Dinaric karst

The designations employed and the presentation of material throughout the publication do not imply the expression of any opinion whatsoever on the part of UNESCO concerning the legal status of any country, territory, city or area or of its authorities, or the delineation of its frontiers or boundaries.

Published in 2010 by the United Nations Educational, Scientific and Cultural Organization 7, place de Fontenoy, 75352 Paris 07 SP Printed by UNESCO © UNESCO 2010 IHP-VII/2010/GW-2

SUSTAINABILITY OF THE KARST ENVIRONMENT DINARIC KARST AND OTHER KARST REGIONS International Interdisciplinary Scientific Conference (Plitvice Lakes, Croatia, 23-26 September 2009) Convened and Organised by: Centre for Karst (Gospi, Croatia) International Scientific Committee Ognjen Bonacci (Croatia), Chairman Franci Gabrovšek (Slovenia) Mladen Jurai (Croatia) Božidar Biondi (Croatia) Wolfgang Dreybrodt (Germany) Arthur Palmer (USA) Derek C. Ford (Canada) David Culver (USA) Andrej Mihevc (Slovenia) Jacques Mudry (France) Daoxian Yuan (China) Nico Goldscheider (Switzerland, Germany) Zoran Stevanovi (Serbia) Mario Parise (Italy) Hans Zojer (Austria) Elery Hamilton - Smith (Australia) Neven Kreši (USA) Bartolomé Andreo (Spain) Local Organizing Committee Jadranka Pejnovi, Chair Željko Župan, Secretary Ivo Lui Neven Boi Aleksandar Luki Ljudevit Tropan Dubravka Kljajo Krešimir ulinovi Ivica Tomljenovi

Foreword The objective of the international interdisciplinary scientific conference “Sustainability of the karst environment - Dinaric karst and other karst regions”, organized by Centre for Karst, Gospi, Croatia, was to give a theoretical and practical contribution to the concept of sustainable development in karst regions, with a special emphasis on the experiences achieved in the Dinaric karst region. The exchange of information and findings obtained in other karst regions worldwide allows for an integral approach to this complex issue, and thereby contribute towards finding reliable solutions. The basic objective of the conference was to apply an interdisciplinary approach to scientifically assess the issues of sustainable development of all forms of karst. The issue was approached from different perspectives, from those of a technical and biological nature, to those addressing the social aspects of environmental issues and life on the karst. The conference itself was held at the Plitvice Lakes (World Heritage Site), one of the most fascinating phenomena on Earth. During the conference, one half-day excursion was organized to visit National Park Plitvice Lakes. Following the conference, an excursion was organised to visit several other significant phenomena of the Dinaric karst in Croatia. Conference themes were: - Geological aspects - Geomorphological aspects - Hydrological and hydrogeological aspects - Coastal and submerged karst - Biological and ecological aspects of karst - Anthropogenic impacts and protecting karst - Sociological, demographic and social aspects of karst - Dinaric karst and other karst regions (China, Alpine, Caribbean karst, etc.) The publication will serve as a contribution to the VIIth Phase of the International Hydrological Programme (IHP 2008-2013) of UNESCO, which has endeavoured to address demands arising from a rapidly changing world. Chairman of Scientific Committee Ognjen Bonacci

International Interdisciplinary Scientific Conference SUSTAINABILITY OF THE KARST ENVIRONMENT DINARIC KARST AND OTHER KARST REGIONS (Plitvice Lakes, Croatia, 23-26 September 2009) TABLE OF CONTENTS International scientific and local organising committees …………………………………......3 Foreword ...................................................................................................................................4 Table of contents (authors in alphabetical order) ……………………………………..………5 BONACCI Ognjen

Sinking, losing and underground karst streamflows …………………..……9 BORDA Daniela, RACOVI Gheorghe, NSTASE-BUCUR Ruxandra, CIUBOTRESCU Christian

Ecological reconstruction of bat cave Roost in western Carpathians ….....17 BRINKMANN Robert

Karst and sustainability in Florida, U.S.A. …………………………..…......25 DELLE ROSE Marco, PARISE Mario

Water management in the karst of Apulia, southern Italy ……………..….33 DÖRFLIGER Nathalie, FLEURY Perrine , BAKALOWICZ Michel , EL HAJJ Hahmad, AL CHARIDEH Abdoul, EKMEKCI Mehmet

Specificities of coastal karst aquifers with the hydrogeological characterisation of submarine springs – overview of various examples in the Mediterranean basin ………………………………………………..…41 DÖRFLIGER Nathalie, PLAGNES Valérie, KAVOURI Konstantina

PaPRIKa a multicriteria vulnerability method as a tool for sustainable management of karst aquifers - Example of application on a test site in SW France ………………………………………………………………….…49 EFTIMI Romeo

Investigation about recharge sources of Bistrica karst spring, the biggest spring of Albania, by means of environmental hydrochemical and isotope tracers …………………………………………………………....57

GANOULIS Jacques, AURELI Alice, KUKURI Neno

Importance of transboundary karst aquifer resources in South Eastern Europe (SEE) ………………………………………………………………....67 GUO FANG Jiang Guanghui

The resources, environment and development in Fengshan Geopark karst area ………………………………………………………………….......75 HUBINGER Bernhard , REHRL Christoph, BIRK Steffen

Linking generic models to site-related models of conduit evolution ………83 JAMES Julia M., SPATE Andy

Sustainability in a karst - the Bungonia Caves, New South Wales, Australia ………………………………………………………………….…...91 KATSANOU Konstantina, NIKOLAOU Euaggelos, SIAVALAS George, ZAGANA Eleni, LAMBRAKIS Nikolaos

Hydrogeological conditions and water quality of the karstified formations of Louros basin, Epirus, Greece ……………………………......97 KNEZ Martin, SLABE Tadej

Karstology and motorway construction ……………………………….…..107 KNEZ Martin, SLABE Tadej

Shilin - lithological characteristics, form and rock relief of the Lunan Stone Forests (South China karst) …………………………………………115 KOVAI Gregor, PETRI Metka

Contribution of time series analysis to the study of the Malenšica karst spring, Slovenia …………………………………………………….…123 MALEKOVI Sanja, TIŠMA Sanja , FARKAŠ Anamarija

Capacity for managing local development in karst areas …………….…..129 MUDARRA Matías, ANDREO Bartolomé

Hydrogeological functioning of the karst aquifer drained by Yedra Spring (Southern Spain) from hydrochemical components and organic natural tracers ………………………………………………………………………......137

NAUGHTON Owen, JOHNSTON Paul, GILL Laurence

The hydrology of turloughs as groundwater dependent terrestrial ecosystems …………………………………………………………………...147 PARISE Mario

Hazards in karst …………………………………………………………….155 PERNE Matija

Modelling of rillenkarren formation ………………………………………163 RUBINIC Josip, KATALINIC Ana, SVONJA Mirjana, GABRIC Ivana, BUSELIC Gordana, CUZE Maja, HORVAT Bojana

Salinization of the Vrana Lake in Dalmatia within the context of anthropogenic influences and climate changes (situation in 2008) ………171 TERZI Josip, PAVII Ante, MARKOVI Tamara, LUKA REBERSKI Jasmina

Protection of the Miljacka karst spring: an underground connection between the rivers Zrmanja and Krka ……………………………………179

Sinking, losing and underground karst streamflows Ognjen BONACCI Faculty of Civil Engineering and Architecture, Split University, 21000 Split, Matice hrvatske 15, Croatia, e-mail: [email protected]

Abstract: Sinking, losing and underground streamflows are typical and relatively frequent karst phenomena. A sinking surface streamflow can be defined as a surface river or stream flowing onto or over karst and which then disappears completely underground through a swallow hole and which may or may not rise again and flow as a resurgent surface river or stream. A losing streamflow can be defined as an open stream or river that loses water as it flows downstream. The level of water in a losing stream is above the water table: in comparison, the level of water in a gaining stream is below the water table. In a losing stream water infiltrates underground, because the water table is below the bottom of the stream channel. Underground or subterranean streamflows are subsurface karst passages with the main characteristics of open rivers or streams. In underground streamflows water flows through caves, caverns, karst conduits and large galleries in karst underground. The paper treats some conceptual aspects of sinking, losing and underground streamflows. Some cases of the special hydrological and hydrogeological behaviour of karst sinking, losing and underground streamflows are explained. Keywords: karst, sinking, losing, underground streamflow

1 Introduction Karst is defined as a terrain, generally underlain by limestone or dolomite, in which the topography is chiefly formed by the dissolving of rock, and which is characterised by sinkholes, sinking streams, closed depressions, subterranean drainage and caves (Field 2002). A wide range of closed surface depressions, a well-developed underground drainage system, and strong interaction the between circulation of surface water and groundwater typify karst. Due to very high infiltration rates, especially in bare karst, overland and surface flow is rare in comparison with non-karst terrains. Carbonate rocks are more soluble than many other rocks. They are subject to a number of geomorphological processes. The processes involved in the weathering and erosion of carbonate rocks are many and diverse. The varied and often spectacular surface landforms are merely a guide to the presence of unpredictable conduits, fissures and cavities beneath the ground. But at the same time these subsurface features can occur even where surface karstic landforms are completely absent. Diversity is considered the main feature of karstic systems, which are known to change over time and in space so that an investigation of each system on its own is required. Interactions between the surface and subsurface in karst are very strong (Bonacci 1987). Groundwater and surface water are hydraulically connected through numerous karst features that facilitate the exchange of water between the surface and subsurface (Katz et al 1997). High and fast oscillations of groundwater levels in karst control the hydrogeological and hydrological regimes of influent and underground streams. An important issue in studying these streams is that subsurface water is highly heterogeneous in terms of the location of conduits, the location of vertically moving water, and flow velocities. Due to the previously mentioned reasons, the occurrence of losing, sinking and underground streamflows is more the rule than an exception. 9

