS have very similar channels. Vila et al. (2008) added an emissionbased component to this rainfall algorithm to account for oceanic rain systems that have little or no ice in them. Figure 3 shows this rain rate product and some of the corresponding TBs that have the emission and scattering signals of rain systems.

Active microwave methods In contrast to passive microwave radiometers, active microwave sensors provide their own source of microwave radiation. Short pulses are transmitted at microwave frequencies; the time delay and strength of the returned echo gives the distance and intensity of the rain from space. Radar is also able to discriminate in range, that is, in altitude from space. Hence, radar measures fine-scale and vertical distribution of rainfall. In order to provide good horizontal resolution and minimize surface clutter effects for off-nadir incidence without the use of large antennas (greater than 2 m), only frequencies greater than

10 GHz are practical (Fujita and Satake, 1997). At such frequencies, attenuation is significant in moderate to heavy rain and affects the radar reflectivity measurement near the surface where rain information is most desired. Over the rain depth of kilometers, the two-way pathintegrated attenuation (PIA) becomes significant even at moderate rain rates and leads to a bias in rain rate estimate, unless reflectivity is corrected for attenuation effects. Such a bias increases with frequency as a result of an increase in PIA. Techniques have been developed to remove the effects of attenuation when they become significant. The study by Hitschfeld and Bordan (1954) was one of the first to propose a method for directly correcting measured reflectivities for attenuation. After correcting attenuation effects, estimates of non-attenuated reflectivity (Z) are related to rain rate (R) by using Z-R relations. The Hitschfeld-Bordan (HB) solution gives reasonable estimates if the attenuation is small and the radar is well calibrated. When these conditions are not met, the HB solution can become unstable. Numerous retrieval methods have been developed that incorporate this early work along with the more recent technique of constraining the retrievals using PIA derived from surface reference technique (SRT) (Iguchi and Meneghini, 1994; Marzoug and Amayenc, 1994; Amayenc et al., 1996). SRT relies on a reference measurement of the radar return from the surface in rain-free conditions. For example, if the surface returns for a particular incidence angle and surface types are fairly stable, then the difference in surface returns for rain-filled and rain-free radar beams can be ascribed to the effects of attenuation by precipitation. The SRT method provides a reasonable way to monitor PIA over the ocean. Over land, particularly at near-nadir incidence where the surface scattering cross section is usually highly variable in both space and time, the method is less reliable (Meneghini et al., 2000). For cm-wavelength radars (such as the TRMM radar), attenuation corrections are relatively modest for most rains. For mm-wavelength radars (such as the radar on CloudSat), attenuation is strong and often becomes a dominant factor responsible for systematic reflectivity differences. Another example of rainfall retrieval technique from spaceborne radar is to use the attenuation directly caused by rainfall (Matrosov, 2007). This method takes advantage of high attenuation in rain and low variability of non-attenuated reflectivities and uses estimates of height derivatives of attenuated reflectivities. These estimates are then related to rain rate. Early result from a 94 GHz CloudSat radar indicates a possibility of rain rate profile retrieval (Matrosov, 2007; Mitrescu et al., 2008).

Current Sensors (a) TRMM PR In orbit since 1997, the Precipitation Radar (PR) on the TRMM satellite (Tropical Rainfall Measurement Mission) is the first instrument designed to measure rain from space (Kummerow et al., 1998).

RAINFALL

b

NOAA–18 Rain Rate (mm/hr) 2008–05–26 13.30 LST

50

40

40

30

20

10

0 –150

c

–140

–130

–0.5

1.6

–120 –110 –100 –90 Longitude (degree) 3.7

5.8

7.9

–80

–70

–60

20

0 –150

10.0 Snowfall

–140

140

d

NOAA–18 MHS 89.0 GHz AT (K) 2008–05–26 13.30 LST

40

–130

160

–120 –110 –100 –90 Longitude (degree) 180

200

220

240

–80

260

280

–70

–60

300

NOAA–18 MHS 157.0 GHz AT (K) 2008–05–6 13.30 LST 50

Latitude (degree)

40

30

20

30

20

10

10

0 –150

30

10

50

Latitude (degree)

NOAA–18 AMSU–A 31.4 GHz AT (K) 2008–05–26 13.30 LST

50

Latitude (degree)

Latitude (degree)

a

645

–140

150

–130

180

–120 –110 –100 –90 Longitude (degree) 210

240

–80

270

–70

–60

300

0 –150

–140

150

–130

180

–120 –110 –100 –90 Longitude (degree) 210

240

–80

270

–70

–60

300

Rainfall, Figure 3 NOAA-18 rain rate product (NOAA/NESDIS) and antenna temperatures. (a) Rain rate and antenna temperature at (b) AMSU-A 31.4 GHz, (c) MHS 89 GHz, and (d) MHS 157 GHz. The black circle in the images indicates a rain system dominated with ice cloud and characterized with gradual increase in scattering signal from 31.4 to 157 GHz and hence decreasing in antenna temperatures. The scattering-based portion of the NESDIS rain rate algorithm captures this rain system very well. The black oval shows a warm rain system in the tropical Pacific Ocean that was detected by the emission-based portion of the NESDIS algorithm. The emission signal is visible in 31.4 and 89 GHz but weak in 157 GHz images. Comparing with the nonprecipitating ocean area, the antenna temperatures in the raining zone are warmer at 31.4 and 89 GHz but have little difference at 157 GHz.

TRMM’s PR provides direct, fine-scale, and vertical distribution of precipitation. PR is a cross-track radar that surveys a swath of 220 km at a horizontal resolution of 4.3 km. It gives the vertical distribution of rain, with the height resolution of 250 m. TRMM is a in a low orbit, 402.5 km above the Earth’s surface. Its subtrack covers latitudes of
35 . The PR rain retrieval algorithm (Iguchi et al., 2000) uses a hybrid of the HB and SRT methods to correct for attenuation. When the SRT is reliable, the radar-rain rate relations are adjusted so that the HB path-attenuated estimate is equal to that of SRT. When the SRT is unreliable, the HB method is usually used along with radarrainfall relation derived from ground-based disdrometer measurements, conditioned on rain type, to provide vertical profiles of rain rate. Figure 4

illustrates an example of PR’s ability to probe vertical rain structure. On September 28, Hurricane Lili was southeast of Cuba, but winds in the tropical storm were blowing at a steady 45 kts. Figure 4a shows PR estimated rain in colors: clouds are shown in white; light rain is shown in blue; extreme rain rates in red. Figure 4b shows the cone-shaped eye, newly developed and surrounded by moderate intensity rain clouds. Also shown is a prominent rain band feeding in toward the storm’s inner core. Figure 4c reveals towering clouds that helped power Lili’s growth, called “chimney clouds.” The chimneys contain intense updrafts that release copious amounts of heat energy inside the storm. Such three-dimensional information on precipitation can only be measured by a spaceborne radar.

646

RAINFALL

profiles of radar reflectivity. Although the main objective of CPR is to provide global information on clouds, it also resolves many precipitation systems. It has been suggested that CloudSat data can be used to estimate light rainfall (L’ecuyer and Stephens, 2002). Early results have shown the potential of 94 GHz cloud radar for the detection of precipitation (Haynes et al., 2007) and rain rate profile retrievals (Matrosov, 2007). Figure 5 shows an example of CloudSat rain profiling retrieval (Mitrescu et al., 2008). Figure 5a shows the CloudSat reflectivity on 31 July 2006. The quasi-uniform cloud layer extends up to 13 km; the bright band is identifiable at 4 km. Figure 5b shows the CloudSat retrieved rain rates from two different retrieval algorithms and a ground radar measurement: solid line is retrievals by Mitrescu et al. (2008), squares are retrievals by Matrosov (2007), and triangles are rain rate estimates from the NEXRAD radar KLIX located in New Orleans. Although overall, the retrieved rain rates are comparable, there are noticeable differences.

Rainfall, Figure 4 TRMM PR swath (red lines) and rain rates for Hurricane Lili on 28 September 2002 overlaid on a cloud image (top). PR vertical cross sections for line A–B (middle) and line C–D (bottom) as indicated in top image. The color shading in all images indicates the relative rain intensities (Image courtesy of NASA/GSFC).

(b) CloudSat Launched in April 2006, CloudSat is the first millimeter-wavelength cloud radar in space (Stephens et al., 2002). The Cloud Profiling Radar (CPR) aboard CloudSat is a 94 GHz near-nadir-pointing radar with 500 m vertical resolution. The design concept of CPR is to provide vertical cross section of cloud liquid and ice water content and particle size. At the altitude of 750 km, the CPR provides vertical

Future sensors (a) GPM-DPR GPM (Global Precipitation Measurement) is a joint US-Japan mission designed to extend TRMM’s observations of precipitation to higher latitudes, with more frequent sampling. The GPM core satellite will carry a dual-frequency, cross-track scanning, Ku/Ka-band (13.6/35.5 GHz) precipitation radar, i.e., the DPR (Iguchi et al., 2002). Anticipated advantages of the DPR over the single-frequency PR include enhanced sensitivity at light rain rates, information on the particle size distribution in rain and snow, and improvements in the identification of vertical layers consisting of mixed phase, frozen, and liquid particles. Particle size estimation follows from Mie scattering effects in that the difference in radar reflectivities (in dB) at the two frequencies provides an estimate of the mean particle size of the distribution (Doviak and Zrnic, 1984; Meneghini et al., 1989, 1992; Kuo et al., 2004). This dual-frequency measurement allows for recovery of at least 2 of 3 parameters needed to describe the bulk distribution of the generalized raindrop size distribution (DSD). Such measurements are important because increased knowledge of DSD factors improves the retrievals of rain rate. (b) EarthCARE EarthCARE is a joint Europe-Japan mission with a spaceborne 94 GHz Cloud Profiling Radar (CPR) and is scheduled to be launched in 2012. CPR on EarthCARE will have better sensitivity to cloud detection than CloudSat. One other important feature of CPR/EarthCARE will be its Doppler capability which is expected to provide vertical Doppler velocity profiles of cloud echoes, which are important for cloud physics and drizzle detection (Ohno et al., 2007).

RAINFALL

647

2006/07/31 - CloudSat Reflectivity Zw −30 −25 −20 −15 −10

−5

0

5

10

15

20 dBz

10 km 100

15 200 300 400 500 600 700 800 900 1000

10 5 0 30.10N, 88.89w 19:19:35 NORTH

30.40N, 88.97w 19:19:40 Time (UTC) 2006/07/31 - CloudSat derived Precipitation Rates

NOGAPS* Pressure (mb)

Height (km)

20

30.69N, 89.05w 19:19:45 Naval Research Laboratory, MRY

Rain Rates (mm/h)

14 12 10 8 6 4 2 0 30.10N, 88.89w 19:19:35 NORTH

30.40N, 88.97w 19:19:40 Time (UTC)

30.69N, 89.05w 19:19:45 Naval Research Laboratory, MRY

Rainfall, Figure 5 CloudSat reflectivity in dBZ (top) and derived precipitation rates in mm h1 (bottom) for 31 July 2006 for a short cross section through a storm system over the coastline of Mississippi. The upper image has the surface temperature and pressure contoured to give an indication of the height of the freezing level. The lower image shows the corresponding rain rate from two different CloudSat retrievals (squares and triangles) and surface radar-derived rain rates (solid line) (The image is courtesy of J. Turk and T. Mirtrescu, Naval Research Laboratory, Monterey, CA, USA).

Multisensor datasets The individual sensor records discussed so far each have limitations which make them unsuitable for use in certain situations. For example, PMW estimates over the ocean might be the more accurate than GEO-IR estimates, but the latter are better suited for studies of the diurnal cycle due to the superior sampling obtained from a GEO satellite. Conveniently, the available remotely sensed estimates of precipitation have different strengths and weaknesses, so that combined datasets can be superior to estimates from individual sensors. Several efforts to intercompare and evaluate various types of precipitation algorithms using remotely sensed information were carried out during the 1990s. The WetNet (Dodge and Goodman, 1994) Precipitation Intercomparison Projects (PIP) evaluated multiple global and near-global precipitation algorithms including merged satellite datasets (Barrett et al., 1994; Kniveton et al., 1994; Smith et al., 1998; Adler et al., 2001). The Global Precipitation Climatology Project (GPCP) similarly sponsored three Algorithm Intercomparison Projects (AIP; Ebert et al., 1996) that compared precipitation estimated from satellite observations against high-resolution observations from rain gauges and radars over limited domains (Arkin and Xie, 1994;

Ebert and Manton, 1998). Among other results, these studies showed that PMW estimates were more accurate than IR estimates on an instantaneous basis, but algorithms which combine PMW and IR estimates were superior. However, these intercomparisons did not show significant differences between individual algorithms of a common type, and it remains the case that several merged satellite products exist without a clear consensus on which is superior, and it is common to see a range of similar datasets used in the literature.

Early combinations: GPCP and CMAP The pre-TRMM (i.e., pre-1997) multisensor combinations of precipitation which included PMW as an input were restricted to combinations of VIS or IR estimates (both geosynchronous and low orbit) and estimates from SSM/ I. Many of the early multisource precipitation estimates failed to progress past the research-only phase, but two datasets have endured and are both still in common use: the Global Precipitation Climatology Project (GPCP; Huffman et al., 1997; Adler et al., 2003) and the Climate Prediction Center Merged Analysis of Precipitation (CMAP; Xie and Arkin, 1997b). GPCP and CMAP are both monthly 2.5 resolution and use several common

648

RAINFALL

GPCP July 2007

GPCP 1979–2007

a

b

0

5

10

15

CMAP (with NWP) July 2007

c

0

5

10

0

5 CMAP (with NWP) 1979–2007

d

0

15

10

5

10

TMPA 15th July 2007 10:30z

TMPA July 2007

f

e

0

5

10

15

0

20

40

60

80

Rainfall, Figure 6 (Continued)

inputs with somewhat similar merge algorithms and are therefore frequently considered to be similar, although they use different gauge analyses and have different corrections over the tropical oceans. Both datasets start in

1979 and have changing inputs over time as superior data records become available, such as the SSM/I estimates in 1987. The current version of the GPCP 2.5 monthly mean dataset is the version two dataset (Adler et al., 2003)

RAINFALL

649

CMORPH 15th July 2007 12z

CMORPH July 2007

g

h

0

5

10

15

0

20

40

60

80

PERSIANN 15th July 2007 12z

PERSIANN July 2007

i

j

0

5

10

15

0

20

40

60

80

Rainfall, Figure 6 Maps of precipitation estimates from (a and b) GPCP Version 2, (c and d) CMAP with NCEP/NCAR reanalysis at high latitudes, (e and f) the TMPA, (g and h) CMORPH, and (i and j) PERSIANN. Maps show mean precipitation for (a, c, e, g, and i) July 2007; (b, d) January 1979–December 2007; (f) 10:30–13:30z, 15 July 2007; and (h and j) 12–15z, 15 July 2007. All units are scaled to mm day1 for comparison.

which improved on the first version with a longer record and the addition of TOVS data for improved estimates at mid- and higher latitudes. Both CMAP and GPCP have problems with high-latitude precipitation due to the lack of reliable data: there are few gauges in these sparsely populated regions and available satellite-derived precipitation estimates are of limited use over ice- or snow-covered surfaces. Some studies suggest that model data might be superior to all other estimates at high latitudes (Serreze et al., 2005; Su et al., 2006; Sapiano et al., 2008), and a version of CMAP includes reanalysis precipitation forecasts from the NCAR NCEP reanalysis (Kalnay et al., 1997) over the high latitudes which is the version shown in Figure 6 (CMAP is also available without the reanalysis data). Figure 6 shows the substantial difference in the magnitude of tropical precipitation between GPCP and CMAP as well as the broad differences in precipitation at higher latitudes, although the broad patterns are very similar as would be expected from the use of many common inputs.

Although the monthly, 2.5 products are most commonly used, higher-resolution products also exist as part of the GPCP suite. A pentad (5 days mean) version of GPCP combines similar satellite inputs as the monthly product with a different gauge dataset (Xie et al., 2003). This experimental dataset is also produced on a 2.5 resolution grid and starts in 1979. A one-degree, daily version of the GPCP dataset is also available which starts in 1996 and combines IR from geosynchronous and low orbits with the GPCP Version 2 monthly product and AIRS and TOVS estimates (Huffman et al., 2001). This dataset does not directly use PMW data, although the SSM/I estimates are implicitly used through the inclusion of the monthly GPCP Version 2 product.

High-resolution precipitation products While GPCP and CMAP are still commonly used for global, monthly, coarse resolution studies (including

650

RAINFALL

validation of numerical model forecasts), recent increases in the availability of PMW data (SSM/I, TMI, AMSU, AMSR) have led to the emergence of several near-global, high-resolution products which have begun to supersede the pentad and one-degree daily products of the GPCP. These products are not global (extending no further than 60  N/S) but have spatial resolutions of at least 0.25 and temporal resolution of at least three hourly and are based on a variety of innovative methods for combining estimates derived from PMW observations with GEO-IR imagery. Generally speaking, these high-resolution precipitation products (HRPPs) use the high spatial and temporal resolution of IR data to resolve deficiencies in resolution of the higher-quality PMW data, although there are substantial differences between the exact methodologies employed. The HRPPs can be categorized into two broad types: adjustment-based techniques where IR data is calibrated using PMW estimates (where the two are often then combined) and motion-based techniques, where the IR data is used to interpolate between successive PMW overpasses. The Pilot Evaluation of High-Resolution Precipitation Products (PEHRPP) was established to intercompare and validate these datasets. PEHRPP included a number of high-resolution datasets: the TRMM Multisatellite Precipitation Analysis (TMPA; Huffman et al., 2007), the CPC Morphing technique (CMORPH; Joyce et al., 2004), the Hydro-Estimator (Scofield and Kuligowski, 2003), the NRL-Blended technique (NRL-Blended; Turk and Miller, 2005), Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks (PERSIANN; Hsu et al., 1997; Sorooshian et al., 2000), and the Global Satellite Mapping of Precipitation (GSMaP) project. Studies to compare these datasets are underway, although there is no technique that outperforms the others on a consistent basis (Gottschalck et al., 2005; Brown, 2006; Ebert et al., 2007; Ruane and Roads, 2007; Tian et al., 2007; Sapiano and Arkin, 2009). Figure 6 shows a single three-hourly and daily accumulation for TMPA (Figure 6f), CMORPH (Figure 6h), and PERSIANN (Figure 6j). These three datasets have different properties; TMPA is a simple combination of PMW and IR starting in 1997 and is gauge adjusted; CMORPH uses IR information to advect PMW information between successive overpasses and has near-global resolution of 8 km, three hourly; PERSIANN uses a neural network (trained with PMW estimates) to translate IR only to rain rates. The different properties of these datasets are clear, with PERSIANN giving somewhat smoother estimates than the other two, which generally agree with each other, although TMPA is closer to GPCP and CMAP over land areas. Despite recent advances there is still much uncertainty in the quality of high-resolution precipitation datasets compared to coarser datasets such as GPCP and CMAP. Reconciliation between the HRPPs and the coarser datasets has still not been achieved, and the algorithm designs employed by GPCP and CMAP remove much of the inherent noise from the inputs, so that they

are considered more reliable than the HRPPs. In short, the reader is advised to use the class of dataset that has closest resolution to their needs and is cautioned against averaging high-resolution data.

Summary There are various techniques to retrieve rainfall from satellites, each with their own set of attributes that are dictated by the particular needs for the information; for short-term, high spatial resolution applications like flash flood forecasting, the IR methods are generally preferred, while for global, climate scales, the PMW are usually preferred. In this sense, no single approach can be defined as the best one; however, in terms of accuracy on the instantaneous time scale (i.e., at the time the satellite is making its measurement), it is generally accepted that the active MW is the most accurate, followed by passive MW (ocean), passive MW (land), IR, and VIS. Many of the early methods were developed using sensors that were not necessarily flown for rainfall retrieval but more for tracking cloud features and monitoring atmospheric temperature and moisture. As was described, more recent and near-term missions are being designed specifically for rainfall monitoring and include AMW sensors (e.g., the GPM mission). Additionally, emerging methods such as the blended techniques or multispectral (including PMW and AMW) will yield improvements to the current accuracy of the remote sensing of rainfall and will likely become the standard retrieval method as we enter into the next decade. Bibliography Adler, R. F., and Negri, A. J., 1988. A satellite infrared technique to estimate tropical convective and stratiform rainfall. Journal of Applied Meteorology, 27, 30–51. Adler, R. F., Kidd, C., Petty, G., Morrissey, M., and Goodman, H. M., 2001. Intercomparison of global precipitation products: the third precipitation intercomparison project (PIP-3). Bulletin of the American Meteorological Society, 82, 1377–1396. Adler, R. F., Huffman, G. J., Chang, A., Ferraro, R., Xie, P., Janowiak, J., Rudolf, B., Schneider, U., Curtis, S., Bolvin, D., Gruber, A., Susskind, J., Arkin, P., and Nelkin, E., 2003. The version-2 global precipitation climatology project (GPCP) monthly precipitation analysis (1979-present). Journal of Hydrometeorology, 4, 1147–1167. Amayenc, P., Diguet, J. P., Marzoug, M., and Tani, T., 1996. A class of single- and dual-frequency algorithms for rain-rate profiling from a spaceborne radar. Part II: tests from airborne radar measurements. Journal of Atmospheric and Oceanic Technology, 13, 142–164. Amorati, R., Alberoni, P. P., Levizzani, V., and Nanni, S., 2000. IR-based satellite and radar rainfall estimates of convective storms over northern Italy. Meteorological Applications, 7, 1–18. Arkin, P. A., 1979. The relationship between fractional coverage of high cloud and rainfall accumulations during GATE over the B-scale array. Monthly Weather Review, 106, 1153–1171. Arkin, P. A., and Janowiak, J., 1991. Analysis of the global distribution of precipitation. Dynamics of Atmospheres and Oceans, 16, 5–16.

RAINFALL

Arkin, P. A., and Meisner, B. N., 1987. The relationship between large-scale convective rainfall and cold cloud over the western hemisphere during 1982–84. Monthly Weather Review, 115, 51–74. Arkin, P. A., and Xie, P. P., 1994. The global precipitation climatology project: first algorithm intercomparison project. Bulletin of the American Meteorological Society, 75, 401–419. Ba, M. B., and Gruber, A., 2001. GOES multispectral rainfall algorithm (GMSRA). Journal of Applied Meteorology, 40, 1500–1514. Barrett, E. C., and Martin, D. W., 1981. The Use of Satellite Data in Rainfall Monitoring. London: Academic. Barrett, E. C., Dodge, J., Goodman, H. M., Janowiak, J., Kidd, C., and Smith, E. A., 1994. The first WetNet precipitation intercomparison project. Remote Sensing Reviews, 11, 49–60. Bauer, P., Moreau, E., and Di Michele, S., 2005. Hydrometeor retrieval accuracy using microwave window and sounding channel observations. Journal of Applied Meteorology, 44, 1016–1032. Bellon, A., Lovejoy, S., and Austin, G. L., 1980. Combining satellite and radar data for the short-range forecasting of precipitation. Monthly Weather Review, 108, 1554–1556. Brown, J. E. M., 2006. An analysis of the performance of hybrid infrared and microwave satellite precipitation algorithms over India and adjacent regions. Remote Sensing of Environment, 101, 63–81. Dodge, J., and Goodman, H. M., 1994. The WetNet project. Remote Sensing Reviews, 11, 5–21. Doviak, J. D., and Zrnic, D. S., 1984. Doppler Radar and Weather Observations. New York: Academic. Ebert, E. E., and Manton, M. J., 1998. Performance of satellite rainfall estimation algorithms during TOGA COARE. Journal of the Atmospheric Science, 55, 1537–1557. Ebert, E. E., Manton, M. J., Arkin, P. A., Allam, R. J., Holpin, C. E., and Gruber, A., 1996. Results from the GPCP algorithm intercomparison programme. Bulletin of the American Meteorological Society, 77, 2875–2887. Ebert, E. E., Janowiak, J. E., and Kidd, C., 2007. Comparison of near-real-time precipitation estimates from satellite observations and numerical models. Bulletin of the American Meteorological Society, 88, 47–64. Ferraro, R. R., 1997. Special sensor microwave imager derived global rainfall estimates for climatological applications. Journal of Geophysical Research, 102, 16715–16735. Ferraro, R. R., and Marks, G. F., 1995. The development of SSM/I rain-rate retrieval algorithms using ground-based radar measurements. Journal of Atmospheric and Oceanic Technology, 12, 755–770. Fujita, M., and Satake, M., 1997. Rainfall rate profiling with attenuation-frequency radar using nonlinear LMS technique under a constraint on path-integrated rainfall rate. International Journal of Remote Sensing, 18, 1137–1147. Gottschalck, J., Meng, J., Rodell, M., and Houser, P., 2005. Analysis of multiple precipitation products and preliminary assessment of their impact on global land data assimilation system land surface states. Journal of Hydrometeorology, 6, 573–598. Griffith, C. G., Woodley, W. L., Grube, P., Martin, D. W., Stout, J., and Sikdar, D. N., 1978. Rain estimation from geosynchronous satellite imagery – visible and infrared studies. Monthly Weather Review, 106, 1153–1171. Grody, N. C., 1991. Classification of snow cover and precipitation using the special sensor microwave/imager (SSM/I). Journal of Geophysical Research, 96, 7423–7435. Haynes, J. M., and Stephens, G., 2007. Tropical oceanic cloudiness and incidence of precipitation: early results from CloudSat. Geophysical Research Letters, 34, L09811, doi:10.1029/ 2007GL029335.

651

Hitschfeld, W., and Bordan, J., 1954. Errors inherent in the radar measurement of rainfall at attenuating wavelengths. Journal of Meteorology, 11, 58–67. Hsu, K., Gao, X., Sorooshian, S., and Gupta, H. V., 1997. Precipitation estimation from remotely sensed information using artificial neural networks. Journal of Applied Meteorology, 36, 1176–1190. Huffman, G. J., Adler, R. F., Arkin, P. A., Chang, A., Ferraro, R., Gruber, A., Janowiak, J., McNab, A., Rudolf, B., and Schneider, U., 1997. The global precipitation climatology project (GPCP) combined precipitation dataset. Bulletin of the American Meteorological Society, 78, 5–20. Huffman, G. J., Adler, R. F., Morrissey, M. M., Bolvin, D. T., Curtis, S., Joyce, R., McGavock, B., and Susskind, J., 2001. Global precipitation at one-degree daily resolution from multisatellite observations. Journal of Hydrometeorology, 2, 36–50. Huffman, G. J., Adler, R. F., Bolvin, D. T., Gu, G., Nelkin, E. J., Bowman, K. P., Hong, Y., Stocker, E., and Wolff, D., 2007. The TRMM multisatellite precipitation analysis (TMPA): quasi-global, multiyear, combined-sensor precipitation estimates at fine scales. Journal of Hydrometeorology, 8, 38–55. Iguchi, T., and Meneghini, R., 1994. Intercomparison of single frequency methods for receiving a vertical rain profile from airborne and spaceborne radar data. Journal of Atmospheric and Oceanic Technology, 11, 1507–1516. Iguchi, T., Kozu, T., Meneghini, R., Awaka, J., and Okamoto, K., 2000. Rain-profiling algorithm for the TRMM precipitation radar. Journal of Applied Meteorology, 39, 2038–2052. Iguchi, T., Oki, R., Smith, E. A., and Furuhama, Y., 2002. Global precipitation measurement program and the development of dual-frequency precipitation radar. Journal of the Communication Research Laboratory, 49, 37–45. Joyce, R. J., Janowiak, J. E., Arkin, P. A., and Xie, P., 2004. CMORPH: a method that produces global precipitation estimates from passive microwave and infrared data at high spatial and temporal resolution. Journal of Hydrometeorology, 5, 487–503. Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., Iredell, M., Saha, S., White, G., Woollen, J., Zhu, Y., Chelliah, M., Ebisuzaki, W., Higgins, W., Janowiak, J., Mo, K. C., Ropelewski, C., Wang, J., Leetmaa, A., Reynolds, R., Jenne, R., and Joseph, D., 1996. The NCEP/NCAR 40-year reanalysis project. Bulletin of the American Meteorological Society, 77, 437–471. Kidder, S. Q., and Vonder Haar, T. H., 1995. Satellite Meteorology: An Introduction. New York: Academic. Kniveton, D. R., Motta, B. C., Goodman, H. M., Smith, M., and LaFontaine, F. J., 1994. The first WetNet precipitation intercomparison project: generation of results. Remote Sensing Reviews, 11, 243–302. Kummerow, C., Olson, W. S., and Giglio, L., 1996. A simplified scheme for obtaining precipitation and vertical hydrometeor profile from passive microwave. IEEE Transactions on Geoscience and Remote Sensing, 34, 1213–1232. Kummerow, C., Barnes, W., Kozu, T., Shiue, J., and Simpson, J., 1998. The tropical rainfall measuring mission (TRMM) sensor package. Journal of Atmospheric and Oceanic Technology, 15, 809–817. Kummerow, C., Hong, Y., Olson, W. S., Yang, S., Adler, R. F., McCollum, J., Ferraro, R., Petty, G., Shin, D., and Wilheit, T. T., 2001. The evolution of the Goddard profiling algorithm (GPROF) for rainfall estimation from passive microwave sensors. Journal of Applied Meteorology, 40, 1801–1819. Kuo, K.-S., Smith, E. A., Haddad, S., Im, E., Iguchi, T., and Mugnai, A., 2004. Mathematical physical framework for retrieval of rain DSD properties from dual-frequency Ku/ Ka-band satellite radar. Journal of the Atmospheric Sciences, 61, 2349–2369.

652

RAINFALL

L’Ecuyer, T. S., and Stephens, G. L., 2002. An estimation-based precipitation retrieval algorithm for attenuating radars. Journal of Applied Meteorology, 41, 272–285. Lovejoy, S., and Austin, G. L., 1979. The delineation of rain areas from visible and IR satellite data from GATE and mid-latitudes. Atmosphere–ocean, 17, 77–92. Marzoug, M., and Amayenc, P., 1994. A class of single frequency algorithms for rain rate profiling from a spaceborne radar 1. Principle and tests from numerical simulations. Journal of Atmospheric and Oceanic Technology, 11, 1480–1506. Matrosov, S., 2007. Potential for attenuation-based estimates of rainfall rate from CloudSat. Geophysical Research Letters, 34, L05817, doi:10.1029/2006GL029161. McCollum, J. R., and Ferraro, R. R., 2003. The next generation of NOAA/NESDIS SSM/I, TMI and AMSR-E microwave land rainfall algorithms. Journal of Geophysical Research, 108, 8382–8404. Meneghini, R., and Kozu, T., 1990. Spaceborne Weather Radar. Norwood, MA: Artech House. Meneghini, R., Nakamura, K., Ulbrich, C. W., and Atlas, D., 1989. Experimental tests of methods for the measurement of rainfall rate using an airborne dual-wavelength radar. Journal of Atmospheric and Oceanic Technology, 6, 637–651. Meneghini, R., Kozu, T., Kumagai, H., and Boncyk, W. C., 1992. A study of rain estimation methods from space using dualwavelength radar measurements at near-nadir incidence over ocean. Journal of Atmospheric and Oceanic Technology, 9, 364–382. Meneghini, R., Iguchi, T., Kozu, T., Liao, L., Okamoto, K., Jones, J., and Kwiatkowski, J., 2000. Use of the surface reference technique for path attenuation estimates from the TRMM precipitation radar. Journal of Applied Meteorology, 39, 2053–2070. Mitrescu, C., Heynes, J., L’ecuyer, T., Turk, J., and Miller, S., 2008. CloudSat precipitation profiling algorithm-model description. In The 88th American Meteorological Society meeting, 20–25 Jan 2008, New Orleans. Mugnai, A., and Smith, E.A., 1984. Passive microwave radiation transfer in an evolving cloud medium. In IRS’84: Current Problems in Atmospheric Radiation. Hampton, VA: A. Deepak Publishing. Ohno, Y., Horie, H., Sato, K., Kumagai, H., Kimura, T., Okada, K., Iida, Y., and Kojima, M., 2007. Development of cloud profiling radar for EarthCare. In Thirty-Third AMS Conference on Radar Meteorology, 6–10 Aug 2007, Cairns. Olson, W. S., 1989. Physical retrieval of rainfall rates over the ocean by multispectral microwave radiometry: application to tropical cyclones. Journal of Geophysical Research, 94, 2267–2280. Rosenfeld, D., and Gutman, G., 1994. Retrieving microphysical properties near the tops of potential rain clouds by multispectral analysis of AVHRR data. Atmospheric Research, 34, 259–283. Ruane, A. C., and Roads, J. O., 2007. 6-hour to 1-year variance of five global precipitation sets. Earth Interactions Journal, 11, 1. Sapiano, M. R. P., and Arkin, P. A., 2009. An inter-comparison and validation of high resolution satellite precipitation estimates with three-hourly gauge data. Journal of Hydrometeorology, 10(1), 149–166. Sapiano, M. R. P., Smith, T. M., and Arkin, P. A., 2008. A new merged analysis of precipitation utilizing satellite and reanalysis data. Journal of Geophysical Research, 113, D22103, doi:10.1029/2008JD010310. Scofield, R. A., 1987. The NESDIS operational convective precipitation technique. Monthly Weather Review, 115, 1773–1792. Scofield, R. A., and Kuligowski, R. J., 2003. Status and outlook of operational satellite precipitation algorithms for extremeprecipitation events. Weather and Forecasting, 18, 1037–1051.

Serreze, M. C., Barrett, A., and Lo, F., 2005. Northern high latitude precipitation as depicted by atmospheric reanalyses and satellite retrievals. Monthly Weather Review, 133, 3407–3430. Smith, E. A., Xiang, X., Mugnai, A., and Tripoli, G. J., 1994. Design of an inversion-based precipitation profile retrieval algorithm using an explicit cloud model for initial guess microphysics. Meteorology and Atmospheric Physics, 54, 53–78. Smith, E. A., et al., 1998. Results of WetNet PIP-2 project. Journal of the Atmospheric Sciences, 55, 148–1536. Sorooshian, S., Hsu, K., Gao, X., Gupta, H. V., Imam, B., and Braithwaite, D., 2000. Evaluation of PERSIANN system satellite-based estimates of tropical rainfall. Bulletin of the American Meteorological Society, 81, 2035–2046. Spencer, R. W., Martin, D. W., Hinton, B., and Weinman, J. A., 1983. Satellite microwave radiances correlated with radar rain rates over land. Nature, 304, 141–143. Spencer, R. W., Goodman, H. M., and Hood, R. E., 1989. Precipitation retrieval over land and ocean with the SSM/I: identification and characteristics of the scattering signal. Journal of Atmospheric and Oceanic Technology, 6, 254–273. Stephens, G., et al., 2002. The CloudSat mission and the A-Train. Bulletin of the American Meteorological Society, 83, 1771–1790. Su, F., Adam, J., Trenberth, K., and Lettenmaier, D., 2006. Evaluation of surface water fluxes of the pan-Arctic land region with a land surface model and ERA-40 reanalysis. Journal of Geophysical Research, 111, D05110, doi:10.1029/2005JD006387. Tian, Y., Peters-Lidard, C., Choudhury, B., and Garcia, M., 2007. Multitemporal analysis of TRMM-based satellite precipitation products for land data assimilation applications. Journal of Hydrometeorology, 8, 1165–1183. Todd, M. C., Barrett, E. C., Beaumont, M. J., and Green, J. L., 1995. Satellite identification of rain days over the upper Nile river basin using an optimum infrared rain/no-rain threshold temperature model. Journal of Applied Meteorology, 34, 2600–2611. Todd, M. C., Barrett, E. C., Beaumont, M. J., and Bellerby, T. J., 1999. Estimation of daily rainfall over the upper Nile river basin using a continuously calibrated satellite infrared technique. Meteorological Applications, 6, 201–210. Turk, F. J., and Miller, S. D., 2005. Toward improved characterization of remotely sensed precipitation regimes with MODIS/ AMSR-E blended data techniques. IEEE Transactions on Geoscience and Remote Sensing, 43, 1059–1069. Vicente, G. A., Scofield, R. A., and Menzel, W. P., 1998. The operational GOES infrared rainfall estimation technique. Bulletin of the American Meteorological Society, 79, 1883–1898. Vicente, G. A., Davenport, J. C., and Scofield, R. A., 2002. The role of orographic and parallax correction on real-time high resolution satellite rainfall rate distribution. International Journal of Remote Sensing, 23, 221–230. Vila, D., Ferraro, R. R., and Joyce, R., 2007. Evaluation and improvement of AMSU precipitation retrievals. Journal of Geophysical Research, 112, D20119, doi:10.1029/2007JD008617. Weinman, J., and Guetter, A., 1977. Determination of rainfall distributions from microwave radiation measured by the Nimbus 6 ESMR. Journal of Applied Meteorology, 16, 437–442. Weng, F., and Grody, N. C., 1994. Retrieval of cloud liquid water using the special sensor microwave imager (SSM/I) data. Journal of Geophysical Research, 99, 25535–25551. Weng, F., Zhao, L., Poe, G., Ferraro, R., Li, X., and Grody, N., 2003. AMSU cloud and precipitation algorithms. Radio Science, 338, 8068–8079. Wilheit, T. T., Chang, A., Rao, M., Rodgers, E., and Theon, J., 1977. A satellite technique for quantitatively mapping rainfall rates over the oceans. Journal of Applied Meteorology, 16, 551–560. Wilheit, T. T., Kummerow, C., and Ferraro, R., 2003. Rainfall algorithms for AMSR-E. IEEE Transactions on Geoscience and Remote Sensing, 41, 204–214.

RANGELANDS AND GRAZING

Woodley, W. L., and Sancho, R., 1971. A first step towards rainfall estimation from satellite cloud photographs. Weather, 26, 279–289. Xie, P., and Arkin, P. A., 1997a. Global precipitation: a 17-year monthly analysis based on gauge observations, satellite estimates, and numerical model outputs. Bulletin of the American Meteorological Society, 78, 2539–2558. Xie, P., and Arkin, P.A., 1997b. Global pentad precipitation analysis based on gauge observations, satellite estimates and model outputs. In Extended Abstracts, American Geophysical Union 1997 Fall Meeting, 8–12 Dec 1997, San Francisco, CA: AGU. Xie, P., Janowiak, J. E., Arkin, P. A., Adler, R., Gruber, A., Ferraro, R., Huffman, G., and Curtis, S., 2003. GPCP pentad precipitation analyses: an experimental dataset based on gauge observations and satellite estimates. Journal of Climate, 16, 2197–2214. Zhao, L., and Weng, F., 2002. Retrieval of ice cloud parameters using the AMSU. Journal of Applied Meteorology, 41, 384–395.

Cross-references Climate Monitoring and Prediction Global Earth Observation System of Systems (GEOSS) Land Surface Emissivity Microwave Radiometers Observational Systems, Satellite Optical/Infrared, Radiative Transfer Water Resources

RANGELANDS AND GRAZING Hunt E. Raymond, Jr. USDA-ARS Hydrology and Remote Sensing Laboratory, Beltsville, MD, USA

Synonyms Grasslands; Grazing lands; Pampas; Prairie; Savanna; Shrublands; Steppe Definition Rangelands. Type of land cover dominated by grasses, grasslike plants, broadleaf herbaceous plants (forbs), and shrubs, where the land is managed as a natural ecosystem for multiple uses including wildlife habitat, biodiversity, recreation, and grazing by livestock. Pasturelands. Type of land cover dominated by grasses, grasslike plants, and broadleaf herbaceous plants, where the land is managed as an agricultural system for livestock production. Plant community. Co-occurring plant species and their relative abundance and which is usually recognized by the dominant plant species or functional type. Ecological site. A combination of soil, climate, and hydrological factors that have the potential for a distinctive climax plant community with a given amount of net primary production. Rangeland health. The degree to which the integrity of the soil is maintained and ecological processes are sustained.

653

Introduction Rangelands encompass many different land cover types throughout the world, including grasslands, savannas, shrublands, tundras, marshes, and meadows (Holechek et al., 2004). Based upon the NASA MODIS land cover product, rangelands cover about 48 % of the Earth’s land surface, not counting Antarctica (Friedl et al., 2010). Most rangeland cover types are found in regions where the annual precipitation is low and the interannual variability of precipitation is high, which makes these regions unsuitable for forestry or agriculture. Livestock grazing is a land use that provides a sustainable means of producing food and fiber in rangeland cover types. Confusingly, many nongovernmental organizations use the term rangelands to denote a land-use category of livestock grazing (Lund, 2007). Inventory, assessment, and management of grazing on rangelands within the USA are based on ecological sites, which are generally recognized and delimited by the historic climax plant community (NRCS, 2003). Another paradigm for management is rangeland health, which is based on prevention of soil erosion and the sustainability of ecosystem processes such as energy flow (primary production), the hydrologic cycle, and biogeochemical cycles (BLM, 2005; NRC, 1994). The concepts of rangeland health and ecological sites were developed in parallel in response to the shortcomings of “Rangeland Condition and Trend” (Briske et al., 2005). Ecological sites Management of rangelands was based initially on the theory of plant succession from Clements, (1916). Starting from bare substrate (primary succession) or bare soil (secondary succession), the dominant species change over time from short-lived to long-lived plants, ultimately reaching a climax plant community determined by climate. Climax plant communities were assumed to have the highest sustainable productivity, the greatest resistance to invasive species and pests, and the best protection against soil erosion; thus, rangeland management was based on maintaining or restoring the climax plant community. However, it was recognized that natural disturbances such as drought, grazing, and fire had to be incorporated into the definition of the climax community (NRCS, 2003; Briske et al., 2005). Ecological sites are identified and delimited by (1) the dominant plant community at the onset of European immigration and settlement (NRCS, 2003) and (2) soils and climate delineated by major land resource areas (NRCS, 2006). An ecological state is comprised of group of plant communities which are in a dynamic equilibrium with the historical amounts and types of disturbances at a given ecological site. If disturbances are substantially greater than historical levels, there may be a transition to a different ecological state, with new groups of plant communities (NRCS, 2003; Briske et al., 2005).

654

RANGELANDS AND GRAZING

Rangelands and Grazing, Figure 1 State-and-transition diagram for an ecological site. The large boxes with solid lines are stable ecological states (A, B, . . ., N) within which dynamic changes in plant communities result from natural disturbances. Changes from one state to another (transition) are usually irreversible without large management inputs. Boxes with dashed lines are used to represent a specific plant community that occurs within a stable ecological state. The historical climax plant community is the ideal management goal for sustainable use. Arrows show possible changes and the symbols represent a mechanism for the possible change. Symbols are NF no fire, PF prescribed fire, BM brush management, HG heavy grazing, PG prescribed grazing, CHG continuous heavy grazing, Seed reseeding, Inv invasion, and IM integrated management.

The relationships among plant communities and ecological states for an ecological site are depicted in a state-and-transition model (Figure 1). In Figure 1, State A includes the various plant communities found at an ecological site with historical levels of disturbance, including the historic climax plant community. Within State A, plant community 2 could be the result of woody shrub establishment allowed by reduced fire frequency and plant community 3 could be an earlier stage of succession after disturbance. However, continuous heavy grazing may result in soil erosion, so the site crosses a threshold into a new State B with plant community 4;

changes of grazing intensity alone are unlikely to restore the site to State A without reseeding (Figure 1). Alternatively, the ecological site could be invaded by a nonnative weed species (State N), requiring integrated management (a combination of biological, cultural, and chemical control) for restoration (Figure 1). Because an ecological site is defined by the potential to have a climax plant community, and because an ecological state may have many different plant communities, rarely will land cover classification from satellite remote sensing have a major role for mapping either ecological sites or states (Hunt et al., 2003; Pickup et al., 1994).

RANGELANDS AND GRAZING

A combination of imagery, geospatial databases of soils and climate, and computer simulation modeling could be used to determine potential vegetation for ecological sites (Jensen et al., 2001). Simple detection of land cover change between two dates is not useful for detecting a state transition because of year-to-year variability; analysis of long-term imagery is important to distinguish a state transition (Washington-Allen et al., 2006).

Rangeland health Rangeland health is monitored by relative rankings of 17 indicators based on (1) soils and soil structure; (2) hydrological, nutrient, and carbon cycling (including net primary production); and (3) functional plant diversity and presence of invasive species (BLM, 2005). Of all of the indicators, ground cover fractions have been demonstrated as having the greatest utility for economical sampling and objective measurement on a landscape scale (Booth and Tueller, 2003; Pickup et al., 1994). Other aspects of rangeland health could be monitored operationally with the appropriate sensors, such as changes in vegetation productivity, invasion of noxious weeds, and cover of senesced and living plant material (Hunt et al., 2003; Palmer and Fortesque, 2004). If it is accepted that unbiased, economical monitoring rangelands is required, and remote sensing provides these methods, then the framework for rangeland assessment needs to be reformulated based on what remote sensing can accomplish (Pickup et al., 1994, p. 499). Rangeland productivity Primary production of rangelands can be estimated using the amount of absorbed photosynthetically active radiation integrated over time, which is the basis of MODIS primary production (MOD17A3) data product (Reeves et al., 2001, 2006; Running et al., 2004). NDVI is approximately equal to the fraction of incident photosynthetically active radiation that is absorbed by the canopy, so NDVI integrated over time is also used to estimate primary production (Paruelo et al., 1997; Piñeiro et al., 2006; Tieszen et al., 1997). Net primary production provides the upper limit on available forage on a sustainable basis and hence determines the stocking rate or carrying capacity. Determination of standing biomass by remote sensing is possible, but grazing based on standing biomass may not be sustainable. A 455 kg cow (one animal unit) consumes about 9.1 kg of forage per day (Holechek et al., 2004); 273 kg of forage is one animal unit month (AUM). The aboveground net primary production divided by 273 kg/month sets the stocking rate in AUM per area; factors such as slope and distance to water reduce the stocking rate from the maximum (Holechek et al., 2004). Many of these factors can be estimated from geospatial databases allowing determination of recommended stocking rates (Hunt et al., 2003; Hunt and Miyake, 2006). Environmental data records of NDVI from AVHRR, MODIS, SPOT

655

Vegetation, and VIIRS (Tucker et al., 2005; Brown et al., 2006) are important because the long time series available can assess variability of net primary production due to variability in rainfall.

Detection of invasive weeds Invasive shrubs and weeds dominate the vegetation of many rangeland plant communities, affecting native biodiversity and reducing the amount of available forage (DiTomaso, 2000). Some invasive plants have spectrally distinctive flowers or leaves that can be detected with imaging spectroscopy (Andrew and Ustin, 2006; Everitt et al., 2002; Hunt et al., 2004; Lass et al., 2002; Parker Williams and Hunt, 2002; Underwood et al., 2003). Imaging spectrometers cover small areas and require extensive expertise for image processing; thus, they may not be useful for operational maps of detectable invasive species (Hunt et al., 2007). However, distribution maps from small areas may be used to test and refine geospatial models for the prediction of sites susceptible to invasion (Bradley and Mustard, 2006; Hunt et al., 2010). Other species such as cheatgrass (Bradley and Mustard, 2006) and salt cedar (Everitt et al., 2002; Groeneveld and Watson, 2008) can be detected using differences in phenology compared to the co-occurring native vegetation. Ecosystem degradation The cover fractions of bare soil, senesced plant matter (plant litter or residue), and green vegetation are key indicators of rangeland health (Booth and Tueller, 2003). Bare soil is very susceptible to erosion, and erosion increases the amount of bare soil visible from aircraft and satellites (Vrieling et al., 2007). Green vegetation cover is usually measured by scaling NDVI to account for soil background reflectance (Jiang et al., 2006; Montandon and Small, 2008; Pickup et al., 1998). Senesced plant matter is spectrally similar to bare soil at visible and near-infrared wavelengths; narrowband reflectances in the shortwave infrared may be used to detect senesced plant matter based on the absorption feature of cellulose at a wavelength of 2.1 mm (Daughtry et al., 2004; Nagler et al., 2000). Spectral unmixing of remotely sensed imagery allows the cover fractions of bare soil, senesced plant matter, and green vegetation to be determined simultaneously (Asner and Heidebrecht, 2002; Roberts et al., 1993). Furthermore, there are a large number of studies that have successfully unmixed multispectral data based on the phenology of vegetation (Arsenault and Bonn, 2005; de Asis and Omasa, 2007; Kuemmerle et al., 2006; Metternicht and Fermont, 1998; Numata et al., 2007). Very large-scale aerial photography Very large-scale aerial (VLSA) photographs have very large map scales from 1:50 to 1:500 (pixel sizes of 1–50 mm) and are an important method of rangeland monitoring for either ecological-state or rangeland-health paradigms. Whereas aerial photography has been used

656

RANGELANDS AND GRAZING

for over 50 years, digital cameras may reduce the cost for image acquisition, processing, and storage. At very large scales, composition of spectrally similar plant species may be identified (Petersen et al., 2005). Individual plants of an invasive species may be detected at the initial stages of establishment (Blumenthal et al., 2007; Naylor et al., 2005; Booth et al., 2010). Landscape digital photographs may be compared quantitatively with digitized historic landscape photographs for change detection (Clark and Hardegree, 2005). With special robotic camera mount, landscape panoramas may be acquired at full magnification, resulting in a seamless image for analysis (Nichols et al., 2009). Analysis of large numbers of images may be very labor intensive without automation; supervised image classification is usually problematic because only three bands are acquired. “SamplePoint” software (Booth et al., 2006) reduces the time required to estimate cover from VLSA photographs. Two types of airborne platforms for VLSA photography are being evaluated: unmanned aerial systems (Hardin and Jackson, 2005; Rango et al., 2006) and light manned aircraft (Booth et al., 2006, 2008). Light aircraft and unmanned aerial systems may cover transects or provide wall-to-wall coverage (Laliberte et al., 2010). Ground cover fractions from VLSA photography are highly correlated to ground measurements (Booth et al., 2008). However, because the area covered by each photograph is small, ground cover is not strongly correlated with satellite vegetation indices (Sivanpillai and Booth, 2008).

Management of economic risk with AVHRR NDVI Crop insurance is used as a risk management tool in which payments are based on actual crop losses. In 2007, USDA Risk Management Agency started a pilot insurance program for pasture, rangeland, and forages (RMA, 2010) in order to manage risks from drought. In nine US states, rangeland managers may choose whether to estimate drought from either a rainfall index or a vegetation index (RMA, 2010). Then, the managers select one or more 3 months time periods when rainfall is important for pasture, rangeland, and forage production in their operations. Grasses and forbs grow when water is available in the soil, so AVHRR NDVI is an indicator of rainfall in rangelands and low NDVI is an indicator of drought (Davenport and Nicholson, 1993; Di et al., 1994; Ji and Peters, 2003; Wang et al., 2003). Currently, 14 days composites of AVHRR NDVI are made and aggregated to an 8 km by 8 km grid cell by the USGS Earth Resources Observation and Science (EROS) Center (Sioux Falls, South Dakota). Based on the historical averages for a grid cell, a trigger grid index is determined. If the final value of NDVI falls below the trigger grid index, payments to the consumer are made based on area insured. Actual production losses are not used to calculate payments (RMA, 2010).

Conclusion Primary production and pasture, range, and forage insurance are based on the frequent coverage provided by AVHRR, MODIS, SPOT Vegetation, and VIIRS. In the future, economic losses from drought may be better determined from primary production estimates using these sensors, to better estimate the forage available for livestock and wildlife. Rangeland management is currently based on ecological sites so very large-scale photography is the best remote sensing method; but it is difficult to cover large areas. Rangeland health needs to be better integrated with rangeland management based on ecological sites (Herrick et al., 2006). With the large area of rangelands globally, monitoring rangeland health will be unbiased and cost-effective using remote sensing, but only if the cost of the data is reasonable (Palmer and Fortesque, 2004). Bibliography Andrew, M. E., and Ustin, S. L., 2006. Spectral and physiological uniqueness of perennial pepperweed (Lepidium latifolium). Weed Science, 54, 1051–1062. Arsenault, E., and Bonn, F., 2005. Evaluation of soil erosion protective cover by crop residues using vegetation indices and spectral mixture analysis of multispectral and hyperspectral data. Catena, 62, 157–172. Asner, G. P., and Heidebrecht, K. B., 2002. Spectral unmixing of vegetation, soil and dry carbon cover in arid regions: comparing multispectral and hyperspectral observations. International Journal of Remote Sensing, 23, 3939–3958. BLM, 2005. Interpreting Indicators of Rangeland Health, Version 4. Technical Reference 1734–6. USDI Bureau of Land Management, National Science and Technology Center. Denver, CO: United States Department of Interior. Blumenthal, D., Booth, D. T., Cox, S. E., and Ferrier, C. E., 2007. Large-scale aerial images capture details of invasive plant populations. Rangeland Ecology & Management, 60, 523–528. Booth, D. T., and Cox, S. E., 2006. Very large scale aerial photography for rangeland monitoring. Geocarto International, 21, 27–34. Booth, D. T., and Cox, S. E., 2008. Image-based monitoring to measure ecological change in rangeland. Frontiers in Ecology and the Environment, 6, 185–190. Booth, D. T., and Tueller, P. T., 2003. Rangeland monitoring using remote sensing. Arid Land Research and Management, 17, 455–467. Booth, D. T., Cox, S. E., and Berryman, R. D., 2006. Point sampling digital imagery using “SamplePoint”. Environmental Monitoring and Assessment, 123, 97–108. Booth, D. T., Cox, S. E., Meikle, T., and Zuuring, H. R., 2008. Ground-cover measurements: assessing correlation among aerial and ground-based methods. Environmental Management, 42, 1091–1100. Booth, D. T., Cox, S. E., and Teel, D., 2010. Aerial assessment of leafy spurge (Euphorbia esula L.) on Idaho’s deep fire burn. Native Plants Journal, 11, 327–338. Bradley, B. A., and Mustard, J. F., 2006. Characterizing the landscape dynamics of an invasive plant and risk of invasion using remote sensing. Ecological Applications, 16, 1132–1147. Briske, D. D., Fuhlendorf, S. D., and Smeins, F. E., 2005. State-andtransition models, thresholds, and rangeland health: a synthesis of ecological concepts and perspectives. Rangeland Ecology & Management, 58, 1–10.

RANGELANDS AND GRAZING

Brown, M. E., Pinzón, J. E., Didan, K., Morisette, J. T., and Tucker, C. J., 2006. Evaluation of the consistency of long-term NDVI time series derived from AVHRR, SPOT-vegetation, SeaWiFS, MODIS, and Landsat ETM+ sensors. IEEE Transactions on Geoscience and Remote Sensing, 44, 1787–1793. Clark, P. E., and Hardegree, S. P., 2005. Quantifying vegetation change by point sampling landscape photography time series. Rangeland Ecology & Management, 58, 588–597. Clements, F. E., 1916. Plant Succession: An Analysis of the Development of Vegetation. Washington, DC: Carnegie Institute of Washington. Carnegie Institute Publication, Vol. 242. Daughtry, C. S. T., Hunt, E. R., Jr., and McMurtrey, J. E., III, 2004. Assessing crop residue cover using shortwave infrared reflectance. Remote Sensing of Environment, 90, 126–134. Davenport, M. L., and Nicholson, S. E., 1993. On the relation between rainfall and the normalized difference vegetation index for diverse vegetation types in east Africa. International Journal of Remote Sensing, 14, 2369–2389. de Asis, A. M., and Omasa, K., 2007. Estimation of vegetation parameter for modeling soil erosion using linear spectra mixture analysis of Landsat ETM data. ISPRS Journal of Photogrammetry and Remote Sensing, 62, 309–324. Di, L., Runquist, D. C., and Han, L., 1994. Modelling relationships between NDVI and precipitation during vegetation growth cycles. International Journal of Remote Sensing, 15, 2121–2136. DiTomaso, J. M., 2000. Invasive weeds in rangelands: species, impacts and management. Weed Science, 48, 255–265. Everitt, J. H., Yang, C., Escobar, D. E., and Davis, M. R., 2002. Using remote sensing to detect and map invasive plant species. Annals of the Arid Zone, 41, 32–342. Friedl, M. A., Sulla-Menashe, D., Tan, B., Schnieder, A., Ramankutty, N., Sibley, A., and Huang, X., 2010. MODIS collection 5 global land cover: algorithm refinements and characterization of new datasets. Remote Sensing of Environment, 114, 168–182. Groeneveld, D. P., and Watson, R. P., 2008. Near-infrared discrimination of leafless saltcedar in wintertime Landsat TM. International Journal of Remote Sensing, 29, 3577–3588. Hardin, P. J., and Jackson, M. W., 2005. An unmanned aerial vehicle for rangeland photography. Rangeland Ecology & Management, 58, 439–442. Herrick, J. E., Bestelmeyer, B. T., Archer, S., Tugel, A. J., and Brown, J. R., 2006. An integrated framework for science-based arid land management. Journal of Arid Environments, 65, 319–335. Holechek, J. L., Piper, R. D., and Herbel, C. H., 2004. Range Management: Principles and Practices, 5th edn. Upper Saddle River, NJ: Prentice Hall. Hunt, E. R., Jr., and Miyake, B. A., 2006. Comparison of stocking rates from remote sensing and geospatial data. Rangeland Ecology & Management, 59, 11–18. Hunt, E. R., Jr., Everitt, J. H., Ritchie, J. C., Moran, M. S., Booth, D. T., Anderson, G. L., Clark, P. E., and Seyfried, M. S., 2003. Applications and research using remote sensing for rangeland management. Photogrammetric Engineering & Remote Sensing, 69, 675–693. Hunt, E. R., Jr., McMurtrey, J. E., III, Parker Williams, A. E., and Corp, L. A., 2004. Spectral characteristics of leafy spurge (Euphorbia esula) leaves and flower bracts. Weed Science, 52, 492–497. Hunt, E. R., Jr., Daughtry, C. S. T., Kim, M. S., and Parker Williams, A. E., 2007. Using canopy reflectance models and spectral angels to assess potential of remote sensing to detect invasive weeds. Journal of Applied Remote Sensing, 1, 013506. Hunt, E. R., Jr., Gillham, J. H., and Daughtry, C. S. T., 2010. Improving potential geographic distribution models for invasive

657

plants by remote sensing. Rangeland Ecology & Management, 63, 505–513. Jensen, M. E., DiBenedetto, J. P., Barber, J. A., Montagne, C., and Bourgeron, P. S., 2001. Spatial modeling of rangeland potential vegetation environments. Journal of Range Management, 54, 528–536. Ji, L., and Peters, A. J., 2003. Assessing vegetation response to drought in the northern Great Plains using vegetation and drought indices. Remote Sensing of Environment, 87, 85–98. Jiang, Z., Huete, A. R., Chen, J., Chen, Y., Li, J., Yan, G., and Zhang, X., 2006. Analysis of NDVI and scaled difference vegetation index retrievals of vegetation fraction. Remote Sensing of Environment, 101, 366–378. Kuemmerle, T., Röder, A., and Hill, J., 2006. Separating grassland and shrub vegetation by multidate pixel-adaptive spectral mixture analysis. International Journal of Remote Sensing, 27, 3251–3271. Laliberte, A. S., Herrick, J. E., and Rango, A., 2010. Acquisition, orthorectification, and classification of unmanned aerial vehicle (UAV) imagery for rangeland monitoring. Photogrammetric Engineering & Remote Sensing, 76, 661–672. Lass, L. W., Thill, D. C., Shafii, B., and Prather, T. S., 2002. Detecting spotted knapweed (Centaurea maculosa) with hyperspectral remote sensing technology. Weed Technology, 16, 426–432. Lund, H. G., 2007. Accounting for the world’s rangelands. Rangelands, 29(1), 3–10. Metternicht, G. I., and Fermont, A., 1998. Estimating erosion surface features by linear mixture modeling. Remote Sensing of Environment, 64, 254–265. Montandon, L. M., and Small, E. E., 2008. The impact of soil reflectance on the quantification of the green vegetation fraction from NDVI. Remote Sensing of Environment, 112, 1835–1845. Nagler, P. L., Daughtry, C. S. T., and Goward, S. N., 2000. Plant litter and soil reflectance. Remote Sensing of Environment, 71, 207–215. National Research Council (NRC), 1994. Rangeland Health, New Methods to Classify, Inventory, and Monitor Rangelands. Washington, DC: National Academy Press. Naylor, B. J., Endress, B. A., and Parks, C. G., 2005. Multiscale detection of sulfur cinquefoil using aerial photography. Rangeland Ecology & Management, 58, 447–451. Nichols, M. H., Ruyle, G. B., and Nourbakhsh, I. R., 2009. Veryhigh-resolution panoramic photography to improve conventional rangeland monitoring. Rangeland Ecology & Management, 62, 579–582. NRCS, 2003. National Range and Pasture Handbook, Revision 1. USDA National Resource Conservation Service, Grazing Lands Technology Institute. Washington, DC: United States Department of Agriculture. NRCS, 2006. Land Resource Regions and Major Land Resource Areas of the United States, the Caribbean, and the Pacific Basin. USDA Agriculture Handbook, USDA Natural Resource Conservation Service. Washington, DC: United States Department of Agriculture, Vol. 296. Numata, I., Roberts, D. A., Chadwick, O. A., Schimel, J., Sampaio, F. R., Leonidas, F. C., and Soares, J., 2007. Characterization of pasture biophysical properties and the impact of grazing intensity using remotely sensed data. Remote Sensing of Environment, 109, 314–327. Palmer, A. R., and Fortesque, A., 2004. Remote sensing and change detection in rangelands. African Journal of Range & Forage Science, 21, 123–128. Parker Williams, A., and Hunt, E. R., Jr., 2002. Estimation of leafy spurge cover from hyperspectral imagery using mixture tuned matched filtering. Remote Sensing of Environment, 82, 446–456.

658

REFLECTED SOLAR RADIATION SENSORS, MULTIANGLE IMAGING

Paruelo, J. M., Epstein, H. E., Lauenroth, W. K., and Burke, I. C., 1997. ANPP estimates from NDVI for the central grassland region of the United States. Ecology, 78, 953–958. Petersen, S. L., Stringham, T. K., and Laliberte, A. S., 2005. Classification of willow species using large-scale aerial photography. Rangeland Ecology & Management, 58, 582–587. Pickup, G., Bastin, G. N., and Chewings, V. H., 1994. Remote-sensing-based condition assessment for nonequilibrium rangelands under large-scale commercial grazing. Ecological Applications, 4, 497–517. Pickup, G., Bastin, G. N., and Chewings, V. H., 1998. Identifying trends in land degradation in non-equilibrium rangelands. Journal of Applied Ecology, 35, 365–377. Piñeiro, G., Oesterheld, M., and Paruelo, J. M., 2006. Seasonal variation in aboveground production and radiation-use efficiency of temperate rangelands estimated through remote sensing. Ecosystems, 9, 357–373. Rango, A., Laliberte, A., Steele, C., Herrick, J. E., Bestelmeyer, B., Schmugge, T., Roanhorse, A., and Jenkins, V., 2006. Using unmanned aerial vehicles for rangelands: current applications and future potentials. Environmental Practice, 8, 159–168. Reeves, M. C., Winslow, J. C., and Running, S. W., 2001. Mapping weekly rangeland vegetation productivity using MODIS algorithms. Journal of Range Management, 54, A90–A105. Reeves, M. C., Zhao, M., and Running, S. W., 2006. Applying improved estimates of MODIS productivity to characterize grassland vegetation dynamics. Rangeland Ecology & Management, 59, 1–10. RMA, 2010. Vegetation Index Insurance Standards Handbook 2011 and Succeeding Crop Years. Washington, DC: United States Department of Agriculture. FCIC-18140 (06–2010) Federal Crop Insurance Corporation, Risk Management Agency. Roberts, D. A., Smith, M. O., and Adams, J. B., 1993. Green vegetation, nonphotosynthetic vegetation, and soils in AVIRIS data. Remote Sensing of Environment, 44, 255–269. Running, S. W., Nemani, R. R., Heinsch, F. A., Zhao, M., Reeves, M., and Hashimoto, H., 2004. A continuous satellite-derived measure of global terrestrial primary productivity. Bioscience, 54, 547–560. Sivanpillai, R., and Booth, D. T., 2008. Characterizing rangeland vegetation using Landsat and 1-mm VLSA data in central Wyoming (USA). Agroforestry Systems, 73, 55–64. Tieszen, L. L., Reed, B. L., Bliss, N. B., Wylie, B. K., and DeJong, D. D., 1997. NDVI, C3 and C4 production, and distributions in Great Plains grassland land cover classes. Ecological Applications, 7, 59–78. Tucker, C. J., Pinzon, J. E., Brown, M. E., Slayback, D. A., Pak, E. W., Mahoney, R., Vermote, E. F., and Saleous, N. E., 2005. An extended AVHRR 8-km NDVI dataset compatible with MODIS and SPOT vegetation NDVI data. International Journal of Remote Sensing, 26, 4485–4498. Underwood, E., Ustin, S., and DiPietro, D., 2003. Mapping nonnative plants using hyperspectral imagery. Remote Sensing of Environment, 86, 150–161. Vrieling, A., Rodrigues, S. C., Bartholomeus, H., and Sterk, G., 2007. Automatic identification of erosion gullies with ASTER imagery in the Brazilian Cerrados. International Journal of Remote Sensing, 28, 2723–2738. Wang, J., Rich, P. M., and Price, K. P., 2003. Temporal responses of NDVI to precipitation and temperature in the central Great Plains, USA. International Journal of Remote Sensing, 24, 2345–2364. Washington-Allen, R. A., West, N. E., Ramsey, R. D., and Efroymson, R. A., 2006. A protocol for retrospective remote sensing-based ecological monitoring of rangelands. Rangeland Ecology & Management, 59, 19–29.

Cross-references Climate Data Records Soil Properties Vegetation Indices Vegetation Phenology

REFLECTED SOLAR RADIATION SENSORS, MULTIANGLE IMAGING David J. Diner Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA

Synonyms Multiangular imaging; Multidirectional imaging; Multiple view angle imaging Definitions Sequential multiangle imaging. The acquisition of images of a scene at different view zenith angles (the angle between the local vertical and the direction to the observer) as a result of observing the scene on different occasions, for example, on different orbits of a satellite. In this case, the time separation between views may be hours, days, or weeks. Simultaneous multiangle imaging. The acquisition of images of a scene at different view zenith angles during a single overpass, obtained by reorienting the sensor in the along-track direction (the direction parallel to the sensor’s flight direction) or by using multiple sensors pointed at different along-track angles to observe the target repeatedly as the view zenith angle changes. Here, the time separation between views is several minutes. Introduction Clouds, aerosol layers, vegetation canopies, soils, snow or ice fields, and other natural scenes reflect solar radiation into angular reflectance patterns that depend on the angle at which the scene is illuminated or viewed. The Bidirectional Reflectance Distribution Function (BRDF) (Nicodemus et al., 1977; Schaepman-Strub et al., 2006) is the ratio of the radiance reflected by a target as a function of view direction, divided by the irradiance illuminating the target at a single incidence angle. The magnitude and angular shape of the BRDF is largely governed by the composition, density, and geometric structure of the reflecting medium. Many satellite instruments having a wide image swath, for example, the multispectral MODerate-resolution Imaging Spectroradiometer (MODIS) on the National Aeronautics and Space Administration (NASA) Terra and Aqua satellites, acquire multiangle data according to the sequential approach defined above. For many applications, the simultaneous multiangle measurement strategy is required, and a number of satellite sensors have been

REFLECTED SOLAR RADIATION SENSORS, MULTIANGLE IMAGING

designed with this objective in mind. Examples of such instruments are described below and provide the focus of this discussion. As noted by Diner et al. (2005), the interpretation of multiangle information can be categorized into two main strategies: geometric and radiometric. The first involves real or apparent differences in the location of observed objects with changing angle of view, resulting from stereoscopic parallax (displacement dependent upon distance from the observer) or actual motion of the target during the time interval between views. The second refers to changes in the brightness, color, contrast, or polarization of the reflected light as a function of view angle.

Satellite sensors Several Earth-orbiting instruments have been designed to make simultaneous multiangle observations at visible and shortwave infrared wavelengths (see Table 1). Airborne counterparts are associated with some of these instruments. In addition, airborne sensors such as the Cloud Absorption Radiometer (CAR) with 13 bands between 503 and 2,289 nm (King et al., 1986) (later upgraded to 14 bands between 340 and 2,302 nm; Gatebe et al., 2003) and the Advanced Solid-State Array Spectroradiometer (ASAS) with 29 bands from 465 to 871 nm (Irons et al., 1991) have provided a legacy on the information content of multiangle imagery. The United Kingdom’s Along-Track Scanning Radiometer (ATSR)-1 (Delderfield et al., 1986; Mutlow et al., 1994) was launched on the European Space Agency (ESA) Earth remote sensing satellite (ERS)-1 in 1991. It was primarily designed to retrieve global sea-surface temperatures. The only solar reflective channel in this instrument was a band at 1.6 mm. An enhanced version, ATSR-2, was equipped with additional visible and near-infrared channels and was launched in 1995 aboard ERS-2. In 2002, the Advanced ATSR (AATSR) was launched on ENVISAT. Each of these instruments contains a conical scanning radiometer that provides curved swaths at two measurement angles, one near nadir and the other at an oblique forward-viewing angle (http://www.atsr.rl.ac.uk/).

659

Contact with the ENVISAT spacecraft was lost in spring 2012, and ESA declared an end to the mission. The Polarization and Directionality of the Earth’s Reflectances (POLDER) instrument (Deschamps et al., 1994), developed by the French Centre National d’Etudes Spatiales (CNES), was launched aboard the Japanese Advanced Earth Observing Satellite (ADEOS) in 1996 and acquired 8 months of global data. A follow-on instrument, POLDER-2, was launched on ADEOS-2 in 2002 and acquired data for nearly a year. Both missions were cut short by spacecraft failures. A third POLDER instrument was launched in 2004 on the French PARASOL satellite as part of the polar-orbiting “A-train,” a constellation of satellites in similar orbits to acquire near-simultaneous observations. In late 2011, PARASOL separated from the A-train orbit, but the instrument is still collecting data. POLDER uses a wide-angle imaging system and an area array detector to acquire measurements at a multitude of along-track and cross-track angles (http://www.icare.univ-lille1.fr/parasol/). An airborne version of this instrument, AirPOLDER, also exists (http://loag5-ct13.univ-lille1.fr/AirPOLDER/data.html). The Multiangle Imaging SpectroRadiometer (MISR) (Diner et al., 1998a), developed in the United States, was launched aboard the NASA Terra spacecraft, part of the Earth Observing System, in late 1999. The MISR instrument uses nine separate cameras to acquire data at nine discrete along-track observation angles. Images are acquired globally in push broom fashion, with four line arrays in each camera filtered to four visible and near-infrared spectral bands (http://misr.jpl.nasa.gov). An airborne instrument making use of a single, gimbaled camera to acquire multiangle imagery – AirMISR – has also been used for aerosol, cloud, and surface investigations (Diner et al., 1998b; http://misr. jpl.nasa.gov/Mission/airMISR.html). The US Department of Energy’s Multispectral Thermal Imager (MTI) (Szymanski and Weber, 2005) contains 15 spectral bands of which 10 are at solar reflective wavelengths. MTI was launched in 2000. The satellite has an agile-pointing capability and multiangle imagery

Reflected Solar Radiation Sensors, Multiangle Imaging, Table 1 Earth-orbiting multiangle remote sensing imagers (A)ATSR

POLDER

Number of along- 2 14 track view angles Maximum view 56 (forward only) 60 (forward + backward) angle (at Earth’s surface) Shortwave spectral 555, 659, 865, 443a, 490, 565, 670a, 763, 765, bands 1610 nm 865a, 910 nm (apolarized) Footprint (nadir) 1  1 km2 6  7 km2 Swath width

512 km

2,200 km

MISR

MTI

CHRIS

CIPS

9

2

5

7

70 (forward + backward)

60 (typically backward)

55 (forward + backward)

71 (forward + bac kward)

446, 558, 672, 866 nm

484,558, 650, 810, Nominally 18 265 nm 874, 940, 1,015, channels from 1,376, 1,646, 415 to 1,050 nm 2,224 nm 2 2 5  5 m , 20  20 m 17  17 m2 1  2 km2

275  275 m2, 1.1  1.1 km2 400 km 12 km

13 km

950 km

660

REFLECTED SOLAR RADIATION SENSORS, MULTIANGLE IMAGING

is acquired by reorienting the satellite during a target overpass. Due to this strategy and the sensor’s narrow field of view, MTI is a targeting instrument (http:// www.lanl.gov/orgs/isr/isr2/). The Compact High Resolution Imaging Spectrometer (CHRIS) instrument (Barnsley et al., 2004) was developed in the United Kingdom and launched in 2001 on an agile spacecraft known as PROBA (Project for On-Board Autonomy), under the auspices of ESA. Like MTI, CHRIS is a targeting instrument and acquires multiangle imagery by pointing the PROBA spacecraft (http://earth. esa.int/missions/thirdpartymission/proba.html). The Cloud Imaging and Particle Size Experiment (CIPS) (McClintock et al., 2009), developed in the United States, was launched aboard NASA’s Aeronomy of Ice in the Mesosphere (AIM) mission in 2007. CIPS uses an array of four cameras in a “+” configuration, two pointing forward and backward, and two pointing to either side, each with a 44  44 field of view. Observations are acquired in a single ultraviolet spectral band to measure the physical characteristics of polar mesospheric clouds (PMCs, also known as noctilucent clouds to ground-based observers) (http:// aim.hamptonu.edu/instrmt/cips.html).

Example applications Aerosols (liquid or solid airborne particulates having sizes ranging from less than 0.1 mm to more than 10 mm) and clouds (consisting of larger-sized liquid water or ice particles) play critical roles in our climate system. Whether a particular cloud or aerosol cools or warms the surface depends on the physical properties of the particles, as well as their heights. For aerosols, injection height into the atmosphere also determines how far the particles can be transported away from their source, with ramifications for downwind air quality. The angular distribution of scattering from particles is different for spherical (liquid) cloud or aerosol droplets and among various irregularly shaped (e.g., dust) or crystal (ice) forms. As a result, multiangle imagery has proven valuable in distinguishing cloud and aerosol types characterized by the particle shape, one of several factors affecting climate impact. Data from ATSR-2, POLDER, and MISR have been used to explore ice crystal habit in cirrus clouds (Baran et al., 1999; Chepfer et al., 2001; McFarlane and Marchand, 2008). Scattering phase functions from CIPS were used to derive particle sizes of polar mesospheric clouds (Bailey et al., 2009). Additional information on cloud particle size as well as shape is contained in multiangular polarization data (Parol et al., 2004). Kalashnikova and Kahn (2006) demonstrated the ability to distinguish among granular, platelike, and spheroidal dust particles using MISR data over ocean. Discriminating between nonspherical and spherical particles is also useful in separating dust from fine particle pollution over land. Liu et al. (2007) showed improved correlations between column aerosol amount and the concentration

of near-surface airborne particulate matter with diameters less than 2.5 mm (PM2.5, a regulated air pollutant) by using multiangle data. Over land, accurate satellite aerosol retrievals are complicated by the large spatial variability in surface reflectance. Bright deserts and urban areas are major aerosol source regions, and separating the surface and atmospheric contributions to the observed top-of-atmosphere radiances is challenging. ATSR-2/AATSR and MISR use multiangle observations to solve this problem, making use of the enhanced atmospheric signal at oblique view angles and differences in the manner in which atmospheric and surface layers scatter radiation (Flowerdew and Haigh, 1996; Martonchik et al., 2002). Enhanced retrieval accuracy is obtained at certain view geometries, a subject explored using MTI (Chylek et al., 2003). Differences in the angular reflectance signature of clouds and snow/ice have proven valuable in improving cloud detection capability in polar regions (Di Girolamo and Wilson, 2003; Shi et al., 2007). Many satellite stereoscopic imagers designed to retrieve land surface topography have been placed into Earth’s orbit. Either the sequential or simultaneous viewing strategy can be used for this purpose. For remote sensing of highly dynamic targets such as clouds, however, the simultaneous multiangle strategy is required. Cloud stereo imagery can be acquired with more than one geostationary satellite viewing the targets at different angles (Seiz et al., 2007), but these equatorial-orbiting satellites do not observe polar latitudes. Alternatively, stereo observations can be obtained from a single instrument in polar orbit by acquiring data at more than one along-track view angle. Stereoscopic imaging from MISR and ATSR has been used to capture global cloud-top heights (CTH) (Moroney et al., 2002; Prata and Turner, 1997; Naud et al., 2006). Unlike thermal infrared-based techniques, which are based on measurements of cloud-top temperature, the stereoscopic approach is purely geometric, thereby making it capable of providing cloud height climatologies independent of confounding factors such as temperature inversions (Garay et al., 2008; Harshvardhan et al., 2009). Efficient, automated computational methods have been developed to “match” the images of clouds seen from different view angles (Muller et al., 2002); knowledge of the spacecraft position and view angles is then used to obtain CTH. Stereo imaging using at least three angles has made possible a novel technique for simultaneously retrieving height-resolved cloud-tracked velocity (Horváth and Davies, 2001; Zong et al., 2002). Imaging at multiple view angles helps quantify the effect of 3D cloud morphology on reflected radiation (Horváth and Davies, 2004; Cornet and Davies, 2008). Stereo imaging is also able to determine aerosol plume heights. Statistical studies of wildfire plumes in North America between 2002 and 2007 using MISR show that (a) a significant fraction of fires inject smoke into the free troposphere, (b) smoke tends to concentrate in stable layers within the atmosphere, and (c) there is a pronounced seasonal cycle of plume

REFLECTED SOLAR RADIATION SENSORS, MULTIANGLE IMAGING

heights over certain vegetation types (Kahn et al., 2008; Val Martin et al., 2010). Angular reflectances of surface covers contain information about their physical structure (e.g., Widlowski et al., 2004). Ice surface roughness provides an indicator of snow accumulation, surface melt, and wind ablation, and studies using the sensitivity of multiangle observations to surface roughness show great potential for distinguishing glacial zones and identifying different types of sea ice, which are indicators of climate change (Nolin et al., 2002; Nolin and Payne, 2007). For terrestrial vegetation, the canopy and understory density and morphology (e.g., height and width) of the plants and trees impact photosynthetic efficiency, nutrient cycling, suitability as wildlife habitats, fresh water availability, and fire risk. Multiangle data are particularly useful for separating canopy and understory/ground reflectance (Canisius and Chen, 2007; Pinty et al., 2008), which is particularly significant for bright surfaces (i.e., snow or desert). CHRIS/ PROBA and MISR have been used for remote sensing of desert shrub cover (Chopping et al., 2003, 2008a), which has been expanding during the last century in the southwestern USA at the expense of grasslands. Geometric-optical methods for using multiangle data to retrieve plant number density, mean radius, crown height, and crown shape in forests are also being developed (e.g., Chopping et al., 2008b). One technique that holds significant promise is the use of active lidar, which provides more direct measurements of tree heights, in conjunction with passive multiangular imagery to retrieve canopy structure (Kimes et al., 2006; Schull et al., 2007).

Conclusions Multiangle imaging is a rich source of information regarding the micro- and macrophysical structure of different types of aerosols, cloud forms, and surface covers. Combined with stereoscopic techniques, it enables construction of 3D scene models and estimation of the total amount of sunlight reflected by Earth’s diverse environments. As with any imaging technology, the range of applications is myriad. Just a few representative examples have been presented here. Many other uses exist, and additional ones likely remain to be discovered. Acknowledgment The research to prepare this contribution was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with NASA. Bibliography Bailey, S. M., Thomas, G. E., Rusch, D. W., Merkel, A. W., Jeppesen, C. D., Carstens, J. N., Randall, C. E., McClintock, W. E., and Russell, J. M., III, 2009. Phase functions of polar mesospheric cloud ice as observed by the CIPS instrument on the AIM satellite. Journal of Atmospheric and Terrestrial Physics, 71, 373–380. Baran, A. J., Watts, P. D., and Francis, P. N., 1999. Testing the coherence of cirrus microphysical and bulk properties retrieved

661

from dual-viewing multispectral satellite radiance measurements. Journal of Geophysical Research, 104, 31673–31683. Barnsley, M. J., Settle, J. J., Cutter, M., Lobb, D., and Teston, F., 2004. The PROBA/CHRIS mission: a low-cost smallsat for hyperspectral, multi-angle, observations of the Earth surface and atmosphere. IEEE Transactions on Geoscience & Remote Sensing, 42, 1512–1520. Canisius, F., and Chen, J. M., 2007. Retrieving vegetation background reflectance from multi-angle imaging spectroradiometer (MISR) data. Remote Sensing of Environment, 107, 312–321. Chepfer, H., Goloub, P., Riedi, J., Haan, J. F. D., Hovenier, J. W., and Flamant, P. H., 2001. Ice crystal shapes in cirrus clouds derived from POLDER/ADEOS-1. Journal of Geophysical Research, 106, 7955–7966. Chopping, M., Su, L., Laliberte, A., Rango, A., Peters, D. P. C., and Kollikkathara, N., 2003. Mapping shrub abundance in desert grasslands using geometric-modeling and multi-angle remote sensing with CHRIS/Proba. Remote Sensing of Environment, 104, 62–73. Chopping, M., Su, L., Rango, A., Martonchik, J. V., Peters, D. P. C., and Laliberte, A., 2008a. Remote sensing of woody shrub cover in desert grasslands using MISR with a geometric-optical canopy reflectance model. Remote Sensing of Environment, 112, 19–34. Chopping, M., Moisen, G., Su, L., Laliberte, A., Rango, A., Martonchik, J. V., and Peters, D. P. C., 2008b. Large area mapping of southwestern forest crown cover, canopy height, and biomass using MISR. Remote Sensing of Environment, 112, 2051–2063. Chylek, P., Henderson, B., and Mishchenko, M., 2003. Satellite based retrieval of aerosol optical thickness: the effect of sun and satellite geometry. Geophysical Research Letters, 30, doi:10.1029/2003GL016917. Cornet C., and R. Davies, 2008. Use of MISR measurements to study the radiative transfer of an isolated convective cloud: implications for cloud optical thickness retrieval. Journal of Geophysical Research, 113, Art. No. D04202. Delderfield, J., Llewellyn-Jones, D. T., Bernard, R., de Javel, Y., Williamson, E. J., Mason, I., Pick, D. R., and Barton, I. J., 1986. The along track scanning radiometer (ATSR) for ERS-1. Proceedings of SPIE, 589, 114–120. Deschamps, P.-Y., Bréon, F. M., Leroy, M., Podaire, A., Bricaud, A., Buriez, J.-C., and Sèze, G., 1994. The POLDER mission: instrument characteristics and scientific objectives. IEEE Transactions on Geoscience & Remote Sensing, 32, 598–615. Di Girolamo, L., and Wilson, M. J., 2003. A first look at banddifferenced angular signatures for cloud detection from MISR. IEEE Transactions on Geoscience & Remote Sensing, 41, 1730–1734. Diner, D. J., Beckert, J. C., Reilly, T. H., Bruegge, C. J., Conel, J. E., Kahn, R., Martonchik, J. V., Ackerman, T. P., Davies, R., Gerstl, S. A. W., Gordon, H. R., Muller, J.-P., Myneni, R., Sellers, P. J., Pinty, B., and Verstraete, M. M., 1998a. Multi-angle imaging spectroradiometer (MISR) description and experiment overview. IEEE Transactions on Geoscience & Remote Sensing, 36, 1072–1087. Diner, D. J., Barge, L. M., Bruegge, C. J., Chrien, T. G., Conel, J. E., Eastwood, M. L., Garcia, J. D., Hernandez, M. A., Kurzweil, C. G., Ledeboer, W. C., Pignatano, N. D., Sarture, C. M., and Smith, B. G., 1998b. The airborne multi-angle imaging spectroradiometer (AirMISR): instrument description and first results. IEEE Transactions on Geoscience & Remote Sensing, 36, 1339–1349. Diner, D. J., Braswell, B. H., Davies, R., Gobron, N., Hu, J., Jin, Y., Kahn, R. A., Knyazikhin, Y., Loeb, N., Muller, J.-P., Nolin, A. W., Pinty, B., Schaaf, C. B., Seiz, G., and Stroeve, J., 2005. The value of multiangle measurements for retrieving structurally

662

REFLECTED SOLAR RADIATION SENSORS, MULTIANGLE IMAGING

and radiatively consistent properties of clouds, aerosols, and surfaces. Remote Sensing of Environment, 97, 495–518. Flowerdew, R. J., and Haigh, J. D., 1996. Retrieval of aerosol optical thickness over land using the ATSR-2 dual-look satellite radiometer. Geophysical Research Letters, 23, 351–354. Garay, M. J., de Szoeke, S. P., and Moroney, C. M., 2008. Comparison of marine stratocumulus cloud top heights in the Southeastern Pacific retrieved from satellites with coincident ship-based observations. Journal of Geophysical Research, 113, D18204. Gatebe, C. K., King, M. D., Platnick, S., Arnold, G. T., Vermote, E. F., and Schmid, B., 2003. Airborne spectral measurements of surface–atmosphere anisotropy for several surfaces and ecosystems over southern Africa. Journal of Geophysical Research, 108, doi:10.1029/2002JD002397. Harshvardhan, Zhao, G., Di Girolamo, L., and Green, R. N., 2009. Satellite-observed location of stratocumulus cloud-top heights in the presence of strong inversions. IEEE Transactions on Geoscience and Remote Sensing, 47, 1421–1428. Horváth, Á., and Davies, R., 2001. Simultaneous retrieval of cloud motion and height from polar-orbiter multiangle measurements. Geophysical Research Letters, 28, 2915–2918. Horváth, Á., and Davies, R., 2004. Anisotropy of water cloud reflectance: a comparison of measurements and 1D theory. Geophysical Research Letters, 31, L01102, doi:10.1029/ 2003GL018386. Irons, J. R., Ranson, K. J., Williams, D. L., Irish, R. R., and Huegel, F. G., 1991. An off-nadir-pointing imaging spectroradiometer for terrestrial ecosystem studies. IEEE Transactions on Geoscience & Remote Sensing, 29, 66–74. Kahn, R., Chen, Y., Nelson, D. L., Leung, F.-Y., Li, Q., Diner, D. J., and Logan, J. A., 2008. Wildfire smoke injection heights: two perspectives from space. Geophysical Research Letters, 35, L04809, doi:10.1029/2007GL032165. Kalashnikova, O. V., and Kahn, R., 2006. Ability of multiangle remote sensing observations to identify and distinguish mineral dust types: part 2. Sensitivity over dark water. Journal Geophysical Research, 111, D11207, doi:10.1029/2005JD006756. Kimes, D. S., Ranson, K. J., Sun, G., and Blair, J. B., 2006. Predicting lidar measured forest vertical structure from multi-angle spectral data. Remote Sensing of Environment, 100, 503–511. King, M. D., Strange, M. G., Leone, P., and Blaine, L. R., 1986. Multiwavelength scanning radiometer for airborne measurements of scattered radiation within clouds. Journal of Atmospheric and Oceanic Technology, 3, 513–522. Liu, Y., Koutrakis, P., Kahn, R., Turquety, S., and Yantosca, R. M., 2007. Estimating fine particulate matter component concentrations and size distributions using satellite-retrieved fractional aerosol optical depth: part 2 – a case study. Journal of the Air & Waste Management Association, 57, 1360–1369. Martonchik, J. V., Diner, D. J., Crean, K. A., and Bull, M. A., 2002. Regional aerosol retrieval results from MISR. IEEE Transactions on Geoscience & Remote Sensing, 40, 1520–1531. McClintock, W. E., Rusch, D. W., Thomas, G. E., Merkel, A. W., Lankton, M. R., Drake, V. A., Bailey, S. M., and Russell, J. M., III, 2009. The cloud imaging and particle size experiment on the aeronomy of ice in the mesosphere mission: instrument concept, design, calibration, and on-orbit performance. Journal of Atmospheric and Terrestrial Physics, 71, 340–355. McFarlane, S. A., and Marchand, R. T., 2008. Analysis of ice crystal habits derived from MISR and MODIS observations over the ARM Southern Great Plains site. Journal of Geophysical Research, 113, D07209, doi:10.1029/2007JD009191. Moroney, C., Davies, R., and Muller, J.-P., 2002. Operational retrieval of cloud-top heights using MISR data. IEEE Transactions on Geoscience & Remote Sensing, 40, 1541–1546.

Muller, J.-P., Mandanayake, A., Moroney, C., Davies, R., Diner, D. J., and Paradise, S., 2002. MISR stereoscopic image matchers: techniques and results. IEEE Transactions on Geoscience and Remote Sensing, 40, 1547–1559. Mutlow, C. T., Zavody, A. M., Barton, I. J., and Llewellyn-Jones, D. T., 1994. Sea-surface temperature-measurements by the along-track scanning radiometer on the ERS-1 satellite – early results. Journal of Geophysical Research, 99, 22575–22588. Naud, C., Muller, J.-P., and Clothiaux, E. E., 2006. Assessment of multispectral ATSR2 stereo cloud-top height retrievals. Remote Sensing of Environment, 104, 337–345. Nicodemus, F. E., Richmond, J. C., Hsia, J. J., Ginsberg, I. W., and Limperis, T., 1977. Geometrical Considerations and Nomenclature for Reflectance. Washington, DC: National Bureau of Standards, U.S. Department of Commerce. NBS Monograph, Vol. 160. Nolin, A. W., and Payne, M., 2007. Classification of glacier zones in western Greenland using albedo and surface roughness from the multi-angle imaging spectroradiometer (MISR). Remote Sensing of Environment, 107, 264–275. Nolin, A. W., Fetterer, F. M., and Scambos, T. A., 2002. Surface roughness characterizations of sea ice and ice sheets: case studies with MISR data. IEEE Transactions on Geoscience & Remote Sensing, 40, 1605–1615. Parol, F., Buriez, J. C., Vanbauce, C., Riedi, J., Labonnote, L. C., Doutriaux-Boucher, M., Vesperini, M., Sèze, G., Couvert, P., Viollier, M., and Bréon, F.-M., 2004. Capabilities of multi-angle polarization cloud measurements from satellite: POLDER results. Advances in Space Research, 33, 1080–1088. Pinty, B., Lavergne, T., Kaminski, T., Aussedat, O., Giering, R., Gobron, N., Taberner, M., Verstraete, M. M., Vobbeck, M., and Widlowski, J.-L., 2008. Partitioning the solar radiant fluxes in forest canopies in the presence of snow. Journal of Geophysical Research, 113, D04104, doi:10.1029/2007JD009096. Prata, A. J., and Turner, P. J., 1997. Cloud top height determination using ATSR data. Remote Sensing of Environment, 59, 1–13. Schaepman-Strub, G., Schaepman, M. E., Painter, T. H., Dangel, S., and Martonchik, J. V., 2006. Reflectance quantities in optical remote sensing – definitions and case studies. Remote Sensing of Environment, 103, 27–42. Schull, M. A., Ganguly, S., Samanta, A., Huang, D., Shabanov, N. V., Jenkins, J. P., Chiu, J. C., Marshak, A., Blair, J. B., Myneni, R. B., and Knyazikhin, Y., 2007. Physical interpretation of the correlation between multi-angle spectral data and canopy height. Geophysical Research Letters, 34, L18405, doi:10.1029/2007GL031143. Seiz, G., Tjemkes, S., and Watts, P., 2007. Multiview cloud-top height and wind retrieval with photogrammetric methods: application to Meteosat-8 HRV observations. Journal of Applied Meteorology and Climatology, 46, 1182–1195. Shi, T., Clothiaux, E. E., Yu, B., Braverman, A. J., and Groff, D. N., 2007. Detection of daytime arctic clouds using MISR and MODIS data. Remote Sensing of Environment, 107, 172–184. Szymanski, J., and Weber, P., 2005. Multispectral thermal imager: mission and applications overview. IEEE Transactions on Geoscience and Remote Sensing, 43, 1943–1949. Val Martin, M., Logan, J. A., Kahn, R. A., Leung, F.-Y., Nelson, D. L., and Diner, D. J., 2010. Smoke injection heights from fires in North America: analysis of 5 years of satellite observations. Atmospheric Chemistry and Physics, 10, 1491–1510. Widlowski, J.-L., Pinty, B., Gobron, N., Verstraete, M. M., Diner, D. J., and Davis, A. B., 2004. Canopy structure parameters derived from multi-angular remote sensing data for terrestrial carbon studies. Climatic Change, 67, 403–415. Zong, J., Davies, R., Muller, J.-P., and Diner, D. J., 2002. Photogrammetric retrieval of cloud advection and top height from the Multi-angle imaging spectroradiometer (MISR). Photogrammetric Engineering and Remote Sensing, 68, 821–829.

REFLECTED SOLAR RADIATION SENSORS, POLARIMETRIC

Cross-references Aerosols Cloud Properties Optical/Infrared, Scattering by Aerosols and Hydrometeors Reflected Solar Radiation Sensors, Polarimetric

REFLECTED SOLAR RADIATION SENSORS, POLARIMETRIC David J. Diner Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA

Synonyms Stokes measurements Definitions Stokes vector. A set of four parameters describing the polarization state of a beam of light, named for the Irish mathematical physicist George Gabriel Stokes (1819–1903). As an electromagnetic wave propagates, the orientation of the tip of the electric field vector traces out an ellipse. Specific manifestations of the polarization ellipse are linear polarization, where the electric field vibrates in a single plane, and circular polarization. The Stokes vector consists of intensity, I; the excess of horizontally over vertically polarized light, Q; the excess of light polarized at 45 over 135 , U; and the excess of right-handed over left-handed circular polarization, V. “Handedness” describes the direction in which the electric field vector rotates. Degree of polarization. The ratio offfi polarized to total pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi =I.ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi The degree of intensity is equal to Q2 þ U 2 þ V 2p ffi linear polarization (DOLP) is equal to Q2 þ U 2 =I, and the degree of circular polarization (DOCP) is equal to V/I. Angle of polarization. The orientation of the major axis of the polarization ellipse. In terms of the Stokes components, the angle of polarization equals to 0.5 arctan (U/Q). Polarimetry. Measurement of the Stokes vector, typically involving the use of polarization analyzers, devices that preferentially transmit certain states of polarization, and retarders or phase plates, which alter the shape of the polarization ellipse. Division-of-amplitude polarimeters optically divide the incoming light and direct the resulting beams toward multiple analyzers and detectors. Divisionof-aperture polarimeters employ multiple analyzers operating side-by-side. Snapshot devices encode the polarization state within a spatially varying signal recorded on an area array detector. Spectrum channeling approaches use thick birefringent crystals plus a polarizer and spectrometer to encode the Stokes components in fringes superimposed on the recorded spectrum. A timemultiplexed polarimeter uses a rotating analyzer or

663

retarder, or an electro-optical retardance modulator, and the measurements are acquired sequentially.

Introduction Sunlight incident on the Earth system is unpolarized, but scattering by clouds, aerosols, and surfaces in general polarizes the light. Polarimeters of myriad designs, operating in various wavelength intervals, have been used in a host of scientific applications. A classic example of the application of polarimetry to planetary astronomy was the identification of sulfuric acid clouds in the upper atmosphere of Venus (Hansen and Hovenier, 1974). This entry presents examples of Earth remote sensing polarimeters that record reflected sunlight photographically or photometrically from suborbital and orbital platforms. With the exception of a brief discussion on rocket-borne observations of mesospheric clouds, the scope is confined to sensors viewing the Earth at angles between the downward (nadir) and horizontal (limb) directions. Polarimetric remote sensing from balloons Photopolarimetry of earthlight was obtained from highaltitude balloons over Arizona and New Mexico by researchers at the University of California, Los Angeles (Rao and Sekera, 1967; Rao, 1969). Photomultiplier tube-based instruments with a 3 field of view (FOV) were scanned on either side of nadir. Initial measurements, from 13 km altitude, used a rotating half-wave retarder, a fixed polarizing prism, and spectral filters at 327, 398, 500, and 612 nm. Later observations (from 28 km) used 362, 401, 501, and 599 nm filters and a rotating calcite prism for polarimetric modulation. Balloonborne (20 km altitude) polarimetry over southwestern France was obtained by Herman et al. (1986) at 850 and 1,650 nm using a rotating Polaroid analyzer and a germanium photodiode. Horizontal scans obtained when the Sun was close to the horizon provided evidence of spherical submicron particles consistent with established stratospheric aerosol models. Downwardlooking observations of tropospheric aerosols, clouds, and the ocean surface were later obtained by adding a scan mirror (Deuzé et al., 1989). The three commonly known neutral points in the polarization of skylight – Arago, Babinet, and Brewster, named after their discoverers – are locations where the degree of polarization vanishes, typically displaced tens of degrees from the solar or antisolar directions. Horváth et al. (2002) confirmed the existence of a suspected fourth neutral polarization point below the antisolar direction (thus invisible from the ground). This was accomplished by flying a Nikon F801 film camera equipped with a fisheye lens and linear polarizers oriented at 0 , 45 , and 90 (Gál et al., 2001) in a low-altitude (0.8–3.5 km) hot air balloon. The data were collected during two flights over central Hungary.

664

REFLECTED SOLAR RADIATION SENSORS, POLARIMETRIC

Polarimetric remote sensing from rockets Noctilucent or polar mesospheric clouds consist of tiny ice particles that reside in the Earth’s upper atmosphere at altitudes near 80 km. Remote sensing from suborbital rockets enables determination of their light-scattering properties and vertical distribution. Witt et al. (1976) at the University of Stockholm used a photomultiplier-based instrument operating at 256 and 536 nm to derive a particle size upper bound of 50 nm. A two-color (410 and 540 nm) polarimeter from the Dudley Observatory (New York) implied somewhat larger sizes (Tozer and Beeson, 1974). More recently, the Ultraviolet Imaging Polarimeter (UVIP) built by the University of Colorado (Lawrence et al., 1994) provided rocket-borne limb imaging using a Reticon detector array. Data at polarizer orientations of 0 , 45 , and 90 were acquired at 265 nm. Polarimetric remote sensing from airplanes Among the earliest non-imaging aircraft polarimeters was a Grumman instrument deployed on a Piper Tri-Pacer (Egan, 1968). The instrument had a 1 FOV and was equipped with filters at 533 and 1,000 nm and employed a rotating Polaroid analyzer. Initial results over forest discriminated red pine and hardwood trees. Hariharan (1969) described an airborne polarimeter for atmospheric radiation studies consisting of polarizing (Glan-Thompson) prisms, four blue-green spectral filters, and a photomultiplier tube. It flew in the National Aeronautics and Space Administration (NASA) CV-990 aircraft and made observations over a uniform cloud deck. A University of Arizona instrument (Fernald et al., 1969) with a Polaroid analyzer oriented at 0 , 90 , and 120 observed sparsely vegetated desert terrain at 429 nm from 300 to 600 m altitude. DOLP varied from 3 % to 18 %, depending on scattering angle, and the angle of polarization was generally perpendicular to the scattering plane. The University of Arizona’s AEROPOL instrument (Coffeen et al., 1975) flew on the NASA CV-990. Observations within a 1.5 FOV were acquired between 1,100 and 3,500 nm using a rotating wire-grid analyzer to study cloud and aerosol polarization. Comparisons with theoretical computations showed sensitivity to cloud phase and particle size distribution (Hansen and Coffeen, 1974). Early airborne imaging polarimeters included the Grumman Digital Photometric Mapper (DPM) (Halajian and Hallock, 1972) and an instrument developed at the University of Cologne (Prosch et al., 1983). The DPM used an image dissector tube and required multiple passes to obtain different polarization orientations. An inverse relationship between water turbidity and DOLP was demonstrated. The Cologne instrument contained three Plumbicon-based video cameras mounted side-by-side with polarizing filters at 0 , 60 , and 120 , spectrally filtered to a single band at 550 nm. Data collected over the Oleftalsperre reservoir indicated a sensitivity to surface wind speed.

The Airborne Polarization and Directionality of Earth’s Reflectances (AirPOLDER) instrument, built by the Laboratoire d’Optique Atmosphérique at the Université de Lille became operational in June 1990 (Deuzé et al., 1993). A rotating wheel provides spectral and polarimetric filter selection in discrete bands between 443 and 910 nm. A silicon charge-coupled device (CCD) detector array and a wide-angle lens provide imagery over a
41 
51 FOV. Successive images obtained during flight enable each point to be observed at 12 view angles. More recently, the Lille group developed a single-view-angle aerosol polarimeter, MICROPOL, which operates in the visible/nearinfrared (VNIR) and shortwave-infrared (SWIR) (Waquet et al., 2005). Three separate optical systems at each wavelength contain polarizers in orientations separated by 60 . A new instrument based on the AirPOLDER design, the Observing System Including polaRization in the solar Infrared Spectrum (OSIRIS), uses separate VNIR and SWIR optical systems to acquire data at 440, 490, 670, 763, 765, 865, 910, 940, 1,020, 1,240, 1,365, 1,600, and 2,200 nm (Auriol et al., 2008). It has been tested aboard a French research aircraft flying at 10 km altitude. Several of the instruments described in this entry are incomplete polarimeters (Chipman, 1994) that measure linear polarization in just two perpendicular orientations, typically providing Q but not U. One such sensor was developed at NASA Ames Research Center to study oil slicks (Millard and Arvesen, 1972). Measurements at 380 nm over controlled spills were acquired with polarizers parallel and perpendicular to the flight direction of a Cessna 401. The difference between the two measurements increased by 25 % over the oil-contaminated water. Another Ames sensor was developed for the high-altitude ER-2 aircraft (Hildum and Spinhirne, 1992). It used a commercial silicon CCD camera to acquire data in support of cloud observation campaigns. Colorado State University’s Scanning Spectral Polarimeter (SSP) also included channels for left and right circular polarization (Stephens et al., 2000). The Research Scanning Polarimeter (RSP), developed by SpecTIR Corporation and operated by the NASA Goddard Institute for Space Studies (Cairns et al., 2003), was designed for high polarimetric accuracy (> 0 Io dx dx

720

dU >0 dx (divergent) dU = small dx δI 35 m/s), however, so increases at a much slower rate with increasing wind speed. Similar saturation is found in the European Advanced Scatterometer (ASCAT), measuring at C-band. Such high wind saturation has also been observed from aircraft flying over hurricanes. When the model function developed over the moderate wind range is applied to the strong winds, an underestimation of wind speed results. Strong efforts have been made to adjust the model function (slope in Figure 3) in strong winds and to find the right channel (combination of polarization, frequency, incident angle) that would be sensitive to the increase of strong winds (e.g., Fernandez et al., 2006). The success would be difficult if flow separation occurs at high winds and the surface roughness and stress

−10

−10

−15

−15 sigma O–H (dB)

sigma O–V (dB)

SEA SURFACE WIND/STRESS VECTOR

−20 −25 −30 −35

765

−20 −25 −30

3

10 20 HWIND speed (m/s)

30

40 50 60

−35

3

10

20

30

40 50 60

HWIND speed (m/s)

Sea Surface Wind/Stress Vector, Figure 6 Normalized radar cross section at two polarizations measured by QuikSCAT for 12 hurricanes as a function of colocated surface wind provided by the National Hurricane Center.

do not increase with winds as discussed in section “Relations between wind and stress.”

Direct stress retrieval Weissman and Graber (1999) made an early attempt to build a GMF to retrieve stress instead of wind. There are many reasons for a GMF-S (stress) to retrieve stress (or U*) directly rather than the present GMF-W to retrieve UN. The first reason is the deficiency of the present GMF-W, which should be developed and calibrated with UN computed from research quality in situ wind measurements, as described in the section “Relations between wind and stress.” Such computation of UN from in situ measurements was performed before credible ocean surface wind products became available from operational NWP centers. Most of the tuning of the revised GMF after Seasat was based on NWP products that are not UN (not corrected for stability dependence). The resultant errors are not reversible and difficult to gauge. The second reason is the uncertainty of the drag coefficient. Ideally, stress could be derived from UN retrieved from the scatterometer, using a neutral drag coefficient. However, if the drag coefficient is not the same as that was used to derive UN for development of the GMF, error will be introduced through the uncertainty of drag coefficient. The other two reasons are related to the directional difference between wind and stress. The procedure to “select” the stress direction should be different from wind direction in two ways. In the first way, we should initialize the ambiguity removal process with “nudging” fields that are more relevant to stress than wind. Where a strong ocean current exists, the stress should point to the direction of the vector difference between wind and current. The second way is to develop a flexible median filter to accommodate the small spatial scale of stress as compared

with winds. One of the problems of the selection is the lack of sufficient current information. One of the reasons usually given for promoting scatterometer as a wind sensor instead of a stress sensor is that there is more wind than stress measurements to develop and calibrate the GMF. Such an explanation is not valid because UN, by definition, has an unambiguous relation with stress and needs stress for computation. To provide each UN for development or calibration of the GMF from measured wind U, stress or U* has to be computed first as discussed in the section “Relations between wind and stress.” The stress derived from wind in such a way is not ideal because it addresses only the stability problem but does not include current information. Such deficiency may be somewhat alleviated through the ambiguity removal process by using more appropriate filter size and nudging with the vector difference between wind and optimal surface current information that is available. The direct retrieval of stress depends not only on the fast and largescale atmospheric circulation but also on the small-scale and slow ocean processes, as reflected in surface current and temperature.

Other sensors Both the microwave altimeter and SAR are similar to the scatterometer in the sense that they are active sensors that send microwave pulses to the Earth’s surface and measure the backscattered power. The microwave radiometer is a passive sensor, observing the radiance emitted by the Earth and its atmosphere. While the scatterometer views at oblique angles, the altimeter views at nadir (very small incident angles). At nadir, the backscattered energy is a result of specular reflection (the wavelets serve as small mirrors), and the backscatter is not sensitive to UN direction. Because the

766

SEA SURFACE WIND/STRESS VECTOR

instrument is not scanning, data are only available at very narrow (2 km) repeated ground tracks. The coverage of all the past altimeters is poor compared with the scatterometer and the microwave radiometers. A SAR looks perpendicular to aircraft path only at one azimuth angle and cannot resolve the UN direction like the scatterometer. SAR has spatial resolutions that are much better than scatterometers, but the high resolution also introduces higher uncertainties in accuracy caused by secondary effects that affect surface roughness. The instrument and the data processing procedure are much more complicated than the scatterometer. The scatterometer GMF can be used to relate the so measured by SAR to UN. However, a particular value of so may correspond to a range of UN, depending on the azimuth angle. Hence, in order to retrieve UN with the GMF, the UN direction must first be specified. Whether the a priori direction information is derived from the orientation of km-scale structure in the SAR image or from operational NWP models, the spatial scales are much coarser than those of so. Ocean surface wind speed has also been derived from the radiance measured by a microwave radiometer. It is generally believed that wind speed affects the surface emissivity indirectly through the generation of ocean waves and foam. Radiometers designed to observe the ocean surface operate primarily at window frequencies, where atmospheric absorption is low. Radiances at frequencies sensitive to sea surface temperature, atmospheric water vapor, and liquid water are also measured; they are used to correct for the slight interference by the atmosphere. It was demonstrated in several airborne experiments (e.g., Yueh et al., 1997) that the polarization properties of the sea surface emission vary not only as a function of the wind speed but also as a function of wind direction. The wind direction measuring capability is being tested by a polarimetric radiometer, WindSat on the Coriolis satellite, developed by the US Navy.

Potential improvements Historically, the European Space Agency used the C-band (5 GHz), but NASA prefers the Ku-band (14 GHz) for their scatterometers. The backscatter at higher frequencies is more sensitive to shorter ocean waves. The Ku-band is more sensitive to weak wind-stress variations but is more subject to atmospheric effects and rain contamination. Attempts have been made to retrieve winds from L-band (1.4 GHz) scatterometer on Aquarius (Yueh et al., 2013). There have been calls for a multifrequency scatterometer that is sensitive to various parts of ocean surface wave spectrum and may reduce atmospheric and rain effects. Present scatterometers are real-aperture systems and the spatial resolution is limited by the antenna size. A larger antenna will, of course, enhance the spatial resolution. Another way to achieve higher resolution is to add

a synthetic aperture capability. One of the drawbacks of present scatterometers is the ambiguity in retrieving wind-stress direction. The backscatter is a cosine function of the azimuth angle. The correlation between copolarized and cross-polarized backscatter is a sine function of azimuth angle. By adding polarized measurement capability to the scatterometer, the directional ambiguity problem could be mitigated. One polar-orbiting scatterometer at a low-altitude (e.g., 800 km) orbit can sample at a location on Earth not more than two times a day. Additional instrument flying in tandem will allow the description of higher temporal variability and the reduction of the aliasing (bias introduced by subsampling) of the mean wind stress. As demonstrated by Liu et al. (2008b), adding ASCAT to QuikSCAT decreases the time required to cover 90 % of the Earth from 24 to 19 h. Adding Oceansat-2 of India resolves the inertial frequency desired by oceanographers and provides 6 hourly repeated coverage required by operational weather forecasters. The adverse effect due to the demise of QuikSCAT should be mitigated by data from WindSat or the scatterometer on Haiyang-2 of China. The scatterometer on Space Station will provide new sampling opportunity. Not all space-based ocean surface wind and stress measurements are comparable in quality. Standardizing the technology requirements for observation accuracies of different research and operational applications and international cooperation are very desirable. Many scientific reports have affirmed the need for high-quality, continuous, and consistent long time series of ocean surface vector winds and stress.

Conclusion The basics of scatterometry and air-sea turbulence transfer are reviewed to bring out the uniqueness of the scatterometer in measuring ocean surface stress in addition to wind. The ubiquitous spatial coherence of scatterometer measurements with ocean surface temperature and current is attributed to the two ocean factors that drive the buoyancy and wind-shear production of turbulence transfer (stress); the factors are less directly influential on wind. The reduced sensitivity of the scatterometer to the increase of wind speed at hurricane scale winds is related to the failure of conventional drag coefficient caused by flow separation. A scatterometer that measures stress better than wind is still important to the estimation of the dynamic forcing and the oceanic feedback that affects the maintenance and intensification of the storm. The feasibility, advantage, and need for a geophysical model function to retrieve stress directly rather than the equivalent neutral wind (the present geophysical product of the scatterometer) are explained. The direct retrieval of stress from scatterometer measurements will enable new science applications with a new perspective, as expounded by Liu et al. (2010).

SEVERE STORMS

Acknowledgments This study was performed at the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration (NASA). It was jointly supported by the Ocean Vector Winds, Ocean Surface Salinity, and Precipitation Measuring Mission Programs of NASA. Wenqing Tang provided valuable assistance. Bibliography Bourassa, M. A., Vincent, D. G., and Wood, W. L., 1999. A flux parameterization including the effects of capillary waves and sea state. Journal of the Atmospheric Sciences, 56, 1123–1139. DeCosmo, J., Katsaors, K. B., Smith, S. D., Anderson, R. J., Oost, W. A., Bumke, K., and Chadwick, H., 1996. Air-sea exchange of water vapor and sensible heat: the humidity exchange over the sea (HEXOS) results. Journal of Geophysical Research, 101, 12001–12016. Donelan, M. A., Haus, B. K., Reul, N., Plant, W. J., Stiassnie, M., Graber, H. C., Brown, O. B., and Saltzman, E. S., 2004. On the limiting aerodynamic roughness of the ocean in very strong winds. Geophysical Research Letters, 31, L18306, doi:10.1029/2004GL019460. Emanuel, K., 1995. Sensitivity of tropical cyclones to surface exchange coefficients and a revised steady state model incorporating eye dynamics. Journal of the Atmospheric Sciences, 52, 3969–3976. Fairall, C. W., Bradley, E. F., Rogers, D. P., Edson, J. B., and Young, G. S., 1996. Bulk parameterization of air-sea fluxes in TOGA COARE. Journal of Geophysical Research, 101, 3747–3767. Fernandez, D. E., Carswell, J. R., Frasier, S., Chang, P. S., Black, P. G., and Marks, F. D., 2006. Dual-polarized C- and Ku-band ocean backscatter response to hurricane-force winds. Journal of Geophysical Research, 111, C08013, doi:10.1029/ 2005JC003048. Large, W. G., and Pond, S., 1981. Open ocean momentum flux measurements in moderate to strong winds. Journal of Physical Oceanography, 11, 324–336. Liu, W. T., and Large, W. G., 1981. Determination of surface stress by Seasat-SASS: a case study with JASIN data. Journal of Physical Oceanography, 11, 1603–1611. Liu, W. T., and Tang, W., 1996. Equivalent Neutral Wind. Pasadena: JPL Publication 96–17, Jet Propulsion Laboratory. 16 pp. Liu, W. T., and Xie, X., 2006. Measuring ocean surface wind from space. In Gower, J. (ed.), Remote Sensing of the Marine Environment, 3rd edn. Bethesda: American Society for Photogrammetry and Remote Sensing. Manual of Remote Sensing, Vol. 6, pp. 149–178. Chapter 5. Liu, W. T., and Xie, X., 2008. Ocean–atmosphere momentum coupling in the Kuroshio extension observed from space. Journal of Oceanography, 64, 631–637. Liu, W. T., Katsaros, K. B., and Businger, J. A., 1979. Bulk parameterization of air-sea exchanges in heat and water vapor including the molecular constraints at the interface. Journal of the Atmospheric Sciences, 36, 1722–1735. Liu, W. T., Xie, X., and Niiler, P. P., 2007. Ocean–atmosphere interaction over Agulhas extension meanders. Journal of Climate, 20, 5784–5797. Liu, W. T., Tang, W., and Xie, X., 2008a. Wind power distribution over the ocean. Geophysical Research Letters, 35, L13808, doi:10.1029/2008GL034172. Liu, W. T., Tang, W., Xie, X., Navalgund, R., and Xu, K., 2008b. Power density of ocean surface wind-stress from international

767

scatterometer tandem missions. International Journal of Remote Sensing, 29, 6085–6090. Liu, W. T., Xie, X., and Tang, W., 2010. Scatterometer’s unique capability in measuring ocean surface stress. In Barale, V., Gower, J. F. A., and Alberotanza, L. (eds.), Oceanography from Space. New York: Springer, pp. 93–111. Chapter 6. Park, K.-A., Cornillon, P., and Codiga, D. L., 2006. Modification of surface winds near ocean fronts: effects of the Gulf Stream rings on scatterometer (QuikSCAT, NSCAT) wind observations. Journal of Geophysical Research, 111, C03021, doi:10.1029/ 2005JC003016. Powell, M. D., Vickery, P. J., and Reinhold, T., 2003. Reduced drag coefficient for high wind speeds in tropical cyclones. Nature, 422, 279–283. Weissman, D. E., and Graber, H. C., 1999. Satellite scatterometer studies of ocean surface stress and drag coefficients using a direct model. Journal of Geophysical Research, 104(C5), 11329–11335. Yueh, S. H., Wilson, W. J., Li, F. K., Nghiem, S. V., and Ricketts, W. B., 1997. Polarimetric brightness temperatures of sea surfaces measured with aircraft K- and Ka-band radiometers. IEEE Transactions on Geoscience and Remote Sensing, 35, 1177– 1185. Yueh, S. H., Tang, W., Fore, A. G., Neumann, G., Hayashi, A., Freedman, A., Chaubell, J., & Lagerloef, G. S. E., 2013. L-band passive and active microwave geophysical model functions of ocean surface winds and applications to Aquarius retrieval. IEEE Transactions on Geoscience and Remote Sensing, 51, 4619–4632. See http://ieeexplore.ieee.org/xpl/articleDetails.jsp? tp=&arnumber=6553597&url=http%3A%2F%2Fieeexplore.ieee. org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D6553597.