A great problem regarding the explanation of the hydrological and hydrogeological behavior of such streamflows is connected with the particularities of karst underground features and especially with karst aquifers. Karst aquifers are some of the most complex and difficult systems to decipher. The highly heterogeneous nature of karst aquifers leads to an inability to predict groundwater flow direction and travel times. The circulation of groundwater in karst aquifers is quite different from water circulation in other non-karstic type aquifers. The hydraulic permeability of karst aquifers is essentially created by flowing water and has an anisotropic character. In karst terrains groundwater and surface water constitute a single dynamic system. Due to this reason one of the almost unavoidable characteristics of open streams, creeks and rivers in karst regions is that they either have partial water loss along their course or completely sink into the underground (Bonacci 1987). Sinking, losing and underground streamflows are more typical, significant and relatively frequent karst phenomena than is reflected in their treatment in the karst literature. A synonym for a sinking and losing stream is an influent stream. Such streams have an integral function in karst hydrology and hydrogeology. Influent and underground streams develop when they cross soluble rocks along their transfer route to base-level rivers or seas (Ray 2005). Challenges to the investigation of influent and underground streams include the concurrent existence of fast turbulent flow through large karst conduits and slow, diffuse laminar flow through small karst fissures, joints, cracks and bedding plains (the karst matrix). Numerous and extremely varied surface and underground karst forms make unexpected water connections possible in karst medium space, which changes over time. Changes of the underground flow path over the time are caused by: 1) Different recharges from different surface areas, mainly due to the by variable distribution of areal precipitation; 2) Different groundwater levels and their rapid changes in time and space; 3) Anthropogenic influence; and 4) Exogenic and endogenic forces (Bonacci 2004). The objective of this paper is to discuss the hydrological aspects of losing, sinking and underground streams that are closely connected with the hydrogeological characteristics of the regions through which they circulate. One of the key issues for the better understanding, protection and management of karst systems is the determination of the influent and underground stream catchment area. Due to very special and complex underground and surface karst forms, there are a wide variety of cases of karst sinking, losing and underground streamflows. An attempt at their conceptualisation is provided in the paper. The main issue in the classification of these kinds of karst rivers is that they can be losing, sinking and underground, all at the same time. A description of specific cases of the special hydrological behaviour of sinking, losing and underground streamflows is given. 2 Losing streamflow A losing streamflow can be defined as an open stream or river that loses water as it flows downstream. A losing streamflow is a surface stream that contributes water to the karst groundwater system in localized areas. It has cracks in its bed that allow water to seep into the groundwater. These losses can be massive in particular river sections, whereas in others they are small and difficult or even impossible to observe without performing especially precise measurements. A direct way surface water becomes groundwater is through the capture of surface streams into subsurface voids through swallets. These features swallow the surface stream and represent a rapid and direct way for groundwater recharge. Losing streams segments are important groundwater recharge zones for underlying karst aquifers. A losing streamflow is one having a bed that allows water to flow directly into the groundwater system. The water level in a losing stream is higher than the water table, as opposed to the water level in a gaining stream which is lower than the water table. The water than infiltrates underground as the water table is lower than the bottom of the stream channel. 10

Losing streamflows are often used in relation to karst aquifers. Aquifers gain the water lost by the losing stream. Due to very rapid rise and fall of groundwater levels in karst terrains, some losing rivers or their losing stretches can intermittently act as gaining streams. Figure 1 presents an attempt at the conceptualisation of losing streamflows. Occasionally, permanent water courses flow beyond the groundwater level, even for 50 m or more. Bonacci (1987, 1999) called these river sections “suspended“ or “perched”. Water infiltrated from these sections can either flow in another catchment or can reappear in the downstream reaches of same river (at the spring B in Figure 1b). Legend:

a)

spring river section without losses suspended river section

B

b)

river flow direction B spring can be permanent or intermittent

sinkhole (swallow hole, ponor) flow direction of infiltrated water through the large karst conduits in the same river catchment flow direction of infiltrated water through the karst matrix in the same river catchment

c)

B

flow direction of infiltrated water through the large karst conduits in an other river catchment flow direction of infiltrated water through the karst matrix in an other river catchment

Figure 1 Conceptualisation of losing streamflows For example “suspended“ or “perched stretches exist on two neighbouring karst rivers Zrmanja and Krka (Dinaric karst of Croatia). While the Zrmanja River dries out, the Krka River never dries out in these sections. The reason why there are no water losses on the Krka “suspended” section of the Krka is in the fact that its riverbed is comprised of fine-grained sediments, which make infiltration impossible. Dye-tracing methods are commonly used to determine groundwater flow paths, relations between surface water and groundwater, and groundwater travel times through the karst underground. It should be stressed that flow paths, connections between certain sinks and springs, very often vary in time and space, mainly due to the varying groundwater conditions in the underground. Complexity of the precise determination of the water losses along open streamflows in karst is discussed by Bonacci (1987). 3 Sinking streamflow A sinking surface streamflow can be defined as a surface river or stream flowing onto or over karst that then disappears completely underground through a swallow-hole (ponor or sinkhole) and which may or may not rise again and flow as a resurgent surface river or stream. Infiltration from sinking streams into the karst groundwater system is the most rapid form of recharge for carbonate aquifers (Hess et al 1989). Sinking streams represent the most direct access to the sensitive and highly vulnerable karst groundwater system. The unique nature of sinking rivers is their development and evolution of conduit flow routes and caves through soluble rocks. The evolution of most of the world’s largest and most significant karst caves and springs are formed as a consequence of large volumes of concentrated recharge from sinking rivers (Ray 2005). 11

Figure 2 presents an attempt at the conceptualisation of sinking streamflows. Sinking stream can reappear at the surface through a typically large karst spring (Figure 2a), though there are some cases when it reappears through many permanent and intermittent karst springs dissipated over a large area.

Legend:

a)

permanent spring intermittent (temporary) spring river flow direction sinkhole (swallow hole, ponor)

b)

flow direction of sinking water through large conduits flow direction of sinking water through karst matrix

Figure 2 Conceptualisation of sinking streamflows Hess et al (1989) explains that the south-Central Kentucky karst aquifer is fed by many sinking streams. Their catchments are made up of an aggregate of many small surface catchments ranging over an area of a few square kilometres. Some of these streams have several surface tributaries, but most of the sinking creeks are short, first order streams. The Lika and Gacka Rivers (Dinaric karst of Croatia) are typical sinking streamflows. These rivers are located in the central part of the Dinaric karst region of Croatia (Figure 3) between 44°17’ and 44°58’N and 15°07’ and 15°48’E. Their precise hydrological catchment areas and boundaries are not known (Bonacci and Andri 2008). The Velebit Mountain (max. altitude 1758 m a. s. l.) separates their catchments from the Adriatic Sea. Water from the both rivers sinks at altitudes between 400 and 450 m a. s. l. and reappears at many permanent and intermittent coastal and submarine karst springs of the Adriatic Sea (Figure 3). 4 Underground streamflow Underground or subterranean streamflows are subsurface karst passages that have the main characteristics of open rivers or streams. In an underground streamflow, water flows through caves, caverns, karst conduits and large galleries in the karst underground. The karst underground system provides access to fragments of the abandoned conduit system, which have hydraulic geometries comparable, though not identical, to those of surface rivers or streams. The Port Miou system (Cassis, France) is a two kilometre long submarine gallery that extends in the limestone series of Calanques (Marseille, France). The two largest karst submarine springs, Port Miou and Bestouan, represent the mouths of two underground karst rivers into the Mediterranean Sea. The average discharge of brackish water flowing from the Port Miou spring is between 2 to 5 m3/s (Potié et al. 2005; Cavalera and Gilli 2009). The roof of the entirely submerged Port Miou gallery lies between 10 and 20 m below sea level to about 800 m from the spring exit. It then goes between 10 to 30 m deeper. At about 2200 m from the entrance, the primarily horizontal karst conduit suddenly drops into a deep vertical shaft. Cave divers were able to explore the conduit to a depth of 179 m below sea level. At

12

that depth, the water is still brackish. The system extends further and deeper, however exploration is limited by the present diving technology.

Figure 3 The Lika and Gacka Rivers (Dinaric karst of Croatia), typical sinking streamflows The Cassidaigne canyon cuts the continental shelf where bathymetric studies have shown the presence of dolines. Caves and speleothems have been observed during submarine explorations on the walls of the canyon. Its presence is related to the several stages of the lowering of the Mediterranean Sea during the Messinian salinity crisis. Cavalera and Gilli (2009) suggest that during the important drop of sea level of the Mediterranean, the underground river of Port-Miou, flowed several hundreds meters below its current position, and excavated the canyon. At the end of the Messinian crisis, the system was flooded by sea water. Karst water now flows through an upper gallery, however the presence of a paleo-drain filled by sea water makes deep marine intrusion into the karst system possible. In order to prevent intrusion of sea water, two submarine dams were constructed in the horizontal conduit, about 500 m from the spring exit. However, the issue of contamination with sea water was not resolved by their construction. Cavalera and Gilli (2009) consider that the saline contamination of Port Miou could be carried out by a sea water inflow through a deep karstic conduit connected to the canyon of Cassidaigne. The Santa Fe River (Florida, USA) flows from an impervious catchment onto karstified Eocene limestones. At the O’Leno Sink, it sinks underground for 5 km before resurfacing at the Santa Fe River Rise. Hisert (1994) conducted a geochemical tracer study to determine which of the numerous karst features occurring between the O’Leno Sink and the Santa Fe River Rise are connected to the underground river. In addition, water temperature measurements were made to distinguish the relative proportions of groundwater and surface water in each water filled karst feature. The results showed that the Santa Fe River Rise is a point of resurgence for a portion of the Santa Fe River flow diverted underground at the O’Leno Sink. The underground river course is singular and sinuous. The flow is conduit and rapid with a velocity of 2.5 km/day (Hisert 1994). The upstream half of the underground river is fairly well delineated, due to the great number of surface sink features. In the downstream 13

section, the underground course is questionable due to the lack of surface karst features that can be used as windows to the karst underground. Figure 4 presents the map of the Disu underground stream system (Yuan 1991), which has a catchment area of 1004 km2. The system has a total length of 241.1 km, and includes a main conduit that is 57.2 km long and 12 tributaries. The Disu underground system is the longest identified subterranean stream in China. In the upstream section, it is about 100 m in depth, with karst conduits usually in a simple fissure-shape, from several meters to 30 m wide, and ten to tens of meters high. The average hydraulic gradient is about 12%. At the middle and lower reaches, it is 30 to 50 m below the bottom of the valleys. The cross-section of the conduit here varies between 145 and 184 m2, and the average hydraulic gradient is 1%. Discharges at the exit of the Disu underground river vary from the minimum 4.03 m3/s in dry season to the maximum 544.9 m3/s (Yuan 1991).