Cross-references Climate Data Records Microwave Surface Scattering and Emission Radar, Scatterometers Radiation, Polarization, and Coherence Severe Storms Tropospheric Winds Water and Energy Cycles

SEVERE STORMS Charles A. III Doswell Doswell Scientific Consulting, Norman, OK, USA

Synonyms Severe convection; Severe local storms; Severe weather Definition Severe storms. Any deep convective storm, usually associated with lightning and thunder, that produces one or more of the following: large hailstones, strong winds, and/or tornadoes. For the United States, the threshold diameter for severe hail is 2.54 cm (1 in.) and the threshold speed for severe winds is 25 m s1 (50 knots), but any tornado is considered severe (tornadoes are ranked in six categories according to the Enhanced Fujita scale (Fujita, 1971), ranging from EF0 to EF5). In many countries,

768

SEVERE STORMS

heavy rainfall is also considered a severe storm, with varying thresholds for rainfall per unit time (Doswell III et al. 2009).

Introduction: deep, moist convective storms Any type of weather can be called severe: extremely high or low temperatures, extended periods without rainfall, heavy snowfall, freezing drizzle or rain, and so on. However, the generally accepted use of the term “severe weather” is for the events produced by deep convective storms (Ludlam, 1963; Doswell III, 2001). Such storms are typically accompanied by lightning and thunder caused by that lightning, but in some cases, the clouds are not sufficiently electrified to produce lightning discharges while still capable of some forms of severe weather. Severe convection is the result of a condition broadly referred to as conditional instability (Schultz et al., 2000). This type of instability results from an excess of latent and sensible heat at low levels, which in turn is due to incoming solar radiation that both heats the surface and enhances evapotranspiration of water into the atmosphere in vapor form. The heated surface conducts its heat to the atmosphere immediately above the surface, which then carries that heat upward by convection. As meteorologists use the term, convection refers to the transport of atmospheric properties in the vertical by the motion of the air itself – it can be considered a form of mixing that is a response to the imbalance created when the air is heated from below. Water vapor plays a critical role in the development of severe convection by releasing latent heat – which became “latent” when the liquid water evaporated into the overlying air or was “exhaled” during transpiration by plants – during the process of condensation into cloud droplets. It is the energy associated with latent heat release that powers deep convective clouds and which then can be manifest in the various forms of severe weather. Condensation of this water vapor into cloud droplets and, ultimately, into various forms of precipitation is the result of air containing sufficient amounts of water vapor being forced upward by some external process, such as a socalled front or by flowing up the side of a mountain. Pressure in the atmosphere decreases upward owing to the effect of gravity, so ascent results in expansion of the ascending air currents (updrafts). Surrounding air is drawn into the developing updraft, producing what is known as inflow. Following the laws of thermodynamics, the air’s expansion during its ascent results in cooling. For all practical purposes, it can be assumed that rising air exchanges no heat with its surroundings during its ascent, which means the process is adiabatic. The rate of cooling with height in the rising air prior to condensation is called the dry adiabatic lapse rate, corresponding to a constant value of 9.8 C km1. During dry adiabatic ascent, the water vapor content of the rising air current remains constant, so that ultimately, the rising air reaches a level at which

the relative humidity becomes high enough for condensation to begin (near 100 % relative humidity, where the air is said to be saturated with water vapor). Condensation can begin slightly before attaining a relative humidity of 100 % owing to the presence of tiny particles called condensation nuclei, which promote condensation because of their hygroscopic properties. Once condensation begins, the released latent heat causes the ascending air to cool at a lesser rate, known as the moist adiabatic lapse rate, but this value is not a constant. Rather, the rate is smallest at low levels (around 6  C km1) and approaches the dry adiabatic rate asymptotically as the air continues to rise. Because of the release of latent heat, the ascending air may become less dense than its surroundings (when temperature decreases with height sufficiently in that environment) and so becomes positively buoyant. If the lapse rate in the environment is between the dry and moist adiabatic values, then that environment is considered conditionally unstable – it is stable for ascent without condensation, but can become unstable when the ascending air becomes saturated. Positive buoyancy results in a strong acceleration of the air’s ascent – it has become unstable and continues to rise on its own after having been lifted initially by some external mechanism. In extreme cases of this updraft instability, the vertical speed can reach values as high as 50 m s1 or more during an ascent to 10 km or beyond. As the air in the updraft continues to rise, it eventually reaches a level where it loses its buoyancy, primarily because most of the moisture has condensed and its lapse rate increases to approach the dry adiabatic value. The height of this level also depends on the vertical temperature profile in the environment. Owing to the inertia of its ascent accumulated during its buoyant acceleration, it will overshoot this equilibrium level before dropping back downward and spreading out in the upper parts of the storm, forming the so-called anvil cloud (Figure 1). The cloud material the storm produces eventually will accumulate at the equilibrium level and ultimately re-evaporate. Various processes within the developing convective cloud cause the conversion of cloud droplets, which are initially very small (perhaps only 20 mm in diameter) into precipitation-size particles (a few mm in diameter), which are large enough to have terminal velocities sufficient to let them fall out of the cloud (a few m s1). The development of precipitation within a convective cloud (Lamb, 2001) signals a transition in the evolution of a convective storm. The presence of precipitation results in the creation of descending air currents, or downdrafts through two physical processes. When precipitation falls into unsaturated environmental air, it begins to evaporate. Evaporation causes that environmental air to cool (by absorbing the latent heat of vaporization), so it can become negatively buoyant and begin to sink. Further, the physical presence of precipitation particles exerts a drag on the surrounding air, causing that air to descend

SEVERE STORMS

769

Severe Storms, Figure 1 A deep convective storm cloud (called a cumulonimbus cloud), with its anvil spreading out at the equilibrium level and being carried by the upper-level winds from left to right. The top of an overshooting updraft is visible above the anvil (compare to Figure 2b) (Photograph © 1999 C. Doswell).

along with the precipitation. Downdrafts of precipitationcooled air eventually reach the surface and spread out, forming pools of cold, stable air. The leading edge of the outflow caused by downdrafts is called a gust front, as it is usually associated with gusty surface winds. Only part of the water vapor condensed into cloud droplets ultimately falls out as precipitation. The fraction of the input water vapor deposited on the ground is known as the precipitation efficiency. Some storms are very inefficient, with only 10 % or less of the input water vapor falling out as precipitation. Others have efficiencies exceeding 50 %. Any deep convective storm is made up of one or more of what are called cells. Each cell begins as a current of ascending air. These are discrete elements in space because of a property of fluids known as mass continuity – when air in one place arises, air from its surroundings must descend to take its place. The understanding of the cellular nature of convective storms was the result of a field observational campaign in the years following World War II, known as the Thunderstorm Project (Byers and Braham, 1949). During those field experiments, thunderstorms were observed by radars (which were then a relatively new technology) and penetrated by aircraft at multiple levels. The life cycle of an individual cell was determined from these observations to be about 20–40 min, corresponding roughly to the time it takes air to ascend from near the surface to the top of the storm. This life cycle is illustrated schematically in Figure 2. Most thunderstorms

encompass multiple cells, and so can have lifetimes exceeding that of any individual cell. Convective storms have updrafts that lift warm, moist air into the upper parts of the troposphere, usually resulting in precipitation. That ascent removes the excess latent and sensible heat at low levels, depositing it in the upper troposphere. Conversely, the storm’s downdrafts take potentially cold air from middle and upper tropospheric levels and cause it to descend – which replaces warm air at low levels with that potentially cold air. Thus, updrafts and downdrafts both serve to alleviate the condition of excess heat at low levels. Once the instability that gave rise to the storm is reduced far enough by overturning within the deep convective storms, those storms will cease to exist. Convective storms can be seen as a response to a disturbance to equilibrium that results from solar heating (or from any anomalously warm underlying surface, such as warm surface currents in the oceans).

Severe convective storms To this point, the processes described are common to both severe and nonsevere deep convective storms. What is it that makes a storm severe? To some extent, because there are thresholds for defining what constitutes a severe weather event involving hail or strong winds (Galway, 1989), the definitions are rather arbitrary. Is a storm producing a 1.8 cm hailstone qualitatively different from

770

SEVERE STORMS

a

Towering Cumulus Stage

Rising “Bubble”

Cloudy “Wake”

b

c

Mature Cumulonimbus Stage

Dissipating Stage

Rising “Bubble” Weakening “Bubble”

Upper winds

Anvil

Upper winds

Anvil

Cloudy “Wake” Precipitation

Remnant Precipitation

Severe Storms, Figure 2 Schematic depiction of the life cycle of a single thunderstorm cell – (a) towering cumulus stage, (b) mature cumulonimbus stage, and (c) dissipating stage (From Doswell III (2001)).

a storm producing a 1.9 cm hailstone? Although any quantitative threshold between categories is arbitrary, as a given event passes a particular threshold, there is an increasing likelihood that important damage will result from that event. The impact of a storm also depends strongly on what it is affecting – a severe storm over open grassland is unlikely to do much damage to that grassland unless the

events become extreme. But within inhabited areas, damage with important consequences to humans could result from convective storm events even slightly below a particular threshold. Ultimately, any deep convective storm producing lightning has the potential to do damage and cause casualties, but lightning does not meet any of the existing formal criteria for a nominally severe storm.

SEVERE STORMS

771

Severe Storms, Figure 3 An example of an event with a line of severe thunderstorms as seen by radar.

Severe weather is strongly associated with the degree to which the convective storms are organized. There are two basic modes by which convective storms become organized: into more or less continuous lines of individual convective cells, or as discrete, relatively isolated systems of cells. The ultimate form of discrete organization is the so-called supercell storm, which is characterized by a systematic rotation of the entire storm system through most of its depth. As will be discussed below, the main factor governing the degree of organization is the variation of winds with height. The weakest degree of organization to a severe storm is the so-called pulse severe storm. Such an event is characterized by a very brief period (only a few minutes duration) of marginally severe weather, followed by rapid dissipation of the storm. Another pulse severe storm might develop in the vicinity, but it would be a distinctly different storm. Marginally severe hail and/or surface winds are the predominant events associated with pulse-type storms. At times, even weak updrafts can produce isolated strong downdrafts. The processes that control updraft strength are different from those controlling downdraft strength, so there is no physical reason to assume a high correlation between updraft and downdraft intensity in any given storm. Storms with weak updrafts might not even produce lightning but may result in small but powerful downdrafts called microbursts that can have

important human consequences, especially for aircraft taking off or landing (Fujita and Caracena, 1977). When multiple storms develop in relatively close proximity, it is common for their surface outflows to merge, forming large pools of relatively cold air spreading outward along. The gust front at the leading edge of these merged outflows often serves to initiate new convective cells that then contribute their outflows to the cold pools, which then initiate another series of new convective cells, and so on. Furthermore, many storms are initiated along lines of ascent (typically, frontal boundaries), which imposes a linear structure on the resulting convective system This linear organization (as in Figure 3) is very efficient at overturning large regions of unstable conditions. Strong surface winds arise when the downdrafts of such storms become intense, sometimes producing so-called bow echo morphology to the convective line (Figure 4). Severe hail is also common with such storms. The size of organized convective lines can vary from 20to 200 km long, or even larger. At times, lines of storms (often referred to as squall lines) can produce swaths of damaging winds over regions encompassing 2,000 km2 or larger – these events, meeting certain arbitrary criteria, are called derechos (Johns and Hirt, 1987). Derechos are an extreme form of convective wind event and if they happen to hit a populated area, the result can be devastating. Some derechos incorporate embedded supercell storms,

772

SEVERE STORMS

Severe Storms, Figure 4 Example of a bow echo as seen on radar, with white lines and arrows indicating the airflow. See text for discussion.

which often are responsible for the most intense winds and the largest hail. Tornadoes can occur with such linear structures, although not typically of the highest intensities unless they occur in association with embedded supercells. Another form of organization arises when the environment favors the development of new convective cells in roughly the same storm-relative position, resulting in an isolated multicellular storm (Marwitz, 1972b). Although it appears to be a single storm on radar, it really consists of a series of individual cells that are forming, maturing, and dissipating in an organized way. In effect, each cell is moving through the storm system as a whole. Such storms can produce strong winds and large hail, and occasional tornadoes. The severity of such events in a few cases can be quite high, but typically, the severe weather is of marginal to moderate intensity. However, such storms are a primary mode for the occurrence of heavy rainfall, when the storm system moves slowly while the individual cells move through the system (Figure 5). The supercell (Browning, 1964; Marwitz, 1972a; Doswell III and Burgess, 1993) is the most organized type of severe storm, and almost all supercells produce some form of severe weather. The most extreme forms of severe weather are associated with supercells (Figure 6). If hailstone diameters reach 5 cm or more, the storm that produced them is likely to be a supercell. The most violent

tornadoes are almost exclusively confined to supercell storms (Davies-Jones et al., 2001). Even supercells possess multicellular characteristics (Foote and Frank, 1983), with an internal evolution cycle governed by the rotation process embedded within them, known as the mesocyclone. Mesocyclones evolve with time as the individual convective cells move through the storm, resulting in the cyclic production of severe weather, sometimes including families of tornadoes. In a few cases, supercells persist for many hours, producing swaths of severe weather that can be 200 km long or even longer.

Ingredients for large hail The processes that result in large hail (Browning, 1977; Knight and Knight, 2001) are not entirely understood, but some conditions clearly are necessary. A strong updraft is required to hold the growing hailstone aloft long enough to attain their size – for a giant hailstone (say, 6 cm or larger), these updrafts must be about 50 m s1 or more. Supercells are unique in that they occur in environments where the winds increase rapidly with height (i.e., strong vertical wind shear). When an updraft develops in an environment with strong vertical wind shear, the updraft can interact with that shear to produce a perturbation pressure gradient force that can augment the effect of positive buoyancy. That augmentation to the updraft’s

SEVERE STORMS

773

a cell I 15

height (km)

cell II 10

10

10

cell III

30

cell IV

50

5 Cell Motion 0

X

b II

I

15 10

height (km)

III 10

30

IV 30

5

V 50

0

X II

c

I 15

height (km)

IV 10

10

III 10

V 50

30

5

0

X

Severe Storms, Figure 5 Schematic illustrating the passage of individual convective cells through an isolated multicellular storm. Individual cells, labeled with Roman numerals I–V, form on the left and move to the right. The outflow’s gust front is denoted by the black line with triangles. Precipitation intensity is indicated by the green shading. The point indicated by the circled “X” is a fixed point on the ground, which experiences repeated episodes of heavy rain coming from a sequence of cells moving through the storm system. The time required for the evolution in (a–c) is on the order of 20–30 min (Adapted from Figure 7 in Doswell et al. 1996).

acceleration can result in a contribution as large as that from buoyancy alone. Thus, supercells generally have the strongest updrafts of any deep convective storms, which enhances their potential for giant hailstones. Not all supercells produce large hail, however. This implies that there are other ingredients for large hail. One likely candidate is the presence of large amounts of supercooled water in the updraft. Liquid water condensing in clouds is typically very pure water and if carried above the level in the storm where the internal cloud temperature is 0  C, such pure

water may not freeze immediately. Very pure water’s temperature can be lowered to about 40  C, the so-called homogeneous nucleation temperature, before it freezes, unless there is a freezing nucleus present. One type of freezing nucleus is water already frozen: a frozen raindrop or a snow particle, for example. In strong updrafts with copious condensed water within them, the presence of supercooled water is likely. When a supercooled water droplet makes contact with ice, it freezes quite rapidly, so hailstones grow by a process of accreting supercooled liquid water – called wet growth. Hailstones also can grow

774

SEVERE STORMS

Supercell Storm 20 km Storm motion Forward Flank Downdraft Edge of anvil cloud

Light rain Rain and small hail Large hail

Rear Flank Downdraft

Surface inflow

Severe Storms, Figure 6 Schematic view of a supercell storm as it might appear on radar, showing the location of important features associated with such storms, with small black arrows indicating the low-level airflow. The dashed black line is the gust front associated with the forward-flank downdraft, while the dash-dot black line is that associated with rear-flank downdraft. The location of a mesocyclonic tornado is shown by the purple inverted triangle.

Severe Storms, Figure 7 A cross section of a large hailstone, showing the “growth rings” produced by alternating wet and dry growth of the hail. On the left, the view is through crossed polarizers, showing the internal crystalline grain structure. On the right, through ordinary transmitted light – it is evident that cloudy ice has smaller grain sizes (Photograph courtesy C. Knight (used with permission)).

by dry growth, which is the conversion of water vapor directly into solid form (sometimes called riming). As the growing hailstone alternates between periods of wet and dry growth, its internal structure exhibits layering

between nearly transparent ice from wet growth and translucent ice from dry growth (Figure 7). The notion of hailstones acquiring this layered structure from repeated ascent to high levels and descent to low levels in the storm

SEVERE STORMS

Heavy precipitation Moderate precipitation Light precipitation

775

Squall Line Cross Section Storm motion

Fro

Rear-to

nt-

-front In

flow Je

t

Trailing Precipitation

to-

rea

rA

sce

ndi

ng

Inf

low

Severe Storms, Figure 8 Schematic cross section of the airflow in a typical mesoscale convective system (MCS). The deep convective cells are along the leading edge of the system, with updrafts depicted in the red lines and downdrafts shown in blue, but airflow also goes in between the cells along an ascending path from front to rear (heavy black line). In some systems, a rear-to-front descending inflow jet (heavy black line) develops, entering the storm at middle levels and can result in a bow echo configuration for the line of storms (as in Figure 4). The box outlined with dotted lines and colored light yellow contains the active deep convective part of the system (cf. the red colors indicating heavy rain and possibly hail in Figures 3 and 4).

has now largely been discredited (Knight and Knight, 2001). Minor height fluctuations and movement within the storm can cause the growth mode to alternate between wet and dry without requiring cycles of substantial ascent and descent through the storm. In order to reach the ground as a large hailstone, once the hailstone grows large enough to begin its descent from the growth zone, it should not experience an extended period of melting once it falls below the melting level (0  C). It turns out that a relatively dry environment inhibits melting during descent, as well. There likely are microphysical factors that promote or inhibit the development of large hail, but these remain unknown at present.

Ingredients for strong winds As discussed earlier, two primary factors promote strong downdrafts, with their associated strong surface winds: negative buoyancy produced by evaporation (or by direct conversion of ice to vapor, known as sublimation), and precipitation loading (Wakimoto, 2001). It is possible that another contribution to strong outflow is perturbation pressure gradient forces. At present, the extent to which such pressure forces are involved is not known comprehensively. The interaction between a downdraft and the surface produces strong vertical deceleration of the updraft, which results in an outward-directed perturbation pressure gradient force that drives the outflow once the downdraft descends far enough to feel the effects of the surface. There may also be an interaction between

the downdraft and the environmental vertical wind shear, but the importance of that is yet to be shown. In supercell storms, the mesocyclone provides an augmentation of the pressure gradient force that causes a swirling airflow. This results in an increase of the magnitude not only of the downdraft’s outflow, but also the updraft’s inflow. Thus, some supercells may have narrow channels of potentially damaging winds in their inflows, as well as within their outflows. In large systems of deep convection, called mesoscale convective systems (or MCSs) (Fritsch and Forbes, 2001), a relatively narrow band of inflow can develop a few kilometers above the surface and enter the storm system from the rear (illustrated schematically in Figure 8). As precipitation falls into this rear inflow, it is cooled by evaporation and can descend to the surface where it can augment the convective storm-scale downdrafts driving the gust front. This process may explain the bowed-shaped morphology of many bow echoes. Because of the relatively narrow jet of strong winds into the rear of an MCS, strong horizontal wind shear on its flanks produces counterrotating vortices on the ends of bow echoes, called “bookend” vortices (cf. Figure 4). Also shown in Figure 8 is the inflow that passes between deep convective cells forming on the leading edge of the outflow and ascends gradually toward the rear of the storm. This carries precipitation toward the trailing side of the storm, where it falls out as moderate rainfall. Generally speaking, the cyclonic vortex (see the next section) can be augmented by the

776

SEVERE STORMS

Coriolis effect (an apparent acceleration of the airflow due to the Earth’s rotation) over the life cycle of an MCS, whereas the anticyclonic vortex is inhibited by the Coriolis effect.

Ingredients for tornadoes Not all tornadoes arise the same way, so the task of describing the factors that promote the development of tornadoes (tornadogenesis) is somewhat challenging. It has been mentioned that tornadoes occasionally form from isolated non-supercell storms (Wakimoto and Wilson, 1989), but even within supercells, it appears that tornadoes can develop in different ways. Hence, the distinction made here is between tornadoes that occur embedded within mesocyclones and those that occur in the absence of, or some distance away from, a mesocyclone (Davies-Jones et al., 2001). The former are mesocyclonic tornadoes, the latter are non-mesocyclonic tornadoes (Brady and Szoke, 1989). For discussing tornadoes, it is useful to define a quantity associated with fluid motion called vorticity. Vorticity is a vector, denoted by V, defined as the curl of the velocity field: O 
V; where the velocity vector V has components (u,v,w) in the eastward, northward, and upward vertical directions, respectively in a rectangular Cartesian coordinate system (x,y,z), with unit vectors (i,j,k). Vorticity follows a “right-hand rule” convention. If the direction of the vorticity vector is parallel to and in the same direction as the thumb on the right hand, the slightly curled finders are aligned with the sense of rotation (counterclockwise) and the vorticity’s magnitude is positive. Negative vorticity is antiparallel to the thumb of the right hand. Vorticity has three components:       @w @v @u @w @v @u    iþ jþ k  ðx;;zÞ: O¼ @y @z @z @x @x @y

The vorticity in a tornado is primarily about a vertical axis (the z-component of the vector vorticity). For a tornado, the peak value of the vertical vorticity is on the order of 1.0 s1, whereas for a mesocyclone, the peak value is of order 1.0
102 s1, two orders of magnitude less (Doswell III and Burgess, 1993). When speaking of cyclones and anticyclones, in the Northern Hemisphere, the vorticity in a cyclone is positive (counterclockwise) and negative (clockwise) in an anticyclone. In the Southern Hemisphere, the vorticity of a cyclone is negative and positive in an anticyclone. The issue of tornadogenesis concerns the mechanism(s) by which the vertical vorticity becomes so large, especially near the surface (Davies-Jones et al., 2001). At one point in the study of tornadoes, it was felt that it was simply a matter of amplification of the mesocyclone’s vorticity by conservation of angular momentum within the updraft. However, observations using Doppler radar have

made it clear that this simple idea is not so easily applied. Mesocyclonic tornadoes typically form near the interface between updraft and downdraft (Lemon and Doswell, 1979), whereas if they were associated only with simple conservation of angular momentum, it would be expected they would form near the center of the updraft. While there can be no doubt that conservation of angular momentum is involved, the specific internal storm and environmental processes that promote tornadogenesis in a few storms and not in most others remain unknown. Only around 20 % (or perhaps less) of supercells produce tornadoes (the percentage for other storm organization levels is much lower than for supercells), so the presence of mesocyclonic vorticity is not by itself sufficient to explain tornadogenesis. Further, the fact that some convective storms without mesocyclones can produce tornadoes means that other, presently unknown factors must be involved. The primary mechanism for producing supercells is reasonably well known. When an environment is characterized by strong vertical wind shear through a deep layer (6 km or more above the surface), with magnitudes of order 3
103 s1 or more, this implies ambient vorticity about a horizontal axis. The two components of this horizontal vorticity vector are Oh ¼ (x,) and for storm environments, that horizontal vorticity is dominated by the two terms involving the vertical wind shear, so that on large scales, x 

@v ; @z



@u ; @z

because the vertical component of the airflow (w) in a convective storm’s environment is at least two orders of magnitude smaller than the horizontal components. This means that the horizontal component of the environmental vector vorticity is very nearly perpendicular to the vertical wind shear vector, S, given by S  @Vh =@z, where Vh ¼ ðu; vÞ. The horizontal vorticity also can be decomposed into components along and perpendicular to the airflow, known as streamwise and crosswise vorticity, respectively. Davies-Jones et al. (2001) demonstrate that an updraft causes crosswise vorticity to be tilted into the vertical to form a pair of counterrotating vortices: one cyclonic and the other anticyclonic. When the vorticity vector is parallel to the airflow, that vorticity is either streamwise or antistreamwise. Supercell environments favor streamwise vorticity through a deep layer, which is what we generally observe, although in some circumstances, supercells can occur with a large component of crosswise vorticity, as well. In such instances, a common observation is the production of counterrotating supercells, the cyclonic one moving to the right of and slower than the mean wind through a deep layer, and the anticyclonic one moving to left of and faster than the mean wind (Figure 9), in the Northern Hemisphere. As usual, things work the opposite way in the Southern

SEVERE STORMS

777

Severe Storms, Figure 9 Tracks of left- and right-moving supercells, as shown by the accumulated rainfall amounts over the life of the echoes. The fast-moving left movers do not accumulate much rainfall, but the relatively slow-moving right movers produce heavier rainfall amounts and so show brighter colors along their tracks.

Hemisphere – the cyclonic storm moves to the left and slower than the mean wind, whereas the anticyclonic supercell moves to the right and faster than the mean wind. Although the physical understanding needed to define the specific ingredients for tornadogenesis is lacking, empirical results suggest two factors that seem to be important (Doswell and Evans, 2003; Thompson et al., 2003). The first is the presence of strong vertical wind shear in the lowest kilometer of the atmosphere – vertical wind shear values approaching 1.0
102 s1 seem be common in tornadic storm environments, be they supercells or not. The second is high relative humidity near the surface, which seems to be related to yet another empirical finding about tornadic supercells (Figure 10): Their downdrafts tend to be relatively warm and may even retain some buoyancy if lifted far enough. A cold, very stable outflow from a storm reduces the likelihood of tornadoes and such outflows are promoted by low relative humidity in the near-surface air. Again, there is as yet no comprehensive physical explanation for these empirical results. Much less is known about tornadoes produced by nonsupercell storms. It appears that in at least some cases, the existence of preexisting maxima of vertical vorticity on a scale larger than that of the convective storm (called misocyclones) means that it may be simple

amplification of that vorticity by developing updrafts through conservation of angular momentum which can explain this type of tornado (Brady and Szoke, 1989). Non-mesocyclonic tornadoes have not been studied very extensively, so much continues to be unexplained. It should be noted that waterspouts are simply tornadoes over the water, and some of them also seem to be associated with supercell storms and others with nonsupercellular deep convection. There are areas around the world that produce frequent waterspouts from primarily non-supercellular storms. Many of these waterspouts arise in deep convection that has very little or no lightning activity, at least prior to waterspout development. An important scientific question relating to tornadoes is their role in the atmosphere. Most meteorological processes have a well-defined purpose; for example, deep, moist convection is a response to the excessive sensible and latent heat at low levels and convective storms result in alleviating the instability that gives them birth. Tornadoes almost certainly have some role to play, because the alternative would be that they are just random events associated with deep convective storms, which seems unlikely. Our inability to predict which storms will produce tornadoes is evidently related to this currently unanswered question. The atmosphere is an example of

778

SEVERE STORMS

Severe Storms, Figure 10 A tornadic supercell, with annotations, showing the location of the rear-flank (white dash-dot line) and forward-flank (white dashed line) gust fronts (cf. Figure 6) at cloud base level. This example is looking roughly toward the northeast at a storm like that illustrated in Figure 6. The rear-flank downdraft is creating the turquoise-colored “hole” in the clouds seen above and toward the left of the tornado (Original photograph © 2005 C. Doswell, annotated version © 2008 C. Doswell).

a nonlinear dynamical system (Lorenz, 1993) and it is known that such systems can be difficult to predict even when the underlying dynamical processes are understood perfectly. Gaps in that understanding make it even more challenging to anticipate such events.

Convective storm observations Because convective storms are small relative to the sparse conventional in situ observing systems, a great deal of what we know about them has been derived from remote sensing – primarily radar. The development of Doppler radar for probing convective storms began in the late 1960s and has become the basis for our existing operational severe weather warning system. Recently, the implementation of polarimetric capability for the operational Doppler radars has begun, which will give additional information for observing the microphysical structure within convective storms. Another important observing capability has been spacecraft with cameras and other active and passive remote sensors. The distribution of deep convective storms in space and time via satellite observations has been another important component in scientific understanding. Further, it is possible to obtain critical information about the distribution of environment variables relevant to deep convection using a number of different spacecraft-borne observing systems, including lightning detection capability. In the past two decades, both spacecraft and ground-based remote sensing tools have

been developed that offer considerable promise to revolutionize our ability to sample the global environment associated with deep convection.

Summary The definition of severe weather as used here is based on arbitrary criteria for hail and strong winds. The storms that produce large hail are clearly those that include strong updrafts, and since supercells develop the strongest vertical motions of any convective storm, the largest hail is produced primarily by supercells. Not all supercells result in large hail; however, predicting the occurrence of large hail is still somewhat problematic. It is likely that the detailed microphysical processes within convective storms are important for the production of hail but much remains to be learned about microphysical processes in deep convective storms. Polarimetric radars, which are just now on the verge of becoming operational, may eventually help with understanding hail production. However, a challenge for using polarimetric radars (and other remote sensing tools) to infer microphysical properties in the atmosphere is that in situ validation of the interpretation of the observations in terms of the microphysics is needed. Are the remote sensing observations being interpreted properly? This continues to be a vexing issue. Strong winds are arguably the easiest type of severe weather event to understand. They are almost entirely associated with strong downdrafts, the physics of which

SEVERE STORMS

is reasonably well understood. It is not always possible to know in detail what a particular storm is likely to do, but the general conditions for producing a strong surface wind are relatively easy to anticipate. One challenge is to know when Doppler radar–observed strong winds are actually making their way to the surface. When a shallow stable layer is present at the surface, the strong winds just above may or may not be observed at levels near the surface where they can do damage. Radars have a “horizon” problem associated with the curvature of the Earth –at its lowest scanning angle, it observes events only above a height that increases with distance. Hence, Doppler radars generally cannot confirm or deny that strong winds observed above the surface are actually reaching down far enough to do damage. At this time, it is not clear just what role tornadoes are playing that other physical processes apparently cannot accomplish. Whatever that purpose might be, the need for it is relatively infrequent – tornadoes are rare events in a global sense. The challenge posed by tornadogenesis is connected to the reason for their existence in the first place – if their role in the atmosphere can be understood definitively, the mechanism(s) by which they are created should become more clear than at present. A major concern for hypothesized tornadogenesis mechanisms is that they need to be able to account for non-mesocyclonic tornadoes, as well. Most of our observations of tornadoes are for mesocyclonic events associated with supercells. Research in the near future is likely to remain concentrated on trying to explain mesocyclonic tornadoes, seeking to understand the physical processes that explain the empirical associations with low-level vertical wind shear and high relative humidities near the surface, but the broader topic is likely to remain problematic for some time, as non-mesocyclonic tornadoes (including waterspouts) seem much harder to predict. New technologies for observing the airflow and thermodynamic characteristics in the lowest kilometer of the atmosphere will need to be developed if much progress is to be made. Another remaining challenge is the development of an accurate understanding of the true distribution of severe weather (Brooks et al., 2003; Doswell et al., 2005). Even in the United States, which has the best infrastructure for documenting severe weather in the world, hail and strong winds are reported as point events, whereas the reality is that both winds and hail occur in swaths. Thus, the reports of severe weather at points are likely to be providing only a biased depiction of the frequency and intensity of severe weather. Such reports depend on having a knowledgeable observer at the right place and time to observe the most intense events, an unlikely scenario. Remote sensing systems (mostly radar) do not make direct observations of severe weather events but may offer the best hope for an accurate spatiotemporal distribution of severe storms. Tornadoes are assigned path lengths and widths, but at the moment, the occurrence of tornadoes (and other forms of severe weather) in sparsely populated regions often goes unreported. Further, there simply are not enough

779

available resources to devote to a careful study of each reported tornado event and the intensity determination is based almost entirely on the degree and type of damage done. When a tornado occurs in sparsely populations regions of grassland, with little or nothing to damage, no way exists at present to determine the tornado’s wind speeds. If reported at all, such tornadoes typically are given a “default” rating as weak events, which can be substantially wrong. Moreover, even when careful damage surveys are done, there is no simple relationship between wind speed and damage. Doing a damage survey to estimate wind speeds requires considerable experience and knowledge, and this is not generally available. Until some way to estimate tornado intensity objectively and remotely is developed, comparable to the Richter scale in estimating earthquake intensity, our knowledge of the distribution and intensity of tornadoes will remain uncertain.

Bibliography Brady, R. H., and Szoke, E. J., 1989. A case study of nonmesocyclone tornado development in Northeast Colorado: similarities to waterspout formation. Monthly Weather Review, 117, 843–856. Brooks, H. E., Doswell, C. A., III, and Kay, M. P., 2003. Climatological estimates of local daily tornado probability for the United States. Weather and Forecasting, 18, 626–640. Browning, K. A., 1964. Airflow and precipitation trajectories within severe local storms which travel to the right of the winds. Journal of the Atmospheric Sciences, 21, 634–639. Browning, K. A., 1977. The structure and mechanism of Hailstorms. In Knight, C. A. (ed.), Hail: A Review of Hail Science and Hail Suppression. Boston: American Meteorological Society. Meteorological Monographs, Vol. 38, pp. 1–39. Byers, H. R., and Braham, R. R., Jr., 1949. The Thunderstorm. Washington, DC: U.S. Government Printing Office, p. 287. [out of print]. Davies-Jones, R., Trapp, R. J., and Bluestein, H. B., 2001. Tornadoes and tornadic storms. In Doswell, C. A. (ed.), Severe Convective Storms. Boston: American Meteorological Society. Meteorological Monographs, Vol. 28, No. 50, pp. 167–221. Doswell, C.A. III, H.E. Brooks, and R.A. Maddox, 1996: Flash flood forecasting: An ingredients-based methodology. Wea, Forecasting, 11, 560-580. Doswell, C. A. III, 2001. Severe convective storms – an overview. In Doswell, C. A. III (ed.), Severe Convective Storms. Boston: American Meteorological Society. Meteorological Monographs, Vol. 28, No. 50, pp. 1–26. Doswell, C. A. III, and Burgess, D. W., 1993. Tornadoes and tornadic storms: a review of conceptual models. In Church, C., et al. (eds.), The Tornado: Its structure, Dynamics, Prediction, and Hazards. Washington DC: American Geophysical Union. Geophysical Monographs, No. 79, pp. 161–172. Doswell, C. A., III, and Evans, J. S., 2003. Proximity sounding analysis for derechos and supercells: an assessment of similarities and differences. Atmospheric Research, 67–68, 117–133. Doswell, C. A., III, Brooks, H. E., and Kay, M. P., 2005. Climatological estimates of daily nontornadic severe thunderstorm probability for the United States. Weather and Forecasting, 20, 577–595. Doswell, C.A. III, Brooks, H.E., and Dotzek, N., 2009. On the implementation of the enhanced Fujita scale in the USA. Atmos. Res., 93, 554–563.

780

SNOWFALL

Foote, G. B., and Frank, H. W., 1983. Case study of a hailstorm in Colorado. Part III: airflow from triple-Doppler measurements. Journal of the Atmospheric Sciences, 40, 686–707. Fritsch, J. M., and Forbes, G. S., 2001. Mesoscale convective systems. In Doswell, C. A. III (ed.), Severe Convective Storms. Boston: American Meteorological Society. Meteorological Monographs, Vol. 28, No. 50, pp. 323–357. Fujita, T. T., 1971. Proposed characterization of tornadoes and hurricanes by area and intensity. Satellite and Mesometeorology Research Project Research Paper No. 91. Chicago, IL: University of Chicago, p. 42. Fujita, T. T., and Caracena, F., 1977. An analysis of three weatherrelated aircraft accidents. Bulletin of the American Meteorological Society, 58, 1164–1181. Galway, J. G., 1989. The evolution of severe thunderstorm criteria within the weather service. Weather and Forecasting, 4, 585–592. Johns, R. H., and Hirt, W. D., 1987. Derechos: widespread convectively induced windstorms. Weather and Forecasting, 2, 32–49. Knight, C. A., and Knight, N. C., 2001. Hailstorms. In Doswell, C. A. III (ed.), Severe Convective Storms. Boston: American Meteorological Society. Meteorological Monographs, Vol. 28, No. 50, pp. 223–254. Lamb, D., 2001. Rain production in convective storms. In Doswell, C. A. III (ed.), Severe Convective Storms. Boston: American Meteorological Society. Meteorological Monographs, Vol. 28, No. 50, pp. 299–321. Lemon, L. R., and Doswell, C. A., III, 1979. Severe thunderstorm evolution and mesocyclone structure as related to tornadogenesis. Monthly Weather Review, 107, 1184–1197. Lorenz, E. N., 1993. The Essence of Chaos. Seattle: University of Washington, p. 227. Ludlam, F. H., 1963. Severe local storms: a review. In Atlas, D. (ed), Severe Local Storms. Boston: American Meteorological Society. Meteorological Monographs, Vol. 5, No. 27, pp. 1–30. Marwitz, J. D., 1972a. The structure and motion of severe hailstorms. Part I: supercell storms. Journal of Applied Meteorology, 11, 166–179. Marwitz, J. D., 1972b. The structure and motion of severe hailstorms. Part II: multi-cell storms. Journal of Applied Meteorology, 11, 180–188. Newton, C. W., 1963. Dynamics of severe convective storms. In Atlas, D. (ed.), Severe Local Storms. Boston: American Meteorological Society. Meteorological Monographs, Vol. 5, No. 27, pp. 33–58. Schultz, D. M., Schumacher, P. N., and Doswell, C. A., III, 2000. The intricacies of instabilites. Monthly Weather Review, 128, 4143–4148. Thompson, R. L., Edwards, R., Hart, J. A., Elmore, K. L., and Markowski, P., 2003. Close proximity soundings within supercell environments obtained from the rapid update cycle. Weather and Forecasting, 18, 1243–1261. Wakimoto, R., 2001. Convectively driven high wind events. In Doswell, C. A. III (ed), Severe Convective Storms. Boston: American Meteorological Society. Meteorological Monographs, Vol. 28, No. 50, pp. 255–298. Wakimoto, R. M., and Wilson, J. W., 1989. Non-supercell tornadoes. Monthly Weather Review, 117, 1113–1140.

Cross-references Cloud Liquid Water Earth Radiation Budget, Top-of-Atmosphere Radiation Lightning Observational Systems, Satellite Ocean-Atmosphere Water Flux and Evaporation

Radiation, Electromagnetic Rainfall Water Vapor Weather Prediction

SNOWFALL Ralf Bennartz Atmospheric and Oceanic Sciences Department, University of Wisconsin-Madison, Madison, WI, USA

Definition Snow. Low-density ice particles. The density is typically in the order of 0.1 g/cm3. Individual snowflakes can exhibit a wide variety of different forms. Other frozen particles include graupel or hail. Graupel particles are medium-density ice particles with a density of about 0.4 g/m2. Graupel is produced when ice particles fall through extensive layers of supercooled cloud liquid water. Hail particles are of a very high density of about 0.9 g/m2. Hail fall is associated with intensive convection and is formed via a series of melting and refreezing processes. Melting layer. The melting layer is the layer in which falling hydrometeors transition from the ice into the liquid phase. The top of the melting layer coincides with the 0 C isothermal. Bright band. The bright band is a band of enhanced radar reflectivity associated with melting precipitation particles. It is typically found in stratiform precipitation. The bright band roughly coincides with the melting layer. Introduction Frozen precipitation exists in a huge variety of different forms depending on weather situation as well as on cloud microphysical processes. Precipitation-sized ice particles can be broadly categorized into three classes: snow, hail, and graupel. All deep convective and frontal precipitation events involve the ice phase even if ice particles will melt before reaching the ground. Snowfall at the surface constitutes a large fraction of polar, boreal, and mid-latitude winter precipitation. In the northern hemisphere, winter snowfall is ubiquitous north of roughly 40  N to 50  N. In most mid- and high-latitude countries, snowstorms can have a major disrupting effect on traffic and the economy. The atmospheric temperature profile of the atmospheres plays a key role in determining whether liquid, frozen, or mixed precipitation reaches the surface. In this regard, the problem of snowfall remote sensing can be separated into three different issues. Firstly, precipitation has to be detected. Secondly, the phase or type of precipitation at the surface has to be determined. Thirdly, the amount of precipitation has to be quantified. This entry is only concerned with snowfall at the surface and concentrates

SNOWFALL

781

Snow frequency (%)

30 Zonally averaged snow frequency

25 20 15 10 5 0

Snowfall rate (mm/day)

1.4 Zonally averaged mean snowfall rate

1.2

Upper-Bound Average Z-S Relationship Lower-Bound

1.0 0.8 0.6 0.4 0.2 0.0

−75

−60

−45

−30

−15

0

15

30

45

60

75

Latitude

Snowfall, Figure 1 Zonally averaged snowfall frequency and mean snowfall rate derived from 1 year of CloudSat radar observations. The uncertainty in the snowfall rate (gray-shaded area) is solely due to assumptions about ice habit (Figure from Hiley et al., 2010).

on the second and third issue. While there are various ground-based remote sensing techniques for snowfall, the main focus here is on global satellite-derived estimates. Subsequently, a distinction will be made between active (radar) and passive (microwave radiometer) remote sensing techniques.

Principles of snowfall remote sensing Frozen versus liquid precipitation The discrimination of frozen from liquid precipitation at the ground is clearly crucial, especially in high- and midlatitude areas where snowfall occurrence varies widely on annual and interannual time scales. From a passive remote sensing standpoint, the determination of the phase of falling precipitation at the ground is an ill-posed problem, since passive measurements typically have broad vertical weighting functions. For instance, if the precipitation melts just a few meters above the surface, it will not have any measurable “liquid” signature that could be detected using a passive sensor. Passive instruments are therefore not directly sensitive to the phase of falling precipitation at any given level. Indirectly, the phase of precipitation at the surface can be determined with some accuracy by assessing the height of the freezing level. If the freezing level is known, the only additional information needed to assess precipitation phase at the surface is the distance it takes for the falling precipitation to melt. A reasonable value for this falling distance is in the order of

500–600 m for snowfall (e.g., Bennartz, 2007). As an example, the upper panel in Figure 1 shows the zonally averaged snowfall frequency based on temperature fields obtained from a numerical weather prediction model and precipitation identification using CloudSat’s 94 GHz Cloud Profiling Radar (Stephens et al., 2008). Some active remote sensing techniques (radar) provide additional means to distinguish frozen from liquid precipitation. The occurrence of a bright band in radar observations is a sign for melting particles (Fabry and Zawadzki, 1995). Thus, if a bright band is identified sufficiently high above the surface, precipitation at the surface can be assumed to be liquid. Unfortunately, the reverse is not true, i.e., even if a bright band is not identified, precipitation at the surface can still be liquid. Other active techniques involve polarimetric measures (Matrosov et al., 2007; Zrnic et al., 1993). While these methods are promising for ground-based observations, Matrosov et al. (2009) point out that polarimetric observations are often not available for operational weather radars. Also spaceborne cloud and precipitation radars do not currently offer polarimetric capabilities.

Snowfall intensity Ground-based active remote sensing techniques have been widely used for many years to measure snowfall intensity. First measurements of snow size distributions together with relationships between radar reflectivity (Z) and

782

SNOWFALL

snowfall rate (S), so-called Z–S relations, were first reported by Gunn and Marshall (1958). In 1970, a widely used size distribution and related Z–S relation was published by Sekhon and Srivasta (1970). Since then, much progress has been made in understanding the relation between radar backscatter and snowfall, as well as its inherent uncertainties. The variability in Z–S relations is largely driven by ice particle habit (shape) and size distribution – both highly variable in space and time. Various recent publications have theoretically assessed single particle and size distribution, averaged scattering properties, and their variability (Hong, 2007; Kim, 2006; Liu, 2004; Matrosov, 2007; Petty and Huang, 2010). With the advent of CloudSat (Stephens et al., 2008) and its 94 GHz Cloud Profiling Radar (CPR), first global remote sensing observations of snowfall have become possible. Based on the theoretical work referenced above, Liu (2008), Kulie and Bennartz (2009), and Hiley et al. (2010) have studied snowfall occurrence and intensity using CloudSat’s CPR. Exemplarily, zonally averaged results are shown in the lower panel of Figure 1. The uncertainty range given is solely due to uncertainties in choice of ice scattering models. Spaceborne passive remote sensing of snow intensity or snowfall rate is hampered not only by the aforementioned uncertainties in ice scattering but also by relatively broad weighting functions and the strong and highly variable contribution of surface emission over heterogeneous, potentially snow- or ice-covered surfaces. Despite these issues in the last few years, some progress has been made regarding passive microwave estimates of snowfall at the ground (Kim et al., 2008; Noh et al., 2006, 2009; Skofronick-Jackson et al., 2004). While these studies show promising results, it is unclear at the point of writing to what extent the various issues related mostly to ice particle scattering and surface emissivity can be disentangled. More studies are needed to establish bounds on the uncertainty of passive remote sensing techniques for precipitation. A potentially interesting research venue might also lie in the combination of active and passive observations.

Summary Over the last years, our capabilities to remotely sense snowfall have significantly increased. Significant progress has been made in three areas. Firstly, new in situ and ground-based remote sensing techniques allow us to better specify snowfall size distribution, fall speed, and particle shapes. Secondly, our understanding of basic interactions between nonspherical frozen particles and the radiation field has advanced considerably. Thirdly, with the advent of spaceborne radars and improved radiometers, it has become feasible to study snowfall globally from space. While initial studies show very promising results, many of the related retrieval techniques are still in their infancy. In the future, significant effort will have to be put into characterizing retrieval accuracies and uncertainties.

Bibliography Bennartz, R., 2007. Passive microwave remote sensing of precipitation at high latitudes. In Levizzian, V., Levizziani, V., Turk, J., and Bauer, P. (eds.), Measuring Precipitation from Space – EURAINSAT and the Future. Dordrecht: Springer, pp. 165–178. 745 p. ISBN ISBN 978-1-4020-5834-9. Fabry, F., and Zawadzki, T., 1995. Long-term radar observations of the melting layer of precipitation and their interpretation. Journal of the Atmospheric Sciences, 52, 838–851. Gunn, K. L. S., and Marshall, J. S., 1958. The distribution with size of aggregate snowflakes. Journal of Meteorology, 15, 452–461. Hiley, M., Kulie, M., and Bennartz, R., 2010. Uncertainty analysis for CloudSat snowfall retrievals. Journal of Applied Meteorology and Climatology, 50(2), 399–418. Hong, G., 2007. Radar backscattering properties of nonspherical ice crystals at 94 GHz. Journal of Geophysical Research-Atmospheres, 112, D22203, doi:10.1029/2007jd008839. Kim, M. J., 2006. Single scattering parameters of randomly oriented snow particles at microwave frequencies. Journal of Geophysical Research-Atmospheres, 111, D14201, doi:10.1029/ 2005jd006892. Kim, M. J., Weinman, J. A., Olson, W. S., Chang, D. E., SkofronickJackson, G., and Wang, J. R., 2008. A physical model to estimate snowfall over land using AMSU-B observations. Journal of Geophysical Research-Atmospheres, 113, D09201, doi:10.1029/2007jd008589. Kulie, M. S., and Bennartz, R., 2009. Utilizing spaceborne radars to retrieve dry snowfall. Journal of Applied Meteorology and Climatology, 48, 2564–2580, doi:10.1175/2009jamc2193.1. Liu, G. S., 2004. Approximation of single scattering properties of ice and snow particles for high microwave frequencies. Journal of the Atmospheric Sciences, 61, 2441–2456. Liu, G. S., 2008. Deriving snow cloud characteristics from CloudSat observations. Journal of Geophysical Research-Atmospheres, 113, D00A09, doi:10.1029/2007JD009766. Matrosov, S. Y., 2007. Modeling backscatter properties of snowfall at millimeter wavelengths. Journal of the Atmospheric Sciences, 64, 1727–1736, doi:10.1175/jas3904.1. Matrosov, S. Y., Clark, K. A., and Kingsmill, D. E., 2007. A polarimetric radar approach to identify rain, melting-layer, and snow regions for applying corrections to vertical profiles of reflectivity. Journal of Applied Meteorology and Climatology, 46, 154–166. Matrosov, S. Y., Campbell, C., Kingsmill, D., and Sukovich, E., 2009. Assessing snowfall rates from X-band radar reflectivity measurements. Journal of Atmospheric and Oceanic Technology, 26, 2324–2339, doi:10.1175/2009jtecha1238.1. Noh, Y. J., Liu, G. S., Seo, E. K., Wang, J. R., and Aonashi, K., 2006. Development of a snowfall retrieval algorithm at high microwave frequencies. Journal of Geophysical ResearchAtmospheres, 111, D22216, doi:10.1029/2005jd006826. Noh, Y. J., Liu, G. S., Jones, A. S., and Haar, T. H. V., 2009. Toward snowfall retrieval over land by combining satellite and in situ measurements. Journal of Geophysical Research-Atmospheres, 114, D24205, doi:10.1029/2009jd012307. Petty, G. W., and Huang, W., 2010. Microwave backscatter and extinction by soft ice spheres and complex snow aggregates. Journal of the Atmospheric Sciences, 67, 769–787, doi:10.1175/2009jas3146.1. Sekhon, R. S., and Srivasta, R. C., 1970. Snow size spectra and radar reflectivity. Journal of the Atmospheric Sciences, 27, 299–307. Skofronick-Jackson, G. M., Kim, M. J., Weinman, J. A., and Chang, D. E., 2004. A physical model to determine snowfall over land by microwave radiometry. IEEE Transactions on Geoscience and Remote Sensing, 42, 1047–1058.

SOIL MOISTURE

Stephens, G. L., Vane, D. G., Tanelli, S., Im, E., Durden, S., Rokey, M., Reinke, D., Partain, P., Mace, G. G., Austin, R., L’Ecuyer, T., Haynes, J., Lebsock, M., Suzuki, K., Waliser, D., Wu, D., Kay, J., Gettelman, A., Wang, Z., and Marchand, R., 2008. CloudSat mission: performance and early science after the first year of operation. Journal of Geophysical Research-Atmospheres, 113, D00A18, doi:10.1029/2008jd009982. Zrnic, D. S., Balakrishnan, N., Ziegler, C. L., Bringi, V. N., Aydin, K., and Matejka, T., 1993. Polarimetric signatures in the stratiform region of a mesoscale convective system. Journal of Applied Meteorology, 32, 678–693.

Cross-references Microwave Radiometers Radars Radiation, Volume Scattering Rainfall Weather Prediction

SOIL MOISTURE Yann Kerr CNES/CESBIO, Toulouse, France

Synonyms Water content; Wetness Definition: what is soil moisture? Soil moisture usually refers to the amount of water stored in the soil. Any soil can absorb a given amount of water before being saturated. It is common knowledge that different types of soils will have different types of behavior. The best example is probably the difference between a sandy beach and a clay patch in their dry and wet states. This is mainly due to the size of the soil’s particles and the way water can lodge itself between them. Actually, water locates itself between the particles of soil, either very close and with a strong link (bound water) or more loosely (free water). Typically, the free water will “move” more easily (percolation, lateral flow, or evaporation) in the ground than the bound water which is more difficult to extract. The two types of status for water in soils will have an influence on the behavior of soils at microwave frequencies. So different soils will have different “water-holding capacities” and will evolve, depending on the outside environment, between dry and wet. The most basic way of qualifying soil moisture is to range either between dry and wet or with respect to its impact on vegetation. The terms used then are the wilting point (when vegetation starts to suffer, meaning water is difficult to extract: there is only bound water left) and the field capacity (saturation, i.e., when water is so abundant that the soil cannot hold any more and water either accumulates over the surface (ponding) or runs away (runoff)).

783

In all of the above, soil moisture refers implicitly to the near-surface soil moisture. Actually, depending on the use of such information, soil moisture may refer to different quantities. The most usual distinction is made between surface soil moisture and root zone soil moisture. Surface soil moisture corresponds to soil moisture in the first centimeters of the soil. It is the soil moisture that is most visible when walking around, which interacts directly with the atmosphere (evaporation) and which drives infiltration and hence runoff in case of rain events. However, most plants have roots near the surface but also in deeper layers (depending on soil depth), so vegetation growth and health is directly linked to the water available in the root zone. The root zone is very close to what is referred to in hydrology as the vadose or unsaturated zone. Finally, there might be another layer of stored water, deeper, the saturated zone, or water table. This layer is used by the deepest roots of trees and by man-made wells. Just to be exhaustive one must remember that, when dealing with mass (water) transfers between the atmosphere and the soil, there are other areas where water is stored and that have an influence:

Water stored in the vegetation, which is pumped from

the soil and can be evaporated into the atmosphere through respiration/transpiration

Water stored above the surface (lakes, rivers, ponds, snow/ice), which can be evaporated (sublimated) or can percolate or even runoff

Water intercepted by vegetation (during rain events or as dew), which may also evaporate, be absorbed by the leaves, or eventually fall to the ground

Introduction: why measure soil moisture? Weather forecast and extreme event prediction Water is one of the key elements that sustain life on Earth. It is used by most of the fauna and flora and most living bodies that are mainly made of water. Consequently, humankind has always been interested in water availability. For eons, life has been organized by the cycles of vegetation and the corresponding rain cycles, with their direct impact on crop yield: predicting whether the next season would be rather wet or dry has been a constant preoccupation. In parallel to the development of techniques and the progress of science, humankind has tried to improve both its measuring skills and its forecasting capacities. If the basic concern about the weather/ water availability for the next season is still a daunting challenge, our knowledge of the factors influencing the mass and energy exchanges between the surface and the atmosphere has made some progress. Models now can simulate these exchanges, taking into account the forcings (wind, solar radiation, rain, etc.) and the state of the surface (soil moisture, vegetation type and state, local slope, roughness, etc.). Thanks to these models and observations, we have now some insights into the various factors that are crucial to

784

SOIL MOISTURE

improve weather forecasts and extreme event prediction. Among them, soil moisture plays an important role as:

A water reservoir (water storage)

The source for water which can be evaporated into the

atmosphere (mass transfer)

A tracer of water which fell onto the surface (rain)

A factor influencing energy budget at the surface

atmosphere interface (evaporation requires energy hence induces a decrease in temperature)

To summarize, knowledge of surface soil moisture is of paramount importance for weather forecast. Extreme events can also be better predicted when accurate values of soil moisture are used in models. An added advantage of knowing soil moisture is that large rain events (for instance, storms), depending on the soil characteristics, can lead to extreme runoff, landslides, or flooding as water infiltration is also governed by the water content of the uppermost layer.

Water resources management Another important use of soil moisture is to get access to important information on water availability. The most obvious example is to know whether a field should be irrigated or not depending on its state, the crop growth stage and its water requirements, and the forecasted weather. This is crucial in arid or semiarid areas where irrigation is very often required, but water is not necessarily available in adequate quantities. How to measure soil moisture Soil moisture is traditionally measured by taking a sample of soil (in a vessel of normalized size), weighing it, drying the sample in an oven to evaporate all the water, and weighing it again. The ratio of the two measurements gives the gravimetric soil moisture. A more commonly used unit is the volumetric soil moisture which corresponds to the ratio of the sampled volume versus the volume of water it contains. The volumetric soil moisture can be inferred from the gravimetric soil moisture by simply multiplying it by the bulk density of the soil. Volumetric soil moisture is favored by many as it relates directly to remote measurements (see microwave part below). To achieve the goals mentioned in paragraph two, it is necessary to have access to soil moisture estimates. Punctually in space and time, this is relatively easy with gravimetric sampling. However, to have measurements representative of a larger area (such as a field), the procedure is already somewhat complex as it involves a dedicated sampling strategy. Moreover, as these measurements are time consuming, regional and, even more so, global coverage are out of question. Provided one uses automatic probes (resistive, capacitive, time domain reflectometry, etc.), it is possible to achieve larger coverage and continuous measurements, but these approaches can only be confined to well-equipped and

manned sites, as they require care and maintenance. Finally, these systems carry their own problems and inaccuracies. So global monitoring of soil moisture can only rely on remote sensing from space approaches. From space, we have access to a global approach. (All the points are measured with the same set of tool and technique, independently of countries, of ease of access, of units, or even of locally available technology.) More to the point, the measurements are by nature area integrated and thus more representative, while ground measurements are by essence very local (gravimetric samples taken a few meters apart may lead to different measurements). Conversely, if ground measurements can be very direct and accurate, measurements from space are bound to be indirect and therefore imply caveats.

Remotely sensed soil moisture, the main approaches A large number of remote sensing approaches have been tested. For surface soil moisture, the first ones were based on shortwave measurements and on the fact that soils get darker when wet. Obviously, due to atmospheric effects and potential cloud cover, as well as vegetation cover masking effects, this approach is bound to fail in most cases. A more promising feature is linked to latent heat effects. Wet soils have a higher thermal inertia and are “cooler” than dry soils. These properties led to various trials (thermal inertia monitoring, rate of heating in the morning, surface temperature amplitude, etc.) to assess soil moisture indirectly. All those approaches proved to be somewhat disappointing due to factors inherent to optical remote sensing (atmospheric effects, cloud masking, and vegetation cover opacity) as well as to the fact that thermal infrared: (1) probes the very skin of soils and (2) the layer probed in thermal infrared is dominated by the exchanges with the atmosphere. Consequently, to infer soil moisture from such measurements, one needs to know exactly the forcings (wind, for instance, will change drastically the apparent temperature of a wet soil). As microwave systems measure the dielectric constant of soils, which is directly related to the water content, research has quickly focused its efforts on assessing soil moisture with radars, scatterometers, or radiometers. These systems offer, when operated at low frequency, the added advantage of being all weather (measurements are not much affected by the atmosphere and clouds) and able to penetrate vegetation. They can operate at night. Finally, in an attempt to be exhaustive, a new approach relies on measurements of the gravity field from space. As the gravity is linked to the mass, one may consider that changes in mass are mainly linked to changes in the total amount of water. In total, one should understand here: water table; water in soil layers, possibly lakes, rivers, snow, and ice; and water in vegetation and in the atmosphere. Gravimetry should thus indicate changes in the total column of water with a spatial resolution of 500 km or more. The first results certainly show a signal but its relationship with water storage has yet to be

SOIL MOISTURE

validated and explicated. The main problem with such measurements is that they require a very large number of corrections which can be very sophisticated (for instance, taking into account the influence of tides or of the postglacial rebound). These corrections being prone to degrade the error budget in a case where the errors and corrections are of the same magnitude as the signal to be measured.

Microwaves as a tool for soil moisture monitoring: current status The most popular approach relies on the use of synthetic aperture radars (SAR). These systems, in use since 1977 with SEASAT, offer all weather measurements with a fine spatial resolution (tens of meters). Operationally, they however suffer – as most high-resolution systems – from a rather low temporal sampling (for instance, 35 days for the European Remote Sensing (ERS) satellite) which is not really compatible with hydrologic requirements or weather forecast models. But the most adverse characteristic of SAR is the coherent nature of the signal itself and the interactions with the scattering medium. SAR images are affected by speckle and by the scattering at the surface. The scattering can be due to the vegetation cover (distribution of water in the canopy) or the soil’s surface (surface scattering when wet and volume scattering). The direct consequence of these perturbations is a signal at least as sensitive to surface roughness as to moisture itself (see also Wigneron et al., 1999), not mentioning vegetation. Obviously, these effects are frequency dependent. All these inherent difficulties might explain why no real soil absolute moisture mapping was done by the several SAR that have flown since 1977. To avoid the roughness and vegetation perturbations, an approach relying on change detection, hence relative, has been used with relative success (Moran et al., 2002). However, temporal coverage is still often an issue. The use of scatterometers offers an interesting trade-off. The spatial resolution is much coarser (tens of kilometers) but with a much wider swath allowing reasonably frequent coverage (every 4–6 days on average). It also offers the added advantage of being less subject to speckle (averaging). Consequently, several authors routinely produce soil wetness maps of many areas of the world with scatterometers. The effect of vegetation is however still significant and actually corresponds to most of the signal, as the currently available frequencies are C-band on (ERS-1) and higher. So the most interesting results were obtained over arid and semiarid regions, where vegetation and soil moisture are very highly correlated anyway. The influence of surface roughness is also significant and it is best dealt with by using change detection. The last possibility in the microwave domain is to use radiometers. The technique is old and well mastered as many sensors, notably sounders, rely on passive microwaves. To infer soil moisture, these systems are bound to offer the best compromise if used at low frequency, as

785

demonstrated in the early 1970s with the very short SKYLAB mission. However, to be efficient, one needs to work in a protected frequency band to avoid unwanted man-made emissions and radio frequency interferences (RFIs) and to be sensitive to soil moisture while atmosphere is transparent and vegetation plays a limited role. At L-band, the emissivity may vary from almost 0.5 for a very wet soil to almost 1 for a very dry one, giving a range of 80–100 K for an instrument sensitivity usually of the order of 1 K. As the signal is not coherent, surface roughness and vegetation structure play a reduced role when compared to active systems. So one may wonder why L-band radiometry was not used extensively before when it has been proved to be most efficient during ground and airborne measurements (Schmugge et al., 1988). This is due to an inherent limitation: The spatial resolution is proportional to the antenna diameter and inversely proportional to the wavelength. At 21 cm, to achieve a 40 km resolution from an altitude of 750 km requires an antenna of about 8 m in diameter which is a very significant technical challenge. So the research was performed with higher frequency systems as available on the scanning multichannel microwave radiometer, SMMR (6.6 GHz) (Kerr and Njoku, 1990); the special sensor microwave imager, SSM/I (19 GHz); and now the advanced microwave scanning radiometer, AMSR-E (6.8 GHz) (Njoku and Li, 1999). Even though the frequency is not adapted, good results were obtained with SMMR (in spite of a very poor resolution due to important side lobes) and now AMSR-E (Njoku and Li, 1999). The limitations are mainly linked to the fact that the vegetation becomes rapidly opaque, the frequency is not protected and thus bound to be polluted by RFIs, and the single-angular measurement makes it difficult – in several cases – to separate vegetation and soil contributions to the signal.

The step forward Considering the necessity to make L-band measurements, several approaches have been tested to overcome the antenna-size issue. The first was initiated in the early 1990s with the idea to apply radio astronomy techniques (very large arrays) and very large baseline interferometers to Earth remote sensing. The one-dimensional concept, electronically scanned thinned array radiometer (ESTAR) was implemented as an aircraft version and proved to fulfill the requirements (Le Vine et al., 1994). It is a system, deployable in space as a sort of large rake, that offers – at the cost of a reduced sensitivity – an acceptable spatial resolution. In parallel, another approach, using inflatable (or umbrellalike) deploying technology, was studied at the Jet Propulsion Laboratory (JPL). Both concepts were proposed without success to space agencies on several occasions. The concepts appeared to be complex to deploy and to run, or to offer too limited measurements (single angle and frequency). By 1991, a small group started to work for ESA on the development of a similar instrument working in two dimensions (Goutoule et al., 1996).

786

SOIL MOISTURE

The concept was named Microwave Imaging Radiometer with Aperture Synthesis (MIRAS) and an airborne prototype was made and operated (Bayle et al., 2002). From then on, the concept evolved into a more tailored instrument which was proposed to the European Space Agency (ESA) (Kerr, 1998) in the framework of the Earth Explorer Opportunity mission under the name of Soil Moisture and Ocean Salinity (SMOS) mission. The mission was selected and is now underway. It is an ESA-led project with contributions from France and Spain. SMOS is scheduled for a launch in 2009 (Kerr et al., 2001). Similarly, a mission proposal was submitted to NASA, the Hydrospheric State Mission (Hydros) (Entekhabi et al., 2004). This mission relied on a deployable and rotating antenna linked to both a radiometer and a scatterometer. It is currently being investigated under the name of Soil Moisture Active and Passive (SMAP). SMOS is a Y-shaped instrument consisting of 69 elementary antennas regularly spaced along the arms providing at each integration step a full image (about 1,000
1,200 km) at either two polarizations or full polarization of the Earth’s surface (Kerr et al., 2001). The average ground resolution is 43 km and the globe is fully imaged twice (ascending and descending orbits) every 3 days at 6 am and 6 pm local solar time (equator crossing time). As the satellite travels on the orbit, any point of the surface is imaged at several angles, giving the angular signature of the pixel. The beauty of the concept is thus that a reasonable spatial resolution is obtained at the cost of a reduced sensitivity. Moreover, the pixels are viewed frequently at different angles and polarizations. The angular information is then used to separate the different contributions (soil – vegetation) to the signal (Wigneron et al., 2000).

Root zone soil moisture The big caveat of the remotely sensed soil moisture is that the direct measurement only concerns the surface layer. For instance, at X-band, a few millimeters and, at L-band, 4–5 cm are probed on average (depending on soil characteristics and condition). It is necessary however, for several applications, to know the available water in the unsaturated zone. Only one direct approach can currently be considered: using even lower frequencies (wavelengths of several meters) so as to reach deeper layers. This leads to large problems in terms of spatial resolution (a few hundred kilometers) as well as problems linked to ionospheric effects. So this option is not feasible now. The indirect approaches could be either to use gravity change approaches (if really validated and provided very coarse resolutions – hundreds of kilometers – are acceptable) or to rely on assimilation techniques, i.e., to use models to infer – from regular surface measurements and forcing conditions – what the root zone soil moisture is. The approach has been validated both using simulations and using ground data. The real limitations are linked to the models’ performances and to the input data quality.

Expression of needs Obviously, apart from the fact that the unsaturated zone is only partly probed, there are some requirements that are not fully fulfilled. The main one is the spatial resolution. Some needs, notably in hydrology, can only be resolved by having a better spatial resolution while still retaining a high temporal sampling. From space, this is not straightforward, but the most promising solution is probably to use external information to redistribute the area’s average moisture within the pixel: the so-called disaggregation. Several studies recently proved the feasibility of the approach with SMOS data (Pellenq et al., 2003; Merlin et al., 2006) and now the real-life validation has only to be performed. Caveats This is not to say all problems have been solved. There are still a number of outstanding issues which will require attention before an accurate and global soil moisture product is routinely delivered. Some problems, such as RFI, can be general issues, especially if protection is reduced in the future, which is a concern. The specific issues identified are currently being tackled and several references in the literature identify them, but, obviously, as long as the actual data (SMOS or any other) are not available, definitive conclusions and/or solutions will not be available and some unexpected issues might arise. Currently, the following issues are well identified. The most stringent is the pixel heterogeneity with components which may have very significant differences in behavior. The presence of free water within the pixel, for instance, has to be very accurately known (better than 2 %) to reach the overall accuracy of 4 % vol. in soil moisture. And water bodies can be variable as a function of season or weather. Vegetation is not totally opaque at L-band, and when the integrated water content is above 4–5 kg/m2, soil moisture retrievals will be difficult and approximate. Therefore, forested canopies will impact the signal. It may be noted at this level that recent studies showed that the main contributor at L-band for forest was the branch which does not evolve so rapidly (Ferrazzoli and Guerriero, 1996). Litter on the ground may behave as blackbody, masking strongly the soil’s signal. During rain events, water interception by the canopy might artificially increase the apparent vegetation’s water content. Topography will induce an altered angular behavior; snow and frozen soils will induce different signals which, if not accounted for, will produce wrong estimates. Urban areas and rocks are not fully assessed in terms of emissivity. Finally, and generally speaking, good retrieval will require some a priori knowledge of the surface cover, and state and the quality of the retrievals will be linked to the quality of the input data. It must also be noted that systems like SMOS will bring inherent specificities and complexities such as image reconstruction which is still a challenging point.

SOIL MOISTURE

Conclusion and perspectives Soil moisture has been for a long time a very specific variable. Though well identified, there were so few global measurements that models do not use it per se. However, after many unsuccessful attempts, a real soil moisture mission is now underway. It should enable the community to have, finally, access to global fields of soil moisture. Until launched and commissioned, the concept still has to be proved but the elements available can make one very confident to such an extent that an operational SMOS follow-on is currently being studied. Nevertheless, if SMOS answers some questions, (it) still does not fulfill all existing needs and ways forward must be sought. The most important is probably to improve the spatial resolution and there the SMOS concept is close to an optimum, as increasing the arm’s length will improve the spatial resolution but degrade significantly the sensitivity, to the point where it would not be useful anymore. So other techniques, such as disaggregation, will have to be found. To be more efficient, a SMOS-like instrument might gain either from being multifrequency or from having a coupled active system. To test those options, we will be expected to test these solutions using existing sensors (Advanced SCATterometer (ASCAT) and AMSR-E) when SMOS is operating. Another approach to improve spatial resolution could be to use even larger antennas, depending on the possibility of deploying them efficiently in space. In that case, to resolve the ambiguities, it will probably be necessary to improve the system by adding other frequencies while keeping the active source. This might lead to addressing the cryosphere as well, another key element in the global water and energy budget of the planet. It was stated in the SMOS proposal that the concept, though challenging, would open a new field with new measurements – soil moisture – made with a new type of sensors, paving the way for operational monitoring of water in soils. With the inception of the SMOS mission, this step is taken, opening a whole avenue of scientific challenges and making the long-awaited tool for water resources and water cycle monitoring a closer possibility. Bibliography Bayle, F., Wigneron, J.-P., Kerr, Y. H., Waldteufel, P., Anterrieu, E., Orlhac, J.-C., Chanzy, A., Marloie, O., Bernardini, M., Sobjaerg, S., Calvet, J.-C., Goutoule, J.-M., and Skou, N., 2002. Two-dimensional synthetic aperture images over a land surface scene. IEEE Transactions on Geoscience and Remote Sensing, 40(3), 710–714. Entekhabi, D., Njoku, E. G., Houser, P., Spencer, M., Doiron, T., Kim, Y., Smith, J., Girard, R., Belair, S., Crow, W., Jackson, T. J., Kerr, Y. H., Kimball, J. S., Koster, R., McDonald, K. C., O’Neill, P. E., Pultz, T., Running, S. W., Shi, J., Wood, E., and van Zyl, J., 2004. The hydrosphere state (hydros) satellite mission: an earth system pathfinder for global mapping of soil moisture and land freeze/thaw. IEEE Transactions on Geoscience and Remote Sensing, 42(10), 2184–2195.

787

Ferrazzoli, P., and Guerriero, L., 1996. Passive microwave remote sensing of forests: a model investigation. IEEE Transactions on Geoscience and Remote Sensing, 34, 433–443. Goutoule, J. M., Anterrieu, E., Kerr, Y. H., Lannes, A., and Skou, N., 1996. MIRAS Microwave Radiometry Critical Technical Development. Toulouse: MMS. Kerr, Y. H., 1998. The SMOS Mission: MIRAS on RAMSES. A Proposal to the Call for Earth Explorer Opportunity Mission. Toulouse: CESBIO. Kerr, Y. H., and Njoku, E. G., 1990. A semiempirical model for interpreting microwave emission from semiarid land surfaces as seen from space. IEEE Transactions on Geoscience and Remote Sensing, 28(3), 384–393. Kerr, Y. H., Waldteufel, P., Wigneron, J.-P., Martinuzzi, J.-M., Font, J., and Berger, M., 2001. Soil moisture retrieval from space: the soil moisture and ocean salinity (SMOS) mission. IEEE Transactions on Geoscience and Remote Sensing, 39(8), 1729–1735. Le Vine, D. M., Griffis, A. J., Swift, C. T., and Jackson, T. J., 1994. ESTAR: a synthetic aperture microwave radiometer for remote sensing applications. Proceedings of the IEEE, 82, 1787–1801. Merlin, O., Chehbouni, A. G., Kerr, Y. H., and Goodrich, D., 2006. A downscaling method for distributing surface soil moisture within a microwave pixel: application to monsoon ‘90 data. Remote Sensing of Environment, 101, 379–389. Moran, M. S., Hymer, D. C., Qi, J., and Kerr, Y., 2002. Comparison of ERS-2 SAR and Landsat TM imagery for monitoring agricultural crop and soil conditions. Remote Sensing of Environment, 79(2–3), 243–252. Njoku, E. G., and Li, L., 1999. Retrieval of land surface parameters using passive microwave measurements at 6–18 GHz. IEEE Transactions on Geoscience and Remote Sensing, 37(1), 79–93. Pellenq, J., Kalma, J., Boulet, G., Saulnier, G.-M., Wooldridge, S., Kerr, Y., and Chehbouni, A., 2003. A disaggregation scheme for soil moisture based on topography and soil depth. Journal of Hydrology, 276(1–4), 112–127. Schmugge, T. J., Wang, J. R., and Asrar, G., 1988. Results from the push broom microwave radiometer flights over the Konza Prairie in 1985. IEEE Transactions on Geoscience and Remote Sensing, 26(5), 590–597. Wigneron, J. P., Ferrazzoli, P., Calvet, J. C., Kerr, Y. H., and Bertuzzi, P., 1999. A parametric study on passive and active microwave observations over a soybean crop. IEEE Transactions on Geoscience and Remote Sensing, 37(6), 2728–2733. Wigneron, J.-P., Waldteufel, P., Chanzy, A., Calvet, J. C., and Kerr, Y., 2000. Two-D microwave interferometer retrieval capabilities of over land surfaces (SMOS mission). Remote Sensing of Environment, 73(3), 270–282.

Cross-references Agriculture and Remote Sensing Climate Data Records Climate Monitoring and Prediction Data Assimilation Global Climate Observing System Irrigation Management Land Surface Temperature Land-Atmosphere Interactions, Evapotranspiration Microwave Dielectric Properties of Materials Microwave Radiometers Microwave Radiometers, Conventional Microwave Radiometers, Interferometers Microwave Surface Scattering and Emission Radar, Scatterometers Radars Radiation, Volume Scattering Radiative Transfer, Theory

788

SOIL PROPERTIES

Rainfall Remote Sensing, Physics and Techniques Soil Properties Surface Water Water and Energy Cycles Water Resources

also play important roles in determining soil properties and additionally contribute to the heterogeneous soil spatial patterns. Remote-sensing systems, with their synoptic, multi-scale, and repetitive coverage of the land surface, provide an ideal means of mapping the spatial variation and biogeophysical properties of soils for environmental and natural resource management purposes.

SOIL PROPERTIES Alfredo Huete Plant Functional Biology and Climate Change Cluster, Faculty of Science, University of Technology, Sydney, NSW, Australia

Synonyms Pedology; Pedosphere Definition Soils. Soils are three-dimensional living bodies, with spatially variable biologic, physical, and chemical properties, that form the outer skin of the Earth’s terrestrial surface. Soil formation. Soils form slowly over time and develop distinguishing properties as a function of climate, geologic and organic parent materials, topography, time, vegetation type, and land use history. Soil profile. The vertical depth of a soil body varies from a few centimeters up to several meters and contains a series of soil horizons. The surface layers are termed the “O” (organic) or “A” (mineral) horizons, while a lower zone of clay accumulation is the “B” horizon, and the lowest zone that interfaces with the parent material is the “C” horizon. Introduction The thin biologically active soil layer of the Earth’s terrestrial surface is an integral part of the biosphere, regulating biogeochemical and hydrologic cycling of matter and energy and affecting land productivity and biodiversity. Soils perform numerous ecologic functions, as they buffer and filter pollutants that impact water quality, serve as source and sink of greenhouse gases, and partition rainfall between infiltration and runoff. Soils are home to the richest biodiversity on the planet that includes countless microorganisms that perform biochemical transformations, vital to ecosystem functioning, such as nitrogen fixation and organic matter decomposition. Reliable information on the spatial distribution of soil properties along with a better understanding of fundamental soil processes are vital to achieve sustainable land management of ecosystem services and to solve many environmental challenges facing society. Soils exhibit great spatial variability as a result of the interactions of climate, topography, parent material, and organisms acting on the soil body over time (Amundson and Jenny, 1997). Human activities and land use history

Soil optical properties Soil optical properties have been extensively studied in laboratory and field measurements and found to be primarily dependent on soil biogeochemical composition, soil surface geometric-optical scattering, and surface moisture status, as reviewed in Baumgardner et al. (1985). Generally, the spectral composition and the strength of the optical signal measured are determined by the minerals and organic compounds that coat the surface layer of soil particles (Stoner and Baumgardner, 1981). In contrast to the strong absorption spectra of pure mineral samples, soil absorption features are only broadly defined as soils consist of aggregate mixtures of many mineral and organic constituents. Many quantitative tools for extracting information about the physical and biochemical characteristics of soils have been developed and recognized as the field of soil spectroscopy (Irons et al., 1989; Ben-Dor et al., 1999). As an example, Ben-Dor and Bannin (1994) utilized a visible and near-infrared analysis scheme to predict a wide variety of soil chemical constituents, including CaCO3, Fe2O3, Al2O3, SiO2, free iron oxides, and K2O, from high-resolution spectra of arid and semiarid soils. In field radiometric measurements of in situ soils, there are strong optical-geometric interactions resulting from the angles in which the sun illuminates, and the sensor views, the surface. In addition, the atmosphere influences the relative proportions of direct (sun) and diffuse (sky) illumination on the soil surface. Particle size distribution and surface-height variation (roughness) are some of the most important factors influencing the reflectance properties of in situ soils. The bidirectional reflectance distribution function (BRDF) describes the manner in which surfaces scatter radiation across all sun-surface-sensor view geometries and can be used, through models, to derive geometric descriptors of a soil, such as size, shape, and orientation of surface “roughness” elements (Irons et al., 1992). There is a general decrease in reflectance with increasing surface roughness as coarse aggregates contain a lot of inter-aggregate spaces and “light traps.” With the shortest wavelengths most affected, optical-geometric interactions alter a soil’s spectral signature and the inferences made of soil biogeochemical properties such as soil color and mineralogy. When integrated across a hemisphere, the BRDF provides “albedo,” the ratio of shortwave (0.4–4 mm) radiant energy scattered in all directions to the incident irradiance on the surface. Albedo is a fundamental variable in energy balance studies, climate modeling

SOIL PROPERTIES

studies, and soil degradation studies, e.g., smooth, crusted, and structureless soils generally reflect more energy and are brighter than well-structured soils. Soil moisture also has a strong influence on the amount and composition of reflected energy from a soil with a general decrease in reflectance, proportional to the thickness and energy status of the adsorbed water. Soil moisture and soil roughness impart strong, or bright, microwave signals, and imagedifferencing techniques can be used to separate the dynamic soil moisture signal from the relatively static soil roughness signal (Sano et al., 1998). More recently developed LiDAR (Light Direction and Ranging) technology enables land surface roughness to be measured, on the order of centimeters, and offers promise in mapping terrain stability and hazard analysis, such as landslides (Jones, 2006).

Space- and airborne imagery Remotely sensed imagery from air- and spaceborne sensors integrates much of the knowledge learned from laboratory and field spectroscopy into the spatial domain and to the much more heterogeneous landscape scale. There is a wide array of airborne and satellite sensor systems currently available to study soils at the landscape level, encompassing a wide range of spatial and temporal resolutions, number of spectral bands and bandwidths, sun-view geometries, and polarization states. The derivation of soil information within landscapes may involve (1) direct measurements of exposed soils; (2) extraction of soil information from partially vegetated, open canopies; and (3) inferring soil properties from measurements of the vegetation layer. In all cases, correction and filtering of the image data are required to minimize atmosphere contamination, clouds, topography effects, BRDF influences, and sensor noise characteristics. Galvão et al. (2008) demonstrated how a subset of exposed soil pixels from airborne AVIRIS and spaceborne ASTER imagery, collected in Campo Verde, Brazil, could be used to study several soil physical and chemical weathering properties and their relationship to landscape position. Spectral soil indices, often in the form of band ratios, and spectral mixture analysis (SMA) have been used to ascertain soil information where vegetation and litter residues mask portions of the soil surface and pure soil pixels are not observable. SMA unmixes pixels into their respective soil, vegetation, and litter signal contributions, useful in soil and vegetation mapping, land degradation, soil erosion, and land cover conversion studies (Adams et al., 1986; Smith et al., 1990; Ustin et al., 1993). As an example, Palacios-Orueta et al. (1999) used a sequential spectral mixture analysis technique that utilized laboratory-derived “training” vectors to extract soil organic matter and iron sources of spectral variation. Alternatively, one may also infer soil type and properties from remote-sensing measures of foliar chemistry, a technique known as geobotany (see Ustin et al., 1999).

789

Landscape soil processes There have been considerable accomplishments made in understanding dynamic soil processes at the landscape scale with remote-sensing data, particularly in the areas of soil degradation, soil erosion, salinization, soil contamination, drought, carbon cycling, and climate and land use impacts. Soils are vulnerable and susceptible to poor land use practices that can yield soils with different physical properties and characteristics and thereby impact on their ability to cycle water and nutrients for sustained plant growth. Soil degradation Soil degradation is a major environmental concern that impacts on critical environmental issues such as food security, water quality, loss of biodiversity, and global climate change. Degraded soils exhibit an appreciable loss of soil organic matter with reduced fertility and plant productivity, deteriorated soil structures, and increases in albedo, thus altering carbon cycling and the land surface energy balance. Remote sensing provides repeatable and verifiable techniques for monitoring and assessing the spatial extent and severity of soil degradation. This is accomplished through satellite-based indicators of soil degradation that measure the loss of vegetative cover, increases in albedo, reduced soil organic matter levels, salinization, wind and water erosion, and soil crusting. Satellite data are increasingly being utilized to monitor the albedo of arid and semiarid lands given the importance of albedo as an indicator of land degradation and as a physical parameter with possible impacts on climate. Spatial heterogeneity indices from satellite images are also sensitive indicators of landscape instability, with increases in variance due to soil degradation and erosion (Pickup and Nelson, 1984). An important threat posed by soil degradation is soil erosion, in which the upper and most fertile soil layers are removed by wind and water forces and deposited into waterways and coastal zones, resulting in loss of arable land and the silting of reservoirs. Windblown atmospheric dust may also pose public health concerns. Soil erosion is evident in remote-sensing imagery when subsoil characteristics are exposed, which contrast with the “intact” surface soil spectral signals. In very fine spatial resolution imagery, such as from QuickBird, detailed information on linear erosion features such as gullies, rills, and sand dune formations can be directly observed. Soil carbon There is great interest in analyzing the importance of soils as a source or sink of greenhouse gases such as carbon dioxide, methane, and nitrous oxide. The soil carbon pool is approximately twice that in the atmosphere and three times that in vegetation (Smith et al., 2008). Although the response of soils to global warming is critical, reliable estimates of carbon reserves in undisturbed soils are not available, and data on rates of carbon loss following

790

SOIL PROPERTIES

disturbances are lacking (Lal, 1997). Soils subject to accelerated decomposition of organic matter tend to release carbon dioxide and thus contribute to the enhanced greenhouse effect, while areas being revegetated and enriched with organic matter can absorb and sequester quantities of carbon that are extracted from the atmosphere through photosynthesis. In the context of global sustainability, it is essential to understand how the source/sink function of soils can be managed and controlled to mitigate the impact of climate change (Post et al., 1982). Soils also play important roles in water quality, sediment and pollutant discharges into water systems, and environmental remediation efforts. The soil has the ability to filter and break down toxic waste materials and many industrial by-products and render them less harmful. New capabilities are also becoming available for soil process studies enabling the integration of environmental process models, climate data, human interactions, and Geographic Information Systems (GIS) with remote-sensing data.

Conclusion In summary, soils exhibit tremendous spatial and dynamic variations across a landscape due to variations in moisture content, biogeochemical composition, surface roughness elements, and ongoing processes of land use, soil degradation, erosion, contamination, and global change. Although not fully developed, remote sensing remains the only viable technique to map, monitor, and manage the fragile soil resource. With over 90 % of the Earth’s terrestrial surface classified as “open canopies,” there is an appreciable soil component that can potentially be remotely sensed. Many of the remote-sensing techniques used to derive soil information, however, remain undeveloped, and there are no currently available operational satellite data products on soil properties. In part, the extraction of soil information from landscape-scale imagery is too complex, hindering many potential end uses. As noted by Ben-Dor et al. (2008), the adoption of imaging spectroscopy techniques, although cost-effective, remains a new frontier in soil science with advances limited by the lack of availability of operational hyperspectral sensors. Nevertheless, satellite data products, such as albedo, BRDF, and vegetation indices, measure important landscape properties that are useful in deriving soil information. Remote sensing is also only sensitive to the immediate surface, which prevents the assessments of important subsoil and root-zone soil properties. However, the soil surface is the most dynamic, biologic, and hydrologic interface that responds immediately to climatic changes and human forcings. Remote sensing can provide quantitative measurements of such changes at the surface, which are needed to address a wide range of environmental and global change issues. Further advancements are needed to infer soil properties with depth through pedotransfer functions that enable the derivation of subsoil biogeophysical products from remotely sensed surface properties.

Bibliography Adams, J. B., Smith, M. O., and Johnson, P., 1986. Spectral mixture modeling, a new analysis of rock and soil types at the Viking Lander 1 site. Journal of Geophysical Research, 91(B8), 8098–8112. Amundson, R., and Jenny, H., 1997. On a state factor model of ecosystems. BioScience, 47, 536–543. Baumgardner, M. F., Silva, L. F., Biehl, L. L., and Stoner, E. R., 1985. Reflectance properties of soils. Advances in Agronomy, 38, 1–44. Ben-Dor, E., and Bannin, A., 1994. Visible and near infrared (0.4–1.1 mm) analysis of arid and semi arid soils. Remote Sensing of Environment, 48, 261–274. Ben-Dor, E., Irons, J. R., and Epema, G., 1999. Soil reflectance. In Rencz, A. N. (ed.), Remote Sensing for the Earth Sciences, Manual of Remote Sensing, 3rd edn. New York: Wiley, pp. 111–188. Ben-Dor, E., Taylor, R. G., Hill, J., Demattê, J. A. M., Whiting, M. L., Chabrillat, S., and Sommer, S., 2008. Imaging spectrometry for soil applications. Advances in Agronomy, 97, 321–392. Galvão, L. S., Formaggio, A. R., Couto, E. G., and Roberts, D. A., 2008. Relationships between the mineralogical and chemical composition of tropical soils and topography from hyperspectral remote sensing data. ISPRS Journal of Photogrammetry and Remote Sensing, 63, 259–271. Irons, J. R., Weismiller, R. A., and Petersen, G. W., 1989. Soil reflectance. In Asrar, G. (ed.), Theory and Application of Optical Remote Sensing. New York: Wiley, pp. 66–106. Irons, J. R., Campbell, G., Norman, J. M., Graham, D. W., and Kovalick, W. M., 1992. Prediction and measurement of soil bidirectional reflectance. IEEE Transactions on Geoscience and Remote Sensing, GE-30, 249–260. Jones, L., 2006. Monitoring landslides in hazardous terrain using terrestrial LiDAR: an example from Montserrat. Quarterly Journal of Engineering Geology & Hydrogeology, 39, 371–373. Lal, R., 1997. Soil processes and the greenhouse effect. In Lal, R., Blum, W. H., Valentin, C., and Stewart, B. A. (eds.), Methods for Assessment of Soil Degradation. Boca Raton, FL: CRC Press. Advances in Soil Science, pp. 199–212. Palacios-Orueta, A., Pinzón, J. E., Ustin, S. L., and Roberts, D. A., 1999. Remote sensing of soils in the Santa Monica Mountains: II. Hierarchical foreground and background analysis. Remote Sensing of Environment, 68, 138–151. Pickup, G., and Nelson, J., 1984. Use of Landsat radiance parameters to distinguish soil erosion, stability and deposition in arid central Australia. Remote Sensing of Environment, 16, 195–204. Post, W. M., Emanuel, W. R., Zinke, P. J., and Stangenberger, A. G., 1982. Soil carbon pools and world life zones. Nature, 298, 156–159. Sano, E. E., Huete, A. R., Troufleau, D., Moran, M. S., and Vidal, A., 1998. Relation between ERS-1 synthetic aperture radar data and measurements of surface roughness and moisture content of rocky soils in a semiarid rangeland. Water Resources Research, 34, 1491–1498. Smith, M. O., Ustin, S. L., Adams, J. B., and Gillespie, A. R., 1990. Vegetation in deserts: 1. A regional measure of abundance from multispectral images. Remote Sensing of Environment, 31, 1–26. Smith, P., Fang, C., Dawson, J. J. C., Moncrieff, J. B., and Smith, P., 2008. Impact of global warming on soil organic carbon. Advances in Agronomy, 97, 1–43. Stoner, E. R., and Baumgardner, M. F., 1981. Characteristic variations in reflectance of surface soils. Soil Science Society of American Journal, 45, 1161–1165. Ustin, S. L., Smith, M. O., and Adams, J. B., 1993. Remote sensing of ecological processes: a strategy for developing and testing ecological models using spectral mixture analysis.

SOLID EARTH MASS TRANSPORT

In Ehleringer, J. R., and Field, C. B. (eds.), Scaling Physiological Processes: Leaf to Globe. San Diego: Academic, pp. 339–357. Ustin, S. L., Smith, M. O., Jacquemoud, S., Verstraete, M. M., and Govaerts, Y. M., 1999. Geobotany: vegetation mapping in earth sciences. In Rencz, A. N. (ed.), Remote Sensing for the Earth Sciences, Manual of Remote Sensing, 3rd edn. New York: Wiley, pp. 189–248.

Cross-references Land Surface Roughness Land-Atmosphere Interactions, Evapotranspiration Lidar Systems Soil Moisture Water and Energy Cycles Water Resources

SOLID EARTH MASS TRANSPORT Erik Ivins Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA

Synonyms Time-variable gravity of the solid Earth Definition Time-variable gravity of the solid Earth. Any measurement of gravity at, or below, the Earth’s surface or from above by airborne or space instruments, wherein the physical origins of the variations are solely, or interactively, caused by movements, deformations, or changes in density of the solid Earth. Generally, solid Earth time variability includes, but is not limited to, fracture or poroelasticrelated crustal volatile and hydrological material transport, changes of state and solid Earth’s response to periodic forcing by ocean and lunisolar tides, earthquake-related stress reorganization, or by nutation and wobble of the Earth’s rotation axis. Large-scale and long-term mass variability is associated with the solid creep of the mantle in response to the surface environmental load changes associated with the last global ice age (100,000–10,000 years ago). Smaller-scale mass variability driven by flow of the upper mantle is also associated with the more recent Little Ice Age surface loading, involving time scales of 50–500 years and surface vertical motions of 1–40 mm/year. Only recently have the last two phenomena become amenable to accurate measurement by modern space gravimetry. Introduction The study of the Earth’s gravity, topography, and seismological structure has provided several of the most basic observational cornerstones that support the dynamic theories for the evolution of Earth’s mountain belts, sedimentary basins, massive continental eruptive and breakup events, ocean ridges, plate motions, slab subduction, and all phenomenon that are generally related to mantle

791

convection (e.g., Richards and Hager, 1984; Forte and Woodward, 1997). While inferences concerning gravity change and mass transport may be made for mantle convection, the minimum relevant time scales, t, are defined by those for plate tectonic reorganization (t  5 million years). The class of solid Earth time-variable gravity that is measurable by satellite remote-sensing methods is shorter by at least a factor of 104 (t 50 years). Experts in near-Earth orbital dynamics developed early measurement concepts focused on simple drifts and periodic patterns in well-tracked orbital elements (Paddack, 1967; Kozai, 1968). The launch of two passive laser-ranging satellites, Starlette and LAGEOS-1 in 1976, marked the beginning of a new era in global gravity remote sensing. The key orbital elements proved to be sensitive to changes in planetary oblateness, and these were best explained by the presence of a slow viscoelastic flow of the mantle in response to the two-way transfer of ocean-to-continentto-ocean ice sheet expansion and contraction that began about 100,000 years before present (year BP) and ended 7,000 year BP. The response phenomenon is known as glacial isostatic adjustment (GIA) of the combined solid and liquid Earth. GIA alone gave a reasonable explanation for the amplitude and sense of the nontidal part of variations in the Earth’s gravity field (Yoder et al., 1982; Peltier, 1982; Rubincam, 1984). An excellent survey of the methods and theoretical representation of the static gravitational field prior to time-variable mapping has been given by Phillips and Lambeck (1980). These methods, including coefficient-normalization conventions, are applicable to time-varying gravity.

Passive satellites in constellation With the passing of two decades since the launch of Starlette and LAGEOS-1, and the accumulation of tracking data for a virtual constellation of more than ten subsequent satellites, it has been possible to refine the bestfitting GIA models and to incorporate realistic scenarios for the mass balance of mountain glaciers, the Greenland and Antarctic ice sheets. A series of papers, each analyzing slightly different tracking-derived solutions for the zonal harmonic drift rates (e.g., Cheng et al., 1997), identified a statistically significant improvement in the residual fit to all zonal harmonic rates of drift by accounting for Antarctic and Greenland ice mass change. The solved-for drift rates are especially sensitive to the latitudinal dependencies of the low-degree zonal harmonics, hence natural antecedents to the study of both GIA and present-day mass balance of the continental cryosphere. James and Ivins (1997), Johnston and Lambeck (1999), Cox et al. (2001), and Tosi et al. (2005) used zonal harmonic rates of order n 9 and showed that Antarctic mass loss at rates of 70 to 340 Gt/year (1 Gt ¼ 109 t) was favored across a broad spectrum of mantle viscosity profiles. A critical aspect of the finely tuned relationships between nontidal satellite orbital element drifts to solid Earth GIA and cryospheric imbalance is that modeled

792

SOLID EARTH MASS TRANSPORT

accelerations due to drag have appropriate error estimates and that the mass changes in the oceans have no large-amplitude, low-order, and long-term secular-like components (cf. Hughes and Stepanov, 2004).

Satellite-to-satellite tracking: a new era for time-variable gravity Plans for a high-resolution global gravity-mapping mission with satellites at mean orbital altitudes below 800 km were well underway by the late 1970s (e.g., Kaula, 1983). The lower orbits of nonpassive satellites required new navigational and tracking strategies. Passively laser-tracked gravity satellites in nearly circular orbit have apogees, a, ranging from 7,173 to 25,001 km and are relatively drag-free. The measurement sensitivity to surface gravity falls off as (Re/a)n +1 where Re is the mean radius of the solid Earth and the harmonic degree and n is related to spatial resolution as the half wavelength Dl Re p/n. Optimistically, the limits of spatial resolution (in latitude only) of the passive satellites are Dl < 2,200 km. The limitation is actually far more severe than this. At full wavelength, 2Dl, say, for example, the orbits of these passive satellites are responsive to mass changes at, or beneath, the combined Central Siberian Plateau and Western Siberian Plain (representing approximately one-fifth of the Eurasian land area). Yet the rate solutions for the harmonic coefficients, @ tCnm, determined by passive-satellite tracking are unable to resolve the latitudinal location of the disturbing mass, for there is equal probability that the observed orbital drifts are related to mass changes in central Canada! This drawback is related to the lack of sensitivity to nonzonal harmonics (generally, nonaxially symmetric), having m 6¼ 0. This is essentially the “Achilles’ heel” of satellite laser-ranging (SLR) strategies for monitoring time-variable gravity from space.

Multiple-satellite tracking and smaller a open a new window to resolution. In fact, the German geoscience satellite launched in 2000, CHAMP, utilizing tracking by the entire constellation of high-orbit Global Positioning System (GPS) satellites, and onboard accelerometer and star-tracking data, was capable of resolving the global static field to degree nmax 72 (Reigber et al., 2003). Although sampling in polar orbit yields a degraded longitudinal resolution, effectively limiting nonzonal resolution to mmax O(nmax)/2, seasonal changes in mass can unambiguously detect mass changes associated, for example, to the combined Central Siberian Plateau and Western Siberian Plain. The role that a unified single tracking reference frame plays in reducing the errors in the end-product gravity mapping should not be underestimated (Rummel et al., 2005). The polar orbit also allowed, for the first time, a global gravity map to be completed that is based on a single instrument and observing scheme, as opposed to the prior generation of merging strategies required to produce global maps. The latter always contained regionally varying quality and had no possibility of representing global time-varying properties. The Gravity Recovery and Climate Experiment (GRACE) satellite pair determines an even higherresolution static and time-variable field. These were launched into polar orbit in 2002 and fly in tandem, separated by about 220–240 km. The two track one another with a Ka-band radar system, in addition to using onboard accelerometers and GPS tracking (Figure 1). Although degraded in resolution by the same anisotropic sampling problem as CHAMP, the resolution recovered by a standard isotropic Gaussian filtering technique (Wahr et al., 1998) is about 400 km, and a variety of more sophisticated filtering methods may reduce this by 50–130 km, depending on the sought-after frequency component and latitude (e.g., Davis et al., 2008).

Solid Earth Mass Transport, Figure 1 GRACE satellite pair with Ka band microwave tracking links. The separation is about 220– 240 km. The altitude is about 550 km and the pair has roughly 15 Earth revolutions per day. Also see http://www.nasa.gov/mov/ 161008main_GraceBeauty1web.mov (Graphic from http://op.gfz-potsdam.de/grace/index_GRACE.html).

SOLID EARTH MASS TRANSPORT

793

environments such as Alaska (Luthcke et al., 2008) with resolution approaching the inter-satellite tracking separation distance (240 km). To illustrate the importance of the cryospheric measurements, adding all of the mass imbalance estimates recovered from GRACE alone since its launch in 2002, including that of Chen et al. (2007) for southern Patagonia, these add to nearly 1.7 mm/year equivalent uniform global sea-level rise, nearly half that observed for the last 15 years (Church and White, 2006) and possibly the lion’s share of that inferred to be caused by all continental sources of water mass imbalance. What then is the role of monitoring solid Earth related gravity change?

Solid Earth Mass Transport, Figure 2 GRACE-based static gravity field and inset photo of the passive gravity satellite Lageos-2 equipped with retroreflectors, weighing 400 kg and having a 60 cm diameter (http://www.nasa.gov/vision/earth/ lookingatearth/earth_drag.html).

The higher-accuracy and complete even-degree harmonic retrieval of the fields allowed for a highly accurate determination of modeled residual drifts in the nodes of the LAGEOS-1 and 2 satellites and the later confirmed the predicted drifts of Einstein’s theory of relativity. The drifts are a result of additional space-time curvature caused by the Earth’s rotation (Ciufolini and Pavlis, 2004). The passive, laser-reflecting, LAGEOS-2 satellite is shown as an inset to the GRACE-gravity anomaly map shown in Figure 2. GRACE and CHAMP determine fields and their time variability at higher latitudes with relatively enhanced fidelity. The mapped fields are inherently less prone to errors due to the greater collection of orbital mapping data. The standard (level-2) data consist of monthly (or submonthly) spherical harmonic coefficients, Cnm, as there exist a collection of well-developed methodologies for such representation and use (e.g., Rapp and Pavlis, 1990; Petrovskaya et al., 2001; Tapley et al., 2004). Important results derived from the harmonic releases for polar latitudes include the detection of huge ice loss in the southeastern part of Greenland (Velicogna and Wahr, 2006a) and in West Antarctica (Velicogna and Wahr, 2006b: Ramillien et al., 2006). A more sophisticated method uses the raw satellite-to-satellite range and range-rate data and accelerometer data to directly infer the location and size of perturbing masses beneath the satellites without reference to global spherical harmonics. The methods use the same basic potential field representation for lunar mascons that was developed in the late 1960s (e.g., Muller and Sjogren, 1968). These direct methods have shown great sensitivity to year-to-year ice mass loss in coastal

Mantle deformation To monitor the hydrological or cryospheric mass changes using a time series constructed from individual GRACE maps, or equivalently harmonic coefficients, Cnm(ti), a coefficient representing water mass must be constructed, WE Cnm(ti.), by finding the deviation at time ti from the average of the sum of maps at all times, ti. In the presence of surface water masses, the instantaneous elastic deformation of the solid Earth can be accounted for using theory for loading of a spherically symmetric radially stratified Earth model (e.g., Farrell, 1972) and an n-dependent factor hen/(1 + ken), where hen and ken are the radial displacement and potential Love numbers, respectively (hen 1.01, 0.3 < ken < 0, for all n > 1) (e.g., Wahr et al., 1998). For example, if surface water mass builds in a region, it depresses the solid surface and the radially stratified density structure of the entire Earth; thus, the accelerations on the satellite are less than those that would correspond the same water mass on a rigid Earth. These elastic corrections for solid Earth deformation are routine and straightforward. The main complexities arise when GIA-related solid Earth motion and gravitational potential changes are locally, or regionally, coincident with the location of the water mass changes. Such is the case in Antarctica, as has been discussed by Velicogna and Wahr (2006b) and Ramillien et al. (2006). However, the coincidence of regional changes in hydrology can also be difficult to separate from GIA, where GIA dominates the secular gravity changes, such as in central Canada and central Scandinavia. Ivins and Wolf (2008) have recently reviewed the current status of overlaps in GIA and hydrology. The ice caps and glaciers of Iceland provide an excellent example of how this coincidence of viscoelastic GIA and ice mass change-related gravity changes are measured. There is now consensus that such ice masses are melting and raising global sea level at an accelerating pace (e.g., Kaser et al., 2006). The main center of cryospheric mass in Iceland is located in the southeastern part of the island, as seen in the cloud-free Terra Satellite image of Figure 3a, where ice- and snow-dominated regions are white in color. A large thermal plume, probably rising from the core-mantle boundary (e.g., Nataf, 2000),

794

SOLID EARTH MASS TRANSPORT

Solid Earth Mass Transport, Figure 3 Iceland true color image in summer using the Moderate Resolution Imaging Spectroradiometer (MODIS) instrument onboard the Terra spacecraft (a) (Credit: http://visibleearth.nasa.gov/view_rec.php? id¼6606). Vatnajoku¨ll ice cap is the larger white area in the southeast. GRACE monthly solution trends from release 04 (February 2003–April 2007) using an isotropic Gaussian smoothing filter applied to the harmonics with radius a ¼ 282.5 km. The units are in millimeter of water-height equivalent (b). Solid Earth uplift predictions at present day on an incompressible viscoelastic mantle halfspace of viscosity 3
10 18 Pa s under an elastic lithosphere of 25 km thickness. The ice cap evolves in accordance with the history simulated by Marshall et al. (2005) and calibrated by regional paleoclimate indicators (e.g., Flowers et al., 2008). This is a simple twodisk model of the Little Ice Age evolution of the Vatnajoku¨ll ice cap that evolves into the late twentieth century with the same observational constraints as in Fleming et al. (2007). GPS data are from UNAVCO archives and up arrows indicate uplift and down subsidence (with formal error estimates in red). These data are from a combination of permanent and episodic GPS tracking stations, references to International Terrestrial Reference Frame ITRF2000 on bedrock and have more than 6 years of observing history.

characterizes the regional mantle. A greatly reduced regional upper mantle viscosity is generally accepted, likely at values close to 3
1018 Pa s (Árnadóttir et al., 2005). This viscosity reduction greatly exacerbates both the prediction of the sensitivity to, and the amplitude of, GIA-related geodetic quantities. Icelandic paleoclimate and glacier history are relatively well known (e.g., Marshall et al., 2005; Flowers et al., 2008), as are recent mass

changes of the Vatnajoküll ice cap and other major Icelandic glaciers (Magnússon et al., 2005). It is possible, therefore, to use a relatively complete ice load history and a parameter search of mantle viscoelastic properties to match the uplift patterns and magnitudes measured by GPS adjacent to the Vatnajoküll ice cap as shown in Figure 3b. The model used to compute uplift in Figure 3b assumes an idealized circular disk to represent the ice cap

SOLID EARTH MASS TRANSPORT

with the computational methods described by Ivins and James (1999) and includes the present-day responses to current ice mass changes (for a detailed description of viscoelastic GIA modeling of Iceland, readers should refer to recent work by Fleming et al., 2007). Ignoring the ice loss, the viscoelastic uplift of bedrock, @ tur, in response to ice mass loss (Figure 3b), the water-equivalent rate of roughly rMC/rw x @ tur, where rMC/rw is the ratio of bedrock to water density, or about 2.8–3.8. The negative water-height rate deduced from GRACE spherical harmonic release 04 is in excess of 20 mm/year and has a local minima at the geographical position of the Vatnajoküll ice cap (Figure 3c). Due to the unique circumstance of having GPS uplift data adjacent to the ice cap, and a wellconstrained ice history, when the GIA phenomenon is correctly accounted for, the ice mass loss rate solution for eastern Iceland is roughly 5–7 Gt/year and would be almost half this value if GIA-related mantle mass variability had not been accounted for. This example of the Vatnajoküll ice cap is for illustrative purposes, as the GRACE errors are quite poorly calibrated at this wavelength. The scales over which this GRACE rate is extracted from harmonic solutions are at the upper limits of resolution, shorter, for example, than the 300–400 km resolution in the example of mantle deformation and mass movement from the coseismic offsets caused by the Mw 9.1 December 2004 Aceh-Andaman earthquake (Han et al., 2006) and from Patagonian ice mass loss (Chen et al., 2007).

Conclusion Future space-gravimetry systems, much like the highly successful polar-orbiting GRACE satellite pair, strive to meet a critical environmental remote-sensing challenge of the twenty-first century: determination of the long-wavelength mass stability of the Earth’s continental fresh water supply (e.g., Velicogna and Wahr, 2006a, b; Ramillien et al., 2006; Syed et al., 2008). Terrestrial remote-sensing data have recently been directed at the severe consequences of the unprecedented warmth that has appeared during the last 100 hundred years and that now appears to be accelerating to paces unprecedented during the past 2,000 years (Mann, 2007). Monitoring the secular trends in both continental water storage and sustained ice sheet change will be of increasing societal concern. Use of space-determined gravity fields is strongly influenced by viscoelastic motions of the mantle and crust, and the study of such mass variability deep in the Earth’s interior will be necessary for fully retrieving accurate and well-resolved trends in changing cryosphere and surface hydrology. Acknowledgment This research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the NASA.

795

Bibliography Árnadóttir, T., Jónsson, S., Pollitz, F. F., Jiang, W., and Feigl, K. L., 2005. Postseismic deformation following the June 2000 earthquake sequence in the south Iceland seismic zone. Journal of Geophysical Research, 110, B12308, doi:10.1029/ 2005JB003701. Chen, J. L., Wilson, C. R., Tapley, B. D., Blankenship, D. D., and Ivins, E. R., 2007. Patagonian icefield melting observed by Gravity Recovery and Climate Experiment (GRACE). Geophysical Research Letters, 34, L22501, doi:10.1029/ 2007GL031871. Cheng, M. K., Shum, C. K., and Tapley, B. D., 1997. Determination of long-term changes in the Earth’s gravity field from satellite laser ranging observations. Journal of Geophysical Research, 102, 22377–22390. Church, J. A., and White, N. J., 2006. A 20th century acceleration in global sea-level rise. Geophysical Research Letters, 33, L01602, doi:10.1029/2005GL024826. Ciufolini, I., and Pavlis, E. C., 2004. A confirmation of the general relativistic prediction of the Lense-Thirring effect. Nature, 431, 958–960. Cox, C. M., Klosko, S. M., and Chao, B. F., 2001. Changes in ice-mass balance inferred from the time variations of the geopotential observed through SLR and DORIS tracking. In Sideris, M. (ed.), Gravity, Geoid and Geodynamics 2000. Berlin: Springer. IAG Symposia Series, Vol. 123, pp. 355–360. Davis, J. L., Tamisiea, M. E., Elósegui, P., Mitrovica, J. X., and Hill, E. M., 2008. A statistical filtering approach for Gravity Recovery and Climate Experiment (GRACE) gravity data. Journal of Geophysical Research, 113, B04410, doi:10.1029/ 2007JB005043. Farrell, W. E., 1972. Deformation of the Earth by surface loads. Reviews of Geophysics and Space Physics, 10, 761–797. Fleming, K., Martinec, Z., and Wolf, D., 2007. Glacial-isostatic adjustment and the viscosity structure underlying the Vatnajokull ice cap. Iceland, Pure and Applied Geophysics, 164, 751–768. Flowers, G. E., Bjornsson, H., Geirsdottir, A., Miller, G. H., Black, J. L., and Clarke, G. K. C., 2008. Holocene climate conditions and glacier variation in central Iceland from physical modeling and empirical evidence. Quaternary Science Reviews, 27, 797–813. Forte, A. M., and Woodward, R. L., 1997. Seismic-geodynamic constraints on vertical flow between the upper and lower mantle: the dynamics of the 670 km seismic discontinuity, Chapter 10. In Crossley, D. J. (ed.), The Earth’s Deep Interior: The Dorboos Memorial Volume. Amsterdam: Gordon and Breach Science, pp. 337–404. Han, S.-C., Shum, C. K., Bevis, M., Ji, C., and Kuo, C.-Y., 2006. Crustal dilatation observed by GRACE after the 2004 SumatraAndaman earthquake. Science, 313, 658–662. Hughes, C. W., and Stepanov, V. N., 2004. Ocean dynamics associated with rapid J2 fluctuations: importance of circumpolar modes and identification of a coherent Arctic mode. Journal of Geophysical Research, 109, C06002, doi:10.1029/ 2003JC002176. Ivins, E. R., and James, T. S., 1999. Simple models for late Holocene and present-day Patagonian glacier fluctuations and predictions of a geodetically detectable isostatic response. Geophysical Journal International, 138, 601–624. Ivins, E. R., and Wolf, D., 2008. Glacial isostatic adjustment: new developments from advanced observing systems and modeling. Journal of Geodynamics, doi:10.1016/j.jog.2008.06.002. James, T. S., and Ivins, E. R., 1997. Global geodetic signatures of the Antarctic ice sheet. Journal of Geophysical Research, 102, 605–633.