Figure 4 The Disu underground stream system in China (Yuan 1991) 5 Discusion True cases of karst losing, sinking and underground rivers are much more complex than any concept can imagine. In reality, very different combinations exist. Some streams can, at the same time, be losing, sinking and underground. The Dobra River (Dinaric karst of Croatia) serves as a good example. Figure 5 represents the longitudinal cross-section of the entire Dobra River, divided into three parts. The first one is a losing and sinking river called Upper Dobra, with a length from the spring to the ula sink-hole of 51.2 km. The second part is an underground karst river flowing from the ula sink-hole to the karst spring zone near the village Gojak. The shortest aerial distance between the ula sink-hole and the Lower Dobra River karst springs zone is 4.6 km. In order to reappear at the karst springs zone, the Lower Dobra River flows through karst caves and conduit system that is 16,296 km long. The longitude of the Lower Dobra River is 52.1 km. There are huge water losses along some sections of the open watercourse of the Upper Dobra River through small karst sinks located at the bottom of its channel. These have changed over 14

time as a function of the groundwater level. During periods with high groundwater levels losing stretches become gaining stretches. The importance of sinking, losing and underground streamflows in karst system functioning is very significant. Their hydrological, hydrogeological and other characteristics are extremely complex. Due to these reasons, it is necessary to apply interdisciplinary approaches, methods and concepts in their investigation. It is obvious that efforts aimed at to their better understanding should be intensified.

"LUKE" GAUGING STATION

"MORAVICE" GAUGING STATION

77.9 350.0

94.7 420.0

107.9km

RIVER SPRING

"HRELJIN" GAUGING STATION 65.6 335.0

100 km

"SV. PETAR" GAUGING STATION "TURKOVICI" GAUGING STATION

80 km

60.7 322.5 62.6 323.0

HEPP GOJAK

ULA SINKHOLE

HEPP LEŠCE

60 km

"TROŠMARIJA" GAUGING STATION

ALTITUDE [m a.s.l.]

800 700

"DANI" GAUGING STATION

900

40 km

"LEŠCE" GAUGING STATION

1000

20 km

"STATIVE" GAUGING STATION

RIVER MOUTH

0 km

~ 840.0

600 500 320.0

400 300

175.0

200

110.0

107.9 840.0

56.3 320.0

48.0 175.0

51.7 200.0

40.1 147.4

34.8 140.0

39.0 147.0

11.1 110.0

L [km]

0.0

ALTITUDE [m a.s.l.]

110.0

100

Figure 5 Longitudinal cross-section of the Dobra River (Dinaric karst of Croatia) References Bonacci O (1987) Karst hydrology with special references to the Dinaric karst. Springer Verlag, Berlin, 184 pp Bonacci O (1999) Water circulation in karst and determination of catchment areas: example of the River Zrmanja. Hydrological Sciences Journal 44(3):373-386 Bonacci O (2004) Hazards caused by natural and anthropogenic changes of catchment area in karst. Natural Hazards and Earth System Sciences 4:655-661 Bonacci O, Andri I (2008) Sinking karst rivers hydrology: case of the Lika and Gacka (Croatia). Acta Carsologica 37(2-3):185-196 Cavalera T, Gilli E (2009) The submarine river of Port Miou (France), A karstic system inherited from the Messinian deep stage. Geophysical Research Abstracts Vol. 11, EGU 2009-5591 Katz BG, DeHan RS, Hirten JJ, Catches JS (2007) Interactions between ground water and surface water in the Suwannee river basin, Florida. Journal of the American Resources Association 33(6):1237-1254 Field MS (2002) A lexicon of cave and karst terminology with special reference to environmental karst hydrology. USEPA, Washington, DC, 214 pp Hess JW, Wells SG, Quinlan JF, White WB (1989) Hydrogeology of the South-Central Kentucky karst. In: WB White, EL White (eds) Karst hydrology concepts from the Mammoth cave area. Van Nostrand Reinhold, New York, pp 15-63 15

Hisert RA (1994) A multiple tracer approach to determine the ground water and surface water relationships in the Western Santa Fe River, Columbia County, Florida. Ph.D. Dissertation, Department of Geology, University of Florida, Gainesville, FL 32601. Potié L, Ricour J, Tardieu B (2005) Port-Mioux and Bestouan freshwater submarine springs (Cassis-France) investigations and works (1964-1978). Proceedings of International Conference “Water resources & environmental problems in karst”, Belgrade and Kotor, pp 266-274 Ray JA (2005) Sinking streams and losing streams. In: DC Culver, WB White (eds) Encyclopedia of Caves. Elsevier, Amsterdam, pp 509-514 Yuan D (1991) Karst of China. Geological Publishing House, Beijing, 232 pp

16

Ecological reconstruction of bat cave Roost in western Carpathians Daniela BORDA1, Gheorghe RACOVI1, Ruxandra NSTASE-BUCUR1, Christian CIUBOTRESCU2 1

“Emil Racovitza” Institute of Speleology, Clinicilor St., no 5, 400006 Cluj-Napoca, Romania, e-mail: [email protected] 2 Speleological Association for Environmental Protection and Karst "Sfinx" – Gârda, Romania e-mail: [email protected]

Abstract: The underground environment represents an extreme and, at the same time, fragile environment because of the particularities of biotic and abiotic factors and of its trophic dependence on surface ecosystems. The high constancy of cave factors makes it one of the most vulnerable environments on Earth. Because of the increased human pressure, in the last decades we witnessed a strong degradation of the underground environment in areas exposed to pollution, as a consequence of restraining, retreating or even extinction of specific fauna. In this context we monitor a show cave where the cave electrification and the wood staircase, which facilitated the tourist’s passage in the upper level, were removed. Also, the artificial entrance to the upper level was reversibly obstructed. Our research focused on analyzing the microclimatic conditions, as well as on the airborne microorganisms from the cave, and on the bat dynamics, after the obstruction of the artificial entrance in the upper level of the cave. The results show a direct relation between cave climate, bats, airborne microorganisms, and the cave visitors. From the climatic perspective Poarta lui Ionel Cave is characterized by a permanent bidirectional thermal circulation and the existence of a convection cell exclusively at the level of the lower gallery. The bat monitoring showed that a nursery colony of Miniopterus schreibersii re-inhabited the cave in a very short time after the show paths were removed. The success of the ecological reconstruction was confirmed by the return of the colony next summer. Bats contribute significantly to the generation, spreading and maintaining of a rich and diversified air microflora. Also, the morphology of the cave and its ventilation system contribute to conducting and concentrating airborne microbial communities toward the upper level of the cave. In the evolution of the cave air microflora a seasonal tendency is evident, according to which a quantitative and qualitative maximum is recorded in springsummer and a minimum in autumn-winter. To conclude, our results reveal important implications for cave and bat management. Keywords: cave, ecological reconstruction, climate, bats, airborne microorganisms

1 Introduction The subterranean habitat represents an extreme environment with unique characteristics like its trophic dependence on surface ecosystems, and to the particularities of biotic and abiotic factors (Biswas 2009). The high constancy of these factors makes cave and their associated faunas one of the most vulnerable environments on Earth (Juberthie 2000). Because of the increased human pressure in the last decades we witnessed a strong degradation of the underground environment followed by the retreating or even extinction of specific fauna (Elliot 2000). Bats are particularly sensible to a persistent human disturbance in maternity sites and hibernacula (Kurta et al. 1993). Disturbances of these roosts are the major cause for the bat depopulation and may induce bats mortality and caves abandonment (Martin et al. 2000). Poarta lui Ionel Cave is an example for the bat depopulation due to improper show exploitation in the last 20 years. A recent effort was realized by the Speleological 17

Association Sfinx Gârda who tried to recover the bat colonies from this cave. The conservation measures, which intended to reduce the human pressure in the bat roost, were accomplished by a climatic study that is still in progress, as well as a mesophilic airborne microorganisms screening. The airborne microbial communities are well represented in the subterranean environment, but not all of them are resident in cave, being carried in from outside by human and animals (Borda and Borda 2006, Borda et al. 2009). Additional work has shown that bats are responsible by an increase of the airborne microorganisms in the cave atmosphere (Borda et al. 2004). Our researches focus on: (i) the analysis of the climate of Poarta lui Ionel Cave after the ecological reconstruction, (ii) determination of the airborne bat-related microorganisms, and (iii) monitoring the seasonal presence of bats in the cave. 2 Materials and methods 2.1 Climatic Data The temperature and relative humidity were measured by Tinytag Dataloggers Plus 2 (-25 + 85oC and 0 100% RH) registering. The data loggers were set to take the measurements at 1-hour intervals. In order to detect the meroclimatic structure of the cave (Racovi  1984), we established 5 sample sites (samples 1-5), located from entrance to the terminal passage of the cave (Fig. 1). The climatic study is still in progress, the records being carried out for at least one year. Therefore, our results are partial, covering the time period from 15 November 2008 to 11 June 2009. 2.2 Air microflora Samples Collecting Airborne microorganisms samples were collected by gravitational sedimentation (Koch’s sedimentation method) in two sample sites: in the visiting passage (samples I) from the lower level of cave, and in the passage not open to public access from the upper level of the cave (samples II) (Fig. 1). Investigations were performed seasonally. The specific media were exposed to the cave air for 30 min. After that, the Petri dishes were stored and transported to the laboratory at 50C, where they were incubated in specific conditions. 2.3 Culture Mediums We used sterile media for the growth of the following groups of air microorganisms: x Beef-extract agar medium - for the total count of aerobic bacteria growth (TAG); x Levine medium - for gram-negative bacteria growth (GNB); x Chapmann medium - for staphylococci growth (SPH); x Holmes medium - for streptococci growth (STP); x The Sabouraud medium – for fungi growth (FUN). The media for the aerobic mesophilic bacteria (TAG, GNB, STP, and SPH) were incubated at 37ºC for 24 hours and the media for fungi was incubated for 3-5 days at 200C, in darkness conditions. The total count of the colony-forming units was calculated using the Omelianski’s formula. The results were expressed per m3 of air (cfu/m3) (Popescu and Borda 2008). 3 Results and discussions 3.1 Cave reconstruction Poarta lui Ionel Cave located near the village Garda de Sus (Bihorului Mountains, Western Carpathians) is easily accessible, being known by natives from very old times. The cave was first mentioned by J. Vass in 1857 (Bleahu 1976) and described by Jeannel and Racovi  (1929). The entrance is represented by a portal impressive by its dimensions (20m 18

height and 15m width) open in the background of a rocky amphitheatre. Thus, the name of “Poarta” (Gate), given to this cave by natives, is very proper. Beyond this portal an active gallery opens, wide of 5-7m and long of 130m that turns twice toward left in right angle (Fig. 1). Except of some fragments on the limestone floor and a big stalagmite flow of 6m height, this sector of the cavity lacks any concretion. In its extremity, on the clayey floor is an excavation in form of a funnel with a diameter of 5m that fills up with water in periods with abundant precipitations. Usually, the subterranean stream appears only at the base of the portal, from a spring situated at the base of the left wall. At high flood the waters appear in the first turn of the gallery among the alluvium accumulated beside the right wall.

Figure 1 Map of Poarta lui Ionel Cave and the samples sites. M I - M II, Airborne Microorganisms sample sites; C 1 – C 5, Climate sample sites During an international expedition in 1988 an upper level of the cave was discovered, fact that determined the inclusion of the entrance of “Poarta lui Ionel” Cave in a more ample touristic circuit, together with the Scri oara Glacier Cave. The galleries that form the upper level of the cavity are ordered on two levels, inter-connected by a couple of not very deep wells. The terminal part is totally closed by a compact wall of limestone, in which an impenetrable fissure can be seen. The first arrangements consisted in installing the wood access stairs toward the upper level and digging an opening of about 150/40cm at the base of the stalagmite wall that blocked the access to the new discovered sector. Starting with 1992, the Speleological Association for Environmental Protection and Karst "Sfinx" from Gârda restored the arrangements, and in 2003, together with the mayoralty of Gârda de Sus village, they realised the electric illumination of the cave. Based on the evidences of chiropterit spots on the ceiling of the cave, and also on the native’s token, the Speleological Association Sfinx Gârda restored the old maternity shelter. 19

The ecological reconstruction of the cave started in April 2008 and lasted a few days. The work consisted in the elimination of the cave electrification and of the wood staircase, restricting thus the tourist’s passage to the upper level. Also, the artificial entrance to the upper level was reversibly obstructed with resident stones from the cave and walled in. 3.2 Cave climate The climatic particularities of the cave are mainly determined by the air currents that move between the exterior and the subterranean void (Racovi  1975). Their origin relies in the thermo-circulation, determined by temperature difference between the subterranean and the surface atmosphere and, implicitly, the air density difference that moves (Andrieux 1970). Because Poarta lui Ionel Cave has a single communication way with the surface, represented by the big entrance (20m/15m), the air changes with the exterior are permanently bidirectional, with moderate external perturbation. In winter time the cold air enters the cave at the floor level and the warm air from inside out at the ceiling level. In summer time the air thermocirculation is inverted. Because the lower gallery is huge and slightly ascendant, the air flow is very weak, functioning as cell convection. Besides, the effects of hibernal thermocirculation upon the subterranean atmosphere materializes by the emergence of an important number of ice formations (stalagmites, stalagmitic domes and parietal crusts), but only in the vicinity of the cave entrance. The temperature values registered in the lower level of the cave (Fig. 2) show following winter averages (15 November – 31 March): -0.3960C (St. Dev. = 2.346) at the entrance of the cave (sample 1), 1.6510C (St. Dev. = 2.215) at the basal level (sample 2), 3.0190C (St. Dev. = 1.546) in the lake gallery (sample 3). Lower Level - Sample Site 1

Humidity 100.0 % RH

20. °C 15. °C

Lower Level - Sample Site 2

Temperature

Temperature Humidity

20. °C

100.0 % RH

15. °C

80.0 % RH

80.0 % RH

10. °C

10. °C

60.0 % RH

60.0 % RH

5. °C

5. °C 40.0 % RH 0. °C 20.0 % RH

-5. °C

-10. °C 0.0 % RH Nov-08 Dec-08 Jan-09 Jan-09 Feb-09 Mar-09 Apr-09 May-09 Jun-09

40.0 % RH 0. °C 20.0 % RH

-5. °C

0.0 % RH -10. °C Nov-08 Dec-08 Jan-09 Jan-09 Feb-09 Mar-09 Apr-09 May-09 Jun-09

Lower Level- Sample Site 3 20. °C

Temperature Humidity 100.0 %RH

15. °C

80.0 %RH

10. °C 60.0 %RH 5. °C 40.0 %RH 0. °C 20.0 %RH

-5. °C

-10. °C 0.0 %RH Nov-08 Dec-08 Jan-09 Jan-09 Feb-09 Mar-09 Apr-09 May-09 Jun-09

Figure 2 The air temperature and relative humidity (RH %) trend in the lower part of the Poarta lui Ionel Cave from November 2008 to June 2009 In upper part of the cave, behind the walled gate, the air temperature is constantly higher then in the basal passages (Fig. 3). Theoretically, the upper level can function as a warm air trap and the temperature should rise to about 20°C (i.e. Wonder’s Room from Huda lui Papar Cave, Trascu Mountains). Before digging the artificial opening, the access to the upper level 20

was made only through the natural window that has a lower section. Therefore, the existence of considerable temperature differences between the two levels can be excluded. Besides, the natural window that opens close to the ceiling remains inaccessible to any cold air current that move at the level of the floor. As a result, the upper gallery shows a more constant temperature slightly higher than the lower level, 7.6850C (St. Dev. = 0.832) behind of the walled gate (sample 4), and 7.4540C (St. Dev. = 0,123) at the end of the cave (sample 5). Due to the huge entrance that is the subject to high temperature fluctuations dependent to the external climate and in accordance with the climatic data, the basal level of the cave is represented by a perturbation meroclimate. The relative humidity ranged 67% near the entrance to constantly 100%, above the temporary lake (Fig. 1). Upper Level - Sample Site 4

Temperature

Upper Level - Sample Site 5

Humidity 20. °C 15. °C

Temperature Humidity

100.0 % RH

20. °C

100.0 % RH

80.0 % RH

15. °C

80.0 % RH

10. °C 60.0 % RH

10. °C 60.0 % RH

5. °C

5. °C 40.0 % RH

40.0 % RH

0. °C -5. °C

0. °C 20.0 % RH

0.0 % RH -10. °C Nov-08 Dec-08 Jan-09 Jan-09 Feb-09 Mar-09 Apr-09 May-09 Jun-09

-5. °C -10. °C

20.0 % RH

0.0 % RH

Nov-08 Dec-08 Jan-09 Jan-09 Feb-09 Mar-09 Apr-09 May-09 Jun-09

Figure 3 The air temperature and relative humidity (RH%) trend in the upper part of the Poarta lui Ionel Cave from November 2008 to June 2009 The upper level is represented by a stability meroclimate, where the temperature is constant all year long and the relative humidity was also constantly saturated (Fig.2). 3.2 Bats Although Jeannel and Racovi  (1927) did not cite the bat colony occurrence when they visited the cave, the spots of chiropterit on the ceiling and parietal leaking, as well as the testimony of natives prove that “Poarta lui Ionel” Cave sheltered in the past significant bat nurseries. In the period of tourist management during our research we did not record any bat colonies. We recorded a few specimens of Myotis spp., Rhinolophus spp, Barbastella barbastellus, Plecotus spp. only in the lower level, at the beginning of the winter. After the pressure induced by tourists was eliminated and the cave was brought back to the natural state, the bat colonies immediately installed in the old summer roost. In May 2008 a nursery colony of Miniopterus schreibersii of more than 150 individuals was found in the upper part of the lower level of the cave (about 20 m high). The bats re-inhabited the cave in a few weeks after the ecological reconstruction of the cave was implemented. The success of the ecological cave reconstruction was proved by the maternity colony of M. schreibersii that returned to the cave in May 2009. Pursuant to the climatic traits, the lower level of the cave is not suitable for bats hibernation. But in the upper level the climate is more constant, with higher temperatures. Therefore, Rhinolophidae prefer that part of the cave for the hibernation period. During the winter 2008-2009 in this part of the cave were recorded Rhinolophus ferrumequinum (13 ind.), R. hipposideros (1 ind.), R. euryale (1 ind.), Myotis myotis/M. oxygnathus (2 ind.), Myotis spp. (1 ind), Miniopterus schreibersi (5 ind). All species encountered in the cave are strictly protected (Annex II, 13/1993 Law), migratory species (Annex II, 13/1998 Law) and also species of European Community interest whose exploitation may be subject to management measures (Annex 3, 57/2007 OU). 21