796

STRATOSPHERIC OZONE

Kaser, G., Cogley, J. G., Dyurgerov, M. B., Meier, M. F., and Ohmura, A., 2006. Mass balance of glaciers and ice caps: consensus estimates for 1961–2004. Geophysical Research Letters, 33, L19501, doi:10.1029/2006GL027511. Kaula, W. M., 1983. Inferences of variations in the gravity field from satellite-to-satellite range rate. Journal of Geophysical Research, 88, 8345–8349. Kozai, Y., 1968. Love’s number of the Earth derived from satellite observation. Publications of the Astronomical Society of Japan, 20, 24–26. Luthcke, S., Arndt, A. A., Rowlands, D. D., McCarthy, J. J., and Larsen, C. F., 2008. Recent glacier mass changes in the Gulf of Alaska region from GRACE mascon solutions. Journal of Glaciology, 54, 767–777. Magnússon, E., Björnsson, H., Dall, J., and Pálsson, F., 2005. Volume changes of Vatnajökull ice cap, Iceland, due to surface mass balance, ice flow, and subglacial melting at geothermal areas. Geophysical Research Letters, 32, L05504, doi:10.1029/ 2004GL021615. Mann, M. E., 2007. Climate over the past two millennia. Annual Review of Earth and Planetary Sciences, 35, 111–136, doi:10.1146/annurev.earth.35.031306.140042. Marshall, S. J., Björnsson, H., Flowers, G. E., and Clarke, G. K. C., 2005. Simulation of Vatnajökull ice cap dynamics. Journal of Geophysical Research, 110, F03009, doi:10.1029/2004JF000262. Muller, P., and Sjogren, W., 1968. Mascons: lunar mass concentrations. Science, 161, 680–684. Nataf, H.-C., 2000. Seismic imaging of mantle plumes. Annual Review of Earth and Planetary Sciences, 28, 391–417, doi:10.1146/annurev.earth.28.1.391. Paddack, S. J., 1967. On the angular momentum of the solid Earth. Journal of Geophysical Research, 72, 5760–5762. Peltier, W. R., 1982. Dynamics of the ice age earth. Advances in Geophysics, 24, 1–146. Petrovskaya, M. S., Nershkov, A. N., and Pavlis, N. K., 2001. New analytical and numerical approaches for geopotential modeling. Journal of Geodesy, 75, 661–672, doi:10.1007/s001900100215. Phillips, R. J., and Lambeck, K., 1980. Gravity fields of the terrestrial planets – long-wavelength anomalies and tectonics. Reviews of Geophysics and Space Physics, 18, 27–76. Ramillien, G., Lombard, A., Cazenave, A., Ivins, E. R., Remy, F., and Biancale, R., 2006. Interannual variations of the mass balance of the Antarctic and Greenland ice sheets from GRACE. Global and Planetary Change, 53, 198–208. Rapp, R. H., and Pavlis, N. K., 1990. The development and analysis of geopotential coefficient models to spherical harmonic degree 360. Journal of Geophysical Research, 95, 21885–21911. Reigber, C., Balmino, G., Schwintzer, P., Biancale, R., Bode, A., Lemoine, J. M., Konig, R., Loyer, S., Neumayer, H., Marty, J. C., Barthelmes, F., Perosanz, F., and Zhu, S. Y., 2003. Global gravity field recovery using solely GPS tracking and accelerometer data from CHAMP. Space Science Reviews, 108, 55–66. Richards, M. A., and Hager, B. H., 2004. Geoid anomalies in a dynamic Earth. Journal of Geophysical Research, 89, 5987–6002. Rubincam, D. P., 1984. Postglacial rebound by Lageos and the effective viscosity of the lower mantle. Journal of Geophysical Research, 89, 1077–1087. Rummel, R., Rothacher, M., and Buetler, G., 2005. Integrated Global Geodetic Observing System (IGGOS) – science rationale. Journal of Geodynamics, 40, 357–362. Syed, T. H., Famiglietti, J. S., Rodell, M., Chen, J., and Wilson, C. R., 2008. Analysis of terrestrial water storage changes from GRACE and GLDAS. Water Resources Research, 44, W02433, doi:10.1029/2006WR005779. Tapley, B. D., Bettadpur, S., Ries, J. C., Thompson, P. F., and Watkins, M. M., 2004. GRACE measurements of mass variability in the Earth system. Science, 305, 503–505.

Tosi, N., Sabadini, R., Marotta, A. M., and Vermeersen, L. L. A., 2005. Simultaneous inversion for the Earth’s mantle viscosity and ice mass imbalance in Antarctica and Greenland. Journal of Geophysical Research, 110, doi:10.1029/2004JB003236. Velicogna, I., and Wahr, J., 2006a. Acceleration of Greenland ice mass loss in spring 2004. Nature, 443, 329–331. Velicogna, I., and Wahr, J., 2006b. Measurements of time-variable gravity show mass loss in Antarctica. Science, 311, 1754–1756. Wahr, J., Molenaar, M., and Bryan, F., 1998. Time variability of the Earth’s gravity field: hydrological and oceanic effects and their possible detection using GRACE. Journal of Geophysical Research, 103, 30205–30230. Yoder, C. F., Williams, J. G., Dickey, J. O., Schutz, B. E., Eanes, R. J., and Tapley, B. D., 1982. J˙2 from LAGEOS and the non-tidal acceleration of Earth rotation. Nature, 303, 757–762.

Cross-references Cryosphere, Climate Change Effects Cryosphere, Measurements and Applications Global Earth Observation System of Systems (GEOSS) Sea Level Rise

STRATOSPHERIC OZONE Michelle Santee Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA

Definition Stratosphere. The region of the atmosphere extending from about 13 to 50 km altitude. Ozone. A naturally occurring trace gas in Earth’s atmosphere. An ozone molecule consists of three atoms of oxygen bound together in a triangular arrangement. Introduction The vast majority of ozone (about 90 %) resides in the stratosphere, where it plays a crucial role in protecting life on Earth from harmful solar ultraviolet radiation. Most of the remaining ozone is found in the troposphere, the part of the atmosphere between the surface and the stratosphere and the region in which most clouds and weather occur. Tropospheric ozone is an important pollutant and a greenhouse gas. In addition to its direct influence on air quality, tropospheric ozone helps to regulate the ability of the atmosphere to “cleanse” itself of many other pollutants. Anthropogenic (human-induced) increases in ozone at ground level, where it is a key component of photochemical smog, can have very deleterious effects. Thus, although all ozone molecules are chemically identical, their environmental impacts depend strongly on where they are situated in the atmosphere. In this entry, we concentrate on ozone in the stratosphere. In recent decades, anthropogenic emissions have severely compromised the stability of the stratospheric ozone layer. Stratospheric ozone depletion is a global problem that has its most dramatic manifestation in the polar regions,

STRATOSPHERIC OZONE

especially over Antarctica, where an “ozone hole” has formed each spring for the last few decades.

Ozone production and distribution The abundance of ozone is relatively low, with peak mixing ratios (the ratio of the concentration of ozone to that of air) of only about 10–12 parts per million by volume (ppmv) (e.g., McPeters et al., 2007). Ozone is commonly measured in terms of its column abundance, the vertically integrated density overhead at a given location. Total ozone column abundances are traditionally reported in Dobson units (DU, named for G.M.B. Dobson, who in the 1920s pioneered routine measurements of ozone column amounts), which correspond to the thickness in millicentimeters that the ozone column would have if all the gas were at standard temperature and pressure. Typical ozone column abundances range from 200 to 500 DU, depending on the latitude and season (e.g., Dessler, 2000). Despite its low concentration, ozone is arguably the most important stratospheric constituent. It is an extremely efficient absorber in the ultraviolet (UV); as a result, an intact ozone layer prevents practically all solar radiation at wavelengths shorter than 300 nm from reaching the Earth’s surface. Excessive amounts of UV-B (280–315 nm) radiation are particularly injurious to humans and other animals, terrestrial and aquatic ecosystems, and physical materials. It is therefore not an exaggeration to say that without Earth’s protective ozone shield, life as we know it could not have evolved. The distribution of stratospheric ozone is controlled by the interplay between chemistry and transport by the prevailing winds. Ozone is produced in the stratosphere when molecular oxygen (O2) is photolyzed (i.e., split apart) by solar UV radiation. The resulting two oxygen atoms (O) each rapidly combine with other oxygen molecules to produce two ozone molecules (O3). The overall result is that, in the presence of sunlight, three oxygen molecules react to form two ozone molecules. Because these reactions take place wherever UV radiation is available, the greatest ozone production occurs in the tropical stratosphere. Interestingly, however, the largest total column ozone amounts are found at middle and high latitudes in late winter and early spring (after several months of total or near-total darkness), whereas the tropics actually contain the smallest column ozone amounts. This lack of correlation between the regions of highest production and those of highest abundance underscores the importance of transport in determining the overall morphology of stratospheric ozone. Winds circulate air in the stratosphere, with rising motion in the tropics, poleward motion aloft, and sinking motion in the polar regions. Known as the Brewer-Dobson circulation, this flow carries air rich in ozone away from the tropics and toward the poles. Although ozone molecules are continuously produced, they are also continuously destroyed in chemical reactions

797

with oxygen atoms and oxides of nitrogen, hydrogen, chlorine, and bromine. These reactions are termed “catalytic” because the species that initiate ozone destruction are not themselves consumed in the process, allowing the cycle to repeat many times. The net effect of the catalytic cycle is the conversion of O3 molecules into O2 molecules. The catalysts in these reactions are free radicals (i.e., molecules with an unpaired electron, making them highly reactive). The radicals are derived from long-lived species (which in most cases have both natural and anthropogenic sources) transported from the troposphere into the stratosphere by the large-scale circulation. Typical mixing ratios of the oxides of nitrogen are in the part per billion (ppbv) range, while those of the reactive hydrogen, chlorine, and bromine species are even lower. Thus, ozone abundances are in large part regulated by radical species present in the stratosphere in even more minute quantities. The atmosphere normally maintains a balance, or steady state, between photochemical production and catalytic destruction of ozone, with the total level in the stratosphere remaining fairly constant. Small variations in total ozone occur because of seasonal changes in the strength of atmospheric transport, the intensity of incident sunlight, stratospheric temperatures, and other factors. Natural variations can arise from other sources as well; for example, fluctuations in UV radiation over the course of the 11 year solar cycle alter ozone production rates slightly (1  2 %) (e.g., Brasseur and Solomon, 2005).

Chlorine-catalyzed ozone depletion in the polar lower stratosphere Over the last several decades, the natural balance between ozone formation and destruction has been perturbed as manufactured chemicals have entered the stratosphere. It is now understood that the severe springtime depletion of the stratospheric ozone layer over Antarctica – the so-called ozone hole – is caused by chlorine and bromine chemistry (known collectively as halogen chemistry) (e.g., Solomon, 1999; Fahey, 2007). The primary source of stratospheric chlorine is chlorofluorocarbons (CFCs), chemical compounds composed of chlorine, fluorine, and carbon. CFCs were introduced in the 1930s and used in a variety of industrial and commercial applications, including as refrigerants, solvents, and aerosol propellants, as well as in foam packaging. CFCs were used so extensively because they have low boiling points and are stable, nonflammable, and unreactive in the lower atmosphere. Other halogen source gases include carbon tetrachloride, methyl chloride, and methyl chloroform, as well as methyl bromide and the bromine-containing “halons” used as fire retardants. Chlorine and bromine are also released at ground level through other human activities and natural processes. Common examples include emission of chlorine gases from swimming pools, wastewater treatment, and household bleach, as well as from volcanic ejecta; chlorine is also present in sea salt produced by evaporation of

798

STRATOSPHERIC OZONE

ocean spray. In all of these cases, however, the chlorine is in forms that are highly reactive or soluble and are thus washed out in rain or ice in the lower regions of the atmosphere before they can be delivered to the stratosphere. In contrast, the very properties that made CFCs so useful turned out to pose a grave environmental threat. Because they are inert and insoluble in water, CFCs spread throughout the troposphere in a matter of months and are then carried along as the overturning circulation lofts air into the upper stratosphere. There, above the protection afforded by the bulk of the ozone layer, CFCs are broken down by high-energy solar UV radiation, liberating chlorine atoms. The transit time between the troposphere and the stratosphere is about 5 years. During each circuit through the Brewer-Dobson circulation, only about 10 % of the mass of the troposphere is exchanged with the upper stratosphere, and only a fraction of CFC molecules is decomposed; thus, most CFCs have long atmospheric lifetimes (about 50–500 years) (e.g., Brasseur and Solomon, 2005). In the mid-1970s, it was recognized that chlorine released from CFCs could destroy stratospheric ozone in a catalytic cycle (Molina and Rowland, 1974), a notion that generated enormous scientific interest. For their work in helping to elucidate the sensitivity of the ozone layer to anthropogenic emissions, Mario Molina and F. Sherwood Rowland were awarded (along with Paul Crutzen) the 1995 Nobel Prize in Chemistry. Even after the connection with chlorine from CFCs was made, however, model calculations indicated that ozone loss would be gradual and moderate (5  20 %), limited mainly to altitudes in the upper stratosphere (near 40 km) above the peak in the ozone profile. Consequently, scientists and the public alike were shocked by the news in 1985 that ground-based measurements from the British Antarctic Survey station at Halley Bay, Antarctica, had revealed a dramatic downturn in October average column ozone over the previous few years (Farman et al., 1985). NASA satellite measurements subsequently confirmed the rapid decline in total ozone and demonstrated that the phenomenon extended over a vast geographic region roughly encompassing the Antarctic continent (Stolarski et al., 1986). Vertical profile measurements from balloon-borne ozonesondes showed the loss to be largely confined to the lower stratosphere, with virtually all ozone in the layer between about 14 and 20 km removed within a period of 4–6 weeks (Hofmann et al., 1987; Solomon, 1999; Fahey, 2007). Such severe loss in the lower stratosphere, where ozone mixing ratios are normally largest, leads to a reduction in total column ozone of as much as 60  70 %. Early models failed to predict the development of the ozone hole because they focused exclusively on homogeneous (gas-phase) chemistry. A series of coordinated field campaigns to Antarctica, together with laboratory experiments and modeling studies, provided the information necessary to establish the cause of the massive ozone loss within a few years of its discovery. The first step is the formation of the “polar vortex.” In the continuous

darkness of winter polar night, temperatures drop very low, leading to a large temperature gradient near the polar terminator (the line delimiting the unilluminated region), which in turn induces a steep pressure gradient. The resulting flow is deflected by the Coriolis force, creating a band of intense westerly (i.e., eastward) winds encircling the pole. This strong wind jet provides a barrier to mixing, effectively isolating the region poleward of it, termed the polar vortex, from lower-latitude air. In the very low temperatures inside the lower stratospheric winter polar vortex, water vapor (H2O) and nitric acid (HNO3) condense to form polar stratospheric clouds (PSCs). PSCs were long known to occur – for example, they were described in 1911 by noted Antarctic explorer Robert Falcon Scott and his party – but they remained little more than scientific curiosities until the ozone hole was reported. PSCs are a crucial link in the chain leading to severe ozone destruction. Their particles provide surfaces on which heterogeneous reactions (i.e., reactions in which one reactant is absorbed onto/into the particle while the other remains in the gas phase) can take place very rapidly. These reactions convert chlorine from relatively benign “reservoir” species such as hydrogen chloride (HCl) and chlorine nitrate (ClONO2) to intermediate forms such as molecular chlorine (Cl2) and hypochlorous acid (HOCl) that are quickly photolyzed. In this manner, vortex air is primed for ozone destruction during the cold, dark winter months. When sunlight returns to the polar regions in spring, the intermediate species are rapidly broken apart to produce highly reactive ozone-destroying forms. The predominant form of reactive chlorine in the stratosphere – the “smoking gun” that signals chlorinecatalyzed ozone destruction – is the chlorine monoxide radical, ClO. In the mid-1980s, two sets of measurements cemented the link between elevated levels of ClO and severe ozone destruction in the lower stratosphere. First, ground-based microwave emission measurements from McMurdo Station, Antarctica, indicated greatly enhanced ClO near 20 km, with mixing ratios of more than 1 ppbv, as much as two orders of magnitude larger than predicted, based on standard gas-phase photochemistry (de Zafra et al., 1987). Second, in situ measurements from the NASA ER-2 high-altitude research aircraft (a converted U2 spy plane) revealed an abrupt increase in ClO closely correlated with a steep decline in ozone along the flight track inside the Antarctic polar vortex (Anderson et al., 1989). The large enhancements in ClO found by these studies could only have arisen through the near-total transformation of chlorine from reservoir to reactive species, and the spatial and temporal variation of the ClO/O3 anticorrelation confirmed that ozone loss was indeed driven by chlorine chemistry. It was not until the launch of the Microwave Limb Sounder (MLS) on NASA’s Upper Atmosphere Research Satellite (UARS) in September 1991 that researchers were able to map the full threedimensional distribution of ClO in the stratosphere and

STRATOSPHERIC OZONE

track its seasonal evolution around the globe. UARS MLS found an almost exact coincidence between enhanced ClO and depleted ozone throughout the Antarctic vortex (Waters et al., 1993). UARS MLS measurements also verified that ClO enhancement in the Arctic polar vortex can be comparable, in terms of both magnitude and areal extent, to that in the Antarctic. Nevertheless, the Arctic has so far been spared an annual ozone hole. This interhemispheric asymmetry in the severity of ozone destruction is attributable to differences in meteorological conditions that arise in large part because of differences in the distribution of oceans and continents, in particular mountains at middle and high latitudes. Winds flowing over mountains generate waves that propagate into the stratosphere and disturb the polar vortex. As a result, the Arctic vortex exhibits much greater interannual variability than its southern counterpart and is typically warmer (by 10–20 K), weaker, more permeable, smaller, and shorter lived, leading to fewer, less persistent PSCs. In the cold, isolated Antarctic vortex, PSC particles grow sufficiently large that they undergo appreciable gravitational sedimentation. As they fall, they irreversibly remove from the stratosphere the HNO3 (and H2O) sequestered in them in a process known as denitrification (dehydration). Unlike the Antarctic, the Arctic generally experiences minimal denitrification. Photolysis of gasphase HNO3 in springtime produces nitrogen dioxide (NO2), which combines with ClO, to reform the reservoir ClONO2. This reaction is the primary pathway for chlorine deactivation in the Arctic. For all of these reasons, the horizontal and vertical extent, duration, and magnitude of ClO enhancement are typically much smaller in the Arctic than in the Antarctic, and ozone loss, although it does occur in many winters, is correspondingly smaller.

Midlatitude ozone depletion Though much less dramatic than those observed in the polar regions, statistically significant downward trends have also been detected in ozone at middle latitudes (e.g., Solomon, 1999). A substantial fraction of the midlatitude depletion is attributable to dilution with processed air exported from the decaying vortex at the end of winter, especially in the Southern Hemisphere, where polar ozone loss is severe. Another factor is in situ chemical destruction. Heterogeneous reactions akin to those that activate chlorine on PSCs also take place (albeit more slowly) on sulfate aerosol droplets, which are ubiquitous throughout the lower stratosphere even under background conditions. Major volcanic eruptions, which inject large amounts of sulfur gases directly into the stratosphere, can greatly enhance stratospheric aerosol surface area density and intensify ozone loss. In fact, substantial decreases in column ozone (on the order of 5  10 %) were recorded over large regions of the globe following the eruption of Mt. Pinatubo in the Philippines in June 1991 (Solomon, 1999; Dessler, 2000). It is important to note, however, that the eruption of Mt. Pinatubo would

799

not have had a significant impact on global ozone had stratospheric chlorine loading not far exceeded natural levels.

Ozone recovery and climate change At its peak around the year 2000 (Fahey, 2007), stratospheric chlorine reached levels more than six times its natural abundance. Since then, however, chlorine has been declining in response to regulations enacted under an international agreement known as the Montreal Protocol on Substances that Deplete the Ozone Layer. The 1987 Protocol, and subsequent more stringent Amendments and Adjustments, established legally binding controls on the production of halogen source gases. Eventually ratified by over 190 nations, these provisions averted a global environmental calamity. The stratospheric ozone layer has now stabilized, but more than 20 years after the signing of the original Protocol, large ozone holes continue to occur, attaining an areal extent of about 25 million km2 (roughly twice the surface area of the Antarctic continent) in an average year. Because of the long residence times of CFCs, it will take several more decades to cleanse the stratosphere of excess chlorine and curtail ozone depletion. Assuming continued worldwide compliance with the Protocol, stratospheric chlorine and bromine loading are expected to return to pre-1980 values by about 2050 (Fahey, 2007). Recovery of the ozone layer does not depend solely on diminishing anthropogenic halogen abundances, however. Earth’s changing climate may have a significant effect: While the surface of the planet is warming in response to increasing greenhouse gases, the stratosphere is cooling. PSC lifetimes are prolonged in a colder lower stratosphere, exacerbating polar ozone depletion. This could delay recovery and trigger development of an ozone hole in the Arctic, where wintertime temperatures often hover near PSC formation thresholds. But assessment of the sensitivity of ozone chemistry to temperature is complicated by concomitant changes in dynamics and transport. Greenhouse-gas-induced acceleration of the Brewer-Dobson circulation is expected to alter the concentrations of water vapor and trace gases in the stratosphere, which in turn influence ozone. Moreover, lower temperatures will slow catalytic ozone destruction cycles outside of the polar lower stratosphere, hastening recovery in those regions. On the other hand, higher surface temperatures may modify emission rates of some naturally occurring halogen source gases. Increased abundances of other pollutants arising from human activities may also perturb the balance between global ozone production and loss. Separating the effects of these competing factors is challenging, making future changes in ozone difficult to predict. Current state-of-the-art models show a range of projections, with most indicating full recovery around the middle of the twenty-first century at middle latitudes and slightly later in the Antarctic (Fahey, 2007). It is important to note, however, that because of changes in

800

SUBSIDENCE

climate and other parameters, the atmosphere is unlikely to return to a state identical to that before the onset of anthropogenic ozone depletion.

Conclusion The last few decades have seen enormous changes in the stratospheric ozone layer so vital to life on Earth. The ozone hole serves as a cautionary tale of how an apparently harmless class of chemical compounds can prove to have disastrous consequences and thus highlights how human activities can change the natural state of our atmosphere in unintended ways. But it is also a story of hope that shows how the scientific community, national governments, and industry can identify a global environmental threat and work together to mitigate it. It is imperative that we as a society continue to monitor the stability of the ozone shield to see how the story ends in light of another environmental challenge: anthropogenic climate change. Acknowledgment This research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the NASA. Bibliography Anderson, J. G., Brune, W. H., and Proffitt, M. H., 1989. Ozone destruction by chlorine radicals within the Antarctic vortex: the spatial and temporal evolution of ClO  O3 anticorrelation based on in situ ER-2 data. Journal of Geophysical Research, 94, 11465. Brasseur, G., and Solomon, S., 2005. Aeronomy of the Middle Atmosphere, 3rd revised and enlarged edn. Dordrecht: Springer. de Zafra, R. L., Jaramillo, M., Parrish, A., Solomon, P., Connor, B., and Barrett, J., 1987. High concentrations of chlorine monoxide at low latitudes in the Antarctic spring stratosphere: diurnal variation. Nature, 328, 408. Dessler, A. E., 2000. The Chemistry and Physics of Stratospheric Ozone. San Diego: Academic. Fahey, D. W., 2007. Twenty questions and answers about the ozone layer: 2006 update, scientific assessment of ozone depletion: 2006. Geneva: World Meteorological Organization. Farman, J. C., Gardiner, B. G., and Shanklin, J. D., 1985. Large losses of total ozone in Antarctica reveal seasonal ClOx/NOx interaction. Nature, 315, 207. Hofmann, D. J., Harder, J. W., Rolf, S. R., and Rosen, J. M., 1987. Balloon-borne observations of the development and vertical structure of the Antarctic ozone hole in 1986. Nature, 326, 59. McPeters, R. D., Labow, G. J., and Logan, J. A., 2007. Ozone climatological profiles for satellite retrieval algorithms. Journal of Geophysical Research, 112, D05308, doi:10.1029/ 2005JD006823. Molina, M. J., and Rowland, F. S., 1974. Stratospheric sink for chlorofluoromethanes: chlorine atom catalyzed destruction of ozone. Nature, 249, 810. Solomon, S., 1999. Stratospheric ozone depletion: a review of concepts and history. Reviews of Geophysics, 37, 275. Stolarski, R. S., Krueger, A. J., Schoeberl, M. R., McPeters, R. D., Newman, P. A., and Alpert, J. C., 1986. Nimbus 7 satellite measurements of the springtime Antarctic ozone decrease. Nature, 322, 808.

Waters, J. W., Froidevaux, L., Read, W. G., Manney, G. L., Elson, L. S., Flower, D. A., Jarnot, R. F., and Harwood, R. S., 1993. Stratospheric ClO and ozone from the microwave limb sounder on the upper atmosphere research satellite. Nature, 362, 597.

Cross-references Trace Gases, Stratosphere, and Mesosphere

SUBSIDENCE Stuart Marsh1 and Martin Culshaw2,3 1 Nottingham Geospatial Institute, The University of Nottingham, Nottingham, UK 2 Honorary Research Associate, British Geological Survey, Keyworth, Nottingham, UK 3 Honorary Visiting Professor, School of Civil Engineering, University of Birmingham, Edgbaston, Birmingham, UK

Synonyms Collapse; Compaction; Ground motion; Ground movement; Settlement Definition The term “subsidence” is often misunderstood and misused, both by specialist geoscientists and by the general public. As such, the term should be avoided and more precise terms used instead. Strictly, the term has been used to mean movement of the ground surface in a generally vertical, downward direction, without reference to the areal extent of movement, the rate at which movement takes place, or the cause of the movement. For example, Keller (1985) defined subsidence as “Sinking, settling, or other lowering of parts of the crust of the earth.” However, Johnson and DeGraff (1988) took some of the factors into account by defining subsidence as “. . . displacement of the ground surface vertically over a broad region or at localized areas. It may be either a gradual lowering or a collapse.” In the public’s mind, subsidence is sometimes associated with a particular cause, for example, mining subsidence, without an understanding that there are many other causes. As remote sensing platform instrumentation can measure both subsidence (sensu stricto) and “negative” subsidence (rise), it is proposed, here, that the term “subsidence” should not be used and be replaced by the term “ground movement.” The definition of this term is proposed as: “The movement of the ground, however caused, in any direction, either vertically up or down, or laterally, or components of these. The areas involved can range from less than 1 m2 to several square kilometers. The amount of movement can vary from a few millimeters to more than a kilometer. The length of time for the movement to be completed can vary from a few seconds to several years.”

SUBSIDENCE

The synonymous term “ground motion” is best reserved for the quantitative amount of ground movement observed and measured repeatedly by remote sensing platform instrumentation over a period of time. Hence, a rate of movement can be determined for different time periods. The movement measured is likely to be in the direction of the “line of sight” of the satellite instrumentation, which may not be vertical.

Importance of ground movement information Ground movement causes death, injury, damage to buildings, structures, and infrastructure, and, hence, financial loss. It occurs on every continent and offshore. However, the amount of injury and loss depends upon the rate of ground movement and its magnitude. Sometimes, rapid and large ground movements will be preceded by slower and smaller movements that can be monitored. Examples of this include the growth of lava domes prior to an eruption and ground level rise/fall prior to an earthquake. Even where ground movement takes place slowly and the amount of movement is only a few tens of millimeters, the financial cost can be great. In the United Kingdom (UK), damage to buildings caused by ground movement can be insured against, and in the last 20 years, losses have averaged between £3 and 400 million per year. Most of these losses were the result of the shrinkage of Mesozoic and Tertiary clay soils in southeast and central England. This does not represent the total loss due to ground movement as losses from ground movement caused by coal mining are not included in this total. (These losses are covered by a public body called the Coal Authority). In France, where ground movement is also covered by insurance (though in a slightly different form), losses averaged over €300 million per year between 1989 and 2003. Again, the cause was the presence of expansive clay soils (Cheetham, 2008). Almost every country in the world has several geohazards that cause ground movement activity for at least part of the year. To summarize, large-magnitude, rapid ground movements tend to cause both significant human and financial loss. On the other hand, small-magnitude, slow ground movements cause little human loss but can still result in substantial financial loss. As a result, the acquisition of ground movement information is very important if losses are to be anticipated. Research into the prediction of large-magnitude events has yet to identify ways of reliably predicting the location, magnitude, and timing of earthquakes and volcanic eruptions. However, frequent monitoring of ground movement, particularly using remote sensing, is likely to provide a major contribution to the anticipation of these events. Smaller and slower ground movements can also be monitored remotely, though the processes are better understood and hence are more predictable. Many of these processes are related to climatic conditions, and hence, prediction of large and/or intense rainfall events and drought is important. Predicted temperature rise is also important, for example, in relation to the thawing of thermokarst. In the future,

801

long-term monitoring of ground movement, however caused, is likely to become increasingly important in reducing losses arising from ground movement events.

Description of ground movement phenomena Ground movement can have a wide range of natural and anthropogenic causes. Bell (2007) specifically referred to subsidence caused by mining, groundwater extraction, and dissolution of more soluble rocks, while Johnson and DeGraff (1988) listed dissolution, collapse of lava tube roofs, groundwater, oil or gas extraction, mining and swelling and shrinking clay soils as potential causes. Bell (1998) added fault movement to the identified causes and settlement as a low-rate and low-magnitude form of subsidence. To the lists of these authors can be added isostatic movement, seismic activity, volcanic activity, mass movement (including landsliding), soil fabric collapse, soil compressibility, and settlement of waste material. Natural causes of ground movement Most of the causes of ground movement discussed below are explained in detail in engineering, geological, and geomorphological textbooks (e.g., Bell, 1999; Keller et al., 2008; McCall et al., 1992; Nuhfer et al., 1993). Earth tides: Movements of the Earth’s surface caused by the gravitational effects of the moon and the sun. These can take place diurnally or over longer periods. Movements are both vertical and horizontal and may amount to up to more than a 100 mm. Isostatic movement Readjustment to a state of equilibrium between the Earth’s lithosphere and asthenosphere. Lack of equilibrium is brought about by factors such as ice loading during glacial periods and movement of tectonic plates. Ground level rise from glacial unloading may amount to several meters over periods of thousands of years. Seismic activity While earthquakes, particularly large ones, can involve considerable and rapid ground movement (“strong ground motion”), usually associated with fault activity, slower- and longer-term ground movements may take place prior to, or as a result of, an earthquake. The “strong ground motions” associated with the earthquake occurrence may have no net ground movement effect. However, permanent uplift or depression of the ground surface over several square kilometers is known. Total change in ground level may amount to a meter, or more. For example, the Limón-Telire earthquake in Costa Rica in April 1991 caused between 0.3 and 1.9 m of uplift in the Limón area on the Caribbean coast. Volcanic activity The movement of volcanic magma toward the surface frequently leads to the uplift and subsidence of the area around an active volcano. Bell (1999) noted that the

802

SUBSIDENCE

summit of Kilauea, Hawaii, has risen by nearly a meter prior to eruption. However, such changes in ground level are not reliable indicators of eruption. Dissolution Rocks such as salt, gypsum, chalk, and limestone are susceptible to dissolution that can cause voids to be formed not far below the ground surface. If these collapse, ground movement is likely to result. The amount of movement depends upon the size of the void; the rate of movement (collapse) is likely to be high. Sometimes dissolution leads to the washing down of overlying soil particles into the dissolving rock, leading to gradual ground movement. The ground movement features are known as sinkholes or dolines and may range in diameter from a few meters to, exceptionally, more than 1,000 m (Waltham et al., 2005). Similar features can also be caused by the collapse of lava tubes where molten lava has run downslope leaving a solidified cover at the surface. Mass movement Landslides, flows, and falls involve the downward and lateral movement of soil, rock, or debris as a result of ground failure of a slope (Turner and Schuster, 1996). Movement can involve thousands of tonnes of material moving distances ranging from a few meters to, exceptionally, more than a kilometer. Melting of permafrost (thermokarst) As frozen ground thaws, settlement will occur as water is expelled by the overburden pressure and any artificial load. If the soil is free draining, the amount of settlement is likely to be greater. The amount of sinking of the ground movement will depend on the soils present and the rate of thawing. Heave can take place if the soil refreezes. Shrinkage of peat As peat dries out, it shrinks causing a lowering of the ground surface. This can amount to several meters. For example, in the Fens of eastern England, long-term drainage of the peat for agricultural purposes has caused a lowering of the ground surface of close to 10 m. Between 10 % and 75 % of the peat volume can be lost, and the change is permanent (Bell, 1999).

a process that takes place gradually as soils age and as natural processes (“weathering”) act upon them. Under natural conditions, settlement will be slow and will amount to a few millimeters or tens of millimeters. Soil fabric collapse Some soils, particularly windblown silts and fine sands (loess), have a “metastable” microfabric caused by the way in which they were deposited. Quartz particles are separated by “bridges” of clay-sized minerals. The soils have high void ratios and low densities. On wetting, the microfabric collapses due to a reduction in the bonds between particles causing a reduction in void ratio and increase in density. The amount of volume change can range from 1 % to over 20 %. In extreme cases, settlements in excess of a meter have been observed (Bell, 1999).

Anthropogenic (artificial) causes of ground movement Mining Mining involves the removal of an economic solid mineral either at the ground surface (pits and quarries) or underground. Partial extraction methods involve leaving some of the mineral to support the roof or walls of the mine. However, over time, the supporting pillars may deteriorate and, ultimately, fail resulting in collapse of the mine workings and lowering of the ground surface. Total extraction methods involve the removal of almost all of the mineral and the controlled collapse of the overlying material into the void created. This results in lowering of the ground surface. The advantage is that the lowering of the ground surface is controlled and predictable. The amount of lowering may range from a few tens of millimeters to ten, or so, meters. Abandoned mine entrances (shafts and adits) may also fail creating either holes at the surface or its lowering.

Swelling and shrinkage of clay soils The main cause of swelling and shrinkage of clay soils is the presence of clay minerals that experience significant volume changes as their moisture content changes. Montmorillonite is a particularly susceptible clay mineral. Both shrinkage (causing downward movement of the ground surface) and swelling (causing upward movement) can result in damage to buildings. Amounts of movement range from a few millimeters to many tens of millimeters.

Fluid withdrawal Shrinkage of peat has been described above. However, subsidence, and consequent lowering of the ground surface, can take place when a fluid, usually groundwater, oil, or brine, is withdrawn from the ground. The overburden pressure is initially borne by the solid particles and the fluid in the pore space. As fluid is withdrawn, the load is transferred to the solid particles, increasing the so-called effective overburden pressure. This, in turn, causes consolidation of the sediment. The amount of subsidence depends upon the increase in effective pressure and the thickness and nature of the sedimentary deposit. In the San Joaquin Valley of California, USA, 9 m of subsidence occurred between 1925 and 1977 due to groundwater extraction for agriculture. Recharge of the fluid into the ground can reduce the rate of subsidence and may result in some recovery, with the ground surface rising. However, full recovery cannot be achieved.

Soil consolidation Soils consolidate (reduce in volume) due to overburden pressure and any artificially imposed load. This is

Fault reactivation Geological faults can be locations where ground movement caused by mining is concentrated. This reactivation

SUBSIDENCE

of the faults may take place during mining and may continue for years afterward. Movements along a fault greater than 2 m can occur, particularly where several seams have been extracted. Fault reactivation is more associated with shallow mine workings (Bell and Donnelly, 2006). Settlement of waste material Though modern waste disposal sites are usually carefully controlled, both in terms of their content and of how the site is constructed, older sites and those where controls are limited may contain mixtures of materials that have not been compacted. Over time, as the waste material alters chemically and as a result of the overburden pressure from dumping of later waste, the tip will settle, lowering the level of its surface. Amounts of settlement may range from millimeters to more than a meter, depending on the nature of the fill material and the degree of control of fluids entering the tip. Engineering compaction Compaction is a civil engineering process by which soil is densified by expelling air from the void space to achieve a stronger material. Usually, this involves artificially placed material (“fill”) used to construct embankments, dams, and other structures. The amount of lowering of the surface might range from millimeters to several tens of millimeters.

Measuring ground movement using remote sensing data Ground movement can be measured using a variety of spaceborne, airborne, and ground-based remote sensing systems. Measuring ground movement requires a minimum of two data acquisitions over a study area: one to establish the baseline ground position and another to capture changes to it after a period of movement. The necessary time separation depends on a trade-off between the rate of movement and the ability of the sensor to resolve the ground movement. This measurement will capture the total ground movement in a specific time period. It can be extended via repeated data acquisitions to cover more time periods, leading to the concept of monitoring ground movement. Such monitoring allows us to measure ground motion and changes to its rate over time. The cost of monitoring is higher than that of measuring but may be justified by the impact of the ground movement on lives and economies. Ground movement is typically measured on a perpixel basis and in the line of site of the sensor. Of course, the ground movement may actually be vertical, horizontal, or a combination of both. There are various ways of accounting for the mismatch between the geometry of the movement and the viewing geometry of the sensor measuring it, to give a true measurement of ground movement in x, y, and z.

803

Spaceborne measurement techniques Satellite altimetry Altimeters are active remote sensing instruments. That is, they operate by sending a signal from the satellite to the surface under study and measuring how long it takes to return to the sensor. The signal may be radar or light from a laser. It is used to build up an image of the surface geometry. Changes to that geometry could, in principle, be measured via a second acquisition. However, the spatial resolution of such instruments is typically insufficient for application to ground movement. They are more commonly used in oceanography applications (Cheney et al., 1986). Monoscopic optical satellites It is possible to work out the horizontal component of ground movement using two passively acquired optical images. Once they are geometrically rectified and registered to each other precisely, image analysis software can be used to assess whether features in the first image, such as prominent outcrops, are in a different location in the second. This gives a simple measurement of horizontal ground movement. It is not widely used, because it cannot capture the vertical component and so has limited application in a 3D science like geology. Stereoscopic optical satellites Stereoscopic optical imagery from satellites like Landsat and SPOT can be used to produce a digital terrain model (DTM) with a resolution on the order of meters to tens of meters in a process known as photogrammetry (Toutin, 2001). The technique uses two images of the same area taken from adjacent orbits or two positions along track. These are compared using photogrammetric software, which takes advantage of parallax between the two viewing geometries to work out the surface geometry, in a similar way to the brain processing images from two eyes to generate stereoscopic vision. If the process is repeated, a second DTM results, and the two can be differenced to quantify changes in the DTM between acquisitions. These changes measure the ground movement. The main constraint on this technique is the need for two cloud-free acquisitions. Satellite radargrammetry The same principle of using two viewing geometries to help extract the geometry of the surface being viewed can be extended to active radar data (Toutin and Chénier, 2009). Radar data penetrate cloud cover and can even be used at night, so the technique compensates for two limitations of optical data. It can be used to generate DTMs at similar resolutions to optical satellites of areas under perennial cloud cover, where optical sensors cannot image the ground easily. The main constraint is that the algorithms are complex for most remote sensing practitioners to implement and rarely feature in off-the-shelf software packages. So, the technique is not widely applied.

804

SUBSIDENCE

Satellite radar interferometry InSAR A more common way of exploiting satellite radar data to extract changes in digital terrain models is InSAR (Ferretti et al., 1999). This also exploits two images, but separated in time rather than space. The technique relies on measuring a change between acquisitions in the number of radar waves from a reflector on the ground to the sensor in orbit. The simplest analogy is the radar gun used by police to measure the speed of approaching vehicles. In practice, the two images are interfered and a series of fringes result, whose width is governed by the radar wavelength. These can be thought of as contours of ground motion. Because the technique measures only the change in ground position, an existing terrain model is used for the baseline digital terrain models, against which to measure the change. This technique is powerful because it directly measures the change in ground position as a fraction of the radar’s wavelength, irrespective of its spatial resolution. So, while the spatial resolution may be 25 m, the sensitivity to ground movement is typically on the order of a few cm, depending on the wavelength used. This matches closely the order of magnitude of many ground movements, and so the technique has been applied to a wide range of such phenomena, including seismic activity, volcanic uplift, swelling and shrinking clays, and groundwater and mineral extraction. It is even more powerful when combined with a geodetic surveying technique such as GPS. InSAR gives a wide area, synoptic view of motion restricted to specific time periods. GPS is a spatially discrete, sparse measurement but can be acquired almost constantly. It can calibrate InSAR and partially fill gaps between InSAR acquisitions. InSAR, inevitably, has several significant weaknesses. Between acquisitions, surface characteristics other than the ground’s position can change; if this change is large, the two images become incomparable (they are referred to as having low coherence) and the technique does not work. Changing vegetation is a particular challenge for short-wavelength, C-band radar, and so InSAR works best in arid countries and urban areas. Longer wavelengths such as L band are less affected by vegetation, but they are also less sensitive to ground motion. Changes occur in atmospheric properties, too, and these must be removed for the technique to work. Finally, properly calibrated measurements are still in the line of sight of the sensor; this matters more for radar data than optical data, because radars are acquired with a look angle of less than 90 to the horizontal. So, the result must be transformed into x, y, and z coordinates, either by using near-contemporary image pairs from both ascending and descending satellite passes with different geometries or by integrating InSAR with data like GPS. Persistent Scatterer Interferometry PSInSAR To get around the coherence limitation, a technique known as PSInSAR exploits strong radar reflectors such as building corners and metal objects that persist over time (i.e., are coherent) (Ferretti et al., 2000). In this technique,

several tens of images are acquired over a decade or more and processed together to extract these persistent scatterers (PS), which are then the focus for the interferometric processing. Rather than fringes over wide areas, this results in ground movement histories for each PS identified. These can then be gridded and interpolated to make an image, typically shown as mean annual velocity or viewed as graphs of ground movement against time for individual PS. Statistics dictate that the use of large numbers of images leads to enhanced resolution, on the order of 2 mm/year for C-band data. Constraints are that the measurement is again in the line of sight of the radar, atmospheric changes must be screened, orbital effects must be removed, and the technique cannot measure ground movements the size of multiple radar wavelengths, as this leads to aliasing. So, applications like mining with larger ground movements require caution. There is also no guarantee that PS will occur where investigators want them to. This can be overcome by placing metal corner reflectors or active transponders in the study area on the features of interest, such as reservoirs or bridges (Xia et al., 2002), to act as well-located persistent scatterers.

Airborne measurement techniques All of the above techniques can be applied from an airborne platform. The biggest advantages that this brings are increased spatial resolution and the ability to target good weather windows. A particularly powerful technique is laser ranging, or lidar (Liu, 2008). The increased resolution that results from flying much closer to the target means that cm resolution lidar DTMs can be acquired and compared. Digital terrain models at tens of cm resolution are routinely generated from stereo aerial photography (Fabris and Pesci, 2005), and repeat airborne surveys can be used to measure changes in DTM in exactly the same way as for stereo optical satellite data. Radar interferometry can also be performed from an airborne platform to give, and compare, 1 m or better DTMs (Madsen et al., 1993). The natural geometric variability of flight lines flown through the lower atmosphere is greater than that for the orbits of satellites, and so this places a high premium on geocorrection and registration. But, at these high resolutions, the major constraints are vegetation and buildings, because the DTM measures the first surface a signal is returned from, such as the top of the tree canopy or the roofs of buildings. Corrections must be applied to remove these features and produce a bare earth model of the ground. Buildings are predictable enough to be removed cleanly, but trees are variable in height and much harder to deal with. Of the three techniques, lidar comes into its own here, because lidar with sufficiently high point densities that record the entire spectrum of returns ensures that at least some of the returns come from the ground, in between the trees. By always taking the last return, a bare earth model can be constructed even in fairly heavily forested terrain (Liu, 2008).

SUBSIDENCE

Ground-based measurement techniques All of the above techniques can also be applied from a ground-based platform. The biggest advantages that this brings are further increases in spatial resolution and even greater flexibility to target weather windows. Such highresolution surveys naturally cover smaller study areas and so are best targeted on ground motion of known extent on a site-specific basis. Terrestrial lidar has been used extensively to monitor slope stability (Jaboyedoff et al., 2012), with the further advantage that vertical surfaces like cliffs that are not seen in airborne or satellite datasets can be imaged easily. When combined with photography, the models produced can provide detailed and realistic views of the surfaces under investigation. This may allow more to be said about the causes of ground movement, rather than simply measuring the motion. Terrestrial radar interferometry is starting to be used in similar ways; its ability to penetrate clouds means that it has particular utility in volcano monitoring (Rödelsperger et al., 2010). One drawback is that, even for restricted study areas, these ground-based sensors generate very large datasets that have a large processing, manipulation, and storage overhead. Interpreting ground movement measured in remote sensing data Remote sensing measures the movement of the ground at the surface. As we have seen, this can have many and even multiple causes. For example, an average ground movement map from PSInSAR over a typical city shows motion that has a variety of causes. There may be settlement of alluvium along major river valleys, subsidence above tunneling and subsidence due to groundwater abstraction. In addition there may be evidence of tectonic motion and almost certainly some swelling and shrinking clays. The remote sensing technique does not discriminate, but merely measures the ground movement in each case. The interpretation of the data requires interaction between the remote sensing expert and the geoscientist, if the causes of motion are to be fully understood. In fact, this interaction is needed from the outset, so the best technique can be selected to measure the particular type of movement that can be expected to be found in the study area. Study of the ground movement should also go beyond measurement of the motion. All of these remote sensing datasets tell us something about the geology exposed at the surface and can be interpreted in terms of faulting, folding, and lithology, for example. Such interpretation helps us understand the observed ground movement. In addition, it may be necessary to use other types of remote sensing that do not measure the ground motion at all, but do provide further insights into the geology, such as multior hyperspectral optical data for lithology and mineralogy. In addition, the remote sensing should be used in combination with other techniques that are not restricted to measuring surface phenomena, such as geophysics, in order to understand ground movement in the full three dimensions in which it occurs.

805

Summary and conclusions Ground movements occur for a variety of reasons including natural geological processes such as landsliding and anthropogenic activities such as mining. They have a serious impact in terms of financial losses, injuries, and even deaths. Remote sensing can measure the movement of the ground in a variety of ways, all of which can be applied from space, aircraft, or in field surveys. These include photogrammetry with both optical and radar data, radar interferometry, and ranging techniques, such as lidar. The measurement of ground movement requires two acquisitions separated in time and can be extended into the monitoring of ground motion by repeat surveys. Bibliography Bell, F. G., 1998. Environmental Geology: Principles and Practice. Oxford: Blackwell Science. 594p. Bell, F. G., 1999. Geological Hazards. Their Assessment, Avoidance and Mitigation. London: E & F N Spon. 648p. Bell, F. G., 2007. Engineering Geology, 2nd edn. Oxford: Butterworth-Heinemann. 581p. Bell, F. G., and Donnelly, L. J., 2006. Mining and Its Impact of the Environment. London: Taylor & Francis. 547p. Cheetham, R., 2008. Subsidence: a gradual catastrophe. Catastrophe Risk. December 2008. Cheney, R., Douglas, B., Agreen, R., Miller, L., Milbert, D., and Porter, D., 1986. The GEOSAT Altimeter Mission: a milestone in satellite oceanography. Eos, Transactions American Geophysical Union, 67(48), 1354–1355, doi:10.1029/EO067i048p01354. http://dx.doi.org/10.1029/EO067i048p01354 Fabris, M., and Pesci, A., 2005. Automated DEM extraction in digital aerial photogrammetry: precisions and validation for mass movement monitoring. Annals of Geophysics, 48(6), 973–988, doi:10.4401/ag-3247. http://www.annalsofgeophysics.eu/index. php/annals/article/view/3247 Ferretti, A., Prati, C., and Rocca, F., 1999. Multibaseline INSAR DEM reconstruction, the wavelet approach. IEEE Transactions on Geoscience and Remote Sensing 37(2), 705–715, doi:10.1109/36.752187. http://ieeexplore.ieee.org/xpl/login.jsp? tp=&arnumber=752187&url=http%3A%2F%2Fieeexplore.ieee. org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D752187 Ferretti, A., Prati, C., and Rocca, F., 2000. Permanent scatterers in SAR interferometry. IEEE Transactions on Geoscience and Remote Sensing, 38(5), 2202–2212, doi:10.1109/36.868878. http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=868878 &url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_ all.jsp%3Farnumber%3D868878 Jaboyedoff, M., Oppikofer, T., Abellán, A., Derron, M.-H., Loye, A., Metzger, R., and Pedrazzini, A., 2012. Use of LIDAR in landslide investigations: a review. Natural Hazards, 61(1), 5– 28, doi:10.1007/s11069-010-9634-2. http://link.springer.com/ article/10.1007/s11069-010-9634-2 Johnson, R. B., and DeGraff, J. V., 1988. Principles of Engineering Geology. New York: Wiley. 497p. Keller, E. A., 1985. Environmental Geology, 4th edn. Columbus, OH: Charles E Merrill Publishing Company. 480p. (Note: 8th Edition 1999). Keller, E. A., Blodgett, R. H., and Clague, J. J., 2008. Natural Hazards: Earth’s Processes as Hazards, Disasters and Catastrophes, Canadianth edn. Toronto: Pearson. 448p. Liu, X., 2008. Airborne LiDAR for DEM generation: some critical issues. Progress in Physical Geography, 32(1), 31–49, doi:10.1177/0309133308089496. http://ppg.sagepub.com/content/32/1/31.abstract

806

SURFACE RADIATIVE FLUXES

Madsen, S. N., Zebker, H. A., and Martin, J., 1993. Topographic mapping using radar interferometry: processing techniques. IEEE Transactions on Geoscience and Remote Sensing, 31(1), 246–256, doi:10.1109/36.210464. http://ieeexplore.ieee.org/ xpl/login.jsp?tp=&arnumber=210464&url=http%3A%2F% 2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber% 3D210464 McCall, G. J. H., Laming, D. J. C., and Scott, S. C. (eds.), 1992. Geohazards Natural and Man-Made. Chapman & Hall: London. 227p. Nuhfer, E. B., Proctor, R. J., Moser, P. H., et al., 1993. The Citizens’ Guide to Geologic Hazards. Arvada, CO: The American Institute of Professional Geologists. 134p. Rödelsperger, S., Läufer, G., Gerstenecker, C., and Becker, M., 2010. Monitoring of displacements with ground-based microwave interferometry: IBIS-S and IBIS-L. Journal of Applied Geodesy, 4(1), 41–54, doi:10.1515/jag.2010.005, ISSN (Online) 1862-9024, ISSN (Print) 1862-9016. http://www. degruyter.com/dg/viewarticle/j$002fjag.2010.4.issue-1$002fjag. 2010.005$002fjag.2010.005.xml Toutin, T., 2001. Elevation modelling from satellite visible and infrared (VIR) data. International Journal of Remote Sensing, 22(6), 1097– 1125, doi:10.1080/01431160117862. http://www.tandfonline.com/ doi/abs/10.1080/01431160117862#.UfKhlbdwah0 Toutin, T., and Chenier, R., 2009. 3-D radargrammetric modeling of RADARSAT-2 ultrafine mode: preliminary results of the geometric calibration. Geoscience and Remote Sensing Letters, IEEE, 6(2), 282, 286, doi:10.1109/LGRS.2008.2010563. http:// ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=4768710&url= http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp% 3Farnumber%3D4768710 Turner, A. K., and Schuster, R. L. (eds.), 1996. Landslides: Investigation and Mitigation. Washington, DC: National Research Council. Transport Research Board Special Publication, Vol. 247. 675p. Waltham, A. C., Bell, F. G., and Culshaw, M. G., 2005. Sinkholes and Subsidence: Karst and Cavernous Rocks in Engineering Construction. Berlin/Chichester: Springer/Praxis Publishing. 382p. Xia, Y., Kaufmann, H., and Guo, X., 2002. Differential SAR interferometry using corner reflectors. In Geoscience and Remote Sensing Symposium, 2002. IGARSS ’02. IEEE International, Vol. 2, 1243–1246, doi:10.1109/IGARSS.2002.1025902. http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber= 1025902&sortType%3Dasc_p_Sequence%26filter%3DAND% 28p_IS_Number%3A22037%29

SURFACE RADIATIVE FLUXES Rachel T. Pinker Department of Atmospheric and Oceanic Science, University of Maryland, College Park, MD, USA

Synonyms Remote sensing of surface radiative fluxes; Shortwave and longwave surface radiative fluxes Definition Radiative fluxes. Electromagnetic radiation received from the sun (shortwave) (SW) or emitted from the atmosphere and/or the Earth surface (longwave) (LW).

Surface radiative fluxes. Electromagnetic radiation received at the surface of the Earth (ocean, land, and cryosphere) in the shortwave or longwave part of the spectrum. Surface radiation balance. The sum of the incoming and outgoing shortwave and longwave radiative fluxes at the surface. Remote sensing of radiative fluxes. Sensing of radiative fluxes by instruments away from the source of radiation such as satellites or aircraft.

Introduction Components of surface radiative fluxes Solar radiation (insolation or shortwave radiation) from the sun (0.3–4.0 mm) that reaches the Earth’s surface (about 50 % of that emitted) is a major source of energy for heating our planet. At the mean distance of the Earth from the sun (1.50
108 km), the incident radiant flux density (irradiance) on a surface perpendicular to the solar beam is known as the solar constant quoted to be about 1,360.8  0.5 W m2 according to measurements from the Total Irradiance Monitor (TIM) on NASA’s Solar Radiation and Climate Experiment (SORCE) (Kopp et al., 2005; Kopp and Lean, 2011). Solar radiation at any location on the Earth is determined by the Earth– Sun distance and the solar zenith angle. On its way from its source, it is partially absorbed and scattered by the atmosphere and clouds and partially absorbed by the surface. It arrives at the surface as direct solar and diffuse sky radiation; the sum is referred to as global radiation. The absorbed energy in the atmosphere reradiates back to the surface and to outer space in the form of longwave (terrestrial) radiation (4.0–100.0 mm), and similarly, part of the absorbed energy at the surface is reradiated back to the atmosphere and space. The Earth climate is a result of the balance maintained between the solar radiation absorbed by the Earth–atmosphere system (gain) and the emission of the terrestrial radiation back to space (loss). This balance is also referred to as net radiation (Q*) or radiation balance composed of net solar radiation and net longwave radiation expressed commonly as Q ¼ SW#  SW" þ LW#  LW"

(1)

where SW (0.3–4.0 mm) and LW (4.0–100.0 mm) are the shortwave and longwave components, respectively, and arrows represent the direction of the flux and Q* is the surface net radiation (0.3–100.0 mm). The net shortwave radiation is a balance between the incoming shortwave and the reflected shortwave; the ratio of reflected shortwave to incoming shortwave radiation is also known as the albedo (A): K ¼ SW#ð1  AÞ

(2)

K* is the surface net shortwave radiation (0.3–4.0 mm).

SURFACE RADIATIVE FLUXES

The net longwave radiation is a balance between the incoming longwave and outgoing longwave; the latter depends on surface temperature and emissivity: L ¼ ðLW#  LW"Þ

(3)

L* is net longwave radiation at the surface (4.0–100.0 mm), and LW↑ can be determined as LW" ¼ esT4

(4)

where e is surface emissivity, s is the Stefan–Boltzmann constant, and T is the surface temperature. Absorption of solar radiation in the atmosphere is mostly from ozone, water vapor, carbon dioxide, oxygen, and clouds. Clouds play a major role in determining the net radiative balance via their optical properties and amount. The effect of clouds on the Earth’s radiation balance is measured as the difference between the clear-sky and total-scene radiation results. This difference is defined as cloud-radiative forcing. Cloud optical depth is a general measure of the capacity of a cloud to control the amount of light that will reach the surface. Greater optical depth means greater blockage of the light and a larger cooling of the Earth–atmosphere system in the SW region of the spectrum. Attempts to estimate the radiative components of the Earth–atmosphere system date as early as Budyko (1958). Annually and globally averaged climatological values as estimated by Kiehl and Trenberth (1997) are 342 W m2 for incoming solar, 107 W m2 for reflected solar (namely, about 30 % of the incoming solar energy is reflected back to space), and 235 W m2 for outgoing longwave. These values are being continuously reevaluated based on improved satellite observations (e.g., Hansen et al., 2005; Trenberth et al., 2009; Ma and Pinker, 2012). For some time, it was believed that about 20 % is absorbed by atmospheric ozone, water vapor, carbon dioxide, oxygen, and clouds. Recent investigations based on field experiments and numerical computations indicate that the atmosphere absorbs more than 25 % of the solar radiation due to several factors, e.g., the neglect of absorption in the wavelength range beyond 2.8 mm (Cess et al., 1997; Valero et al., 2003). The effect of clouds on atmospheric absorption is small, about 3 W m2. In reality, the energy balance distribution over the globe is not uniform, with a net energy gain in lower latitudes and a loss at higher latitudes. One major objective of remote sensing from satellites is to obtain the best possible information on the radiation balance of the Earth–atmosphere system (Trenberth et al., 2009).

Need for information on radiative fluxes Accurate information on the various components of the radiative fluxes at global scale is needed for a wide range of climate research applications such as modeling the hydrological cycle, representing interactions and feedbacks between the atmosphere and the surface, estimating

807

global terrestrial and oceanic net primary productivity, and validating climate and numerical weather prediction models. Each component has significance for a different type of application. For example, net all-wave radiation (0.3–100.0 mm) is needed for hydrological modeling, photosynthetically active radiation (PAR) (0.4–0.7 mm) for modeling net primary productivity and CO2 budgets, and surface albedo for climate change studies. Satellite observations provide a systematic source of information at global scale needed for inferring components of the radiation balance. Recent advances in satellite technology and development of inference schemes have been conducive for obtaining information on the various components of the radiation balance. Reviewed will be basic principles of methodologies used to derive the Surface Radiation Budget by methods of remote sensing, feasibility to derive such fluxes and linkage to international programs, evaluation efforts, and future prospects.

Type of satellites used for estimating radiative fluxes Information on radiative fluxes is being derived from dedicated satellites as well as from operational weather satellites. Both observational systems are composed of two types of satellites: geostationary and polar orbiting. The U. S. Geostationary Operational Environmental Satellites are known as GOES, while the Polar-Orbiting Environmental Satellites are referred to as POES. GOES satellites circle the Earth in a geosynchronous orbit, namely, they orbit the equatorial plane of the Earth at about 35,800 km above the Earth at a speed matching the Earth’s rotation. The United States normally operates two geostationary satellites in orbit over the equator, one is positioned at 75 W while the other at 135 W longitudes. Currently, the United States is operating GOES-15 (or GOES-West) and GOES-13 (or GOES-East). Complementing the geostationary satellites are two polar-orbiting ones constantly circling the Earth in an almost north-south orbit, passing close to both poles. The orbits are circular, with an altitude between 830 and 870 km, and are sun synchronous. Environmental satellites are also operated by other nations and international organizations such as METEOSAT Second Generation (MSG) operated by EUMETSAT; LANDSAT, a high-resolution Earth observation satellite managed by NASA and the USGS; METOP, the European polar orbiter; and MTSAT, a Japanese geostationary satellite. Radiative fluxes by remote sensing: feasibility Satellites can provide information on surface radiative fluxes at various temporal and spatial scales, and attempts to do so have been ongoing for about 40 years. Steering has been provided by the World Climate Research Programme (WCRP)/World Meteorological Organization (WMO) and by several national and international projects, such as the Global Energy and Water Budget Experiment (GEWEX), the International Biosphere/Geosphere Programme (IGBP), and the GEWEX Hydroclimatology Panel (GHP) which coordinates scientific issues related

808

SURFACE RADIATIVE FLUXES

to the development and implementation of Regional Hydroclimate Projects. Specifications of requirements on information on radiative fluxes are provided in the Intergovernmental Panel on Climate Change (IPCC, 2007). In the United States, the Interagency Committee on Earth and Environmental Sciences has identified clouds and the hydrological cycle to be of highest scientific priority for global change research. Under this framework efforts are in progress to derive clouds and radiative fluxes from satellite observations. These include the Surface Radiation Budget (SRB) activity, the International Satellite Cloud Climatology Project (ISCCP) (Rossow and Schiffer, 1999), the Clouds and the Earth’s Radiant Energy System (CERES) (Wielicki et al., 1996), the Geostationary Earth Radiation Budget (GERB) Experiment on the EUMETSAT METEOSAT Second Generation satellites (Harries et al., 2005), and the Moderate Resolution Imaging Spectroradiometer (MODIS) (Platnick et al., 2003; King et al., 2003). The instrument packages on these satellites are known as imagers (sounders are used for inferring the vertical structure of the atmosphere) and usually have about 2–36 channels in the spectral interval of 0.3–12.0 mm that are relevant for estimating both for shortwave and longwave fluxes. Examples of channels on selected Geostationary Operational Environmental Satellites (GOES) are presented in Table 1. The channels of the Advanced Very High Resolution Radiometer (AVHRR) on polar-orbiting satellites are presented in Table 2. METEOSAT-5 and METEOSAT-7 channels are centered at 0.75, 6.4, and 11.5 mm, while the newer SEVIRI instrument on METEOSAT-8 has 12 channels as given in Table 3. The objective of NASA’s Earth Observing System (EOS) Program is to study the Earth from space using a multiple-instrument, multiple-satellite approach. The Clouds and the Earth’s Radiant Energy System (CERES) is one of the highest priority scientific satellite instruments developed for EOS (Wielicki et al., 1996; Loeb et al., 2009). CERES products include both solar-reflected and Earth-emitted radiation. The CERES instrument builds upon previous similar missions such as the Earth Radiation Budget Experiment (ERBE). CERES instruments were launched aboard the Tropical Rainfall Measuring Mission (TRMM) in November 1997 and on the EOS Terra satellite in December 1999. Two additional instruments fly on the EOS Aqua spacecraft since 2002. Each CERES instrument has three channels: a shortwave to measure reflected sunlight, a longwave to measure emitted thermal radiation in the 8.0–12.0 mm “window” region, and a total channel to measure all wavelengths of radiation (0.30–100.0 mm). The CERES instrument is managed by the NASA Langley Research Center in Hampton, Virginia. The Geostationary Earth Radiation Budget (GERB) instrument is similar to CERES for the EUMETSAT’s METEOSAT Second Generation (MSG) satellites intended to make accurate measurements of the Earth Radiation Budget from geostationary orbit. It was

Surface Radiative Fluxes, Table 1 Summary of spectral bands on several GOES imagers Wavelength Central wavelength Meteorological Band range (mm) (mm) objectives 1

0.55–0.75

0.65

2

3.8–4.0

3.9

3 4

6.5–7.0 5.8–7.3 10.2–11.2

6.75 (GOES-8/11) 6.48 (GOES-12) 10.7

5

11.5–12.5

12.0 (GOES-8/11)

6

12.9–13.7

13.3 (GOES-12)

Daytime cloud cover and surface features Low cloud/fog and fire detection Upper-level water vapor Surface or cloud-top temperature Surface/cloud-top temperature and low-level water vapor CO2 band: cloud detection

Surface Radiative Fluxes, Table 2 Summary of spectral bandwidths (mm) of several AVHRR sensors Channel TIROS-N

NOAA-6, 8, -10

NOAA-7, -9, -11, NOAA-12, -14 13

1

0.55–0.90

0.58–0.68

0.58–0.68

2

0.725–1.10

0.725–1.10

0.725–1.10

3

3.55–3.93

3.55–3.93

3.55–3.93

4

10.5–11.5

10.5–11.5

10.3–11.3

5

Same as Same as 11.5–12.5 channel 4 channel 4

0.58– 0.68 0.725– 1.0 3.55– 3.93 10.3– 11.3 11.4– 12.4

Surface Radiative Fluxes, Table 3 Summary of the bands on the SEVIRI Band

Wavelength (mm)

Central wavelength (mm)

1 2 3 4 5 6 7 8 9 10 11 12

0.56–0.71 0.74–0.88 1.50–1.78 3.48–4.36 5.35–7.15 6.85–7.85 8.30–9.10 9.38–9.94 9.80–11.80 11.00–13.00 12.40–14.40 HRV: broadband (about 0.4–1.1)

0.635 0.81 1.64 3.90 6.25 7.35 8.70 9.66 10.80 12.00 13.40

SURFACE RADIATIVE FLUXES

CALIPSO

809

CloudSat

PARASOL Aqua

Aura

Surface Radiative Fluxes, Figure 1 The Afternoon Constellation (“A-Train”) consisting of five US and international satellites (Credit: NASA).

launched in January 2004 on the first METEOSAT Second Generation satellite (MSG-1), renamed to METEOSAT-8. It complements the MSG SEVIRI instrument (Schmetz et al., 2002). The development, integration, and flight of GERB on MSG are supported by EUMETSAT and the European Space Agency (ESA). GERB has capabilities to measure the shortwave and longwave radiation from the Earth every 15 min and carries a black body for thermal calibration and a solar diffuser for shortwave calibration. A recent addition to the suite of satellites relevant for Earth Radiation Budgets is the CloudSat mission (Stephens et al., 2002). CloudSat is a NASA Earth System Science Pathfinder mission to provide observations necessary to advance our understanding of cloud abundance, distribution, structure, and radiative properties. CloudSat is complimented with the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) (launched in April 2006) (Winker et al., 2012) and subsequently joined with several satellites already in orbit to form a constellation of satellites known as the Afternoon Constellation or A-Train (Figure 1). It consists of five satellites flying in close proximity: Aqua, CloudSat, CALIPSO, PARASOL, and AURA. The orbits are sun synchronous and provide near-simultaneous and colocated observations conducive to improved information on the Earth Radiation Budget and Atmospheric Chemistry (primarily from Aura). The Advanced Baseline Imager (ABI) is being developed as the future imager on the Geostationary Operational Environmental Satellite (GOES) series, slated to be launched approximately in 2015 with GOES-R (Schmit et al., 2005). The ABI will begin a new era in US environmental remote sensing with more spectral

bands, faster imaging, and higher spatial resolution than the current imagers (Schmidt et al., 2005). ABI spatial resolution will be nominally 2 km for the infrared bands and 0.5 km for the 0.64 mm visible band. ABI has 16 spectral bands; 5 are similar to the 0.6, 4.0, 11.0, and 12.0 mm windows and the 6.5 mm water vapor band on several GOES imagers (e.g., 8/-9/-10/-11) (Menzel and Purdom, 1994) and another is similar to the 13.3 mm on the GOES-12-13-14-15 imagers. Indeed, multiple satellites are needed to provide adequate temporal sampling since clouds and radiative fluxes vary throughout the day.

Physical principles of inference schemes SW radiation An important physical principle for deriving SW budgets from satellite observations is the close linear coupling between SW (0.3–4.0 mm) reflected radiance at the top of the atmosphere (albedo) and the surface irradiance (Schmetz, 1989). Cloud absorption (transmittance) and albedo are linearly related since atmospheric constituents do not emit radiation at solar wavelengths. There is a dependence on solar zenith angle, gaseous and aerosol absorption and scattering, surface reflectivity, and clouds. A plausible scenario of a retrieval methodology of shortwave radiative fluxes both at the surface and at the top of the atmosphere (TOA) can be summarized as follows. TOA downward flux (Ftd) can be calculated from the extraterrestrial solar spectrum, accounting for variation in Sun–Earth distance and position of the sun in the sky relative to local vertical (solar zenith angle). Downward flux at the surface (Fsd) can be obtained by determining what fraction of Ftd reaches the surface as radiation is

810

SURFACE RADIATIVE FLUXES

transferred through the atmosphere. This fraction is referred to as flux transmittance (T). It can be determined from a combination of the reflected part at the radiative flux at the TOA that the satellite observes and from modeling the radiative effects of atmospheric constituents (such as ozone, water vapor, molecular scattering, and optical properties of clouds and aerosols). It also depends on the length of the path the radiation travels through the atmosphere (determined by the solar zenith angle), and to a lesser degree, on the albedo of the surface. Once T is known, surface downward flux is obtained as Fsd ¼ T Ftd. Once Fsd is known, upward flux at the surface (Fsu) is calculated as Fsu ¼ As Fsd, where As is surface albedo. Similarly, flux reflected to space by the Earth– atmosphere system (TOA upward flux, Ftu) can be obtained from the product of Ftd and the TOA albedo (At), namely, Ftu ¼ At Ftd. For example, the Clouds and the Earth’s Radiant Energy System (CERES) instrument onboard several Earth Observing System (EOS) satellites provides a direct measurement of TOA broadband (At).

LW radiation Downwelling longwave radiation LW# at the surface originates from radiatively active gases in the atmosphere and depends on the vertical profiles of temperature, gaseous absorbers, and clouds. Gaseous emission is present in specific wavebands in the range 0.3–30 mm, whereas emission from clouds corresponds to black body radiation at the temperature of the cloud base. The intensity of the thermal emission from a cloud varies with its temperature and the optical thickness of the cloud. Retrieval of LW# is difficult since only atmospheric window radiances at the top of the atmosphere give information on the near-surface radiation field. For the remainder of the longwave spectrum, the radiation regimes at the top of the atmosphere and at the surface are decoupled; more than 80 % of the clear-sky longwave flux reaching the surface is emitted within the lowest 500 m of the atmosphere. Cloud contributions are mainly from the atmospheric window region (8.0–13.0 mm), and the relevant cloud parameters are cloud base height location and temperature, emittance, and cloud amount. Relative importance of cloud contribution decreases with moister atmospheres since the transparency of the window decreases due to water vapor continuum absorption. From cloudless skies, more than half the longwave flux received at the ground comes from gases in the lowest 100 m and roughly 90 % from the lowest kilometer. Consequently, LW# is often estimated as a function of a bulk atmospheric temperature (in Kelvin) approximated by air temperature at the ground and an estimated broadband atmospheric emissivity (Prata, 1996; Dilley and O’Brien, 1998). It is often assumed that the longwave emissivity of most natural surfaces is unity, so that L↑ can be estimated from knowledge of the surface temperature. Longwave techniques could be characterized as being of distinct types:

flux calculation techniques, which use radiative transfer models to calculate the LW# directly from retrieved profiles of temperature and water vapor and estimates of cloud fraction and cloud base altitude (e.g., Breon et al., 2002; Gupta et al., 2010), flux inference techniques which use direct inference from TOVS or HIRS radiance observations, and statistical inference schemes based on radiance–flux relationship developed with detailed radiative transfer models and large number of observed atmospheric soundings (for clear-sky and overcast conditions) with some built in cloud height and base information. It is often assumed that the longwave emissivity of most natural surfaces is unity, so that LW↑ can be estimated from knowledge of the surface temperature (Prata, 1996).

Methodologies to derive radiative fluxes Inference schemes to derive radiative fluxes from satellite observations have been implemented at a wide range of complexities, both in the SW and LW part of the spectrum. Review type of summaries can be found in Schmetz (1989, 1991), Pinker et al. (1995), and Whitlock et al. (1995); methods to derive PAR are described in Frouin and Pinker (1995); recent approaches for LW methods can be found in Zhou et al. (2007), Nussbaumer and Pinker (2011, 2012), and Niemela et al. (2001). Methods to derive SW fluxes have been implemented on different spatial and temporal scales. Used are observations from geostationary satellites (METEOSAT, GOES, GMS) and from polar-orbiting satellites using instruments such as AVHRR, MODIS, and CERES (Stuhlman et al., 1990; Pinker and Laszlo, 1992; Gupta et al., 1999; Chou et al., 1995; Brisson et al., 1994; Li et al., 1993; Zhang et al., 1995; Ceballos et al., 2004; Hollmann et al., 1999; Zhang et al., 2004; Charlock et al., 2006; Vardavas and Taylor, 2007; Wang and Pinker, 2009; Ma and Pinker, 2012). For obtaining estimates at global scale at three hourly intervals, the International Satellite Cloud Climatology Project (ISCCP) (Rossow and Schiffer, 1991) has merged information from several satellites that has been widely used and implemented with independent algorithms. Initial work on the use of these observations started at the NASA Goddard Institute for Space Studies (GISS) (Rossow and Lacis, 1990). The validation results for the GISS ISCCP surface fluxes show that it has an uncertainty of 10–15 W m–2 at monthly time scales (Zhang et al., 2006, 2007). The ISCCP data in their various versions have been widely used by many groups (Stackhouse et al., 2002; Ma and Pinker, 2012) and serve as the basis for the WCRP-sponsored Radiative Flux Assessment intercomparison activity under which various satellite products (global and regional) are being evaluated. Many of the current geostationary satellites are limited in their capability to accurately detect cloud or aerosol optical properties that are important elements of the

SURFACE RADIATIVE FLUXES

90 °N

90 °N

45 °N

45 °N

0°

0°

45 °S

45 °S

90 °S

0

811

180 °W

100

0°

200

90 °S

180 °E

300

400

500

0

180 °W

100

0°

200

180 °E

300

400

500

Surface Radiative Fluxes, Figure 2 Left: Monthly averaged surface LW radiation for March 2001 as derived from MODIS observations (Nussbaumer and Pinker, 2012); Right: monthly averaged surface SW radiation for March 2001 as derived from MODIS observations (Wang and Pinker, 2009).

radiation budget. Polar-orbiting satellites tend to have higher spatial resolution than the geostationary satellites, as well as more spectrally resolving bands. The MODIS instrument onboard the Terra and Aqua satellites, a stateof-the-art sensor with 36 spectral bands and onboard calibration of both solar and infrared bands, has the capability to measure atmospheric and surface properties with higher accuracy and consistency than previous Earth observation imagers. Moreover, geostationary satellites cover only part of the globe up to about 50 north and south. Polar-orbiting satellites provide coverage twice a day over the entire globe, including the high latitudes. A configuration of satellites, both geostationary and polar orbiting, is needed to cover the entire globe with observations that depict the diurnal cycle. In Figure 2 shown are mean surface SW radiative fluxes (W m–2) averaged for March 2001 (Wang and Pinker, 2008) and the mean LW# fluxes for the same month (Nussbaumer and Pinker, 2012) as obtained from the MODIS instrument on Terra and Aqua. Real-time implementation of inference schemes of shortwave radiative fluxes is now a reality. Flux estimates are produced at hourly time scale by the National Oceanic and Atmospheric Administration (NOAA)/National Environmental Satellite, Data, and Information Service (NESDIS), using observations from GOES (Pinker et al., 2003). Similar efforts are ongoing in other countries such as at the Nowcasting Satellite Application Facility (NWC SAF) in Europe. SAFs are dedicated centers for processing satellite data. They have been developed by a consortium of EUMETSAT Member States and Cooperating States trying to benefit from METEOSAT Second Generation (MSG) and Polar Platform Satellites.

Surface Radiative Fluxes, Figure 3 Evaluation of daily mean downwelling SW flux estimated by UMD/SRB_MODIS model (Wang and Pinker, 2009) against BSRN measurements over land (January 2003–December 2005). Cases eliminated: 1.6 %.

Evaluation of satellite estimates using surface measurements Evaluation of surface fluxes depends on the quality of the surface measurements (Shi and Long, 2002; Philipona et al., 2001). Historic surface SW measurements were

collected under the auspices of the WCRP at the World Radiation Data Centre (WRDC), A.I. Voeikov Main Geophysical Observatory, St. Petersburg, Russia. This database includes sites from all continents and contains measurements in the form of daily and monthly averages. The Swiss Federal Institute of Technology provides the Global Energy Balance Archive (GEBA) (Ohmura and Gilgen, 1993) which includes data from the WRDC. Globally distributed surface measurements over the last two decades are improving with the advent of the GEWEX Baseline Surface Radiation Network (BSRN) (Ohmura et al., 1998). BSRN is dedicated to providing long-term high-quality measurements at locations around the globe critical for satellite validation. Regional-scale networks of surface measurements are also utilized in validation (Augustine et al., 2000; Michalsky et al., 1999) as well as data from special campaigns such as the Clouds and the Earth’s Radiant Energy System (CERES) program/Atmospheric Radiation Measurement (ARM) Validation Experiment (CAVE) (Rutan et al., 2001). Examples of evaluation of SW and LW results from MODIS observations against the BSRN network are shown in Figures 3 and 4. Radiative fluxes are also measured over the oceans. For instance, in situ measurements from the Pilot Research Moored Array in the Tropical Atlantic (PIRATA) moorings in the tropical Atlantic (Bourle`s et al., 2008) and the Tropical Atmosphere Ocean/Triangle Trans-Ocean Buoy Network (TAO/TRITON) moorings in the tropical Pacific Ocean (McPhaden et al., 1998) are frequently used for comparisons with the satellite estimates (Figure 5). Surface radiative fluxes are needed at accuracies suitable to predict transient climate variations and long-term climate trends. The WCRP GEWEX Surface Radiation Budget (SRB) project is specially tasked to produce, validate, and assess long-term surface and atmospheric radiative budgets at global scale. Intercomparison between satellite products with surface sites is part of the GEWEX SRB (Stackhouse et al., 2002) and CERES-SARB programs (Charlock et al., 2006). Reported is agreement with ground observations within 10 Wm-2 on a monthly time scale. At hourly time scales the reported differences are in the range of -50 to +50 Wm-2. More comprehensive evaluation results are reported on in Trenberth et al. (2009), Ma and Pinker (2012), and Nussbaumer and Pinker (2012).

Future prospects A summary of current and future satellite observations of relevance for SRB research is presented in Wielicki et al. (1996). It is believed that SRB activities could benefit from better information on the vertical structure of clouds and meteorology of the lower layer; information on cloud types; improved calibration and stability of all instruments used; better understanding of validation limitations; improvement in inference techniques; and information on aerosol optical properties, better characterization of

UMD/MODIS Calculated LW Flux (W m−2)

SURFACE RADIATIVE FLUXES

500

400

300

200

CC = 0.97 Bias (W m−2) = 0.06 RMSE (W m−2) = 17.01

100

0

0

100

200

300

400

500

Surface Radiative Fluxes, Figure 4 Evaluation of daily mean surface LW flux estimated by UMD/SRB_MODIS_LW at 10 spatial resolution against 18 BSRN stations for 2007.

PIRATA N = 8478 Mean of Obs. = 232 300 Estimated Flux (W/m^2)

812

200

100 Corr. Coef. = 0.89 RMSE = 25 (11 %) BIAS = 3 (1 %) 0

0

100

200

300

Observed Flux (W/m^2)

Surface Radiative Fluxes, Figure 5 Evaluation of daily mean downwelling SW estimated by UMD/SRB_MODIS model (Wang and Pinker, 2009) (2003–2005) against PIRATA buoy observations. Cases eliminated: 1.1 %.

land surface properties, accurate information on snow distribution, and distinction of partially cloud-filled satellite fields of view and the 3-D nature of clouds. In the short term, the SRB inference methods will benefit from activities like the CloudSat, the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations, and the entire constellation of A-Train satellites.

SURFACE RADIATIVE FLUXES

Acknowledgments We thank NASA for the support of our work under grant NNX08AN40A from the Science Mission Directorate-Division of Earth Science and grant NNX13AC12G from the Energy and Water Cycle Study (NEWS) program. The work also benefited from the support under NOAA grant NA09NES4400006, the Cooperative Institute for Climate and Satellites (CICS) at the University of Maryland/ESSIC. Bibliography Augustine, J. A., DeLuisi, J. J., and Long, C. N., 2000. SURFRADA national surface radiation budget network for atmospheric research. Bulletin of the American Meteorological Society, 81, 2341–2357. Bourle`s, B., et al., 2008. The Pirata Program: history, accomplishments, and future directions. Bulletin of the American Meteorological Society, 89, 1111–1125. Breon, F. M., Buriez, J. C., Couvert, P., et al., 2002. Earth’s atmosphere, ocean and surface studies, Book series: Advances in Space, 30(11), 2383–2386. Article Number: PII S0273-1177(02) 00078-9, doi:10.1016/S0273-1177(02)80282-4. Brisson, A., Le Borgne, P., Marsouin, A., and Moreau, T., 1994. Surface irradiance calculated from Meteosat sensor data during SOFIA-ASTEX. International Journal of Remote Sensing, 15, 197–203. Budyko, M. I., 1958. The heat balance of the earth’s surface. Washington, DC: Dept. of Commerce, Weather Bureau, 1958 – Earth temperature, pp. 259. Ceballos, J. C., Bottino, M. J., and de Souza, J. M., 2004. A simplified physical model for assessing solar radiation over Brazil using GOES 8 visible imagery, Journal of Geophysical Research: Atmospheres, 109(D2), Art. No. D02211. Cess, R. D., Zhang, M. H., Potter, G. L., Alekseev, V., Barker, H. W., Bony, S., Colman, R. A., Dazlich, D. A., DelGenio, A. D., Deque, M., Dix, M. R., Dymnikov, V., Esch, M., Fowler, L. D., Fraser, J. R., Galin, V., Gates, W. L., Hack, J. J., Ingram, W. J., Kiehl, J. T., Kim, Y., LeTreut, H., Liang, X. Z., McAvaney, B. J., Meleshko, V. P., Morcrette, J. J., Randall, D. A., Roeckner, E., Schlesinger, M. E., Sporyshev, P. V., Taylor, K. E., Timbal, B., Volodin, E. M., Wang, W., Wang, W. C., and Wetherald, R. T., 1997. Comparison of the seasonal change in cloud-radiative forcing from atmospheric general circulation models and satellite observations. Journal of Geophysical Research, 102, 16593–16603. Charlock, T. P., Rose, F. G., Rutan, D. A., Jin, Z., and Kato, S., 2006. The global surface and atmospheric radiation budget: An assessment of accuracy with 5 years of calculations and observations. In Proceedings 12th Conference on Atmos. Radiation, Madison, WI, July 10–14. Chou, M. D., Ridgway, W. L., and Yan, M. M. H., 1995. Parameterizations for water vapor IR radiative transfer in both the middle and lower atmospheres. Journal of the Atmospheric Sciences, 52(8), 1159–1167, doi:10.1175/1520. Dilley, A. C., and O’Brien, D. M., 1998. Estimating downward clear sky long-wave irradiance at the surface from screen temperature and precipitable water. Quarterly Journal of the Royal Meteorological Society, 124, 1391–1401. Frouin, R., and Pinker, R. T., 1995. Estimating Photosynthetically Active Radiation (PAR) at the Earth’s surface from satellite observations. Remote Sensing of Environment, 51. Gupta, S. K., Ritchey, N. A., Wilber, A. C., Whitlock, C. H., Gibson, G. G., and Stackhouse, P. W. Jr., 1999. A climatology

813

of surface radiation budget derived from satellite data. Journal of Climate, 12(8), 2691–2710, Part 2. Gupta, S. K., Kratz, D. P., Stackhouse, P. W., Jr., Wilber, A. C., Zhang, T., and Sothcott, V. E., 2010. Improvement of surface longwave flux algorithms used in CERES processing. Journal of Applied Meteorology and Climatology, 49(7), 1579–1589, doi:10.1175/2010JAMC2463.1. Hansen, J. E., et al., 2005. Earth’s energy imbalance: confirmation and implications. Science, 308, 1431–1435, doi:10.1126/ science.1110252. Harries, J. E., Russell, J. E., and Hanafin, J. A., et al., 2005. The geostationary Earth Radiation Budget Project. Bulletin of the American Meteorological Society, 86(7), doi:10.1175/ BAMS-86-7-945. Hollmann, R., Feng, J., Leighton, H. G., Mueller, J., and Stuhlmann, R., 1999. ScaRaB as a valuable tool for BALTEX and MAGS: Satellite applications for energy budgets and the hydrological cycle. Advances in Space Research, 4(7), 955–958. Intergovernmental Panel on Climate Change (IPCC), 2007. Climate change 2007: The physical science basis. Contribution of working group I to the fourth assessment report of the intergovernmental panel on climate change, edited by Solomon, S. et al. Cambridge, UK: Cambridge University Press. Kiehl, J. T., and Trenberth, K. E., 1997. Earth’s annual global mean energy budget. Bulletin of the American Meteorological Society, 78, 197–208, doi:10.1175/1520. King, M. D., et al., 2003. Cloud and aerosol properties, perceptible water, and profiles of temperature and humidity from MODIS. IEEE Transactions on Geoscience and Remote Sensing, 41, 442–458, doi:10.1109/TGRS.2002.808226. Kopp, G., and Lawrence, G., 2005. The Total Irradiance Monitor (TIM): instrument design. Solar Physics, 230, 1–2, doi:10.1007/s11207-005-7446-4. Kopp, G., and Lean, J. L., 2011. A new, lower value of total solar irradiance: evidence and climate significance. Geophysical Research Letters, 38, L01706, doi:10.1029/2010GL045777. Kopp, G., Lawrence, G., and Rottman, G., 2005. The Total Irradiance Monitor (TIM): science Results, Solar Physics, 230, 129–140. Li, Z., Leighton, H. G., Masuda, K., and Takashima, T., 1993. Estimation of SW flux absorbed at the surface from TOA reflected flux. Journal of Climate, 6, 317–330. List, R. J. (ed.), 1966. Smithsonian Meteorological Tables, 6th edn. Washington, DC: Smithsonian Institution. Loeb, N. G., Wielicki, B. A., Doelling, D. R., Smith, G. L., Keyes, D. F., Kato, S., Manalo-Smith, N., and Wong, T., 2009. Toward optimal closure of the Earth’s top-of-atmosphere radiation budget. Journal of Climate, 22, 748–766, doi:10.1175/2008JCLI2637.1. Ma, Y., and Pinker, R. T., 2012. Modeling shortwave radiative fluxes from satellites. Journal of Geophysical Research, 117, D23202, 1–19, doi:10.1029/2012JD018332. McPhaden, M. J., et al., 1998. The Tropical Ocean-Global Atmosphere observing system: a decade of progress. Journal of Geophysical Research, 103, 14,169–14,240. Menzel, W. P., and Purdom, J. F., 1994. Introducing GOES-I: the first of a new generation of geostationary operational environmental satellites. Bulletin of the American Meteorological Society, 75, 757–782. Michalsky, J., Dutton, E., Rubes, M., et al., 1999. Optimal measurement of surface shortwave irradiance using current instrumentation. Journal of Atmospheric and Oceanic Technology, 16(1), 55–69, doi:10.1175/1520-0426. Niemela, S., Raisanen, P., and Savijarivi, H., 2001. Comparison of surface radiative flux parameterizations. Part I: longwave radiation. Atmospheric Research, 58, 1–18. Nussbaumer, E. A., and Pinker, R. T., 2011. Estimating surface long-wave radiative fluxes at global scale. Quarterly Journal of Royal Meteorological Society, 137, October 2011 A.

814

SURFACE RADIATIVE FLUXES

Nussbaumer, E. A., and Pinker, R. T., 2012. Estimating surface longwave radiative fluxes from satellites utilizing artificial neural networks. Journal of Geophysical Research, 117, D07209, doi:10.1029/2011JD017141. Ohmura, A., and Gilgen, H., 1993. Re-evaluation of the global energy balance. Geophysical Monograph, 75, IUGG, Vol. 15, pp 93–110. Ohmura, A., Dutton, E. G., Forgan, B., Frohlich, C., Gilgen, H., Hegner, H., Heimo, A., Konig-Langlo, G., McArthur, B., Muller, G., Philipona, R., Pinker, R. T., Whitlock, C. H., Dehne, K., and Wild, M., 1998. Baseline surface radiation network (BSRN/ WCRP): new precision radiometry for climate research. Bulletin of the American Meteorological Society, 79, 2115–2136. Philipona, R., et al., 2001. Atmospheric longwave irradiance uncertainty: pyrgeometers compared to an absolute sky-scanning radiometer, atmospheric emitted radiance interferometer, and radiative transfer model calculations. Journal of Geophysical Research, 106, 28,129–28,141, doi:10.1029/2000JD000196. Pinker, R. T., and Laszlo, I., 1992. Modeling surface solar irradiance for satellite applications on a global scale. Journal of Applied Meteorology, 31, 194–211. Pinker, R. T., Laszlo, I., Whitlock, C. H., and Charlock, T. P., 1995. Radiative flux opens new window on climate research. EOS, 76(15), 145–160. Pinker, R. T., Tarpley, J. D., Laszlo, I., Mitchell, K. E., Houser, P. R., Wood, E. F., Schaake, J. C., Robock, A., Lohmann, D., Cosgrove, B. A., Sheffield, J., Duan, Q., Luo, L., and Higgins, R. W., 2003. Surface radiation budgets in support of the GEWEX continental scale international project (GCIP) and the GEWEX Americas prediction project (GAPP), including the North American land data assimilation system (NLDAS). Journal of Geophysical Research-Atmospheres, 108(D22), Art. No. 8844 19 Nov 2003. Platnick, S., King, M. D., Ackerman, S. A., Menzel, W. P., Baum, B. A., Riedi, J. C., and Frey, R. A., 2003. The MODIS cloud products: algorithms and examples from Terra. IEEE Transactions on Geoscience and Remote Sensing, 41, 459–473. Prata, A. J., 1996. A new long-wave formula for estimating downward clear-sky radiation at the surface. Quarterly Journal of the Royal Meteorological Society, 122, 1127–1151. Rossow, W. B., and Lacis, A. A., 1990. Global, seasonal cloud variation from satellite radiance measurements, 2, Cloud properties and radiative effects. Journal of Climate, 3, 1204–1253. Rossow, W. B., and Schiffer, R. A., 1991. ISCCP cloud data products. Bulletin of the American Meteorological Society, 72(1), 2–20. Rossow, W. B., and Schiffer, R. A., 1999. Advances in understanding clouds from ISCCP. Bulletin of the American Meteorological Society, 80, 2261–2287, doi:http://dx.doi.org/10.1175/15200477(1999)0802.0.CO;2. Rutan, D. A., Rose, F. G., Smith, N. M., and Charlock, T. P., 2001. Validation data set for CERES surface and atmospheric radiation budget (SARB), WCRP/GEWEX Newsletter, 11(1), 11–12. Schmetz, J., 1989. Towards a surface radiation climatology: retrieval of downward irradiance from satellites. Atmospheric Research, 23, 287–321. Schmetz, J., 1991. Retrieval of surface radiation fluxes from satellite data. Dynamics of the Atmospheres and Oceans, 16(1–2), 61–72. Schmetz, J., Pili, P., Tjemkes, S., Just, D., Kerkmann, J., Rota, S., and Ratier, A., 2002. An introduction to Meteosat Second Generation (MSG). Bulletin of the American Meteorological Society, 83, 977–992. Shi, Y., and Long, C. N., 2002. Techniques and methods used to determine the best estimate of radiation fluxes at SGP Central Facility. Paper presented at 12th ARM Science Team Meeting Proceedings, St. Petersburg, FL, 8–12 Apr 2002. Shmit, T. J., Gunshor, M. M., Menzel, W. P., Gurka, J. J., Li, J., and Bachmeier, A. S., 2005. Introducing the next generation

advanced baseline imager on GOES-R. Bulletin of the American Meteorological Society, 1079–1096. Stackhouse, P. W., Jr., Gupta, S. K., Cox, S. J., Mikovitz, J. C., and Chiaachio, M., 2002. New results from the NASA/GEWEX Surface Radiation Budget Project: evaluating El Nino effects at different scales. 11th Conference on Atmospheric Radiation, American Meteorological Society, Ogden, UT, 3–7 Jun 2002. Stephens, G. L., et al., 2002. The CloudSat mission and the A-Train. Bulletin of the American Meteorological Society, 83, 1771– 1790, doi:http://dx.doi.org/10.1175/BAMS-83-12-1771. Stephens, G. L., Vane, D. G., Boain, R. J., Mace, G. G., Sassen, K., Wang, Z., Illingworth, A. J., O'Connor, E. J., Rossow, W. B., Durden, S. L., Miller, S. D., Austin, R. T., Benedetti, A., Mitrescu, C., and the ClouSat Science Team, 2002a. The CloudSat mission and the A-train: a new dimension of spacebased observations of clouds and precipitation. Bulletin of the American Meteorological Society, 83, 1771–1790. Stephens, G. L., Vane, D. G., Boain, R. J., and The ClouSat Science Team, 2002b. The Cloudsat mission and the A-Train: a new dimension of space-based observations of clouds and precipitation. Bulletin of the American Meteorological Society, 12, 1171–1790. Stephens, G. L., Li, J., Wild, M., et al., 2012. An update on Earth’s energy balance in light of the latest global observations. Nature Geoscience, 5(10), 691–696, doi:10.1038/NGEO1580. Stuhlmann, R., Rieland, M., and Raschke, E., 1990. An improvement of the IGMK model to derive total and diffuse solar radiation at the surface from satellite data. Journal of Applied Meteorology, 29, 586–603. Trenberth, K. E., Fasullo, J. T., and Kiehl, J., 2009. Earth’s global energy. Bulletin of the American Meteorological Society, 90, 311–323, doi:10.1175/2008BAMS2634.1. Valero, F. P. J., Pope, S. K., Bush, B. C., Nguyen, Q., Marsden, D., Cess, R. D., Simpson-Leitner, A. S., Bucholtz, A., and Udelhofen, P., 2003. Absorption of solar radiation by the clear and cloudy atmosphere during the Atmospheric Radiation Measurement Enhanced Shortwave Experiments (ARESE) I and II: observations and models. Journal of Geophysical ResearchAtmospheres, 108, doi:10.1029/2001DJ001384. Vardavas, I. M., and Taylor, F. W., 2007. Radiation and climate. International Series of Monograph on Physics, 138. 512 pp. Wang, H., and Pinker, R. T., 2009. Shortwave radiative fluxes from MODIS: model development and implementation. Journal of Geophysical Research, 114, D20201, doi:10.1029/2008JD010442. Whitlock, C. H., et al., 1995. First global WCRP shortwave surface radiation budget dataset. Bulletin of the American Meteorological Society, 76, 905–922. Whitlock, C. H., Charlock, T. P., Staylor, W. F., Pinker, R. T., Laszlo, I., Ohmura, A., Gilgen, H., Konzelman, T., DiPasquale, R. C., Moats, C. D, LeCroy, S. R., Wielicki, N. A., Barkstrom, B. R., Harrison, E. F., Lee, R. B., Louis Smith, G., and Cooper, J. E., 1996. Clouds and the Earth’s Radiant Energy System (CERES): an earth observing system experiment. Bulletin of the American Meteorological Society, 77, 853–868, doi:10.1175/1520-0477(1996)077. Wielicki, B. A., Barkstrom, B. R., Harrison, E. F. Lee, R. B., Louis Smith, G., and Cooper, J. E., 1996. Clouds and the Earth’s Radiant Energy System (CERES): an Earth observing system experiment. Bulletin of the American Meteorological Society, 77, 853–868, doi:10.1175/1520-0477(1996)077. Winker, D. M., Tackett, J. L., Getzewich, B. J., Liu, Z., Vaughan, M. A., and Rogers, R. R. (2012). The global 3-D distribution of tropospheric aerosols as characterized by CALIOP. Atmospheric Chemistry and Physics. Discussion, 12, 24847– 24893, doi:10.5194/acpd-12-24847-2012. Zhang, Y. C., Rossow, W. B., and Lacis, A. A., 1995. Calculation of surface and top of the atmosphere radiative fluxes from physical

SURFACE TRUTH

quantities based on ISCCP data sets. 1. Method and sensitivity to input data uncertainties. Journal of Geophysical Research, 100 (D1), 1149–1165, doi:10.1029/94JD02747. Zhang, Y.-C., Rossow, W. B., Lacis, A. A., Oinas, V., and Mishchenko, M. I., 2004. Calculation of radiative flux profiles from the surface to top-of-atmosphere based on ISCCP and other global datasets: refinements of the radiative transfer model and the input data. Journal of Geophysical Research, 109, D19105, doi:10.1029/2003JD004457. Zhang, Y., Rossow, W. B., and Stackhouse, P. W., Jr., 2006. Comparison of different global information sources used in surface radiative flux calculation: radiative properties of the near-surface atmosphere. Journal of Geophysical Research, 111, D13106, doi:10.1029/2005JD006873. Zhang, Y., Rossow, W. B., and Stackhouse, P. W., Jr., 2007. Comparison of different global information sources used in surface radiative flux calculation: radiative properties of the surface. Journal of Geophysical Research, 112, D01102, doi:10.1029/2005JD007008. Zhang, T., Stackhouse, P. W. Jr., Gupta, S. K., et al., 2013. The validation of the GEWEX SRB surface shortwave flux data products using BSRN measurements: a systematic quality control, production and application approach. Journal of Quantitative Spectroscopy & Radiative Transfer, 122, Special Issue: 127–140, doi:10.1016/j.jqsrt.2012.10.004. Zhou, Y., Kratz, D. P., Wilber, A. C., Gupta, S. K., and Cess, R. D., 2007. An improved algorithm for retrieving surface downwelling longwave radiation from satellite measurements. Journal of Geophysical Research, 112, D15102, doi:10.1029/ 2006JD008159.

SURFACE TRUTH Christopher Ruf Department of Atmospheric, Oceanic and Space Sciences, University of Michigan, Ann Arbor, MI, USA

Synonyms Ground truth Definition and overview The purpose of a scientific remote sensing instrument is to estimate geophysical properties of a target from indirect measurements of its electromagnetic radiation, propagation, and scattering. Geophysical properties, such as ocean surface temperature or atmospheric abundance of a certain molecule, are deduced from their electromagnetic properties. The procedure typically requires approximations, assumptions, and the use of measurements that contain random and systematic errors. As a result, the accuracy and precision of the geophysical estimates are usually validated by some independent means. Surface truth refers to some form of independent knowledge about the geophysical property being estimated. It is used to characterize the quality of the remote sensing data product, often using such statistical measures as the root-mean-square or the average difference between the remote sensing estimate and the surface truth.

815

Point-to-point comparisons In most cases, surface truth comparisons with remote sensing estimates are made on a point-by-point basis. For example, a surface truth measurement may be compared to the remote sensing estimate made from a satellite as it passes overhead. Individual remote sensing estimates usually represent an instantaneous spatial average of the geophysical parameter. For surface imagers, the average is over the horizontal resolution spot size of the sensor. For atmospheric sounders, the average is over a resolution volume that includes both horizontal and vertical resolution boundaries. Different types of surface truth measurements are averaged in different ways. Surface truth for some ocean parameters – including sea surface temperature, wind speed and direction, and sea level height – is often provided by buoys and tide gauges (Wentz et al., 2000; Mitchum, 1998; Goodberlet et al., 1989). Buoy measurements are made at a single point in space but can be averaged over time. Surface truth for atmospheric profiles of temperature and humidity is often provided by radiosondes, that is, instrumented weather balloons released from the ground (Buehler et al., 2004; Ruf et al., 1994). Radiosondes make point measurements along the line of ascension of the balloon during its ascent time, which typically lasts 30–60 min through the bulk of the troposphere. For both buoys and radiosondes, their spatial and temporal sampling characteristics do not match those of the remote sensing estimate. It is therefore not necessarily the case that perfect measurements by each should agree. Allowance for the sampling differences is made when they are compared. Another class of point-to-point comparisons makes use of numerical weather forecast models such as those maintained by the US National Weather Service’s National Center for Environmental Prediction (NCEP), the European Centre for Medium-Range Weather Forecasts (ECMWF), and the US Navy’s Fleet Numerical Meteorology and Oceanography Center (FNMOC). These models produce global maps of atmospheric and surface parameters that are updated at least several times per day. The spatial resolution of the models is generally coarser than that of the remote sensing measurement, so there is once again a mismatch between their spatial sampling characteristics. In addition, the times at which a model is updated generally do not correspond exactly with the time when the remote sensing measurement is made, so temporal sampling is also mismatched. In spite of these differences, numerical forecast models are often used as surface truth. Relative to in situ sources of surface truth, such as buoys and radiosondes, they are particularly useful for assembling a large intercomparison database in a short period of time that is representative of the global variability of a particular geophysical parameter. One important limitation is the fact that these models tend to underestimate large deviations of the geophysical parameters from their mean values, and they generally perform less accurately in extreme weather conditions such as heavy precipitation and very high winds.

816

SURFACE WATER

Statistical properties as surface truth Some statistical properties of geophysical properties can also be used as a form of surface truth. To be useful, the property must be independently predictable, and reliable estimates of it must be obtainable from the available population of remote sensing measurements. One example is the global average wind speed and prevailing wind direction, both of which can be accurately predicted from either a large, globally distributed database of buoys or from a collection of numerical forecast models of the wind field (Frielich and Challenor, 1994; Ebuchi, 1999). Another example is the lower bound on microwave thermal emission by the ocean surface, which is determined by the temperature dependence of the emissivity of water (Ruf, 2000). These statistics tend to be very stationary over time and so are good candidates for statistical surface truth. If the statistics derived from surface truth and remote sensing do not agree, it is usually an indication that either the remote sensing instrument calibration is biased or the relationship between electromagnetic and geophysical properties that is assumed by the remote sensing estimation algorithm is in error. One important difference between point-to-point and statistical surface truth comparisons is that, with the statistical approach, the measurements with which the two statistics are derived need not coincide in space or in time. This can significantly simplify assembly of a satisfactory surface truth database, and it can ameliorate many of the effects of sampling differences between the surface truth and remote sensing measurements. However, these benefits come at the cost of a less direct type of comparison that does not as specifically quantify the performance of the remote sensing estimates as does the point-to-point method. Conclusions Surface truth represents knowledge about the geophysical properties being estimated by remote sensing that can be used to assess the quality of those estimates. The knowledge may result from an independent measurement of the geophysical property by some other means, such as a direct, in situ temperature or humidity probe, or it may result from statistical processing of a population of independent measurements. Surface truth is often used both to calibrate and validate a remote sensing data product and to characterize its quality after calibration is complete. Bibliography Buehler, S. A., Kuvatov, M., John, V. O., Leiterer, U., and Dier, H., 2004. Comparison of microwave satellite humidity data and radiosonde profiles: a case study. Journal of Geophysical Research, 109(D), 13103. Ebuchi, N., 1999. Relationship between directional distribution of scatterometer-derived winds and errors in geophysical model functions. Journal of Advanced Marine Science and Technology Society, 5(1,2), 19. Frielich, M. H., and Challenor, P. G., 1994. A new approach for determining fully empirical altimeter wind speed model functions. Journal of Geophysical Research, 99(C12), 25051.

Goodberlet, M. A., Swift, C. T., and Wilkerson, J. C., 1989. Remote sensing of ocean surface winds with the special sensor microwave/imager. Journal of Geophysical Research, 94, 14547. Mitchum, G., 1998. Monitoring the stability of satellite altimeters with tide gauges. Journal of Atmospheric and Oceanic Technology, 15, 721. Ruf, C. S., Keihm, S. J., Subramanya, B., and Janssen, M. A., 1994. TOPEX/POSEIDON microwave radiometer performance and in-flight calibration. Journal of Geophysical Research, 99 (C12), 2491. Ruf, C. S., 2000. Detection of calibration drifts in spaceborne microwave radiometers using a vicarious cold reference. IEE Transactions on Geoscience and Remote Sensing, 38(1), 44. Wentz, F. J., Gentemann, C. L., Smith, D. K., and Chelton, D., 2000. Satellite measurements of sea surface temperature through clouds. Science, 288, 847.

Cross-references Calibration and Validation Geophysical Retrieval, Overview

SURFACE WATER Michael Durand School of Earth Sciences, The Ohio State University, Columbus, OH, USA

Definition Remote sensing of surface water can be defined as collecting information about lakes, wetlands, floodplains, and rivers when viewed from above using electromagnetic radiation (Rees, 2001) or gravity field fluctuations (e.g., Crowley et al., 2006). Introduction Spaceborne and airborne instruments are used to characterize hydrologic features of surface water bodies: lakes, wetlands, floodplains, and rivers. Such remote sensing measurements are an increasingly important complement to traditional in situ surface water hydrology measurements. Alsdorf et al. (2007a) summarize science questions related to surface water and provide a detailed review of surface water remote sensing methods. Primary quantities of interest for surface water quantity are the spatial extent of surface water bodies (inundated area, hereafter) and the vertical elevation of surface water bodies (water surface elevation, hereafter). From these primary quantities, changes in storage and hydrologic fluxes (e.g., river discharge) can be inferred. Primary quantities of interest for surface water quality relate to the biogeochemical constituents present in the water. This review of surface water remote sensing is organized around the following three broad fields: inundated area mapping, water surface elevation estimation, and water biogeochemical constituent characterization. For each of these three quantities, the remote sensing instruments and methods are explained, and some representative

SURFACE WATER

applications of the geophysical retrieval (see entry Geophysical Retrieval, Overview) of surface water characteristics are described.

Inundated area mapping The radiometric reflectivity and emissivity and radar backscatter behavior of liquid water allows for surface water hydrologic features to be distinguished from other geophysical entities at a number of wavelengths in the electromagnetic spectrum, which allows for the characterization of inundated area. For instance, the reflectance of clear water at visible and near-infrared wavelengths is less than that of most naturally occurring materials (Rees, 2001). Similarly, the microwave backscatter coefficient from open water is relatively low, permitting discrimination from land (Smith, 1997). Moreover, the passive microwave emissivity from inundated land is generally low, and the emissivity polarization difference is generally large (Prigent et al., 2007). Both passive and active remote sensing instruments have been used to map inundated area for lakes, wetlands, floodplains, and rivers. For instance, Smith et al. (2005) assessed the total area and number of Siberian lakes using passive optical/infrared systems (see entry Optical/Infrared, Radiative Transfer): Landsat-1 in 1973 and the Russian RES URS-1 satellite in 1998. The comparison allowed for the quantification of a widespread decline in lake number and area in the 25 years between the two assessments. Prigent et al. (2007) present an excellent review of visible, infrared, and microwave passive methods and synthetic aperture radar (SAR) (see entry Calibration, Synthetic Aperture Radars) methods for characterizing inundated area for wetland studies. Prigent et al. (2007) obtained monthly estimates of global time series of inundated area at a 0.25 spatial resolution by merging disparate but complementary data streams: Advanced Very High Resolution Radiometer (AVHRR) visible and near-infrared (NIR) reflectances, the Special Sensor Microwave/Imager (SSM/I) passive microwave radiometer (see entry Microwave Radiometers) measurements, and radar scatterometer (see entry Radar, Scatterometers) measurements from the European Remote Sensing satellite. Hess et al. (2003) used the Japanese Earth Resource Satellite-1 (JERS-1) SAR data to map inundated area at 100 m resolution in the Amazon Basin for two dates representing high water and low water. These inundated estimates have been useful in assessing the contribution of wetlands to biogeochemical processes in an area where obtaining in situ data is far from trivial. The Advanced Land Observing Satellite (ALOS) launched in January 2006 represents a significant improvement over JERS-1 and will permit more detailed floodplain and wetland characterization (Rosenqvist et al., 2007). Many rivers show significant width (and, thus, inundated area) increase with increasing discharge, although this is not always the case for channelized rivers.

817

For instance, Smith and Pavelsky (2008) demonstrated that for the braided Lena River in Siberia, Moderate Resolution Imaging Spectroradiometer (MODIS) estimates of effective river width show strong correlation to river discharge. In addition to making inference on fluvial processes, inundated area is crucial for characterizing river flood events and evaluating flood inundation models (e.g., Horritt and Bates, 2001).

Water surface elevation estimation Water surface elevations can be characterized by a variety of active (as opposed to passive) remote sensing instruments: laser altimetry, profiling radar altimetry, interferometric SAR, and swath radar altimetry. Laser altimeter systems function by emitting a pulse of visible or NIR radiation, detecting the echo of the pulse, and measuring the time between emission and detection. Traditional (profiling) radar altimetry (see entry Radar, Altimeters) works on much the same principles and provides a (spatially limited) profile of elevation estimates along the flight line (Elachi, 1987). Repeat-pass interferometric SAR (see entry Ocean Applications of Interferometric SAR) utilizes multiple SAR images to estimate elevation changes (e.g., Alsdorf et al., 2000). Imaging (swath) radar altimeters provide spatially continuous elevation estimates; SAR processing is very attractive for imaging radar altimeters (Elachi, 1987). These methods are used for assessing surface water elevations in lakes, wetlands, floodplains, and rivers. For lakes and wetlands, water surface elevations represent a primary quantity of interest. For instance, Birkett (1994) used the Geosat profiling radar altimeter to characterize water surface elevations for lakes. Elevations of surface water stored in man-made reservoirs have been characterized using Shuttle Radar Topography Mission (SRTM) imaging SAR altimeter (Kiel et al., 2006). Yi et al. (2006) characterized surface water elevations in wetlands in Louisiana, USA, using TOPEX/POSEIDON profiling altimeter data. Changes in water surface elevation can be used to estimate changes in surface water storage. River floodplain water elevations represent a combination of upstream hydrologic conditions and complex fluvial hydraulic processes. Alsdorf et al. (2007b) used JERS-1 SAR measurements to perform repeat-pass interferometric processing, which allowed for characterization of change in water elevation between the two satellite overpasses. This temporal change measurement provided new insight into the hydraulic complexity of the Amazon River floodplain. Limitations of the repeat-pass interferometric SAR include the following: (1) It does not characterize water height at a given time, but change in water height between two satellite overpasses, and (2) it requires flooded vegetation to return the radar pulse to the antenna. Birkett et al. (2002) used TOPEX/POSEIDON profiling radar altimetry data to estimate water surface elevations for channelized rivers (as well as floodplains

818

SURFACE WATER

and lakes) across the Amazon Basin. A major limitation of profiling radar altimetry for fluvial hydrology applications is the very limited spatial coverage and coarse spatial resolution. LeFavour and Alsdorf (2005) used SRTM water surface elevations to calculate river slope and then to estimate discharge along the Amazon River using simple flow hydraulics analysis. A limitation of the SRTM dataset is the temporal coverage: The entire mission was a total of 11 days in duration. The Gravity Recovery and Climate Experiment (GRACE) maps changes in Earth’s gravity field. These gravity changes are a result of several factors, including changes in surface water storage. GRACE thus differs from all other remote sensing methods mentioned herein, since the latter rely on observed electromagnetic radiation. For instance, Crowley et al. (2006) characterized water storage anomalies over the Congo River Basin for a 5 year period. A major limitation of GRACE is the coarse spatial resolution, which is on the order of hundreds of kilometers. To overcome the limitations of existing remote sensing instruments, Alsdorf et al. (2007a) describe the need for an imaging (or wide swath) interferometric SAR altimeter to characterize water surface elevations and inundated area. The proposed satellite mission has been named Surface Water and Ocean Topography (SWOT) and has been recommended by the National Research Council Decadal Survey (NRC, 2007) to measure ocean topography as well as WSE over land; the proposed launch date time frame is between 2013 and 2016.

Water biogeochemical constituent characterization In some cases, water biogeochemistry can be inferred from remote sensing when biogeochemical constituents alter the interaction of water with electromagnetic radiation. For example, turbid water generally has a higher spectral reflectance than clear water at visible and NIR wavelengths. Wang and Shi (2008) mapped a massive algae bloom in China’s Lake Taihu using MODIS data. The MODIS spectral signature was used to quantify chlorophyll-a concentrations. Similarly, Chen et al. (2007) used the visible reflectances from the sea-viewing wide fieldof-view sensor (SeaWiFS) to characterize a composite pollution index for the Pearl River estuary in China. Summary Spaceborne and airborne remote sensing measurements have been used to characterize inundated area, water elevations, and water biogeochemical constituents. Surface water is easily distinguished from surrounding terrain by active and passive sensors at a variety of wavelengths, permitting for the characterization of inundated area. Ranging instruments have been used to characterize water surface elevations, permitting the estimation of storage changes and river discharge, which are two crucial terms in the water balance. Remote sensing instruments have also been used to map water quality in terms of biogeochemical

constituents. The recent launch of ALOS PALSAR in January 2006 is expected to greatly increase the ability to effectively map floodplains and wetlands from space. The proposed launch of SWOT in 2013–2016 has the potential to allow for more detailed characterization of surface water storage changes and discharge than has so far been possible from a spaceborne platform.

Bibliography Alsdorf, D. E., Melack, J. M., Dunne, T., Mertes, L. A. K., Hess, L. L., and Smith, L. C., 2000. Interferometric radar measurements of water level changes on the Amazon flood plain. Nature, 404, 174. Alsdorf, D. E., Rodríguez, E., and Lettenmaier, D. P., 2007a. Measuring surface water from space. Reviews of Geophysics, 45, RG2002. 2006RG000197. Alsdorf, D. E., Bates, P., Melack, J., Wilson, M., and Dunne, T., 2007b. Spatial and temporal complexity of the Amazon flood measured from space. Geophysical Research Letters, doi:10.1029/2007GL029447. Birkett, C. M., 1994. Radar altimetry: a new concept in monitoring lake level changes. Eos, Transactions American Geophysical Union, 75(24), 273. Birkett, C. M., Mertes, L. A. K., Dunne, T., Costa, M. H., and Jasinski, M. J., 2002. Surface water dynamics in the Amazon Basin: application of satellite radar altimetry. Journal of Geophysical Research, doi:10.1029/2001JD000609. Chen, C., Tang, S., Pan, Z., Zhan, H., Larson, M., and Jönsson, L., 2007. Remotely sensed assessment of water quality levels in the Pearl River estuary, China. Marine Pollution Bulletin, 54, 1267. Crowley, J. W., Mitrovica, J. X., Bailey, R. C., Tamisiea, M. E., and Davis, J. L., 2006. Land water storage within the Congo basin inferred from GRACE satellite gravity data. Geophysical Research Letters, doi:10.1029/2006GL027070. Elachi, C., 1987. Spaceborne Radar Remote Sensing: Applications and Techniques. New York: IEEE Press. Hess, L. L., Melack, J. M., Novo, E. M. L. M., Barbosa, C. C. F., and Gastil, M., 2003. Dual-season mapping of wetland inundation and vegetation for the central Amazon basin. Remote Sensing of Environment, 87, 404. Horritt, M. S., and Bates, P. D., 2001. Predicting floodplain inundation: raster-based modeling versus the finite element approach. Hydrological Processes, 15, 825. Kiel, B., Alsdorf, D. E., and LeFavour, G., 2006. Capability of SRTM C- and X-band DEM data to measure water surface elevations in Ohio and the Amazon. Photogrammetric Engineering and Remote Sensing, 72(3), 1. LeFavour, G., and Alsdorf, D., 2005. Water slope and discharge in the Amazon River estimated using the shuttle radar topography mission digital elevation model. Geophysical Research Letters, doi:10.1029/2005GL023836. National Research Council, 2007. Earth Science and Applications from Space: National Imperatives for the Next Decade and Beyond. Washington, DC: The National Academies Press. 418 pp. Prigent, C., Papa, F., Aires, F., Rossow, W. B., and Matthews, E., 2007. Global inundation dynamics inferred from multiple satellite observations. Journal of Geophysical Research, doi:10.1029/2006JD007847. Rees, W. G., 2001. Physical Principles of Remote Sensing. Cambridge: Cambridge University Press. Rosenqvist, A., Shimada, M., Ito, N., and Watanabe, M., 2007. ALOS PALSAR: a pathfinder mission for global-scale monitoring of the environment. IEEE Transactions on Geoscience and Remote Sensing, 45, 3307.