We consider that the main perturbation factor that contributed to bat colonies extinction in the Poarta lui Ionel Cave was represented by human disturbance in the vulnerable periods of their biological cycle. To this, we can add the system of electric illumination of the cave and the gate with vertical bars, placed at the artificially digging opening from the base of the upper level. 3.3 Airborne microorganisms The results concerning the airborne microorganisms diversity and concentrations in Poarta lui Ionel Cave are showed in Fig. 4. Total aerobic germs showed the highest values, corresponding with the beginning of the bats maternity season. The morphology of the cave imprints a stability climate at this level, with higher temperatures than in the lower level, favourable to the development of a mesophilic air microflora. Besides, the air circulation favours the airborne microorganisms to remain captive in the upper level. The presence of bats and guano, corroborated with morphological particularities of the cave (Borda et al. 2004) explains the high incidence of the five groups of microorganisms with hygienic significance and the high number of colony forming units in the upper level compared with the lower level. The presence of fungi is less significant, compared with that of other microorganisms, their occurrence being in an increased number in the lower level, close to the entrance. Usually the fungi originate from the exterior (Caumartin 1966, Koilraj et al. 1999), their number decreasing from the entrance to the profound areas of the cave (Borda and Borda 2006). TAG STP SPH GRN FUN

Airborne Microorganisms in the Upper Level Autumn Winter Spring Summer 0

1000

2000

3000

4000

5000

6000

7000

UFC/m3 TAG

Airborne Microorganisms in the Lower Level

STP SPH

Autumn

GRN FUN

Winter Spring Summer 0

1000

2000

3000

4000

5000

6000

7000

UFC/m3

Figure 4 Seasonal variations of airborne microorganisms in Poarta lui Ionel Cave

22

The seasonal tendency of air microflora indicates a maximum of the mesophilic bacteria in the warm season, during the summer for the lower level and during the spring for the upper level. The absence of fungi in the spring can be explained by their biological cycle. All these data are important for the future cave management plans that should take into account the needs of the bats, the effects of human disturbance upon the bat populations and also the airborne microorganism biohazards. 3 Conclusions After the ecological reconstruction, the Poarta lui Ionel Cave was re-populated by a maternity colony of Miniopterus schreibersii. The climatic particularities of the cave support the bats colonies. This is characterised by a permanent bidirectional thermo-circulation and by the existence of a cell convection that closes at the extremity of the lower gallery, without affecting the upper level of the cave. Bats significantly contribute to the spreading and maintaining of a rich and diversified air microflora. The morphology and ventilation of the cave contributes to the directing and concentrating the airborne microbial communities toward the upper level. A seasonal tendency of airborne microorganisms is evident, being assured by the temperature of the external environment and by the presence of tourists. Therefore, we consider these investigations relevant for cave conservation and for the protection of various types of bat colonies that inhabit these caves, even in the conditions of tourist exploitation. Acknowledgments We thank the staff of the Apuseni Natural Park Administration, who helped us and contributed with logistics in the field. We are particularly thankful to Negrea Avram for field assistance. This study was supported by The ID_2325 Grant from CNCSIS. References Andrieux C (1970) Contribution a l’étude du climat des cavités naturelles des massifs karstiques. II. –Aérodiynamique souterraine. Ann. Spéléol. XXV(2):491-529 Bleahu M, Decu V, Negrea S, Ple a C, Povar I, Viehmann I (1976) Caves from Romania, Ed. tiin . Enc. 415 pp Borda D, Borda C, Tma T (2004) Bats, climat, and air microorganisms in a Romanian Cave. Mammalia 68(4):337-343 Borda C and Borda D (2006) Airborne microorganisms in show caves from Romania. Trav. Inst. Spéol. “Émile Racovitza” 43-44: 65-74 Borda D, Bucur-Nstase R, Borda C, Gorban I (2009) The assessment of the airborne microorganismes in subterranean environment, Bulletin UASVM, Veterinary Medicine 66 (1) (in press) Caumartin V (1966) Principes de repartition des associations d’organismes microscopiques en caverns. Bull. Sci. Bourgone 24:39-56 Elliot W R (2000) Conservation of the North American cave and karst biota. In: Wilkens H, Culver D C, Humphreys W (Eds.) Ecosystems of the World - Subterranean Ecosystems, Elsevier. pp 665-689 Jayant B (2009) The biodiversity of Krem Mawkhyrdop of Meghalaya, India, on the verge of extinction, Current Science 96(7):904-910 Jeannel R, Racovi  E G (1927) Enumération des Grottes visitées, 1918-1927 (VII-e série). Archives de Zoologie Expérimentale et Générale 68(2):293-608 Juberthie C (2000) Conservation of subterranean habitats and species. In: Wilkens H, Culver D C, Humphreys W (Eds.) Ecosystems of the World - Subterranean Ecosystems, Elsevier, pp 691-700 23

Koilraj A J, Marimuthu G, Natarajan K, Saravanan S, Maran P., Hsu M J (1999) Fungal diversity inside caves of Southern India. Current Science – Bangalore 77(8):1081-1083 Kurta A, King D, Teramino J A, Stribley, J M, Williams K J (1993) Summer roosts of the endangered Indiana bat (Myotis sodalis) on the northern edge of its range. Am Mid Nat, 129:132-138 Martin K V, Puckette W L, Hensley S L, Leslie D M (2000) Internal Cave Gating as a Means of Protecting Cave-Dwelling Bat Populations in Eastern Oklahoma. Proc. Okla. Acad. Sci. 80:133-137 Popescu S, Borda C (2008) Igiena Animalelor i Protec ia Mediului. Lucrri Practice. Editura Napoca Star Cluj-Napoca, 167 pp Racovi  G (1975) La classification topoclimatique des cavités souterraines. Trav. Inst. Spéol. „E. Racovitza” 14:197-216. Racovi  G (1984) Sur la structure méroclimatique des cavités souterraines. Theor. Appl. Karstol. 1:123-130

24

Karst and sustainability in Florida, U.S.A. Robert BRINKMANN Karst Research Group, Department of Geography, University of South Florida, Tampa, FL 33620, USA, e-mail: [email protected]

Abstract: The State of Florida, which consists of one of the most extensive karst regions in the Americas, is also one of the fastest growing regions in the United Starts. The population increases, which are driven by both migrations into Florida from within the United States and from many areas of the world, particularly Latin America, put significant pressure on the karst systems within Florida’s fragile subtropical environment. Within this cultural and environmental context, various stakeholders are attempting to address a variety of sustainability issues unique to Florida’s subtropical environment. It is evident that karst systems impact in some way key sustainability sectors such as water, food and agriculture, building, energy, and greenhouse gas management. Florida’s karst waters are continually under thread due to over pumping and pollution. Improvements in the systematic management of water eliminated some of these problems. Agricultural production in the state often impacts the state’s karst systems through water extraction, irrigation, and associated fertilizer pollution of the aquifer. Local limestone and marine sand is used in the construction of concrete block building materials. Offshore oil and gas reserves, often found in limestone, are not utilized and the state relies on external sources for energy. Nevertheless, there is a growing interest in developing local policies to address greenhouse gas emissions, some of which involve carbon sequestration within karst systems. A policy review of these five themes reveals that Florida provides strong examples how sustainability can be thematically approached within a subtropical karst environment. The United States Government, until recently, has not provided guidance on a variety of sustainability issues. But at the local level, a number of government and non-government organizations are addressing these important topics. The development of local approaches to enhancing environmental sustainability in karst environments requires examination of regional environmental settings and how human activity impacts them. Keywords: karst, sustainability, Florida

1 Introduction The state of Florida has undergone tremendous population growth in the last 100 years as its population doubles approximately every 20-30 years (U.S. Census 2009). However, it is also one of the most fragile karst systems on the planet with regional interconnected ground water systems, dozens of high-flow springs, hundreds of caves, and unique sinkholes and wetlands (Brinkmann et al. 2007, Brinkmann and Reeder 1994, Fleury, Carson, and Brinkmann 2008, Florea 2006, Scott et al 2004, Screaton et al. 2004, Tiahnsky 1999). The high growth coupled with the natural vulnerability of karst landscapes (North, van Beynen, and Parise 2009), provides challenges for environmental managers. In the last several years, indices that measure community-scale sustainability were developed to benchmark sustainability efforts at the county and local levels in Florida (Florida Green Building Coalition 2009, Myfloridaclimate 2009, Upadhyay 2009). These matrices evolved, in part, due to the complex political landscape of the state that encourages economic development while preserving natural resources. Within this context, the Governor of Florida, Charlie Crist, challenged local governments and state agencies to develop strategies to reduce greenhouse gases and improve the overall environmental sustainability of their organizations 25