SURFACE WATER

Smith, L. C., 1997. Satellite remote sensing of river inundation area, stage, and discharge: a review. Hydrological Processes, 11, 1427. Smith, L. C., and Pavelsky, T. M., 2008. Estimation of river discharge, propagation speed, and hydraulic geometry from space: Lena River, Siberia. Water Resources Research, doi:10.1029/ 2007WR006133. Smith, L. C., Sheng, Y., MacDonald, G. M., and Hinzman, L. D., 2005. Disappearing arctic lakes. Science, 308, 1429. Wang, M., and Shi, W., 2008. Satellite-observed algae blooms in China’s Lake Taihu. Eos, Transactions American Geophysical Union, 89(22), 201.

819

Yi, Y., Lee, H., Ibaraki, M., and Shum, C., 2006. Louisiana wetland monitoring using TOPEX/POSEIDON altimetry. Eos, Transactions American Geophysical Union, 87(52), 1464.

Cross-references Geophysical Retrieval, Overview Microwave Radiometers Radar, Altimeters Radar, Scatterometers Radar, Synthetic Aperture

T

TERRESTRIAL SNOW Son V. Nghiem1, Dorothy K. Hall2, James L. Foster3 and Gregory Neumann1 1 Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA 2 Cryospheric Sciences Laboratory, Code 615, NASA/Goddard Space Flight Center, Greenbelt, MD, USA 3 Hydrological Sciences Laboratory, Code 617, National Aeronautics and Space Administration, Goddard Space Flight Center, Greenbelt, MD, USA

Definitions Among key parameters characterizing seasonal snow cover in cold land regions are snow area, extent, depth, water equivalent, accumulation onset, melt onset, ice layer, melt duration, and season length. Within an area p(i,j) that has a fractional snow cover of fS(i,j), the actual area that is fully covered by snow is sA(i,j) ¼ fS(i,j) · p(i,j) at a location determined by indices i and j on the Earth’s surface. Then, the total snow area SA is the summation of all sA(i,j). Depending on the sensitivity, accuracy, and resolution of a satellite sensor, the existence of snow cover in p(ij) can be detected when sA(i,j) is sufficiently large. The composition of all areas p(i,j) where snow is detectable constitutes the snow extent SE, which is smaller than SA unless snow fully covers all snow areas sA(i,j). Over area p(i,j), the snow volume is calculated as vS(i,j) ¼ da(i,j) · p(i,j) where da(i,j) is the average snow depth in area p(i,j), including the snow-free fraction [1  fS(i,j)] in p(i,j) with zero snow depth. Rather than snow volume, it is more useful for hydrologists to use snow water equivalent (SWE), which is equal to the depth of water that ensues when the snow volume melts on p(i,j). Let ra(i,j) be the average bulk density of snow and rw the

water density, then SWE(i,j) ¼ da(i,j) · ra(i,j)/rw is the average SWE over area p(i,j). The first snowfall in area p(i,j) may be short lived because air and land temperatures are not cold enough to sustain it. As the landscape freezes, subsequent snowfalls may accumulate and contribute the snowpack during a snow season. The day when snow starts to consistently accumulate to build up the seasonal snowpack is called the snow accumulation onset date (TS). During a snow season, an anomalous occurrence of snowmelt may create excessive meltwater, which refreezes in the snowpack, forming an ice layer that contains large icy objects (ice clumps, lens, columns, etc.). As winter transitions into spring and then summer, snowmelt occurs more consistently and continuously (as opposed to isolated anomalous melt in winter such as that caused by warm air advection). Snowmelt onset date (TM) is defined as the day when a consistent snowmelt starts to occur. Snow disappearance date (TF) is the day when area p(ij) becomes snow-free as snow completely melts away. Then, snowmelt duration is DM ¼ TM  TF and snow season length is DS ¼ TS  TF. These key snow parameters can be obtained using remote sensing data from multispectral visible infrared and microwave satellite sensors for each pixel p(ij).

Significance of snow cover Climate change at high latitudes is strongly influenced by the albedo-temperature feedback process. The snow cover influences the global heat budget (Robinson and Kukla, 1985; Foster and Chang, 1993) because it affects the planetary albedo and the Earth’s outgoing longwave radiation (Groisman et al., 1994). Snow is regarded as one of the key variables in global change monitoring (Walsh, 1991; Hall et al., 2005). While large changes have been observed in many cold land environments (sea ice, ice sheet, glacier, tundra, and

E.G. Njoku (ed.), Encyclopedia of Remote Sensing, DOI 10.1007/978-0-387-36699-9, © Springer Science+Business Media New York 2014

822

TERRESTRIAL SNOW

permafrost), Bartlett et al. (2005) indicate no significant trends in the mean snow onset or duration over North America from 1950 to 2002, though their averaging approach over a hemispheric scale may suppress regional snow change signatures. Other studies (Chapin et al., 2005; Schwartz et al., 2006; Bonsal and Prowse, 2003) suggest that the Arctic snow season may have already decreased by more than 2.5 days per decade and that snow is melting earlier (Foster et al., 1992; Robinson et al., 1993; Frei and Robinson, 1999; Brown, 2000). This discrepancy illustrates a need for advanced satellite measurements of snow cover from local to regional and hemispheric scales. Consistent with many climate change projections, significant changes in snow regimes are occurring throughout the Arctic. These changes have important consequences for humans, wildlife, and ecosystems. Recent changes in temperature and precipitation have altered the timing of the snow season and the snow-cover characteristics. The amount of snow (depth and SWE) has probably changed as well (Groisman and Davies, 2001), though this is a difficult quantity to measure accurately over large areas (Benson, 1982; Goodison et al., 1998; Yang et al., 2005; Liston and Sturm, 2004) compared to satellite measurements of snow extent. Furthermore, winter rain is likely more prevalent (Putkonen and Roe, 2003). It is critically important to comprehensively document these spatial and temporal changes with satellite data. Moreover, satellite remote sensing of snow has immediate applications to snowmelt, river discharge, and flood forecasting. The timing and magnitude of river discharge in the Arctic is largely controlled by cold season snow mass storage and subsequent melt. In northern watersheds, snowmelt accounts for most of the annual river discharge and often produces the highest discharge rates of the year (Lo and Serreze, 2002; Bowling et al., 2000). Heavy snowstorms with rapid snowmelt in the spring of 1997 accounted for the “flood of the century” in the US northern Great Plains (Nghiem and Tsai, 2001), causing more than four billion dollars in flood-related damages (National Climatic Data Center, 1997). The ability to identify and track changes in snow regimes with increasing accuracy is an important asset for decision makers.

Seasonal snow extent Satellite multispectral sensors have collected global data since the late 1960s, and algorithms for snow applications have been developed and improved. Satellite visible-band radiometers detect snow by its high albedo and characteristic patterns (Wiesnet and Matson, 1979) and by its spectral signature. An archive of quality-controlled National Oceanic and Atmospheric Administration (NOAA) weekly and now daily snow products is available from the Rutgers Global Snow Lab at Rutgers University (http://climate.rutgers.edu/snowcover/). The Earth Observing System (EOS) Terra and Aqua spacecraft, launched in 1999 and 2002, respectively, each carry an

instrument called the Moderate Resolution Imaging Spectroradiometer (MODIS). MODIS has 36 channels covering visible to thermal infrared parts of the spectrum (0.4–14.4 mm) with a pixel size of 250 m–1 km and a swath width of 2,330 km, providing global coverage daily in most seasonally snow-covered regions. MODIS data are used to obtain snow-cover products, including snow extent (SE) and snow fraction (fS), from automated algorithms at the Goddard Space Flight Center (GSFC) in Greenbelt, Maryland (http://modis-snow-ice. gsfc.nasa.gov). MODIS snow products are archived and distributed at the National Snow and Ice Data Center (NSIDC) in Boulder, Colorado. Global snow products are provided at 500 m, 5 km, or 25 km resolution as daily, 8 day composite, and monthly validated snow maps (Hall et al., 2002; Hall and Riggs, 2007). An example of a monthly, global MODIS snow-cover map is presented in Figure 1 for February 2004. The MODIS map shows snow extent from mid- to high latitudes over the Northern Hemisphere with smaller snow fraction at midlatitudes. Note that Western Europe (UK, France, Denmark, Germany, etc.) was either snow-free or did not have much snow cover compared to other regions at the same latitudes (Figure 1). At very high latitudes, such as in the Canadian Arctic Archipelago and Taymyr Peninsula in Siberia, there were difficulties in identifying snow due to winter darkness and cloud cover. MODIS snow-cover maps complement existing hemispheric-scale snow maps that are available today (e.g., the NOAA maps (Ramsay, 1998)) and provide advances in spatial resolution and snow-cloud discrimination capabilities. The data archiving and distribution system provides free web-based access to the MODIS results. While MODIS has several technological advantages, the Advanced Very High Resolution Radiometers (AVHRR), with four to six channels (0.58–12.5 mm), have over three decades of data providing long-term measurements of seasonal snow extent (Foster et al., 2008a). Furthermore, active and passive microwave data can be used to supplement visible and near-infrared data over areas under clouds and darkness (Foster et al., 2008b).

Snow depth and snow water equivalent The microwave brightness temperature emanating from the ground beneath a snow cover and from the snow itself is sensitive to snow depth, SWE, size of individual snow crystals, and other physical properties of snow. This interaction can be detected by spaceborne passive microwave radiometers. In the past three decades, considerable progress has been made in estimating regional and global snow depth and SWE using instruments such as the Scanning Multichannel Microwave Radiometer (SMMR) and the Special Sensor Microwave Imager (SSM/I). For more information on this topic, see results by Chang et al. (1987), Foster et al. (1997), Chen et al. (2001), and Pulliainen and Hallikainen (2001). Further progress has also been made in developing consistent multisensor

TERRESTRIAL SNOW

823

Terrestrial Snow, Figure 1 MODIS monthly snow-cover product (MOD10CM, Collection 5) for February 2004 showing snow extent and fraction of snow cover with a spatial resolution of 0.05 or about 5 km resolution. The color key is for fractional snow cover from 1 % to 100 %. Gray represents areas of persistent cloud cover according to the cloud mask.

passive microwave SWE datasets for climate studies (Armstrong and Brodzik, 2001). The Advanced Microwave Scanning Radiometer – EOS instrument (AMSR-E) on the Aqua satellite continued the heritage of earlier passive microwave Earth observations but included enhanced spatial resolution and an expanded range of frequency bands (6.925–89.0 GHz). The AMSRE swath width was approximately 1,450 km. Research conducted in support of the implementation of AMSR-E for snowpack monitoring (Chang and Rango, 2000; Chang and Kelly, 2002; Kelly et al., 2003) has led to an automated algorithm for global SWE, at a pixel resolution of 25 km. The map in Figure 2 for February 10, 2001, shows an area of high SWE values in Labrador, Canada, which is consistent with the snow climatology in this region. However, due to the high sensitivity of the microwave response to varying snow properties, retrieval algorithms often produce large random estimation errors. Shallow or discontinuous snowpack, complex stratigraphy, large average snow grain size, or the presence of liquid water can adversely affect the algorithm performance and thus complicate the identification of changes in snow conditions. Backscatter measurements from active microwave sensors complement the passive microwave data. Backscatter data from satellite scatterometers can also be useful for estimating snow depth and SWE. The SeaWinds scatterometer operating at Ku-band (13.4 GHz) aboard

the QuikSCAT satellite (QSCAT) was launched in June 1999. This sensor collects backscatter data with a resolution from 12 to 25 km over a swath of 1,400 km for the horizontal polarization (H) and 1,800 km for the vertical polarization (V). QSCAT already obtained global data for over a decade. Ku-band backscatter is sensitive to snow accumulation (Nghiem et al., 2001; Cline et al., 2004). For example, QSCAT backscatter signatures and snow parameters at a location in the Yellowstone National Park (44.65  N, 111.1  W) in Montana, USA, are shown in Figure 3. A regression analysis between QSCAT H backscatter and SWE data from the National Operational Hydrologic Remote Sensing Center (NOHRSC) National Snow Analysis (NSA) yields a positive correlation coefficient of 0.970, with a change of 108 mm per dB for SWE up to 200 mm. For SWE > 200 mm, the correlation is weaker and the relationship is less accurate because of attenuation effects in backscatter data and uncertainties in NOHRSC SWE results. Based on the rating curve derived from NOHRSC data (Nghiem et al., 2004), QSCAT data were used to estimate SWE values (bottom panel in Figure 3), corresponding to a range of snow density ratio (ra/rw) from 22 % to 56 % at Yellowstone. The Yellowstone case represents the “scattering regime” where backscatter increases with snow depth and SWE. In the “attenuation regime” (see below “Anomalous melt events and ice layers”), backscatter decreases

824

TERRESTRIAL SNOW

0

50 100 150 200 250 Snow water equivalent on February 10, 2001

300

Terrestrial Snow, Figure 2 Map of snow water equivalent (SWE) in millimeter (mm) derived from satellite microwave radiometer data (NASA/GSFC).

with snow accumulation. Both scattering and attenuation regimes can exist in a given area (Figure 4), so identifying the correct regime is important to the retrieval of snow depth and SWE. Temporal backscatter data can help improve results obtained from passive microwave snow observations. A new composite algorithm combining both active and passive data needs to be developed for more accurate SWE estimates.

Snowmelt pattern With a wavelength of 2.2 cm (13.4 GHz), QSCAT backscatter is highly sensitive to snow wetness. Wave propagation and attenuation in an inhomogeneous medium like snow are characterized with a complex effective permittivity, which is determined by a mixture of air, ice grains, and liquid water (in the case of melting snow). The attenuation of Ku-band electromagnetic waves is controlled by the imaginary part of the snow effective permittivity. For dry snow, the imaginary part for pure dry water ice is in the order of 0.002e0 (Tiuri et al., 1984) where e0 is the permittivity of free space. In wet snow, liquid water (no salinity) has an imaginary part of about 38e0 (19,000 times larger than that of dry water ice) (Klein and Swift, 1977). Thus, a small amount of wetness can significantly change the imaginary part of the snow effective permittivity and, consequently, decrease the backscatter. This snow-scattering physics has been modeled by a number of researchers (Tsang et al., 1985; Ulaby et al., 1981; Nghiem et al., 1995). This sensitivity allows the backscatter to be used for snowmelt detection.

From experimental measurements with artificial piling of snow, a backscatter decrease of more than 5 dB is observed for 3 % wetness (e.g., Stiles and Ulaby, 1980). In a natural snow field, the Alaska Experiment (ALEX 1999) for snowmelt measurements was carried out in Ft. Wainwright, Alaska. In ALEX, a tower-mounted Ku-band scatterometer similar to QSCAT monitored a taiga snowpack, starting with freezing conditions and continuing through the snowmelt process until snow completely melted away (Nghiem et al., 2000). Taiga snow is a major snow class, covering about half of Alaska and Canada (Sturm et al., 1995). Results from ALEX show that diurnal backscatter can change as much as 15 dB (more than 30 times) between cold morning and warm afternoon when the heat makes more meltwater in the snowpack. These measurements confirm the high sensitivity of the backscatter to snowmelt. The diurnal difference method was developed by Nghiem et al. (2001) to monitor the snowmelt process. The diurnal backscatter difference is a relative quantity between morning and afternoon measurements in half a day. QSCAT data from the early morning (ta) in an ascending orbit pass and from late afternoon (tp) in a descending pass are collocated for each day. The diurnal backscatter change is defined as the backscatter difference in the decibel (dB) domain as DsVV ¼ sVV(tp)  sVV(ta) (Nghiem et al., 2001), where sVV is the V backscatter and all quantities are in dB. We use sVV to take advantage of the larger swath and coverage. In the case of snowmelt on the Greenland ice sheet, the criteria for the melt detection is based on DsVV greater than 1.8 dB for melt and less

−8

a

825

H ascending

−10 −12 −14

3 2 .

1 0

Snow Depth (cm)

. 120

c

−1

NCDC/GSOD point measurement

80 .

40 0

. 50 25 0

.

d

−50

NOHRSC SWE for 2003–2004 QuikSCAT SWE for 1999–2003

e

08 9/9 /2 9 10 8/9 /2 9 12 7/9 /2 9 02 6/9 /2 9 04 4/0 /2 0 06 4/0 /2 0 08 3/0 /2 0 10 2/0 /2 0 12 1/0 /2 0 02 0/0 /1 0 04 8/0 /1 1 06 9/0 /1 1 08 8/0 /1 1 10 7/0 /1 1 12 6/0 /1 1 02 5/0 /1 1 04 3/0 /1 2 06 4/0 /1 2 08 3/0 /1 2 10 2/0 /1 2 12 1/0 /1 2 02 0/0 /0 2 04 8/0 /0 3 06 9/0 /0 3 08 8/0 /0 3 10 7/0 /0 3 12 6/0 /0 3 02 5/0 /0 3 04 3/0 /0 4 06 3/0 /0 4 08 2/0 /0 4 1/ 04

500 400 300 200 100 0

−25

.

06

/2

SWE (mm)

V Diurnal (dB)

4

b

Temperature (C)

Backscatter (dB)

TERRESTRIAL SNOW

Terrestrial Snow, Figure 3 QSCAT signatures and snow parameters at Yellowstone (44.65  N and 111.1  W) in Montana, USA: (a) H backscatter from ascending orbits, (b) Diurnal V backscatter difference, (c) snow depth from the National Climatic Data Center (NCDC) Global Summary of the Day (GSOD) data, (d) minimum, average, and maximum air temperatures, and (e) SWE from QSCAT and from NOHRSC data. Snow depth and temperature data are from NCDC GSOD dataset. This area consists of 65 % coniferous forest, 23 % grasslands, 5 % woody savannas, and 4 % shrub lands (3 % unknown). Vertical blue lines indicate snow accumulation onsets, and red vertical lines mark snowmelt onsets.

than 1.0 dB for reduced melt or refreezing conditions (Steffen et al., 2004). This algorithm is adapted for snowmelt detection on land. This method has several advantages: independence from the scatterometer long-term gain drift, independence from the cross-calibration between QSCAT and future satellite scatterometers, independence from absolute backscatter from different snow classes and snow conditions, detection of both snowmelt and refreezing, and daily coverage over most cold land regions. A unique feature of this method is that the diurnal backscatter difference stays near 0 dB before and after snowmelt. This feature holds for different snow classes under different physical conditions in different areas or regions. Therefore, a consistent and accurate method for snowmelt mapping can be applied over the continental scale. Because snowmelt is

a temporal process, each pixel is treated not as an independent measurement in time for a given snapshot, but as a part of an ongoing time-dependent process represented by a multiyear dataset for each pixel p(i,j). The conglomerate of land pixels is analyzed to map snowmelt patterns at the continental scale. The diurnal difference algorithm is applied to monitor snowmelt from mid-February to mid-June (Figure 5). The snowmelt map on March 21, 2007 (spring equinox), reveals a band of snow with reduced melt or refrozen conditions (light blue) stretching from the west of the Great Lakes toward the Pacific coast on North America. Below and adjacent to this reduced melt band was a stretch of land where the snowmelt process was completed (light brown). In Eastern Europe and Siberia, several snow regions can be observed at midlatitudes (red). Overall,

826

TERRESTRIAL SNOW

Backscatter (dB)

−11 −13

Snowmelt

Snow season

−9

A

−15

N

S

−17 −19

a

Oct 10

Jun 13

May 19

Sep 16 Oct 03

4

Jun 01

May 10

Sep 26 Oct 31

Jun 09

May 17

Sep 26

May 27

May 06

Oct 01

Jun 11

May 27

Sep 14

Jun 11

May 26

V Diurnal (dB)

6

Oct 13

8

2 0

b

10

/0 1 12 /99 /3 0 03 /99 /2 9 06 /00 /2 7 09 /00 /2 5 12 /00 /2 4 03 /00 /2 4 06 /01 /2 2 09 /01 /2 0 12 /01 /1 9 03 /01 /1 9 06 /02 /1 7 09 /02 /1 5 12 /02 /1 4 03 /02 /1 4 06 /03 /1 2 09 /03 /1 0 12 /03 /0 9 03 /03 /0 8 06 /04 /0 6 09 /04 /0 4 12 /04 /0 3 03 /04 /0 3 06 /05 /0 1 08 /05 /3 0 11 /05 /2 8/ 05

−2

Terrestrial Snow, Figure 4 QSCAT signatures at Tunalik (70.2  N, 161.1  W) in Alaska, USA. The upper panel is for V backscatter and the lower panel is for diurnal V backscatter change. Vertical blue lines indicate snow accumulation onsets (TS), red vertical lines mark melt onsets (TM), and green lines are for snow departure dates (TF). In the upper panel, horizontal blue arrows span the snow season length (DS) and the red arrows mark the snowmelt period (DM). The red dots on the horizontal red line (2 dB) in the QSCAT diurnal plot indicate anomalous melt events occurring in a snow season before melt onset. In the upper panel, symbol S denotes the scattering regime, A represents the attenuation regime, and N is for the neutral regime.

most of the Northern Hemispheric snowpack was in freezing conditions by the spring equinox. One month later (April 21, 2007), a very extensive melt band occurred at higher latitudes from Labrador across the Canadian Shield to Alaska, while snowmelt in Eastern Europe was completed and large areas of active melt were seen in eastern Siberia. By May 21, 2007 (a month before the summer solstice), most of the snowpack was melted except at very high latitudes in the Canadian Arctic, the Alaskan North Slope, and the Taymyr Peninsula in north Siberia. It is fortuitous that the QSCAT dataset covers this decade when significant changes have occurred, and this dataset is therefore invaluable for observations of changes in the Arctic snowmelt pattern.

Anomalous melt events and ice layers Passive microwave emission is dependent on temperature and moisture in snow (Hallikainen et al., 2002). Melting

and refreezing of snow results in wet and dry snow metamorphism and changes in snow grain size (Colbeck, 1986; Colbeck et al., 1992). Refreezing of liquid water, from either melt or rain, can form an ice layer with large scatterers. These changes modify the passive microwave signature and may affect its sensitivity to snow accumulation. For QSCAT backscatter, the scattering at Ku-band frequency follows the Rayleigh scattering law (Nghiem and Tsai, 2001), which dictates the scattering cross section to be proportional to the sixth power of grain size. Thus, QSCAT data can be used as a highly sensitive indicator for changes caused by snow metamorphism from processes such as melt and refreezing. For example, an anomalous event occurred at Tunalik on October 31, 2003 (Figure 4), when in situ meteorological data showed melt and rain, followed by refreezing. This event is characterized by a sharp peak in the QSCAT backscatter at the time indicated by the second red dot on the red horizontal line in the QSCAT diurnal plot in Figure 4. Collocated and

TERRESTRIAL SNOW

827

The sharp increase in QSCAT backscatter on October 31, 2003, and the decreasing trend in backscatter afterward indicate the formation of an ice layer in snow (Nghiem et al., 2005). Such detection allows a determination of the formation timing and extent of the ice layer. Moreover, QSCAT backscatter decreased as snow accumulated after the ice layer formation (October 31, 2003), indicating that the strong backscatter from the ice layer was attenuated by the subsequent overlaying snow accumulation (attenuation regime). In the previous years when there was no anomalous melt to form an ice layer, backscatter increased as snow accumulated from the start of each snow season, indicating that scattering increased with snow accumulation (scattering regime). In the snow season from September 2004 to May 2005, QSCAT backscatter did not show a clear trend with snow accumulation after the anomalous melt on October 3, 2004 (third red dot in the lower panel, Figure 4), indicating a neutral regime with a balance between scattering and attenuation effects. The detection of the switch between scattering and attenuation regimes or the identification of the neutral regime is useful for a better interpretation of passive microwave data for applications to SWE retrieval.

Terrestrial Snow, Figure 5 QSCAT snowmelt patterns on the 21st day of March, April, and May in 2007. White represents frozen snow, red is for active snowmelt, blue for snow with reduced melt or refrozen conditions, light brown for areas where the snowmelt process is completed, dark brown for land where QSCAT cannot detects snowmelt, and dark blue for water (oceans and lakes).

contemporaneous data from AMSR-E and SSM/I revealed that brightness temperature started to decrease on the same date; however, the minimum brightness temperature would not occur until 1 week later.

Timing of snow accumulation, melt onset, disappearance, and season length Snow accumulation onset date (TS), melt onset date (TM), and disappearance date (TF) are denoted, respectively, with blue, red, and green vertical lines in Figure 4 for the location at Tunalik in Alaska. Snow accumulation onset corresponds to the minimum in backscatter at the start of a snow season (Nghiem and Tsai, 2001). Snowmelt onset causes a large change in QSCAT diurnal signature consistently maintained in time without the long subsequent refreezing in the case of anomalous melt in winter. Snow disappearance occurs when backscatter is minimized and diurnal signature returns to a value near 0 dB. At Tunalik, snowmelt duration (DM) varied from 15 to 25 days, while snow season length (DS) changed as much as a month in the 6 years (1999–2005) of the QSCAT data. In 2002, snowmelt onset was the earliest, followed by the longest snow-free season (Figure 4). Snowmelt typically results in an increase in grain size at the snow surface due to preferential melting of smaller grains or due to multigrain coalescence. MODIS has sufficient sensitivity to observe the spectral signatures resulting from the change in surface grain size and has been used to distinguish grain size categories (e.g., small, medium, and large) where there is continuous snow cover. A combination of AMSR-E, QSCAT, and MODIS signatures (satellite data collocated in space and closest in time) can better constrain the identification of snow onset, melt events, and snow season length. Rain on snow-free land may have a similar diurnal signature in microwave data and thus cause a misidentification of snowmelt over such area. Here, MODIS data can help prevent the misidentification

828

TERRESTRIAL SNOW

Terrestrial Snow, Figure 6 Blended snow product (April 20, 2003) with a combination of snow extent (white) obtained from MODIS data and active snowmelt areas (red) derived from QSCAT data in North America (land inside the red boundary).

because of the capability of MODIS data to determine snow-covered and snow-free areas better than microwave data for thin snow areas.

Blended products Because each type of satellite sensor, such as MODIS, AMSR-E, or QSCAT, has different sensitivities to different snow parameters, combinations of measurements from different sensors can be used to complement and improve the accuracy of snow measurement in a way that cannot be achieved by using a single sensor. Moreover, for many practical applications, combining different snow measurements from disparate products based on different satellite datasets with various resolutions, acquisition times, coverages, and formats is indisputably challenging for many users and decision makers. In this view, development of a snow mapping product that blends together different snow measurements is beneficial. Efforts to blend multiple snow datasets into a single, global, daily, and user-friendly product, such as the Air Force Weather Agency/National Aeronautics and Space Administration Snow Algorithm (ANSA), and to improve these products were supported in a research project (see Foster et al., 2008b for further details). A simple example of a blended product is presented in Figure 6. This product combines snow extent measured by MODIS together with actively melting areas obtained from QSCAT data. The map of blended MODIS-QSCAT product (Figure 6) shows melt areas that consistently

occur within the limit of snow extent on April 20, 2003, in North America. An extensive melt band was observed across North America from Newfoundland, Canada, to Alaska, USA, between 50 and 70 latitudes. There are snow areas at lower latitudes or lower elevations that are not detected by QSCAT. This can be due to the low sensitivity of QSCAT backscatter to thin snow in such areas. Future developments for blended products will combine more snow products and will utilize more complex algorithms to account for temporal and spatial differences.

Summary Satellite remote sensing of key parameters characterizing terrestrial snow cover, including snow extent, depth, water equivalent, snowmelt patterns, and seasonal snow dynamics (timing of accumulation onset, melt onset, disappearance, and length of snow season), have been presented. Data from optical infrared sensors (e.g., MODIS and AVHRR), passive microwave radiometers (e.g., SSM/I and, in the recent past, AMSR-E), and active microwave scatterometers (e.g., QSCAT) are used to measure snow parameters from local to regional and global scales on a daily to weekly basis. Techniques to blend different snow products into a single user-friendly global product are under development. Furthermore, different signatures corresponding to different scattering and attenuation regimes are identified based on snow physics. Such information will be valuable in the development of future satellite missions for global snow measurements, such as the

TERRESTRIAL SNOW

Cold Regions Hydrology High-resolution Observatory (CoReH2O) Mission (Rott et al., 2007) and the Snow and Cold Land Processes (SCLP) Mission (National Research Council, 2007).

Acknowledgments The research carried out at the Jet Propulsion Laboratory, California Institute of Technology, was supported by the National Aeronautics and Space Administration (NASA) Terrestrial Hydrology Program and by the US Air Force (USAF) under an agreement with NASA. The research at the NASA Goddard Space Flight Center was also supported by the USAF. The authors would like to thank Janet Y. L. Chien for preparing Figure 1 and D. Cline of NOAA NOHRSC for the NSA SWE data at Yellowstone. Bibliography Armstrong, R. L., and Brodzik, M. J., 2001. Recent northern hemisphere snow extent: a comparison of data derived from visible and microwave satellite sensors. Geophysical Research Letters, 28, 3673–3676. Bartlett, M. G., Chapman, D. S., and Harris, R. N., 2005. Snow effect on North American ground temperatures, 1950–2002. Journal of Geophysical Research, 110, F03008, doi:10.1029/ 2005JF000293. Benson, C. S., 1982. Reassessment of winter precipitation on Alaska’s Arctic Slope and measurements on the flux of wind blown snow. Geophysical Institute, University of Alaska Report UAG R-288, 26 pp. (Available from Geophysical Institute, University of Alaska, P.O. Box 757320, Fairbanks, AK 99775–7320). Bonsal, B. R., and Prowse, T. D., 2003. Trends and variability in spring and autumn 0 C-isotherm dates over Canada. Climatic Change, 57, 341–358. Bowling, L. C., Lettenmaier, D. P., and Matheussen, B. V., 2000. Hydroclimatology of the Arctic drainage basin. In Lewis, L. (ed.), The Freshwater Budget of the Arctic Ocean. New York: Springer. Chap. 4. Brown, R. D., 2000. Northern hemisphere snow cover variability and change, 1915–1997. Journal of Climate, 13(13), 2339–2355. Chang, A. T. C., and Kelly, R. E. J., 2002. Description of snow depth retrieval algorithm for ADEOS II AMSR. EORC Bulletin: Technical Report No. 9, ISSN: 1346–7913, pp. 70–78. Chang, A. T. C., and Rango, A., 2000. Algorithm Theoretical Basis Document (ATBD) for the AMSR-E Snow Water Equivalent Algorithm, Version 3.1. Greenbelt, MD: NASA/GSFC. Chang, A. T. C., Foster, J. L., and Hall, D. K., 1987. Microwave snow signatures (1.5 mm to 3 cm) over Alaska. Cold Regions Science and Technology, 13(2), 153–160. Chapin, F. S., Sturm, M., Serreze, M. C., McFadden, J. P., Key, J. R., Lloyd, A. H., McGuire, A. D., Rupp, T. S., Lynch, A. H., Schimel, J. P., Beringer, J., Chapman, W. L., Epstein, H. E., Euskirchen, E. S., Hinzman, L. D., Jia, G., Ping, C. L., Tape, K. D., Thompson, C. D. C., Walker, D. A., and Welker, J. M., 2005. Role of land-surface changes in arctic summer warming. Science, 310, 657–660. Chen, C. T., Nijssen, B., Guo, J., Tsang, L., Wood, A. W., Hwang, J., and Lettenmaier, D. P., 2001. Passive microwave remote sensing of snow constrained by hydrological simulations. IEEE Transactions on Geoscience and Remote Sensing, 39, 1744–1756. Cline, D., Yueh, S. H., Nghiem, S. V., and McDonald, K., 2004. Ku-band radar response to terrestrial snow properties.

829

Eos Transactions, AGU, 85(47), Fall meeting supplement, Abstract H23D–1149. Colbeck, S. C., 1986. Classification of seasonal snow cover crystals. Water Resources Research, 22(9), 59S–70S. Colbeck, S., Akitaya, E., Armstrong, R., Gubler, H., Lafeuille, J., Lied, K., McClung, D., and Morris, E., 1992. The International Classification for Seasonal Snow on the Ground. Hanover, NH: U.S. Army CRREL, p. 23. Council, N. R., 2007. Earth Science and Applications from Space: National Imperatives for the Next Decade and Beyond. Washington, DC: The National Academies Press. Foster, J. L., and Chang, A. T. C., 1993. Global snow cover. In Gurney, R. J., Parkinson, C. L., and Foster, J. L. (eds.), Atlas of Satellite Observations Related to Global Change. Cambridge: Cambridge University Press, pp. 361–370. Foster, J., Winchester, J., and Dutton, E., 1992. The date of snow disappearance on the Arctic tundra as determined from satellite, meteorological station and radiometric in situ observations. IEEE Transactions on Geoscience and Remote Sensing, 30(4), 793–798. Foster, J. L., Chang, A. T. C., and Hall, D. K., 1997. Comparison of snow mass estimates from a prototype passive microwave snow algorithm, a revised algorithm and snow depth climatology. Remote Sensing of Environment, 62, 132–142. Foster, J. L., Robinson, D. A., Hall, D. K., and Estilow, T. W., 2008a. Spring snow melt timing and changes over Arctic lands. Polar Geography, 31(3–4), 145–157. Foster, J., Hall, D. K., Eylander, J., Riggs, G., Kim, E., Tedesco, M., Nghiem, S. V., Kelly, R. E. J., and Choudhury, B., 2008b. A new blended global snow product using visible, passive microwave, and scatterometer data, 12.1. In Nineteenth Conference on Hydrology, 88th American Meteorological Society Annual Meeting, New Orleans, LA. Frei, A., and Robinson, D. A., 1999. Northern hemisphere snow extent: regional variability 1972–1994. International Journal of Climatology, 19(14), 1535–1560. Goodison, B. E., Louie, P. Y. T., and Yang, D., 1998. WMO soil precipitation measurement intercomparison. Final Report, WMO/TD 872. Geneva: World Meteorological Organization, 212 pp. Groisman, P. Y., and Davies, T. D., 2001. Snow cover and the climate system. In Jones, H. G., Pomeroy, J. W., Walker, D. A., and Hoham, R. W. (eds.), Snow Ecology. Cambridge: Cambridge University Press, pp. 1–44. Groisman, P. Y., Karl, T. R., Knight, R. W., and Stenchikov, G. L., 1994. Changes of snow cover, temperature, and radiative heat balance over the northern hemisphere. Journal of Climate, 7(11), 1633–1656. Hall, D. K., and Riggs, G. A., 2007. Accuracy assessment of the MODIS snow-cover products. Hydrological Processes, 21(12), 1534–1547, doi:10.1002/hyp. 6715. Hall, D. K., Riggs, G. A., Salomonson, V. V., DiGirolamo, N. E., and Bayr, K. J., 2002. MODIS snow-cover products. Remote Sensing of Environment, 83, 181–194. Hall, D. K., Kelly, R. E. J., Foster, J. L., and Chang, A. T. C., 2005. Estimation of snow extent and snow properties. In Anderson, M. (ed.), Encyclopedia of Hydrologic Sciences. Chichester: Wiley, pp. 811–830. Chap. 55. Hallikainen, M., Halme, P., Takala, M., and Pulliainen, J., 2002. Effects of temperature and moisture of snow and soil on SSM/I response to snow. IEEE Transactions on Geoscience and Remote Sensing. In Proceedings of International Geoscience and Remote Sensing Symposium, 1, 680–682. Kelly, R. E., Chang, A. T., Tsang, L., and Foster, J. L., 2003. A prototype AMSR-E global snow area and snow depth algorithm. IEEE Transactions on Geoscience and Remote Sensing, 41(2), 230–242.

830

THERMAL RADIATION SENSORS (EMITTED)

Klein, L. A., and Swift, C., 1977. An improved model for the dielectric constant of sea water at microwave frequencies. IEEE Transactions on Antennas and Propagation, AP-25(1), 104–111. Liston, G. E., and Sturm, M., 2004. The role of winter sublimation in the Arctic moisture budget. Nordic Hydrology, 35(4), 325–334. Lo, R., and Serreze, M. C., 2002. Using snow water equivalent to understand the hydrology of the Arctic drainage system. American Geophysical Union, Fall meeting 2002, Abstract H51A–0762, San Francisco, CA. National Climatic Data Center, 1997. The spring of 1997, reviewing four significant weather events. Asheville, NC: National Environmental Satellite, Data, and Information Service. Technical Report, TR 97–03. Nghiem, S. V., and Tsai, W. Y., 2001. Global snow cover monitoring with spaceborne Ku:band scatterometer. IEEE Transactions on Geoscience and Remote Sensing, 39(10), 2118–2134. Nghiem, S. V., Kwok, R., Yueh, S. H., Kong, J. A., Tassoudji, M. A., Hsu, C. C., and Shin, R. T., 1995. Polarimetric scattering from layered media with multiple species of scatterers. Radio Science, 30(4), 835–852. Nghiem, S. V., Sturm, M., Perovich, D. K., Tsai, W., Neumann, G., Taras, B., and Elder, B., 2000. Large-scale snow cover monitoring with Sea Winds/QuikSCAT scatterometer. Eos Transactions, AGU, 81(48), Fall meeting supplement, Abstract H62G–04. Nghiem, S. V., Steffen, K., Kwok, R., and Tsai, W.-Y., 2001. Detection of snow melt regions on the Greenland ice sheet using diurnal backscatter change. Journal of Glaciology, 47(159), 539–547. Nghiem, S. V., Yueh, S. H., Cline, D., Sturm, M., and Neumann, G., 2004. Ku-band radar signature and snow accumulation inversion. In NASA Terrestrial Hydrology Meeting, Maryland. Nghiem, S. V., Steffen, K., Neumann, G., and Huff, R., 2005. Mapping of ice layer extent and snow accumulation in the percolation zone of the Greenland ice sheet. Journal of Geophysical Research, 110, F02017, doi:10.1029/ 2004JF00234. Pulliainen, J., and Hallikainen, M., 2001. Retrieval of regional snow water equivalent from space-borne passive microwave observations. Remote Sensing of Environment, 75, 76–85. Putkonen, J., and Roe, G., 2003. Rain-on-snow events impact soil temperatures and affect ungulate survival. Geophysical Research Letters, 30(4), 1188, doi:10.1029/2002GLO16326. Ramsay, B., 1998. The interactive multisensor snow and ice mapping system. Hydrological Processes, 12, 1537–1546. Robinson, D. A., and Kukla, G., 1985. Maximum surface albedo of seasonally snow covered lands in the northern hemisphere. Journal of Climate and Applied Meteorology, 24, 402–411. Robinson, D. A., Dewey, K. F., and Heim, R. R., 1993. Global snow-cover monitoring: an update. Bulletin of the American Meteorological Society, 74(9), 1689–1696. Rott, H., Cline, D., Nagle, T., Pulliainen, J., Rebhan, H., and Yueh, S., 2007. CoRe-H2O – a dual frequency SAR mission for hydrology and climate research. In Proceedings of International Geoscience and Remote Sensing Symposium. Barcelona, pp. 1204–1206. Schwartz, M. D., Rein, A., and Aasa, A., 2006. Onset of spring starting earlier across the northern hemisphere. Global Change Biology, 12, 343–351, doi:10.1111/j.1365-2486.2005.01097. Steffen, K., Nghiem, S. V., Huff, R., and Neumann, G., 2004. The melt anomaly of 2002 on the Greenland ice sheet from active and passive microwave satellite observations. Geophysical Research Letters, 31(20), L20402, doi:10.1029/2004GL020444. Stiles, W. H., and Ulaby, F. T., 1980. The active and passive microwave response to snow parameters. 1. Wetness. Journal of Geophysical Research, 85(C2), 1037–1044.

Sturm, M., Holmgren, J., and Liston, G., 1995. A seasonal snow cover classification system for local to global applications. Journal of Climate, 8(5), 1261–1283. Tiuri, M. E., Sihvola, A. H., Nyfors, E. G., and Hallikainen, M. T., 1984. The complex dielectric constant of snow at microwave frequencies. IEEE Journal of Oceanic Engineering, OE-9(5), 377–382. Tsang, L., Kong, J. A., and Shin, R. T., 1985. Theory of Microwave Remote Sensing. New York: Wiley. Ulaby, F. T., Moore, R. K., and Fung, A. K., 1981. Microwave Remote Sensing: Active and Passive. Massachusetts: Artech House. Walsh, J. E., 1991. Operational satellites and the global monitoring of snow and ice. Global and Planetary Change, 90(1–3), 219–224. Wiesnet, D. R., and Matson, M., 1979. The satellite-derived northern hemisphere snowcover record for the winter of 1977–78. Monthly Weather Review, 107, 928–933. Yang, D., Kane, D., Zhang, Z., Legates, D., and Goodison, B., 2005. Bias-corrections of long-term (1973–2004) daily precipitation data over the northern regions. Geophysical Research Letters, 32, 119501, doi:1029/2005GL02405.

Cross-references Cryosphere, Climate Change Effects Cryosphere, Climate Change Feedbacks Cryosphere, Measurements and Applications Data Archival and Distribution Microwave Radiometers Radars Snowfall

THERMAL RADIATION SENSORS (EMITTED) Simon Hook Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA

Synonyms Thermal infrared sensors Definitions Radiant energy. The energy of electromagnetic waves or sometimes of other forms of radiation. The quantity of radiant energy may be calculated by integrating radiant flux (or power) with respect to time, and, like all forms of energy, its SI unit is the joule. The term is used particularly when radiation is emitted by a source into the surrounding environment (Source: Wikipedia). Thermal radiation. Electromagnetic radiation emitted from the surface of an object which is due to the object’s temperature (Source: Wikipedia). Infrared radiation. Electromagnetic radiation whose wavelength is longer than that of visible light but shorter than that of terahertz radiation and microwaves (Source: Wikipedia).

THERMAL RADIATION SENSORS (EMITTED)

Sensor. Device that measures a physical quantity and converts it into a signal which can be read by an observer or by an instrument (Source: Wikipedia).

Introduction In this entry, we consider sensors that remotely measure the emitted thermal radiation from the surface of the Earth and the associated atmosphere at wavelengths between 3 and 14 mm. These measurements are typically made from three vantage points: (1) in situ measurements, where the sensor is placed within a few meters of the surface; (2) airborne measurements, where the sensor is mounted on an aircraft flying at 100 m to many kilometers above the surface of the Earth but still within the Earth’s atmosphere; and (3) spaceborne measurements, where the instrument is on a satellite platform orbiting the Earth. The emitted radiation that is sensed depends on the temperature and composition of the surface and the atmosphere and thereby provides a means for measuring changes in composition or temperature of the surface and/or atmosphere. These same sensor types are used to measure emitted thermal radiation from other planetary bodies over extended wavelength regions and can provide additional information on the composition of the surface depending on the composition of the overlying atmosphere. On Earth, within the 3–14 mm wavelength region, there are two atmospheric “windows.” These are regions where the energy from the surface is not as strongly blocked by the atmosphere; these are referred to as the mid infrared (MIR) and thermal infrared (TIR). In this entry, we consider the MIR window to extend between 3 and 5 mm and the TIR window between 8 and 12 mm. The terms MIR and TIR are used loosely and therefore it is best to define the wavelength region that is being considered when the term is first used. Since the signal from the surface of the Earth is strongest in the MIR and TIR, sensors that measure the emitted radiation in these wavelength regions are typically used to study the surface of the Earth. In contrast, studies of the Earth’s atmosphere will typically make measurements over the entire wavelength region since the effect of gases within the Earth’s atmosphere is pervasive throughout the region albeit stronger at selected wavelengths. Remote measurement of emitted radiation Remote sensing within the 3–14 mm wavelength region is based on the measurement of the radiance emitted from the surface and modified by the atmosphere; its basic goal is to recover the kinetic temperature and emissivity from the measured at-sensor radiance of the surface and/or the atmosphere. The at-sensor radiance (Ls) for a given wavelength (l) in this wavelength region, excluding any reflected solar contribution which occurs below 5 mm, can be written as
 (1) Lsl ¼ el Lbbl ðT Þ þ ð1  el ÞLskyl tl þ Latml

831

where el is the surface emissivity at wavelength l, Lbbl ðT Þ is the spectral radiance from a blackbody at surface temperature T, Lskyl is the sky radiance (spectral downwelling radiance incident upon the surface from the atmosphere), tl is the transmittance (spectral atmospheric transmission), and Latml is the path radiance (spectral upwelling radiance from atmospheric emission and scattering that reaches the sensor). The effect of tl is to reduce the amount of groundemitted radiance measured at the sensor; Latml adds a component unrelated to the ground and Lskyl serves to reduce the spectral contrast since the deeper an emissivity feature, the more it is “filled in” by reflected downwelling radiation due to Kirchhoff’s Law. Any algorithm that is used to produce a fundamental geophysical parameter such as surface radiance, temperature, or emissivity must compensate for these atmospheric effects. Once the atmospheric effects have been removed, the surface radiance is given by Ll ¼ el LBBl ðT Þ

(2)

Or LBBl ¼

C h C21 i l p eðlT Þ  1 5

(3)

where Ll ¼ emitted radiance L BBl ¼ blackbody radiance (Wm2 (m  Dl)1 sr1) l ¼ wavelength of channel (m) T ¼ temperature of blackbody (Kelvin) C 1 ¼ first radiation constant ¼ 3.7415  1016 (Wm2) C 2 ¼ second radiation constant ¼ 0.0143879 (m deg) el ¼ surface emissivity (0–1) The surface temperature (T) is not an intrinsic property of the surface; it varies with external factors such as the irradiance history and meteorological conditions. Conversely, emissivity is a characteristic of the material making up the surface and is independent of the temperature. For most terrestrial surfaces, 200 K < T < 340 K, although a more restricted range of 270–310 K probably brackets most of the temperatures outside the polar regions and deserts. The emissivities for most natural surfaces in MIR and TIR wavelength typically range vary 0.7–1 (Prabhakara and Dalu, 1976) with surfaces with emissivities 80 % 30 % N/A

40–80 K 40 K Room temp

High Very low Low

Fast Fast Slow

12 mm 12 mm Flat response

Land surface temperature (LST) and emissivity are key variables for explaining the biophysical processes which govern the balances of water and energy at the land surface. This table summarizes the science requirements for the LSTE-ESDR

wavelength increments over very small distances on the surface of the Earth. Unfortunately no technology currently exists that meets these requirements, and as a result, most sensors provide data at very small wavelength increments (high spectral resolution) from large areas of the Earth’s surface (low spatial resolution) or data at low spectral resolution over small distances on the Earth’s surface (high spatial resolution). These tradeoffs are necessary so the detector(s) in the sensor can receive sufficient radiation to enable measurement. There are three main detector types in operation today. These are Mercury Cadmium Telluride (MCT) detectors, microbolometers, and Quantum Well Infrared Photodetectors (QWIPs). MCT detectors and microbolometers are more widely used than the newer QWIP technology. Some of the characteristics of these detectors are summarized in Table 1. More detail on the types of detectors is available in Miles and Chow (1996). Microbolometer detectors do not require cooling which can provide a large power saving (cooling takes power), but they respond slowly. As a result, they are used with instruments that can either stare and integrate the radiation over long periods of time or, if used in an application that requires rapid imaging, provide measurements with low spatial and spectral resolution. MCT detectors require cooling but that cooling can be either passive (radiative coolers) or active (mechanical coolers). Radiative coolers are more desirable since they require less power but typically can only reach temperatures around 80 K. At these temperatures, either high spectral resolution and low spatial resolution measurements or low spectral resolution and high spatial resolution measurements are possible but not both, that is, high spatial and high spectral resolution. In order to reach the temperatures required for both high spatial and high spectral resolution, the detector needs to be cooled to temperatures around 40 K, typically requiring active cooling. Within the last decade, spaceborne mechanical coolers have become more available and there are two common types: either sterling cycle or pulse tube and modern sensors that push the current measurement capability use mechanically cooled MCT or QWIP detectors. All three types of detectors are available as one- or two-dimensional arrays. Arrays allow multiple points on the surface to be imaged at once at several wavelengths which provide more signal to each detector (longer integration time) when the time available to make the measurement is short and

consequently detector arrays are widely used today. QWIP detector arrays are more uniform than MCT arrays, which are much more desirable to enable consistent quantitative detection. These different trade-offs have the result that future high spectral and spatial resolution airborne and spaceborne sensors, where the time available to make the measurement is short, will likely use mechanically cooled detector arrays, either MCT or QWIPs. The so-called megapixel arrays (1,000  1,000 elements) are now available for both types of detector. Other detector types continue to be developed which may replace MCT and QWIPs, but the adoption of new detector types such as strained layer super lattice detectors tends to be very gradual. Table 2 provides a summary of a selection of the currently available spaceborne sensors and plans for future spaceborne sensors. These sensors represent the current and planned spaceborne state of the art however; they do not include any high spatial and high spectral resolution sensors. Such sensors do exist but currently only operate from airborne platforms. Airborne platforms are ideal for developing new spaceborne sensors since they provide a large platform with ample power and ready access to the instrument in flight. Many airborne instruments provide example datasets for future spaceborne instruments as well as more detailed spectral and spatial information as spaceborne instruments, allowing the performance of the spaceborne instruments to be validated.

Calibration and validation Most airborne and spaceborne emitted radiation sensors include some form of onboard calibration. Typically this involves one or more blackbodies. A blackbody has an emissivity of 1.0, and therefore, the radiation emitted from it depends entirely on its temperature. This known temperature is used as a reference to bracket the scene temperatures together with a space look which also provides a blackbody measurement. A typical example of a spaceborne instrument is the Moderate Resolution Imaging Spectrometer (MODIS) which acquires data at several wavelength intervals (channels) between 3 and 14 mm (Barnes et al., 1998; Salomonson et al., 1989). MODIS scans 55 from nadir and provides daytime and nighttime imaging of any point on the Earth every 1–2 days with a continuous duty cycle. MODIS data are quantized in

THERMAL RADIATION SENSORS (EMITTED)

833

Thermal Radiation Sensors (Emitted), Table 2 Examples of current and planned spaceborne emitted thermal radiation sensors with their spatial and spectral resolution ranges Subproduct

Spatial resolution

Temporal resolution

Current data sources

Future data source

Global

10–20 km

Hourly

Regional

1–5 km

2–4 times daily

AIRS GOES MSG MODIS AVHRR ATSR

Local

30–100 m

Once every 8–16 days

CrIS GOES MSG VIIRS AVHRR ATSR

ASTER Landsat

Atmospheric Infrared Radiation Sounder (AIRS), Geostationary Operational Environmental Satellites (GOES), Meteosat Second Generation (MSG), Moderate Resolution Imaging Spectroradiometer (MODIS), Advanced Very High Resolution Radiometer (AVHRR), Along Track Scanning Radiometer (ATSR), Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), Cross-track Infrared Sounder (CrIS), Visible Infrared Imager/Radiometer Suite (VIIRS)

12 bits and have a spatial resolution of 1 km at nadir. They are calibrated with a cold space view and full aperture blackbody viewed before and after each Earth view (Salomonson et al., 1989; Barnes et al., 1998). There are many different types of blackbody; MODIS, for example, uses a flat blackbody with a V-groove. Flat blackbodies with a V-groove are often used instead of cone blackbodies since they occupy less physical space. However, from a scientific standpoint, the most accurate blackbodies are cone blackbodies. These are used, for example, on the European Along Track Scanning Radiometer series of instruments. Great care is taken to calibrate emitted thermal radiation sensors used for research such as the two sensors mentioned. This allows measurements from the sensors made over long periods of time (decades) to be tracked to observe small changes in incoming radiation and enable a variety of research studies such as climate change. While onboard calibration is essential, an independent measurement is needed to ensure that the onboard system is working correctly. This independent measurement is used to validate the onboard calibration system. Typically the validation measurement involves measuring the radiation from a sensor close to the ground or in an aircraft and then propagating that measurement through the Earth’s atmosphere with a radiative transfer model (Berk et al., 1989) to obtain the radiance that the satellite instrument should observe and comparing it with what the satellite instrument measures. In the last decade a validation site was established at Lake Tahoe CA/NV, USA, where these measurements are made on a continuous basis and data from the site have been used to validate several sensors including ASTER, MODIS, Landsat (TM and ETM), and ATSR (Barsi et al., 2003; Hook et al., 2003, 2005, 2006; Tonooka et al., 2005).

Conclusion Over the last few decades, multiple emitted radiation sensors have been launched into orbit around the Earth as well as other planetary bodies. Data from these sensors have

been used in a wide variety of studies from mapping the surface composition of a region to studying climate change and forecasting the weather. The instruments that have been launched were typically based on prototypes deployed from aircraft. Our ability to make even higher spectral and spatial measurements from space is limited by the available technology. As the technology evolves, so does the measurement capability and science that the sensors can be used to address. One area of concern is the lack of future sensors capable of making high spatial resolution thermal infrared measurements such as those currently provided by ASTER and Landsat (Yamaguchi et al., 1998; Barsi et al., 2003). Currently the research community is calling for such measurements, but it remains to be seen whether such measurements will be available in the future.

Acknowledgments The research described in this entry was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration. Bibliography Barnes, W. L., Pagano, T. S., and Salomonson, V. V., 1998. Prelaunch characteristics of the moderate resolution imaging spectroradiometer (MODIS) on EOS AM1. IEEE Transactions on Geoscience and Remote Sensing, 36, 1088–1100. Barsi, J. A., Schott, J. R., Palluconi, F. D., Helder, D. L., Hook, S. J., Markhum, B. L., Chander, G., and O’Donnell, E. M., 2003. Landsat TM and ETM + thermal band calibration. Canadian Journal of Remote Sensing, 29(2), 141–153. Berk, A., Bernstein, L. S., and Robertson, D. C., 1989. MODTRAN: A Moderate Resolution Model for LOWTRAN 7. Technical Report GL-TR-89-0122. Bedford, MA: Geophysics Laboratory. Hook, S. J., Prata, A. J., Alley, R. E., Abtahi, A., Richards, R. C., Schladow, S. G., and Pálmarsson, S. Ó., 2003. Retrieval of lake bulk-and skin-temperatures using along track scanning radiometer (ATSR) data: a case study using Lake Tahoe, CA. Journal of Atmospheric and Oceanic Technology, 20(2), 534–548.

834

TRACE GASES, STRATOSPHERE, AND MESOSPHERE

Hook, S. J., Clodius, W. B., Balick, L., Alley, R. E., Abtahi, A., Richards, R. C., and Schladow, S. G., 2005. In-flight validation of mid and thermal infrared data from the multispectral thermal imager (MTI) using an automated high altitude validation site at Lake Tahoe CA/NV USA. IEEE Transactions on Geoscience and Remote Sensing, 43, 1991–1999. Hook, S. J., Vaughan, R. G., Tonooka, H., and Schladow, S. G., 2006. Absolute radiometric in-flight validation of mid and thermal infrared data from ASTER and MODIS using the lake Tahoe CA/NV, USA automated validation site. IEEE Transactions on Geoscience and Remote Sensing, 45(6), 1798–1807. Miles, R. H., and Chow, D. H., 1996. Chap. 7. In Razeghi, M. (ed.), Long Wavelength Infrared Detectors. Singapore: Gordon and Breach. Prabhakara, C., and Dalu, G., 1976. Remote sensing of surface emissivity at 9 mm over the globe. Journal of Geophysical Research, 81(21), 3719–3724. Salomonson, V., Barnes, W., Maymon, P., Montgomery, H., and Ostrow, H., 1989. MODIS: advanced facility instrument for studies of the Earth as a system. IEEE Transactions on Geoscience and Remote Sensing, 27, 145–153. Tonooka, H., Palluconi, F., Hook, S., and Matsunaga, T., 2005. Vicarious calibration of ASTER thermal infrared bands. IEEE Transactions on Geoscience and Remote Sensing, 43, 2733–2746. Yamaguchi, Y., Kahle, A. B., Tsu, H., Kawakami, T., and Pniel, M., 1998. Overview of advanced spaceborne thermal emission reflectance radiometer. IEEE Transactions on Geoscience and Remote Sensing, 36, 1062–1071.