through Executive Orders 126, 127, and 128 (Crist 2007a-c). Due to this challenge, dozens of communities adopted specific policies to enhance sustainability in their operations. While many of these approaches are still evolving, several strategies directly impact karst systems. The new policies can be grouped into strategies for water, food, building, energy, and greenhouse gas management. Prior to discussing each of these in detail, it is worth reviewing the nature of the karst systems in Florida. To many, Florida is a flat, featureless plain. Throughout its approximate 150,000 square kilometers in area, the maximum elevation is 105 meters and the local relief in many portions of the state is difficult to discern. However, upon closer examination, Florida is home to a diverse karst landscape. Its carbonate platform underwent a series of marine transgressive cycles that modified existing carbonate and created new carbonate and sandy marine sediments and rocks. As a result, the state contains older carbonate ridges that were exposed for longer periods than the surrounding lowlands. In addition, in many areas, marine sands blanket the limestone, thereby creating a covered karst landscape. The landforms in the lowlands are what one would expect to see in recently emerged karst plains. Circular dolines dominate the landscape and streams are uncommon. Most of the drainage is to the subsurface were hundreds of kilometers of flooded underground caves are found. In addition, springs are commonly found in these lowland areas. Many form spring runs that lead directly to the coast. In contrast, the karst landscapes on the older ridgelands are more complex. Here, karst depressions are extensive and their forms are complicated. These uvalas are often sites of airfilled cave entrances. The karst landscape is continually forming in Florida. It is evident through the composition of spring water exiting karst aquifers that solution of limestone is occurring at a rapid rate (Scott et al. 2004). There are hundreds of homeowners’ insurance claims in the state each year due to sinkhole damage (Eastman et al. 1995). In addition, human activity, such as over pumping of aquifers, enhances depression formation. The karst plain in Florida, due to its active nature is quite vulnerable because of the 18 million residents that live on top of it. Nevertheless, there are some interesting approaches that have been taken in recent years that improve the overall environmental sustainability in Florida’s karst systems 2 Approaches to environmental sustainability The approaches to environmental sustainability will be discussed within five major themes: water, food and agriculture, building, energy, and greenhouse gas management. Certainly there are others that could be explored, such as population sustainability and disaster resilience. Nevertheless, karst systems have the greatest impact within these five sustainability themes in Florida. 2.1 Water Florida experiences high variation in monthly precipitation. During the summer and early fall, rainfall is quite high due to sea-breeze induced convectional thunderstorms, and occasional tropical storms and hurricanes. Precipitation can exceed several centimeters daily and depressions are filled with storm water runoff via overland flow. Ephemeral rivers begin to flow and the discharge in perennial streams and springs increases. In dry months, many springs and streams either dry out or decrease their discharge significantly. Karst wetlands, lakes, and pond may dry. Within this environment, water managers must provide drinking water to millions of people. Unfortunately, the production of water in Florida in the late 20th century caused a reduction of the regional groundwater table and concomitant drying of wetlands, lakes, and rivers. Collapses of the land surface into underground voids increased. This occurred largely due to the decision to manage water in the state within local water management districts. Thus, a region like Miami must find water within their local region and 26

cannot import it from wetter areas of the state. They must use whatever sources they can within their region. While all efforts are made to protect the environment, withdrawals occur during extreme drought periods in late winter and early spring. Thus, there have been significant impacts to the karst systems in Miami, Tampa, and Orlando due to ground water withdrawal (Rand 2003). However, this sustainability tenet of “using local” water supplies has let to innovative approaches to water management. For example, several governments in the Tampa Bay area formed Tampa Bay Water, an organization charged with providing drinking water to the region. Because the region cannot withdraw more from the aquifer without further damaging the karst systems, Tampa Bay Water developed a 25 million gallons of water per day desalination facility and built a 15 billion gallon reservoir that supplements surface and groundwater sources. While one may question the carbon footprint of the desalination plant and reservoir within a sustainability context, there is no doubt that these innovative projects prevented further damage to the region’s karst environment as the region’s water demand increased. 2.2 Food and agriculture Florida is one of the most productive agricultural states in the U.S (USDA 2008). It is a major beef and dairy producer, although it is probably best known for its citrus groves, strawberry fields, and fresh produce farms. It is also home to niche agricultural markets such as caladium bulbs, orchids, and tropical fish production. Farms and food processing use the greatest amount of water than any sector in Florida (USGS 2009). Water withdrawals in some agricultural areas trigger sinkhole collapses and local well failure (Tehansky 1999). For example, in 1997, strawberry farmers induced sinkholes in rural Polk County Florida when they sprayed millions of liters of water on their crops to prevent them from freezing. Of special concern in the state is nitrogen and other fertilizers and pesticides that enter the groundwater system and rapidly disperse within the interconnected karst aquifers. Nitrogen pollution has steadily increased in springs in the state (Katz 2004, Scott et al. 2004) as a result of increased fertilizer use not only on agricultural lands but also on lawns and golf courses that are ubiquitous on the Florida landscape. New approaches to organic farming, community sponsored agriculture, and community gardening can reduce the impact in some settings. In the last few years, there has been a rapid increase in these efforts in the state. In the Tampa Bay area, for example, several hydroponics and soil-based community gardens started, several organic farms began operation on the edge of the urbanized area, and there is great interest in local governments in encouraging community sponsored agriculture. In addition, many golf courses are using “green” golf course management protocols that reduce fertilizer runoff and officials are encouraging replacement of lawns with native vegetation and trees. However, Florida’s unique growing season makes it a significant fresh food source for many parts of the world and the impacts of agribusiness cannot be discounted. These large food-producing organizations are also trying to do their part to enhance the region’s sustainability. For example, many farms are developing drip irrigation or are transforming fields to hydroponic operations. Others are using hi-tech irrigation schemes that reduce their overall groundwater use. In addition, many agricultural fields take the solid wastes from sewage treatment plants in urban areas, thereby reducing their need for fertilizers. There is also great interest in Florida in developing biodiesel fuels from agricultural crops that need little to no fertilization. The state has invested in research into a variety of crops and there are some biodiesel plants currently in operation. Associated with this is the emerging area of using vegetation in Florida for carbon capture. Large landholders are examining the profitability of transforming pasture or croplands into forested land within the carbon trading market. With the new administration within the executive branch of the U.S. government and 27

with the associated congressional movement on climate change, it is likely Florida will be a high-interest area in the carbon market due to its long growing season. While many have the image of Florida as a highly urbanized region, there are large areas of the state with low population densities where vast agricultural regions are found. 2.3 Building One of the key tenets of modern urban sustainability is green building. In Florida, there are special challenges compared to the rest of the United States due to its unique subtropical climate and maritime setting. While much of U.S. green building policy focuses on heating, in Florida, a larger focus is on cooling and water storage. In fact, green building on the karst landscape in Florida provides special challenges for land stability, cooling, and deriving local building materials. Because the setting is so unique, an organization evolved, the Florida Green Building Coalition (FGBC), which certifies green building and design using matrices developed for Florida. The certification process is similar to the well-known LEED (Leadership in Energy and Environmental Design) process, but focuses on the particularities of Florida (Florida Green Building Coalition 2009). Commercial buildings, developments, high rises, homes and local governments can be certified. For homes, the FGBC, in similar fashion to LEED, will designate a home as a “certified green” home at the Bronze, Silver, Gold, or Platinum levels depending upon how many points one earned on a matrix. The points can be earned in eight categories: energy, water, lot choice, health, materials, disaster mitigation, and other. There are many ways to earn points within each category. However, within the matrix, there is a strong emphasis on the use of technology and design in reducing energy and water use. A significant aspect of the matrix is the focus on local conditions, particularly the high rainfall and coolant needs of homes. Within a karst context, this is significant as the karst environment is highly reactive to surface conditions. For example, if a home is using nonnative vegetation that requires fertilization and extensive irrigation, it is likely that the fertilizer will drain into the aquifer at the same time the aquifer is being depleted. Therefore, points can be gained in the FGBC system for using native vegetation in landscaping and for reducing or eliminating the need for lawn irrigation. There are some particular areas of green building and design that impact karst systems more directly. Some certified green buildings in the state use air-cooled in subsurface karst voids for cooling interiors. This geothermal cooling approach has not seen widespread application to date, but there is high potential for reducing energy use in cooling Florida homes through this innovative technology. Another important aspect associated with sustainable building is the use of local products in construction. In Florida, most of the homes and buildings are constructed out of concrete blocks on a poured cement foundation. In most cases, the cement and concrete are derived from local sources. Local limestone is crushed and processed and mixed with the local marine sands to produce these products. However, the limestone itself is rarely used as a building material. It is an extremely soft, vuggy stone that is not practicable for building except as a decorative element. Instead, the local limestone is sometimes used as decorative landscape features. 2.4 Energy Unlike many karstic areas of the world, Florida is not exploiting its known oil and gas reserves. Due to the importance of tourism to the state’s economy, the majority of the population in Florida does not want oil and gas producing rigs and platforms within view of its popular beaches. Instead, Florida relies on imported energy sources for the vast majority of its power. While Floridians spend less than any other state on energy (USDE, 2009), the state is a one of the country’s top three energy users. The reasons that energy costs are so low 28

in the state are because of the proximity to major energy sources, importance of regional pipelines, and the extensive energy port and distribution system in major urban areas. Most of the energy sources in the state are petroleum, natural gas, and coal (see Figure 1).

Figure 1 Energy sources in Florida (USDOE 2009) However, the state is trying to develop nuclear, wind, solar, and biofuel energy sources to try to reduce the need for fossil fuels and to meet ambitious goals set forward by Florida’s Governor, Charlie Crist. He has challenged the state to reduce greenhouse gas emissions to year 2000 levels (Crist 2007a-c). Regardless, there has been tremendous pressure from a variety of stakeholders to try to develop more locally-derived petroleum-based energy sources in Florida through offshore drilling (Manheim 2004). However, many state leaders resist this effort due to the concerns over the impact to tourism in the state. With Venezuelan and Gulf of Mexico petroleum products cheap and accessible through one of the states several petroleum ports, it seems unlikely that offshore drilling will commence any time soon. While petroleum exploitation would enhance the understanding of Florida’s karst systems, it is more likely that greater emphasis will be placed on conservation measures such as fuel economy and mass transit before these reserves will be tapped. It must be noted that coal use in the area has received a great deal of attention in recent years due to power plant emissions and mountain top mining. Much of the coal that is used in the region comes to Florida from the Appalachian Mountain landscape where mountain top removal coal mining is sometimes employed. Mountain top removal is a process where an entire mountain is removed from the landscape in the mining process. Coal is separated from other rocks. The tailings remain where the mountain once stood. Environmental activists are critical of this mining technique because it destroys ecosystems, viewsheds, and any associated Appalachian karst systems in the process. Florida’s insistence in not developing its own energy resources therefore deleteriously impacts distant karst systems. 2.5 Greenhouse gas management In recent years, Floridians have become more and more concerned with sea level rise due to global warming. Many scientists forecast sea level increases in the coming decades that would decimate beaches, some urban areas, and many low-lying areas including the Florida Everglades. The United States Government, until recently, did not aggressively approach this problem, leaving states and local governments to figure out how to try to protect 29

their lands. Unfortunately, this created a somewhat scattered approach with states like California adopting strong regulations to reduce greenhouse gases while others remain uncommitted to adopting regulatory guidelines. Fortunately, there are several organizations helping state and local governments develop guidelines and benchmarks. The U.S. Conference of Mayors, for example, developed a “Climate Commitment” that emphasizes meeting or beating the Kyoto Protocols within their own city (U.S. Conference of Mayors 2009). To date, over 900 mayors have signed the agreement. There are similar agreements in place for American universities (American College and University Presidents Climate Commitment 2009) and other private and public organizations. While the Federal government has been relatively absent in policy development at the national level, the gap in leadership has been filled at the grassroots. One key aspect of local sustainability approaches is greenhouse gas inventories and management. There are two aspects to conducting a greenhouse gas inventory for any organization: measuring greenhouse gas outputs and greenhouse gas credits that mitigate the output. The outputs are in the form of energy used in transportation and electricity production, as well as gases emitted through cooling, agricultural practices, and waste. Greenhouse gas credits vary considerably. They could be in the form of alternative energy production, development of biostorage of carbon through significant landscape conversion, or through other carbon storage practices. In karst landscapes, there is the potential for mitigating the greenhouse gases emitted through carbon storage in karst voids and through enhancing the formation of carbonate deposits. There is ongoing research on storing carbon in deep aquifers in the state. The idea is to take carbon dioxide and pump it deep within the earth into karst voids or other host rocks. Another approach is to enhance carbon storage in forest or other vegetative reserves and by constructing offshore artificial reef systems. However, the coastal zone in many urbanized areas in Florida is polluted (Fink and Charlier 2003), thereby limiting the capacity of offshore ecosystems to produce carbonate reef materials. In addition, the enhanced carbon dioxide in the atmosphere has caused ocean acidification in some coastal areas of the United States thereby inhibiting the growth of some reef-building animals (Feely et al. 2008). It is unclear how effective the ocean’s ecosystems will be at sequestering carbon as the acidification process advances. 3 Conclusions In summary, Florida’s sustainability policies provide examples of how intricate social and environmental interactions can be modified to enhance environmental sustainability. The karst system is part of the global carbon cycle and it is important to evaluate its role in a variety of emerging sustainability practices. It is evident that Florida’s approaches to water, food, building, and greenhouse gas management can be employed in other karst settings. The United States may be a unique test case for examining sustainability in that most measurable progress was initiated at the local and state level and not within the national government. This bottom-up approach may have assisted the expansion of sustainability efforts since the U.S. Federal Government has been known to develop standard rules that make it difficult to address local complex problems. Therefore, the sustainability efforts in Florida firmly include karst systems in many of the policies, plans, and activities. In this way, Florida may serve as a testing ground for developing and assessing sustainable management on karst terrains, particularly in the areas of water, food and agriculture, and greenhouse gas management. Florida does not provide particularly helpful guidance on energy. While it is admirable that residents have resisted local oil exploration, Florida continues to rely on external sources for its energy and has not developed significant alternatives to traditional energy production to date.

30

References American College and University Presidents Climate Commitment (2009) American college and university presidents climate commitment. July, 2009. http://www.presidentsclimatecommitment.org/html/commitment.php Brinkmann R, Reeder P (1994) The influence of sea-level change and geologic structure on cave development in west-central Florida. Physical Geography 15:52-61 Brinkmann R, Wilson K, Elko N, Seale LD, Florea L, Vacher HL (2007) Sinkhole distribution based on pre-development mapping in urbanized Pinellas County, Florida. In: Parise M and Gunn J (eds) Natural and anthropologic hazards in karst areas: Recognition, Analysis, Mitigation, Geological Society of London Special Publications 279:5-11 Crist C (2007a) State of Florida Office of the Governor Executive Order Number 07126. http://www.flgov.com/pdfs/orders/07-126-actions.pdf Crist C (2007b) State of Florida Office of the Governor Executive Order Number 07127. http://www.flgov.com/pdfs/orders/07-127-emissions.pdf Crist C (2007c) State of Florida Office of the Governor Executive Order Number 07128. http://www.flgov.com/pdfs/orders/07-128-actionteam.pdf Eastman KL, Butler AM, Lilly III CC (1995) The effect of mandating sinkhole coverage in Florida homeowners insurance policies. CPCU Journal 9:165-176 Feely RA, Sabine CL, Hernandez-Ayon, JM, Ianson D, Hales B (2008) Evidence for upwelling corrosive “acidified” water onto the continental shelf. Science 320(5882):14901492 Fink CW, Charlier RH (2003) Sustainability of subtropical coastal zones in southeaster Florida: Challenges for urbanized coastal environments threatened by development, pollution, water supply, and storm hazards. Journal of Coastal Research 19(4):934-943 Fleury S, Carson S, Brinkmann R (2008) Testing reporting bias in the Florida sinkhole database: An analysis of sinkhole occurrences in the Tampa Metropolitan Statistical Area. Southeastern Geographer 48(1):38-52 Florea LJ (2006) The Karst of West-Central Florida. PhD. Diss., Department of Geology, University of South Florida. Florida Green Building Coalition (2009) Standards. July, 2009. http://www.floridagreenbuilding.org/db/?q=node/5357 Katz BG (2004) Sources of nitrate contamination and age of water in large karstic springs of Florida. Environmental Geology 46:689-706 Manheim FT (2004) US offshore oil industry: New perspectives on an old conflict. Geotimes 2004:26 Myfloridaclimate (2009) Executive Orders and Partnership Agreements. July, 2009. http://www.myfloridaclimate.com/2007_climate_summit/executive_orders_partnership_agree ments North LA, van Beynen, PE, Parise M (2009) Interregional comparison of karst disturbance: West-central Florida and southeast Italy. Journal of Environmental Management 90(5):1770-1781 Rand H (2003) Water Wars: A story of people, politics, and power. Exlibris, 282 p. Scott TM, Means GH, Meegan RP, Means RC, Upchurch SB, Copeland RE, Jones J, Roberts T, Willet A (2004) Springs of Florida. Florida Geological Survey Bulletin 66. Florida Geological Survey, Tallahassee, Florida, 377 pp Screaton E, Martin JB, Ginn B, Smith L (2004) Conduit properties and karstification in the unconfined Floridan Aquifer. Ground Water 42:338-346 Tihansky AB (1999) Sinkholes, West-Central Florida. In: Galloway D, Jones DR, and Ingebritsen, SE Land Subsidence in the United States. Reston, Virginia: USGS

31

Upadhyay, NS (2009) Green local governments in Florida: An analysis of sustainability and green building policies. Unpublished master’s thesis. Department of Geography. University of South Florida. U.S. Census (2008) U.S. Census Data July, 2009 http://www.census.gov U.S. Conference of Mayors (2009) Mayors leading the way on climate protection. July 2009. http://www.usmayors.org/climateprotection/revised/ USDA (2008) 2008 State agricultural overview. July 2009. http://www.nass.usda.gov/Statistics_by_State/Ag_Overview/AgOverview_FL.pdf USDE (2009) Florida. July 2009. http://www.energy.gov/florida.htm USGS (2009) Historical water use in Florida. July 2009. http://fl.water.usgs.gov/WaterUse/hwu_FL.htm

32

Water management in the karst of Apulia, southern Italy Marco DELLE ROSE, Mario PARISE National Research Council, IRPI, Bari, via Amendola 122-I, Bari, Italy, e-mails: [email protected]; [email protected]

Abstract: Among the peculiarities of karst environment, distinguishing it from any other natural settings, the very limited surface runoff and the slightly defined surface watersheds play a significant role. Notwithstanding such features, even in flat karst areas as is the case for most of Apulia (south-east Italy), the surface hydrographic lines were a very important element in the karst landscape, that greatly controlled location and spreading of the first human settlements in the region. In the centuries, the many interventions carried out by man have caused heavy changes in the original hydrographic network: swallets have been covered and/or clogged, water lines diverted, and a complex network of artificial channels progressively took the place of the original surface runoff. The artificial channels are still today used to discharge the urban and industrial wastewaters in many areas of the region. All these situations on the occasion of extreme rainfall determine floods, extending over wide areas, and, as indirect consequence, spreading of pollutants in the fields. Water management in the karst environment of Apulia is discussed in this paper. Two examples are used to describe both history of the anthropogenic actions, and the main effects they caused: Castellana-Grotte, in the Murge plateau, and Nardò, in the Salento Peninsula. Keywords: flood, water management, karst

1 Introduction In karst environments water rapidly infiltrates within the epikarst through the network of fissures and conduits, and the swallow holes. This feature, combined with the slightly defined surface watersheds, distinguishes clearly karst from any other natural setting. Nevertheless, even in flat karst areas as is the case for most of Apulia (south-east Italy), the surface hydrographic lines were a very important element in the landscape, and greatly controlled location and spreading of the first human settlements in the region (Lopez et al. 2009, Parise 2009). In the centuries, the many interventions carried out by man caused heavy changes in the original hydrographic network: swallets were covered and/or clogged, water lines diverted, and a complex network of artificial channels took progressively the place of the original surface runoff. The artificial channels are still today used to discharge the urban and industrial wastewaters in many areas of the region. Due to geological and morphological settings, Apulia is frequently affected by flooding (Carrozzo et al. 2003, Parise and Pascali 2003). Although generally non destructive in character, these events cause serious economic losses. During the last centuries, flood events have been prevented or mitigated by a methodical maintenance of karst sinkholes and drainage canals. The last decades, however, have been characterized by several anthropogenic modifications, often realized illegally, that resulted in the partial or total destruction of many natural karst features, of great importance for the hydrologic regime. These changes are at the origin, on the occasion of heavy to extreme rainfalls, of a two-fold hazard: floods, extending over wide areas; and, as indirect consequence of the flood events, spreading of pollutants into the fields and underground. Once the artificial channels are not able to transport the unusual amount of water, and the swallow holes become clogged, in fact, the consequence is flooding over wide areas, locally involving also the built-up environment, especially when the urban 33

areas are located at the lowest part of endorheic basins, a situation which is quite common in Apulia. In the next years it has been estimated that, due to global warming, the soil erosion will greatly increase in semi-arid regions as Apulia, whereas the extreme meteoric events are predicted to be more frequent and strong (IPCC 2007). As regards management of Apulian endorheic catchments, this should result in higher risks of sinkhole clogging and flood problems (Delle Rose 2007). Water management in the karst environment of Apulia is discussed in this paper. Two examples are used to describe both history of the anthropogenic actions, and the main effects they caused: Castellana-Grotte, in the Murge plateau, and Nardò, in the Salento peninsula. 2 Geological setting Apulia (Fig. 1) approximately coincides with a block of the Adria Microplate built up by Mesozoic platform carbonates overlain by Tertiary-Quaternary bioclastic deposits. Salento forms the structurally less elevated sector and is bounded to the N by the Murge plateau along an E-W tectonical deformation strip. Horst and graben structures characterize both Murge and Salento, as a result of mainly distensive stresses that dislocated the carbonate bedrock starting from the Upper Cretaceous (Ciaranfi et al. 1988). The Mesozoic platform carbonates had been intensively shaped by sub-aerial conditions during large part of the Paleogene. Salento hosts continental, transitional and marine Oligocene to Miocene deposits (Fig. 1). These calcareous terrains form a second order sequence which lies on the Mesozoic carbonates, locally with the interposition of continental and residual deposits.