Cross-references Calibration and Validation Calibration, Optical/Infrared Passive Sensors Climate Data Records Crop Stress Data Processing, SAR Sensors Emerging Technologies Irrigation Management Land Surface Emissivity Land Surface Temperature Observational Platforms, Aircraft, and UAVs Observational Systems, Satellite Ocean-Atmosphere Water Flux and Evaporation Optical/Infrared, Radiative Transfer Remote Sensing, Physics and Techniques Resource Exploration Volcanism Water and Energy Cycles Water Resources

TRACE GASES, STRATOSPHERE, AND MESOSPHERE Nathaniel Livesey Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA

Synonyms Atmospheric chemistry; Atmospheric composition; Middle atmosphere (common term used to refer collectively to the stratosphere and mesosphere)

Definitions Trace gases. Minor constituents in Earth’s atmosphere typically having abundances from as large as hundreds of parts per million (e.g., CO2) to parts per trillion (e.g., BrO and OH) or less. Stratosphere. The region of Earth’s atmosphere between 15 and 50 km characterized by temperatures that generally increase with increasing altitude. Mesosphere. The region of Earth’s atmosphere between 50 km and 80 km characterized by temperatures that generally decrease with increasing altitude. Introduction Earth’s atmosphere is divided vertically into regions defined by the vertical gradient of temperature. The lowermost region, the troposphere, ranging from the surface to the “tropopause” at 10–15 km altitude, is characterized by temperatures that mainly decrease with increasing altitude. Above is the stratosphere, where temperatures generally increase with height, up to 50 km where the “stratopause” separates the stratosphere from the mesosphere, within which temperatures generally decrease with altitude again (up to the “mesopause” around 80 km). Collectively, the stratosphere and mesosphere are often referred to as the “middle atmosphere.” Trace gases play disproportionately important roles in these regions, both chemically and radiatively, despite their small abundances ranging from hundreds of parts per million to as low as parts per trillion or less. For example, the absorption of solar ultraviolet radiation by the stratospheric ozone layer, in addition to shielding terrestrial organisms from harmful exposure, also heats the stratosphere, giving rise to its increase in temperature with increasing altitude. This positive vertical temperature gradient results in vertical atmospheric stability, leading to the stratification for which the region is named. Thermal emission by carbon dioxide (in addition to methane and other minor trace gases) is the main vehicle for radiative cooling in the stratosphere and mesosphere. Despite being the most abundant species in the atmosphere, molecular oxygen and nitrogen play relatively minor roles in the radiative balance and chemistry of these regions. In addition to trace gases, aerosols also play a significant role in the chemistry of the middle atmosphere. Most notably, key chemical reactions giving rise to the Antarctic ozone hole take place on the surface of polar stratospheric cloud (PSC) particles (see the companion article on stratospheric ozone). In addition, sulfate aerosol is ubiquitous in the lower stratosphere at small background levels, with occasional significant enhancements resulting from injection of SO2 directly into the stratosphere by large volcanic eruptions (e.g., Mt. Pinatubo in 1991). In situ observations of the middle atmosphere, while having provided valuable information, are by their nature limited to those available from high-altitude aircraft and balloons, with sounding rockets the only source of in situ

TRACE GASES, STRATOSPHERE, AND MESOSPHERE

data in the upper stratosphere and mesosphere. Remote sounding has accordingly formed a cornerstone in scientific investigation of the composition of the stratosphere and mesosphere.

Principals and techniques Information on stratospheric and mesospheric composition can be obtained from observations over a large range of electromagnetic wavelengths, from the microwave (mainly associated with molecular rotational transitions) to the infrared (mainly vibrational transitions) to the visible and ultraviolet (mainly electronic transitions). Signals observed may arise from thermal emission by the atmosphere itself, from atmospheric absorption and/or scatting of solar, lunar, or stellar radiation or of radiation actively emitted, for example, by laser. Observations are made from the ground and from air- or spaceborne platforms. Variations in viewing angle and/or wavelength observed can give information on the vertical distribution of trace gases. Many space-based measurements have been pioneered by balloon or aircraft precursor instruments, and such instruments are frequently used (in conjunction with in situ instrument) for ongoing scientific investigations and to “validate” satellite observations. Molecular spectral lines are broadened in the atmosphere by a combination of the ensemble of Doppler shifts from the thermal motion of the molecules (“Doppler” broadening) and by collisions with other molecules (“collision” or “pressure” broadening). The latter generally dominates line widths of infrared and microwave signals in the stratosphere, while for visible and ultraviolet wavelengths, Doppler broadening dominates throughout the middle atmosphere. Accurate knowledge of molecular spectroscopy, including line positions, strengths, and broadening parameters, underpins the accuracy of middle atmosphere composition remote sounding, and fundamental laboratory spectroscopy measurements are an integral part of atmospheric composition research programs. Pressure broadening of spectral lines can provide valuable information on the vertical distribution of trace gases, with frequencies further from line centers conveying information on lower regions of the atmosphere where lines are broad enough to contribute to the observed signals. For wavelengths where pressure broadening is insignificant, vertical distribution information can still be obtained in some viewing geometries by observing in multiple spectral regions having different atmospheric absorptions (and thus penetration depths). In the stratosphere and lower mesosphere, the atmosphere is generally in local thermodynamic equilibrium (LTE), that is to say, molecular collisions are frequent enough that the kinetic temperature of the air is in equilibrium with the rotational and vibrational temperature of the emitting and absorbing molecules. At lower pressures (typically smaller than 0.1 hPa, 60 km), where the mean time between collisions increases, some molecular transitions (particularly infrared vibrational transitions) can be

835

in non-LTE, that is to say, their “vibrational temperature” differs from the underlying atmospheric kinetic temperature. This complicates the interpretation of observed signals. Most observations give a measure of number density for a particular molecule, either total column or vertical profile. For profile measurements, knowledge of atmospheric density (i.e., temperature and pressure) is needed to convert these to mixing ratio. Typically, this is inferred by also measuring the number density of a well-mixed gas whose abundance is known well (e.g., O2 or CO2).

Direct solar absorption and related approaches Observation of the absorption of solar radiation by the atmosphere provides some of the most precise remotely sensed measurements of stratospheric and mesospheric composition. The strong solar illumination gives a good signal-to-noise ratio for quantification of atmospheric absorption features. Such observations are possible in all spectral regions having significant solar emission, from the infrared to the ultraviolet. Dobson’s ground-based observations of absorption of solar UV by ozone are one of the earliest examples of remote sounding of atmospheric composition. For ground-based solar absorption observations, vertical distribution information can be obtained from measurements at different sun angles (assuming temporal variations are small). Low-Earth-orbiting satellites can make “occultation” observations of atmospheric absorption of solar radiation during each satellite sunrise and sunset (giving 30 measurements per day). Observing variations in absorption as the sun rises/sets through the atmosphere gives high vertical resolution (1 km) information on trace gas distribution. However, horizontal coverage of such measurements is limited. The latitudes of the sunrise and sunset observations drift slowly from day to day, as dictated by the spacecraft orbit, with (in many but not all cases) all latitudes covered in about a month. Increased coverage can be obtained by making lunar and/or stellar occultation observations. However, such measurements have a poorer signal to noise than solar occultation. There is a rich history of stratospheric and mesospheric composition measurements from solar occultation instruments, with a continuous record available from the 1984 launch of the Stratospheric Aerosol and Gas Experiment II (SAGE II) instrument (McCormick et al., 1989), followed by SAGE III, the Polar Ozone and Aerosol Measurement (POAM) series of instruments (e.g., Lucke et al., 1999), the Halogen Occultation Experiment (HALOE) (Russell et al., 1993) on the Upper Atmosphere Research Satellite (UARS), and, most recently, the Atmospheric Chemistry Experiment Fourier Transform Spectrometer (ACE/FTS) (Bernath et al., 2005), the only such instrument still currently operating. Additional occultation observations have been made by the Atmospheric Trace Molecule Spectroscopy (ATMOS) instrument flown on the Space Shuttle ATLAS program

836

TRACE GASES, STRATOSPHERE, AND MESOSPHERE

(Gunson et al., 1996) and the Improved Limb Atmospheric Spectrometer instruments (ILAS I and II) on the Japanese Advanced Earth-Orbiting Satellites (ADEOS I and II) (e.g., Sasano et al., 2001).

Viewing geometries: nadir, limb, and zenith In addition to occultation, two other geometries have been extensively employed to observe middle atmosphere composition from low-Earth orbit. “Nadir sounding” instruments view the atmosphere directly below the orbiting spacecraft (or in some swath either side of vertical). “Limb sounding” instruments observe the atmosphere “edge on,” tangent to the Earth, scanning their field of view vertically across the atmosphere, offering better vertical resolution (typically 2–5 km) than nadir observations. In addition, the longer path length associated with limb sounding results in stronger signals for trace gases than for nadir sounding. However, this same long path length results in poorer horizontal resolution for limb observations that is, in principle, possible for nadir sounding. For groundbased and airborne observations, zenith (upward looking) observations of scattered sunlight or thermal emission are commonly employed. UV/visible scattering measurements Satellite observations of backscattered UV/visible sunlight constitute the longest continuous record of stratospheric composition observations, most notably ozone, from space. The 1978 launch of the Nimbus-7 Total Ozone Mapping Spectrometer (TOMS) and Solar Backscatter Ultraviolet (SBUV) instruments (Heath et al., 1975) heralded the start of nearly continuous observations using this technique. In addition to measuring total ozone column (e.g., TOMS) and stratospheric ozone profile (e.g., SBUV), column measurements of SO2 were made from TOMS, along with a measure of aerosol amount (mainly indicating tropospheric aerosols). More recent instruments include the Global Ozone Monitoring Experiment (GOME) on ERS-2 (Burrows et al., 1999), the SCanning Imaging Absorption SpectroMeter for Atmospheric CHartographY (SCIAMACHY) instrument on Envisat (Bovensman et al., 1999), the Ozone Monitoring Instrument (OMI) (Levelt et al., 2006) on Aura, and GOME-2 on METOP (Munro et al., 2006). These continue the TOMS record and also measure column abundances of additional species (e.g., NO2, HCHO). However, these additional gases are mainly of interest in the troposphere (see the related entry on tropospheric composition observations in this volume). By their nature, solar backscatter observations cannot be made at night (months at high latitude during polar night). Most such observations are made from sun-synchronous low-Earthorbiting spacecraft (where observations are made at fixed local time). While most UV/visible stratospheric observations have been made in nadir (or swath), a limited number of instruments have observed the scattering of solar UV/visible

(and near-infrared) radiation in a limb geometry. However, the accuracy of such observations is highly dependent on knowledge of spacecraft attitude. UV/visible limb sounding instrument includes SCIAMACHY, OSIRIS on Odin (Llewellyn et al., 2003; Murtagh et al., 2002), and the profiling component of the “Ozone Mapping and Profiling Suite” (OMPS) on the NPOESS (National Polar-orbiting Operational Environmental Satellite System) Preparatory Project (NPP).

Thermal emission observations In contrast to UV/visible observations, measurements of atmospheric composition from observations of thermal emission in the infrared and microwave can be made both day and night. While nadir sounding at these wavelengths can yield useful information on stratospheric and mesospheric composition (e.g., for ozone), the majority of space-based thermal emission observations of stratospheric and mesospheric composition have been made in a limb-viewing geometry. Pressure broadening of spectral lines in principle enables instruments viewing limb emission to independently measure both composition and atmospheric pressure for a given limb view, making the overall measurement system less sensitive to knowledge of spacecraft attitude than for UV/visible limb sounding. Limb sounding instruments in the thermal infrared have included the Limb Radiance Inversion Radiometer (LRIR) flown on Nimbus-6, the Nimbus-7 Limb Infrared Monitor of the Stratosphere (LIMS) (Gille and Russell, 1984), Stratospheric and Mesospheric Sounder (SAMS) (Drummond et al., 1980) instruments, the Improved Stratospheric and Mesospheric Sounder (ISAMS) (Taylor et al., 1993) and the Cryogenic Limb Array Etalon Spectrometer (CLAES) on UARS (Roche et al., 1993), the Michelson Infrared Passive Atmospheric Sounder (MIPAS) (Fischer et al., 2008) on Envisat, the Sounding of the Atmosphere using Broadband Emission Radiometry (SABER) instrument (Russell et al., 1999) on the Thermosphere, Ionosphere, Mesosphere Energetics and Dynamics (TIMED) spacecraft, and the High Resolution Dynamics Limb Sounder (HIRDLS) (Gille et al., 2008) on Aura. Atmospheric composition instruments observing thermal microwave emission in the limb include the Microwave Limb Sounder (MLS) instruments on the UARS and Aura spacecraft (Barath et al., 1993; Waters et al., 2006), the Microwave Atmospheric Sounder (MAS) flown on the space shuttle ATLAS program (Croskey et al., 1992), the Submillimeter Radiometer (SMR) on Odin (Murtagh et al., 2002), and the planned Superconducting Submillimeter-Wave Limb-Emission Sounder (SMILES) (Ozeki et al., 2001) to be attached to the International Space Station. Observations in the microwave region of the spectrum are unaffected by scattering or emission by stratospheric aerosols. By contrast, infrared limb (or occultation) instruments can readily observe stratospheric aerosols

TRACE GASES, STRATOSPHERE, AND MESOSPHERE

(notably PSCs or strong volcanic enhancements), but the accuracy of their trace gas observations may be reduced by strong aerosol emission and scattering.

Active techniques: LIDAR While direct measurements of atmospheric absorption between a source and a distant detector are frequently made at ground level, such approaches are impractical in the middle atmosphere (although many in situ instruments work on the same principal but over short distances, often augmented by multipath reflection optics). However, ozone and water vapor in the middle atmosphere can be measured by observing the atmospheric absorption of backscattered laser light from a Light Detection and Ranging (LIDAR) instrument, either ground based or airborne. The backscatter can be either elastic (Rayleigh scattering) or inelastic (Raman scattering). Observations are typically made at multiple wavelengths that have different levels of absorption by the molecule of interest. Observing differences in the backscattered signal received at the different wavelengths gives a measure of the total molecular absorption between the laser source and the atmospheric layer at which the backscatter occurred. In addition, by observing spectral differences, measurements are insensitive to changes in laser output power. High-speed detection of the backscattered signals enables high vertical resolution (typically hundreds of meters) for LIDAR compared to other techniques. Notable findings from remote sounding observations Remote sounding observations of trace gases have provided profound insights in to the behavior of the middle atmosphere, in particular, the complex interplay of chemical, dynamical, and radiative processes that govern the budget of stratospheric ozone. For example, the Antarctic “ozone hole” was originally observed in long-term records of ultraviolet absorption over Halley Bay in Antarctica (Farman et al., 1985). Also, ground-based microwave emission sounding at McMurdo Sound confirmed the role of the chlorine monoxide (ClO) radical in the ozone hole (de Zafra et al., 1987), while UARS MLS measurements showed that similar levels of ClO were present in the Arctic stratosphere (Waters et al., 1993). Remote sounding of the atmospheric distribution of gases with relatively long lifetimes gives valuable insights into transport of air into, out of, and within the stratosphere and mesosphere. Such transport is fundamental to the spatial distribution of stratospheric ozone and also plays an important and complex role in polar chemical ozone loss (see Sect. 6.2 of Solomon (1999) and references therein). Observations of long-lived trace gas distributions are an important means to quantify the slow overturning circulation of air within the middle atmosphere (the so-called Brewer-Dobson Circulation), which may be impacted by changes in climate. Remote sounding observations have also been essential to our understanding

837

of how air is transported from the troposphere into the stratosphere (e.g., Mote et al., 1995). While space-based observations have provided valuable global insights into middle atmosphere composition and processes that affect it, the limited lifetime of spacebased instruments makes the extraction of long-term trends problematic. The majority of the information on long-term trends and variations in stratospheric and mesospheric composition has come from ground-based observations. Notable among these are those in the “Network for the Detection of Atmospheric Composition Change” (NDACC, formerly the “Network for the Detection of Stratospheric Change,” NDSC). This network brings together a variety of highly accurate measurement techniques, including LIDAR, and passive observations in the ultraviolet, infrared, and microwave (along with in situ sonde observations) at locations across the globe.

Outlook The number of space-based instruments observing middle atmosphere composition has decreased significantly in recent years. At the time of writing, this overall decreasing trend is set to continue. All of the spacecraft currently making stratospheric and mesospheric observations are now or will shortly be operating beyond their mission design lifetimes. The only confirmed mission to continue observations is the Japanese Superconducting Submillimeter-Wave Limb-Emission Sounder (SMILES) instrument, to be attached to the International Space Station (ISS) in 2009. Plans for future middle atmosphere composition observations are unclear, with a long gap anticipated between NASA’s Aura satellite and the expected Global Atmospheric Composition Mission (GACM) and a similar gap likely between ESA’s Envisat and expected Sentinel-5 missions. Although concept missions to fill this gap have been proposed or are in formulation, the only confirmed observations are stratospheric ozone profile observations by OMPS on NPP. Conclusion Our understanding of the chemical, dynamical, and radiative processes that govern the behavior of the stratosphere and mesosphere has advanced dramatically over the last two decades, thanks in large part to the wealth of remote sounding composition observations. Notable findings from remote sounding include the discovery of the ozone hole, insights into its origins (in conjunction with in situ observations), and understanding of the large-scale transport processes that operate in the middle atmosphere. Acknowledgment This research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the NASA.

838

TRACE GASES, TROPOSPHERE - DETECTION FROM SPACE

Bibliography Barath, F. T., et al., 1993. The upper atmosphere research satellite microwave limb sounder experiment. Journal of Geophysical Research, 98, 10751–10762. Bernath, P. D., et al., 2005. Atmospheric Chemistry Experiment (ACE): mission overview. Geophysical Research Letters, 32, L15S01, doi:10.1029/2005GL022386. Bovensman, H., Burrows, J. P., Buckwitz, B., Frerick, J., Noël, S., Rozanov, V. V., Chance, K. V., and Goede, A. P. H., 1999. SCIAMACHY: mission objectives and measurement modes. Journal of the Atmospheric Sciences, 56, 127–150. Burrows, J. R., et al., 1999. The Global Ozone Monitoring Experiment (GOME): mission concept and first scientific results. Journal of the Atmospheric Sciences, 56, 151–175. Croskey, C. L., et al., 1992. The Millimeter Wave Atmospheric Sounder (MAS): a shuttle-based remote sensing experiment. IEEE Transactions on Microwave Theory and Techniques, 40, 1090–1100. de Zafra, R. J., Jaramillo, M., Parrish, A., Solomon, P., Connor, B., and Barrett, J., 1987. High concentrations of chlorine monoxide at low latitudes in the Antarctic spring stratosphere: diurnal variation. Nature, 328, 408. Drummond, J. R., Houghton, J. T., Peskett, G. D., Rodgers, C. D., Wale, M. J., Whitney, J. G., and Williamson, E. J., 1980. The stratospheric and mesospheric sounder on nimbus 7. Philosophical Transactions of the Royal Society of London, 296, 219–241. Farman, J. C., Gardiner, B. G., and Shanklin, J. D., 1985. Large losses of total ozone in Antarctica reveal seasonal ClOx/NOx interaction. Nature, 315, 207–210. Fischer, H., et al., 2008. MIPAS: an instrument for atmospheric and climate research. Atmospheric Chemistry and Physics, 8, 2151–2188. Gille, J. C., and Russell, J. M., III, 1984. The limb infrared monitor of the stratosphere: experiment description, performance and results. Journal of Geophysical Research, 89, 5125–5140. Gille, J., et al., 2008. High Resolution Dynamics Limb Sounder (HIRDLS): experiment overview, recovery and validation of initial temperature data. Journal of Geophysical Research, 113, D16S43, doi:10.1029/2007JD008824. Gunson, M. R., et al., 1996. The Atmospheric Trace Molecule Spectroscopy (ATMOS) experiment: deployment on the ATLAS space shuttle missions. Geophysical Research Letters, 23, 2333–2336. Heath, D. F., Krueger, A. J., Roeder, H. A., and Henderson, B. D., 1975. The solar backscatter ultraviolet and total ozone mapping spectrometer (SBUV/TOMS) for nimbus G. Optical Engineering, 14, 323. Levelt, P. F., van den Oord, G. H. J., Dobber, M. R., Mälkki, A., Visser, H., de Vries, J., Stammes, P., Lundell, J. O. V., and Saari, H., 2006. The ozone monitoring instrument. IEEE Transactions on Geoscience and Remote Sensing, 44, 1093–1101. Llewellyn, E. J., et al., 2003. First results from the OSIRIS instrument on-board Odin. Sodankyla Geophysical Observatory Publications, 92, 1–47. Lucke, R. L., et al., 1999. The Polar Ozone and Aerosol Measurement (POAM) III instrument and early validation report. Journal of Geophysical Research, 104, 18785–18799. McCormick, M. P., Zawodny, J. M., Velga, R. E., Larsen, J. C., and Wang, P. H., 1989. An overview of SAGE I and II ozone measurements. Planetary and Space Science, 37, 1567–1586. Mote, P. W., Rosenlof, K. H., Holton, J. R., Harwood, R. S., and Waters, J. W., 1995. Seasonal variation of water vapor in the tropical lower stratosphere. Geophysical Research Letters, 22, 1093–1096. Munro, R., Eisinger, M., Anderson, C., Callies, J., Corpaccioli, E., Lang, R., Lefebvre, A., Livschitz, Y., and Albiñana, A. P.,

2006. GOME-2 on MetOp. In Proceedings of the 2006 EUMETSAT Meteorological Satellite Conference, Helsinki, June 12–16, 2006, EUMETSAT p. 48. Murtagh, D., et al., 2002. An overview of the Odin atmospheric mission. Canadian Journal of Physics, 80, 309–319. Ozeki, H., et al., 2001. Development of superconducting submillimeter-wave limb emission sounder (JEM/SMILES) aboard the International Space Station. In Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, vol. 4540, pp. 209–220, SPIE. Roche, A. E., Kumer, J. B., Mergenthaler, J. L., Ely, H. A., Uplinger, W. H., Potter, J. F., James, T. C., and Sterritt, L. W., 1993. The Cryogenic Limb Array Etalon Spectrometer (CLAES) on UARS: experiment description and performance. Journal of Geophysical Research, 98, 10763–10775. Russell, J. M., III, et al., 1993. The halogen occultation experiment. Journal of Geophysical Research, 93, 10777–10798. Russell III, J. M., Mlynczak, M. G., Gordley, L. L., Tansock, J., and Esplin, R.: An overview of the SABER experiment and preliminary calibration results, Proc. SPIE, 3756, doi:10.1117/ 12.366382, 277–288, 1999. Sasano, Y., Yokota, T., Nakajima, H., Sugita, T., and Kanzawa, H., 2001. ILAS-II instrument and data processing system for stratospheric ozone layer monitoring. Proceedings of SPIE, 4150, 106–114. Solomon, S., 1999. Stratospheric ozone depletion: a review of concepts and history. Reviews of Geophysics, 37, 275–316. Taylor, F. W., et al., 1993. Remote sensing of atmospheric structure and composition by pressure modulation radiometry from space: the ISAMS experiment on UARS. Journal of Geophysical Research, 98, 10799–10814. Waters, J. W., Froidevaux, L., Read, W. G., Manney, G. L., Elson, L. S., Flower, D. F., Jarnot, R. F., and Harwood, R. S., 1993. Stratospheric ClO and ozone from the microwave limb sounder on the upper atmosphere research satellite. Nature, 362, 597–602. Waters, J. W., et al., 2006. The Earth Observing System Microwave Limb Sounder (EOS MLS) on the Aura satellite. IEEE Transactions on Geoscience and Remote Sensing, 44, 1075–1092.

Cross-references Stratospheric Ozone

TRACE GASES, TROPOSPHERE - DETECTION FROM SPACE Pieternel F. Levelt1,2, J. P. Veefkind1,3 and K. F. Boersma1,4 1 Koninklijk Nederlands Meteorologisch Instituut (KNMI), De Bilt, The Netherlands 2 Delft University of Technology, Eindhoven, The Netherlands 3 Eindhoven University of Technology, Eindhoven, The Netherlands 4 Technical University Eindhoven (TUE), Eindhoven, The Netherlands

Introduction The increase of human activity over the last 100 years has resulted in an enormous increase in atmospheric

TRACE GASES, TROPOSPHERE - DETECTION FROM SPACE

concentrations of various gases of anthropogenic origin, polluting the atmosphere and changing climate on Earth. The detection of the stratospheric ozone hole raised awareness that human activities can lead to dramatic changes in atmospheric composition, with large consequences for life on Earth. The satellite maps of air pollution and greenhouse gases in the troposphere (this is the lower 10–16 km of the atmosphere) over the last 5–10 years made people realize that our lifestyle does not only influence the ozone layer but also our daily living environment and the air we breathe.

Trace gases in the troposphere The atmosphere consists of nitrogen (78 %) and oxygen (21 %), with only a minor contribution from other gases (1 %), which are called trace gases due to their low abundance. Although their abundance is low, trace gases play a key role in climate change and air quality. Ozone (O3) is one of the most abundant minor trace gases. Ozone resides mainly in the stratosphere – the so-called ozone layer – protecting us from the harmful ultraviolet radiation from the Sun because it efficiently absorbs this high-energy radiation. Ozone also plays an important role in the troposphere, where it is a greenhouse gas as well as an air pollutant. Other important tropospheric trace gases are carbon dioxide (CO2), water vapor (H2O), methane (CH4), nitrogen oxides (NOx), sulfur oxides (SOx), carbon monoxide (CO), NH3 and VOCs (volatile organic compounds) like formaldehyde (HCHO). CO2 and H2O are important greenhouse gases. H2O plays a major role in the hydrological cycle and is the gas that in reaction with O3 forms the highly reactive radical OH. OH is the key species in cleansing the atmosphere of pollutants. All other mentioned gases play a role as greenhouse gas or air pollutant and/or as a precursor for a greenhouse gas, aerosols, or air pollutant. For example, NO2 is an air pollutant and an important precursor for the greenhouse gas ozone and aerosols, as the air pollutant SO2 is also an important precursor for aerosols. Other nongaseous atmospheric constituents that are of paramount importance for climate change and air pollution are aerosols and clouds. The trace gases CO2, CH4, NO2, SO2, HCHO, NH3 and CO, which can be observed from space and are increasing due to human activities, will therefore be the focus of this entry. H2O is a natural gas and will therefore not be considered. OH in the troposphere is undetectable from space. History of detection of trace gases from space In the early 1970s, the first measurements of (stratospheric) ozone were made from space by the SBUV satellite instrument (Hilsenrath et al., 1995), later followed by the TOMS instrument (McPeters et al., 1996). These were the first measurements of the chemical composition of the atmosphere from space. In addition to stratospheric ozone, TOMS also observed SO2 from volcanic eruptions. GOES was used to retrieve aerosol optical depth (Fraser et al., 1984). Fishman et al. (1987) used ozone columns from

839

TOMS to investigate an episode with ozone smog over the eastern United States. An important step forward was made with the launch of the GOME instrument in 1995 (Burrows et al., 1999), which could measure several of the minor trace gases using the backscatter technique. This instrument was followed by SCIAMACHY (2002, Bovensmann et al., 1999), OMI (2004, Levelt et al., 2006), and GOME-2 (2006, Callies et al., 2000). Apart from the solar backscatter technique, measurements of tropospheric trace gases can also be made in the thermal infrared. MOPITT (1999, Deeter, 2009), AIRS (2002, Aumann et al., 2003), TES (2004, Beer, 2006), and IASI (2007, Blumstein et al., 2004) are instruments that use this wavelength range. All instruments that have been observing tropospheric trace gases were on board satellites in a Sun-synchronous orbit, providing at most one (backscatter techniques) or two (thermal infrared) measurements at the same local time per day.

Retrievals and validation What can be detected from space on tropospheric trace gases? Since the species of interest reside in the troposphere, and the instruments are orbiting the Earth at around 500–800 km altitude, observing in nadir, it is quite a challenge to detect the minor constituents in the lower 10–15 km of the atmosphere from space. There are basically two methods used to measure the troposphere, both based on passive remote sensing techniques: the solar backscatter technique and the thermal infrared technique. In the solar backscatter technique, the instrument measures the solar radiation that has been absorbed and scattered by the atmosphere. This so-called Earth radiance spectrum contains the specific absorption features of the molecules of interest (due to electronic-vibrationalrotational transitions in the molecule, see Figure 1). In the thermal infrared technique, the thermal emission of the Earth-atmosphere system is measured, revealing the specific absorption features of the trace gases (see Figure 2). Due to the specific spectroscopic “fingerprint” of the molecule, it is possible to determine the trace gas concentration, using very sensitive detectors and optimally designed optical instruments. The solar backscatter instruments usually provide tropospheric columns of the trace gases. The technique has the advantage to be sensitive to the surface, since the atmosphere is transparent in the visible wavelength range. With the thermal infrared technique, also some vertical information can be obtained, approximately two layers in the troposphere, but the sensitivity to the surface is less so that accurate total column amounts are more challenging. The retrieval of tropospheric trace gas concentrations from satellite measurements is a so-called ill-posed problem. For ill-posed problems, the information from the measurement is not enough to derive all unknowns independently. This means that to retrieve trace gas concentrations from satellite measurements, assumptions have to be

840

TRACE GASES, TROPOSPHERE - DETECTION FROM SPACE

O3 SO2 NO2 HCHO O2-O2

absorption optical thickness

10

1 O3 0.1 NO2 0.01 SO2

O2-O2 O3

HCHO

O2-O2 0.001 270

300

350

400 wavelength (nm)

450

500

Trace Gases, Troposphere - Detection from Space, Figure 1 Total absorption optical thickness for tropospheric trace gases in the UV–vis wavelength range, as calculated for a mid-latitude atmosphere in summer. Note that a logarithmic scale is used, thus the absorption by ozone decreases by six orders of magnitude in the wavelength range from 270 to 360 nm (Image courtesy: P. Stammes (KNMI)).

made on, for example, the reflectivity of the surface and the approximate vertical distribution of the trace gas. One of the most widely used retrieval techniques is optimal estimation (Rodgers, 2000). For the solar backscatter technique, also differential optical absorption spectroscopy (DOAS) retrieval is widely used (Platt, 1994). Satellite measurements of tropospheric trace gases are hampered by clouds that reflect radiation to space, effectively screening the gas concentrations in the most polluted layer below the clouds. For fully clouded scenes algorithms are therefore fundamentally limited to retrieving above-cloud concentrations. But in situations with partial cloud cover, radiance signals still contain information on trace gas concentrations in the lowest layers, and tropospheric column retrievals are possible. By reducing the size of the ground pixel, the probability of encountering a completely cloud-free pixel increases. Therefore, recent developments have focused on limiting the pixel size, thereby increasing the amount of measurement samples in order to obtain more cloud-free observations while retaining global coverage. In order to check their accuracy, satellite measurements are validated by comparisons with ground-based and aircraft measurements and atmospheric models. Groundbased and aircraft campaigns are often performed for validation for specific atmospheric conditions and are thus limited in scope. Networks of ground-based instruments of known quality are thus a valuable tool for validation of satellite measurements over the whole mission lifetime.

Extensive networks exist only for a limited set of trace gases (for instance, ozone and CO2). For other trace gases, like NO2, SO2, and HCHO, the development of consistent ground-based networks is much called for, but only starting.

The advantages of satellite observations The capability of satellite instruments to measure the tropospheric pollution first became apparent by measurements of GOME on NO2 (Leue et al., 2001) and MOPITT on CO (Edwards et al., 2004). These instruments enabled for the first time global measurements of pollutants in the troposphere in the form of maps of monthly or yearly averaged concentrations. Up to that point, only models provided such information, and independent data to validate these models was sparse. SCIAMACHY extended the measurements of the troposphere to greenhouse gases like methane (Frankenberg et al., 2005). The GOSAT satellite (Hamazaki et al., 2004) was launched in 2009 is dedicated to measure CH4 and CO2 with very high accuracy. Recent results suggest that GOSAT CO2 retrievals are useful to constrain surface fluxes of CO2 (Butz et al., 2011). In 2014 NASA’s OCO satellite (Crisp et al., 2004) will be launched dedicated to measure CO2. Because of its small pixels and daily global coverage, OMI provides much more measurements of the troposphere than previous instruments were able to. Instead of monthly averaged maps, OMI is able to provide daily maps. In Figure 3, four consecutive frames are shown of

TRACE GASES, TROPOSPHERE - DETECTION FROM SPACE

841

Trace Gases, Troposphere - Detection from Space, Figure 2 IASI-normalized radiance spectrum (top panel, recorded by IASI/ MetOp, over west of Australia, on December 20, 2006) and transmittance simulated molecular contributions to identify the main absorbing gases (middle panels) and the weaker (bottom panel) (From Clerbaux et al., 2009).

OMI measurements. The figure clearly shows the Sunday dip in NO2 concentration due to reduction of traffic during the weekend resulting in less NO2 emissions. These instruments clearly showed the unique capability of satellite measurements to obtain global coverage and consistent quality of the measurements. For instance, ground-based networks for most of the tropospheric trace gases are sparse and the quality of the data is sometimes station dependent. The chemistry of the troposphere is complex and involves many trace gases and chemical reactions. Combining the observations of trace gases from satellite instruments is therefore a great advantage. The added value of satellite observations of key species is illustrated in Figure 4. This figure shows the prominent global impact of human activities on the composition of the troposphere. Figure 4 shows satellite maps of concentrations of NO2, CO, CH4, HCHO, and particulate matter. Also shown is a map of the population density. The sources for NO2 and CO are fossil fuel combustion by power plants, industry, and traffic. For CH4, important sources are wetlands (including rice agriculture), energy, livestock, landfills and waste, and biomass burning. Formaldehyde sources include biogenic emissions, as

well as biomass burning and fossil fuel burning. Most of the particulate matter is formed in the atmosphere from the trace gases.

Research themes and operational applications Measurements of tropospheric trace gases are important for air quality monitoring and forecasting. Especially, observations of longer-lived trace gases such as CO have proven useful in analyzing the impact of distant sources to local air pollution levels (e.g., Gloudemans et al., 2009). Air quality prediction systems increasingly use satellite observations to improve their forecasting capability. Satellites provide top-down constraints on emission inventories, which traditionally rely on rapidly outdated bottom-up estimates, and generally go unchecked by measurements. As a result of the unique global character of satellite data and their consistency (one retrieval algorithm for all measurements), satellite measurements provide an exceptional tool for checking emission databases. Examples include Lamsal et al. (2011) who combined a top-down approach with a bottom-up a priori inventory to produce an optimal a posteriori inventory for NOx emissions. Mijling et al. (2012) recently applied a new

842

TRACE GASES, TROPOSPHERE - DETECTION FROM SPACE

OMI mean tropospheric NO2 Friday 13 April 2007

Saturday 14 April 2007

0

10

20

0

10

20

50

50

20

50

10

50

0

0

Sunday 15 April 2007

10

20

Monday 16 April 2007

10

50

0

10

20

0

10

20

50

0

20

50

10

50

0

20

No2 tropospheric column [1015 molecules / cm2] 0

1

2

3

4

6

8

11

15

20

Trace Gases, Troposphere - Detection from Space, Figure 3 Four consecutive frames of OMI tropospheric NO2 measurements. The figure clearly shows the Sunday dip in NO2 concentration due to reduction of traffic resulting in less NO2 emissions (Image courtesy: R. Noordhoek (KNMI); data courtesy: KNMI/NASA/NSO and www.temis.nl).

inversion technique to use daily satellite observations for fast updates of emission estimates at high spatial resolution (25  25 km2) for eastern Asia. VOC emissions were first derived from GOME retrievals of HCHO over North America using an inversion based on the relationship between HCHO columns and VOC emissions scaled by their HCHO yields (Palmer et al., 2003). Important developments include simultaneous multispecies inversions to infer CO and NOx emissions from MOPITT and GOME information (Müller and Stavrakou, 2005), and recently, Beirle et al. (2011) estimated megacity NOx emissions and lifetimes directly from the decay-with-distance of NO2 concentrations downwind of megacities. The satellite data can also be used to evaluate models. Comparison of air quality models – traditionally focusing on the lower troposphere – to satellite measurements, for instance, has shown the need to accurately represent the free troposphere in these models (Blond et al., 2007).

Using SCIAMACHY NO2 columns, Bertram et al. (2005) were able to improve the model description of microbial soil NOx emissions that depend on humidity, fertilizer application, and temperature. Satellite data are also very suitable to extent our knowledge on the atmospheric composition by analyzing the long-term records. Richter et al. (2005) showed the increase of NO2 pollution over eastern Asia over the last 10 years. Van der A et al. (2008) extended this work and analyzed spatially resolved trends in NO2 concentrations for the entire globe. Recently, Castellanos and Boersma (2012) showed that reductions in NO2 concentrations over Europe reflect a combination of environmental policy and economic activity. Especially the 2008–2009 global economic recession triggered a distinct change in anthropogenic activity in Europe, indicated by sharp downturns in grossdomestic product and industrial production, and, consequently, in air pollution.

TRACE GASES, TROPOSPHERE - DETECTION FROM SPACE

843

Trace Gases, Troposphere - Detection from Space, Figure 4 Global concentration maps. Color scale concentrations range from red, via yellow and green, to blue, which represents very high, high, medium, and low values, respectively. For population (lower right) white represents small, and dark brown, red represents large population numbers. From top-left to bottom-right: (a) OMI tropospheric column NO2, average for 2007, (b) POLDER/PARASOL (Tanre´ et al., 2011) fine-mode aerosol optical thickness, JuneAugust 2006, (c) MOPITT CO average mixing ratio between 0-2km and 7 km altitude, 7 year average (2000–2007), (d) SCIAMACHY methane column- averaged mixing ratio, Jan 2003 – Oct 2005, (e) SCIAMACHY formaldehyde, average for year 2004, (f) World population map from the Center for International Earth Science Information Network (CIESIN), Columbia Univesity, and Centro Internacional de Agricultura Tropical. (Image courtesy Henk Eskes, KNMI and http://esamultimedia.esa.int/docs/SP1313-6_TRAQ.pdf).

For air quality prediction, usually data assimilation systems are used, where satellite data and ground-based data are integrated with the model data in order to improve the forecast. This technique is only effective when the data are available within a few hours of measurement, also called near real time (NRT). Nowadays, measurements from most backscatter instruments can be obtained within this time frame (e.g., Boersma

et al., 2007). In the thermal infrared, the IASI instrument is also delivering in NRT. The NRT SO2 measurements from volcanoes are used for aviation control, relaying aircraft after a volcanic outburst. Figure 5 shows a SO2 plume from a volcanic outburst as observed by OMI in June 2009. The observations clearly indicate the locations that should be avoided by intercontinental flights.

844

TRACE GASES, TROPOSPHERE - DETECTION FROM SPACE

130

–150

55

–170

150 170

45

SO2 column [DU] 0

1

2

3

4

5

6

7

8

9

10

Trace Gases, Troposphere - Detection from Space, Figure 5 The image shows average column sulfur dioxide concentrations between 11 and 17 June 2009, based on data from OMI. The high SO2 amounts originate from an eruption of the Sarychev Volcano (Kuril Islands, northeast of Japan) on June 12, 2009 (Images like these are used to redirect aircrafts. Image courtesy: S. Carn (Michigan Technology University)).

Future To improve the measurements of the tropospheric trace gases, smaller ground pixels are essential to reduce cloud contamination and obtain highly resolved information on suburban scales. New instruments will target at a 5  5 km2 spatial resolution. In 2015 the TROPOMI instrument will be launched on ESA’s sentinel-5 precursor (Veefkind et al., 2012) with a spatial resolution of 7  7 km2. From 2020 onward this instrument will be followed by ESA’s sentinel 5 instrument on METOP-SG as part of the EU Copernicus program (Ingmann et al., 2012). Special cloud-observing capabilities are in development so that extra cloud information can be measured and the accuracy of the trace gas retrievals for partly cloudy pixels can be improved. By developing combined retrieval techniques that take advantage of backscatter as well as the thermal infrared spectral measurements, vertical profiles of tropospheric trace gases with more than two pieces of information in the troposphere are anticipated. Such combined techniques can be used when both instruments are located on the same satellite. Adding a dedicated aerosol instrument to such a platform would allow for simultaneous observations of all anthropogenic influences on the troposphere for the first time. The METOP-SG satellite is expected to have this combination of instruments on board (Berger et al., 2012)

Apart from a Sun-synchronous orbit, also non-Sunsynchronous as well as geostationary orbits could be used. The Sun-synchronous orbit will provide full global measurements with one or two observations a day. From the geostationary orbit, hourly observations can be made for a relatively small part of the world (e.g., Europe or North America). The non-Sun-synchronous-inclined orbit (see e.g., http://esamultimedia.esa.int/docs/SP13136_TRAQ.pdf) combines global coverage (apart from the polar regions) and frequent measurements (about five per day). Two to three of these non-Sun-synchronous satellites are needed to provide high temporal measurements throughout the year for the whole globe, except the poles. Which orbit to use is dependent on the specific scientific or operational purposes of the satellite. For climate observations, the global coverage of the measurements is of paramount importance. For regional air quality, the high temporal sampling is more important than global coverage. In the timeframe 2017-2020 three geostationary missions are planned: the Sentinel 4 (Ingmann et al., 2012) as part of the EU Copernicus programme over Europe, the South Korean GEMS (Kim et al., 2012) over Southeast Asia, and the U.S. TEMPO (Fishman et al., 2012) over North America. These instrument form a constellation of geostationary air quality missions, can be linked to the Low-Earth-Missions like the sentinel 5 precursor and

TRACE GASES, TROPOSPHERE - DETECTION FROM SPACE

METOP-SG that provide the global coverage and thus interconnect the regional missions.

Summary In the last 15 years, observing trace gases in the troposphere from space has been a tremendously fast developing field. Sophisticated retrieval techniques have been developed in order to obtain tropospheric trace gas concentrations, and these techniques have been validated successfully. Satellite instruments now provide daily observations of the chemical composition of the troposphere with spatial detail as fine as urban scale. Satellite observations have been successfully used to provide constraints on emissions of pollutants, to evaluate and improve atmospheric models, and to identify trends in air pollution. In the future, pixel sizes will even decrease further, and combined measurements from backscatter and thermal infrared will provide better vertical profiles in the troposphere. Bibliography Aumann, H. H., Chahine, M. T., Gautier, C., Goldberg, M. D., Kalnay, E., McMillin, L. M., Revercomb, H., Rosenkranz, P. W., Smith, W. L., Staelin, D. H., Strow, L. L., and Susskind, J., 2003. AIRS/AMSU/HSB on the aqua mission: design, science objectives, data products, and processing systems. IEEE Transactions on Geoscience and Remote Sensing, 41(2), 253–264, doi:10.1109/TGRS.2002.808356. Beer, R., 2006. TES on the aura mission: scientific objectives, measurements, and analysis overview. IEEE Transactions on Geoscience and Remote Sensing, 44, 1102–1105. Beirle, S., Boersma, K. F., Platt, U., Lawrence, M. G., and Wagner, T., 2011. Megacity emissions and lifetimes of nitrogen oxides probed from space. Science, 333, 1737–1739, doi:10.1126/science.1207824, 2011. Berger, M., Moreno, J., Johannessen, J. A., Levelt, P. F., and Hanssen, R. F., 2012. ESA’s sentinel missions in support of Earth system science. Remote Sensing of Environment, 120, 84–90, doi:10.1016/j.rse.2011.07.023. Bertram, T. H., Heckel, A., Richter, A., Burrows, J. P., and Cohen, R. C., 2005. Satellite measurements of daily variations in soil NOx emissions. Geophysical Research Letters, 32, L24812, doi:10.1029/2005GL024640. Blond, N., Boersma, K. F., Eskes, H. J., van der A, R. J., Van Roozendeal, M., De Smedt, I., Bergametti, G., and Vautard, R., 2007. Intercomparison of SCIAMACHY nitrogen dioxide observations, in situ measurements and air quality modeling results over Western Europe. Journal of Geophysical Research, 112, doi:10.1029/2006JD007277. Blumstein, D., Chalon, G., Carlier, T., Buil, C., Hébert, P., Maciaszek, T., Ponce, G., Phulpin, T., Tournier, B., Siméoni, D., Astruc, P., and Clauss, A., 2004. IASI instrument: technical overview and measured performances. In Proceedings SPIE Conference, Denver, Aug 2004, vol. 5543, pp. 196–207 doi: 10.1117/12.560907. Boersma, K. F., Eskes, H. J., Veefkind, J. P., Brinksma, E. J., van der A, R. J., Sneep, M., van den Oord, G. H. J., Levelt, P. F., Stammes, P., Gleason, J. F., and Bucsela, E. J., 2007. Near-real time retrieval of tropospheric NO2 from OMI. Atmospheric Chemistry and Physics, 7, 2103–2118. Bovensmann, H., Burrows, J. P., Buchwitz, M., Frerick, J., Noel, S., Rozanov, V. V., Chance, K. V., and Goede, A. P. H., 1999.

845

SCIAMACHY: mission objectives and measurement modes. Journal of Atmospheric Sciences, 56, 127–155. Burrows, J. P., Weber, M., Buchwitz, M., Rozanov, V. V., Ladstatter-Weissenmayer, A., Richter, A., De Beek, R., Hoogen, R., Bramstedt, K., Eichmann, K. W., Eisinger, M., and Perner, D., 1999. The Global Ozone Monitoring Experiment (GOME): mission concept and first scientific results. Journal of Atmospheric Sciences, 56, 151–175. Butz, A., Guerlet, S., Hasekamp, O., Schepers, D., Galli, A., Aben, I., Frankenberg, C., Hartmann, J.-M., Tran, H., Kuze, A., Keppel-Aleks, G., Toon, G., Wunch, D., Wennberg, P., Deutscher, N., Griffith, D., Macatangay, R., Messerschmidt, J., Notholt, J., and Warneke, T., 2011. Toward accurate CO2 and CH4 observations from GOSAT. Geophysical Research Letters, 38(14). L14812. doi:10.1029/2011GL047888. Callies, J., Corpaccioli, E., Eisinger, M., Hahne, A., and Lefebvre, A., 2000. GOME-2 – MetOp’s second generation sensor for operational ozone monitoring. ESA Bulletin, 102, 28–36. Castellanos, P., and Boersma, K. F., 2012. Reductions in nitrogen oxides over Europe driven by environmental policy and economic recession. Scientific Reports, 2(2), 265, doi:10.1038/srep00265. Clerbaux, C., Boynard, A., Clarisse, L., George, M., Hadji-Lazaro, J., Herbin, H., Hurtmans, D., Pommier, M., Razavi, A., Turquety, S., Wespes, C., and Coheur, P.-F., 2009. Monitoring of atmospheric composition using the thermal infrared IASI/ MetOp sounder. Atmospheric Chemistry and Physics, 9, 6041– 6054, doi:10.5194/acp-9-6041-2009. Crisp, D., Atlas, R. M., Breon, F.-M., Brown, L. R., Burrows, J. P., Ciais, P., Connor, B. J., Doney, S. C., Fung, I. Y., Jacob, D. J., Miller, C. E., O'Brien, D., Pawson, S., Randerson, J. T., Rayner, P., Salawitch, R. J., Sander, S. P., Sen, B., Stephens, G. L., Tans, P. P., Toon, G. C., Wennberg, P. O., Wofsy, S. C., Yung, Y. L., Kuang, Z., Chudasama, B., Sprague, G., Weiss, B., Pollock, R., Kenyon, D., and Schroll, S., 2004. The Orbiting Carbon Observatory (OCO) mission. Advances in Space Research, 34(4), 700–709. Deeter, M. N., 2009. MOPITT (Measurements of Pollution in the Troposphere) Validated Version 4 Product User’s Guide. Available from http://www.acd.ucar.edu/mopitt/v4_users_guide_val.pdf. Edwards, D. P., Emmons, L. K., Hauglustaine, D. A., Chu, D. A., Gille, J. C., Kaufman, Y. J., Pétron, G., Yurganov, L. N., Giglio, L., Deeter, M. N., Yudin, V., Ziskin, D. C., Warner, J., Lamarque, J.-F., Francis, G. L., Ho, S. P., Mao, D., Chen, J., Grechko, E. I., and Drummond, J. R., 2004. Observations of carbon monoxide and aerosols from the Terra satellite: northern hemisphere variability. Journal of Geophysical Research, 109, D24202, doi:10.1029/2004JD004727. Fishman, J., Vukovich, F. M., Cahoon, D., and Shipman, M. C., 1987. The characterization of an air pollution episode using satellite total ozone measurements. Journal of Applied Meteorology, 26, 1638–1654. Fishman, J., and co-authors, 2012. The United STATES’ next generation of atmospheric composition and coastal ecosystem measurements: NASA’s geostationary coastal and Air pollution events (GEO-CAPE), mission. Bulletin of the American Meteorological Society, 93, 1547–1566. http://dx.doi.org/10.1175/ BAMS-D-11-00201.1. Fraser, R. S., Kaufman, Y. J., and Mahoney, R. L., 1984. Satellite measurements of aerosol mass transport. Atmospheric Environment, 18, 2577–2584. Gloudemans, A. M. S., de Laat, A. T. J., Schrijver, H., Aben, I., Meirink, J. F., and van der Werf, G. R., 2009. SCIAMACHY CO over land and oceans: 2003–2007 inter annual variability. Atmospheric Chemistry and Physics, 9, 3799–3813, doi: doi:10.5194/acp-9-3799-2009. Hamazaki, T. Kuze, A., Kondo, K., 2004. Sensor system for Greenhouse Gas Observing Satellite (GOSAT). In Proceedings SPIE

846

TRAFFICABILITY OF DESERT TERRAINS

Conference, Denver, Aug 2004, vol. 5543, pp. 275–282 doi: 10.1117/12.560589. Hilsenrath, E., Cebula, R. P., Deland, M. T., Laamann, K., Taylor, S., Wellemeyer, C., and Bhartia, P. K., 1995. Calibration of the NOAA-11 Solar Backscatter Ultraviolet (SBUV/2) ozone data set from 1989 to 1993 using in-flight calibration data and SSBUV. Journal of Geophysical Research, 100, 1351–1366. Ingmann, P., Veihelman, B., Langen, J., Lamarre, D., Stark, H., and Courreges-Lacoste, G. B., 2012. Requirements for the GMES Atmospheric Service and ESAs implementation concept: Sentinels-4/-5 and -5p. Remote Sensing of the Environment, 120, 58–69, doi:10.1016/j.rse.2012.01.023. Kim, J., 2012. GEMS(Geostationary Environment Monitoring Spectrometer) onboard the GeoKOMPSAT to Monitor Air Quality in high Temporal and Spatial Resolution over Asia-Pacific Region, EGU General Assembly 2012, held 22–27 April, 2012 in Vienna, Austria., p. 4051, EGU2012–4051. Lamsal, L. N., Martin, R. V., Padmanabhan, A., van Donkelaar, A., Zhang, Q., Sioris, C. E., Chance, K., Kurosu, T. P., and Newchurch, M. J., 2011. Application of satellite observations for timely updates to global anthropogenic NOx emission inventories. Geophysics Research and Letters, 38. L05810, doi:10.1029/2010GL046476. Leue, C., Wenig, M., Wagner, T., Klimm, O., Platt, U., and Jähne, B., 2001. Quantitative analysis of NOx emissions from global ozone monitoring experiment satellite image sequences. Journal of Geophysical Research, 106, 5493–5505. Levelt, P. F., van den Oord, G. H. J., Dobber, M. R., Mälkki, A., Visser, H., de Vries, J., Stammes, P., Lundell, J., and Saari, H., 2006. The ozone monitoring instrument. IEEE Transactions on Geoscience and Remote Sensing, 44(5), 1093–1101, doi:10.1109/TGRS.2006.872333. McPeters, R. D., Bhartia, P. K., Krueger, A. J., Herman, J. R., Schlesinger, B. M., Wellemeyer, C. G., Seftor, C. J., Jaross, G., Taylor, S. L., Swissler, T., Torres, O., Labow, G., Byerly, W., Cebula, R., McPeters, R. D., Bhartia, P. K., Krueger, A. J., Herman, J. R., Schlesinger, B. M., Wellemeyer, C. G., Seftor, C. J., Jaross, G., Taylor, S. L., Swissler, T., Torres, O., Labow, G., Byerly, W., and Cebula, R. P., 1996. Nimbus-7 Total Ozone Mapping Spectrometer (TOMS) Data Products User’s Guide. Washington, DC: NASA. NASA Reference Publication, Vol. 1384. Mijling, B., and Van der A, R. J., 2012. Using daily satellite observations to estimate emissions of short-lived air pollutants on a mesoscopic scale. Journal of Geophysics and Research, 117, D17302, doi:10.1029/2012JD017817. Müller, J.-F., and Stavrakou, T., 2005. Inversion of CO and NOx emissions using the adjoint of the IMAGES model. Atmospheric Chemistry and Physics, 5, 1157–1186. Palmer, P. I., Jacob, D. J., Fiore, A. M., Martin, R. V., Chance, K., and Kurosu, T. P., 2003. Mapping isoprene emissions over North America using satellite observations of formaldehyde columns. Journal of Geophysical Research, 108, 4180, doi:10.1029/ 20052JD002153. Platt, U., 1994. Differential optical absorption spectroscopy (DOAS). In Sigrist, M. W. (ed.), Air Monitoring by Spectroscopic Techniques. New York: Wiley, pp. 27–84. Richter, A., Burrows, J. P., Nüß, H., Claire Granier, C., and Niemeier, U., 2005. Increase in tropospheric nitrogen dioxide over China observed from space. Nature, 437, 129–132. Rodgers, C. D., 2000. Inverse Methods for Atmospheric Sounding: Theory and Practice. River Edge: World Science. Tanré, O., Bréon, F. M., Deuzé, J. L., Dubovik, O., Ducos, F., François, P., Goloub, P., Herman, M., Lifermann, A., Waquet, F., 2011. Remote sensing of aerosols by using polarized, directional and spectral measurements within the A-Train: the

PARASOL mission. Atmospheric Measurement Techniques, 4, 1383–1395. doi:10.5194/amt-4-1383-1395-2011. van der A, R. J., Eskes, H. J., Boersma, K. F., van Noije, T. P. C., Van Roozendael, M., De Smedt, I., Peters, D. H. M. U., Kuenen, J. J. P., and Meijer, E. W., 2008. Identification of NO2 sources and their trends from space using seasonal variability analyses. Journal of Geophysical Research, 113, D04302, doi:10.1029/ 2007JD009021. Veefkind, J. P., Aben, I., McMullan, K., Förster, H., de Vries, J., Otter, G., Claas, J., Eskes, H. J., de Haan, J. F., Kleipool, Q., van Weele, M., Hasekamp, O., Hoogeveen, R., Landgraf, J., Snel, R., Tol, P., Ingmann, P., Voors, R., Kruizinga, B., Vink, R., Visser, H., and Levelt, P. F., 2012. TROPOMI on the ESA Sentinel-5 Precursor: A GMES mission for global observations of the atmospheric composition for climate, air quality and ozone layer applications. Remote Sensing of Environment, 120, 70–83, doi:10.1016/j.rse.2011.09.027.

Cross-references Aerosols Air Pollution Climate Monitoring and Prediction Data Assimilation Geophysical Retrieval, Inverse Problems in Remote Sensing Optical/Infrared, Atmospheric Absorption/Transmission, and Media Spectral Properties Stratospheric Ozone Trace Gases, Stratosphere, and Mesosphere Ultraviolet Remote Sensing Ultraviolet Sensors

TRAFFICABILITY OF DESERT TERRAINS Charles Hibbitts Applied Physics Laboratory, Laurel, MD, USA

Definition Trafficability (of desert terrains): The ability to traverse a surface in a cart vehicle. The trafficability of desert terrain is controlled by soil strength, surface roughness, and the propensity for a surface to generate dust (dust loading) that would obscure vision or impair engine performance. This entry focuses on the most important and treacherous factor in desert trafficability, soil strength. Surface roughness and dust loading are mentioned in context with specific terrain types. Finally, methods for remotely estimating these factors are briefly discussed. Perhaps the most difficult to recognize, and likely the most important factor for trafficability in desert terrains, is soil strength, which itself is dependent upon several factors that can be measured or estimated: soil moisture, grain (or clast) size, and composition. It is the combination of soil strength and vehicle footprint (the area over which the weight of the vehicle is distributed) that dominates a surface’s trafficability.

TRAFFICABILITY OF DESERT TERRAINS

Contributing factors to soil strength Soil grain size In general, as the amount of fine-grained material (