Figure 1 Geologic sketch of Apulia and location of case studies. Explanation: 1) Mesozoic platform carbonates; 2) Oligocene-Miocene deposits; 3) Pliocene and Quaternary deposits During the Messinian Salinity Crisis, Apulia formed a chain subject to intensive dismantling and karst processes. The subsequent Lower Pliocene carbonates locally passes into Middle Pliocene marlstones that are overlain by Upper Pliocene-Lower Pleistocene calcarenite deposits. The latter are bounded by a disconformity from a glauconitic fossiliferous deposit that is overlain by clayey marls, capped in turn by sandy deposits formed until the Middle Pleistocene. The whole marine Cretaceous-Middle Pleistocene Murge and Salento succession is overlain by a series of discontinuous terraces, formed because of the interaction between regional tectonic raising and the glacio-eustatic sea level changes. Karst processes have produced a dense network of cavities and conduits which characterize large portions of Apulia. Many karst caves and sinkholes act to collect and 34

transport underground the surface waters; these doline-type landforms have different names, according to the different parts of the region: pulo, gurgo, vora, àviso, etc. (Parise et al. 2003). 3 Case study no. 1 – Castellana-Grotte The oldest part of Castellana-Grotte lies at the bottom of a karst valley (Fig. 2), which main morphological features are represented by flat bottom valleys filled with alluvial and residual deposits.

Figure 2 The karst valley of Castellana-Grotte (after Parise 2003). Explanation: 1) watershed divide; 2) morphological saddle; 3) water lines; 4) lame (karst valleys); 5) urban area Low permeability of infilling materials determines high surface runoff on the occasion of intense rainstorms. Location of the historical part of town in the lowest sector of the valley (indicated in ancient maps with the name of Lago, meaning lake; Colamonico 1917), and progressive clogging and/or closure of the main swallow holes there present, due to urban expansion, are at the origin of the several flooding events that repeatedly hit the town (Orofino 1990). The most tragic was that occurred on November 9, 1896, when severe damage were recorded, including 4 casualties, loss of livestock and 600 houses uninhabitable. The 1896 flood is the best-documented event in the historical series of floods at CastellanaGrotte, since it was described in the work by a local attorney, Antonio Sgobba, published few weeks after the disaster (Sgobba 1896). According to this testimony, the rainfall started in the morning, after a few days of antecedent rainfall. From 19:00 LT the main event began: there was a strong increase in the intensity of the rain and, after four hours of continuous rainfall, the lowest part of town was completely inundated by a huge lake that reached about one third of the entire town, and a maximum height of 5.4 m. 35

A historical analysis, performed through scrutiny of several data sources (state-wide archives, newspaper clips, historical documents, local works, scientific publications) resulted in some 23 events of hydrological disasters at Castellana-Grotte since the XIII century (Parise 2003). The highest number of events was registered in a 30-year time span, from 1874 to 1905, at the turn of the XX century. After this impressive series of floods, many hydraulic engineering works were designed to facilitate the infiltration of water underground in the lowest sector of the town. Unfortunately, lack of findings from the local administration resulted in delaying the works, until a new project was designed at the expense of the Italian State. Again, however, there was a delay in the realization, due to the occurrence of a major natural catastrophe, the 1909 earthquake in the Messina Strait that took for some years all the funds from the Italian Government. Eventually, in 1911 the work began, to be completed two years later (Viterbo 1913). The system consisted of an artificial tunnel, dug at depth around 10 m, connecting natural shafts located at the extremities of the lowest sector of town. After its realization, only minor floods, causing very limited damage to the built-up environment, were recorded. 4 Case study no. 2 – Nardò The Asso catchment basin can be defined as a graben-polje (Gams 1978) and is part of a complex hydrological system located within a wide tectonic depression, enclosing a network of natural and artificial channels directed toward swallow holes, the most important of which is Vora Colucce (Fig. 3). It covers ten municipalities within the Lecce province, and an overall area greater than 165 km2 (Reina 1972). Different stratigraphic units crop out in the catchment: clayey marls, locally overlain by sands and silts; Plio-Quaternary calcarenites; Cretaceous limestones. In the Vora Colucce the stratigraphic passage between these latter and the above calcarenites has been individuated. This contact represented in late Quaternary the karstic base level that controlled the speleogenesis of the karst system that, in the mature stage, resulted in the formation of the Vora Colucce collapse sinkhole. The Asso catchment has very low gradients, whilst the hydrographic network shows tendency to silting of the channels and instability of the banks, both being contrasted with discontinuous maintenance works. The swallets, too, require periodic interventions, aimed at cleaning them from the accumulated sediments. These works have been carried out in 1994, 2000 and 2008 by the local caving group, Gruppo Speleologico Neretino. During the XIX century, large sectors of the catchment remained flooded for long times during the spring (De Giorgi 1884); on the other hand, greater wet areas coverage are documented in the previous centuries by botanical (Medagli et al. 1990) and archaeological studies (Arthur 1999). The hydrographic network of the Asso catchment (Fig. 3) was originated from the evolution of natural water lines during Middle Pleistocene; in its final sector, in the northern part of the catchment, it was the result of reclamation works realized in the 1930s to solve both flooding and epidemiological problems. Today the channel network functions to discharge wastewater into the subsoil (Delle Rose 2007, Delle Rose and Marras 2008), and as a tool to reduce seawater intrusion which affects the groundwater resource (Masciopinto 2006). The system consists also of a draining channel to the sea (E in Fig. 3), realized in the ’70s as a defence tool against flooding in the Nardò area. Recent chemical-physical analyses have highlighted high bacteria contents, together with high percentage of nitrates and presence of ammonia and nitrites as well (Delle Rose 2007). Therefore, the terminal sinks of the Asso karst system work at present: i) to mitigate the flood hazard; ii) to remediate seawater intrusion and iii) to dispose wastewater. In terms of safeguard of the environment and the public health, these functions are quite in contrast. 36

Whilst the first aim should be better obtained through an increase of the absorbing capability of the swallow holes, the second and third would be negatively impacted by such an option that would inevitably cause an easier spreading of contaminants into the aquifer. Thus, water management of the system is with no doubt a hard task.

Figure 3 Hydrographic network of the Asso catchment (after Reina 1972, mod.). Explanation: 1) watershed divide; 2) natural and man-made water lines; 3) ponors; 4) urban/industrial wastewater discharges; 5) draining channel to the sea; 6) flooding-prone towns In spite of the efforts made by Public Bodies, the knowledge related to speleogenesis and the hydraulic properties of the swallow holes is disregarded by the current water resource management. The geomorphological studies so far carried out allow us to distinguish natural, partially modified and man-realized water sinks. Some of the natural swallow holes can be described as collapse sinkholes (such Vora Colucce), but a number of them present different origin and development as percolation shafts. The Asso has discharges greater than 5 m3/sec, and generally it is not able to discharge this amount of water through the existing swallow holes, which often results in inundation of both the nearby urban and rural areas. When events over 30 m3/sec, with runoff time on the order of 12-15 hours are recorded, the situation may become really critical. The few available historical data indicate that 100 mm rainfall in 9-12 hours can be already considered as extreme rainfall, with negative effects on the whole catchment (Province of Lecce 2006). The last calamitous event occurred on November 2004, when several districts in the Nardò area were flooded with severe economic losses to the town and surrounding rural areas. Despite the efforts of local cavers to clean Vora Colucce from the sediments (Fig. 4), in order to facilitate its function as a swallowing site, due to the limited capacity of the site to absorb water, since September 2008 the draining channel is mostly used to discharge storm water. Recent events, with flooding of the field around Vora Colucce have again been registered on December 2008 and January 25, 2009 (Fig. 5). The Civil Protection Plan considers the following critical situations for the Asso catchment: limiting the road traffic, especially along the extra-urban areas; possibility of flooding for the houses below the ground; diffusion on the soil or in the water table of contaminants; possibility of drowning for people living in houses located below the ground surface, especially for people with a limited movement ability (Province of Lecce 2006). 37

Figure 4 Cleaning works at Vora Colucce (August 2008, photo: Massimiliano Beccarisi)

Figure 5 Flooded fields in the area surrounding Vora Colucce (December 2008) Despite all of this, the Asso catchment does not have a monitoring system of water discharge, neither a rain gauge network. 38

5 Conclusions When designing storm water management in karst areas, it is common that the existing sinkholes which receive overland flow will be improved to safely and, hopefully permanently, handle surface water flow into the subsurface. These “designer sinks“ are essentially cleaned sinkholes surrounded with gabion baskets to both maintain the integrity of the sinkhole and purify the flows (Fischer and Fischer 1997). As other means to control of stormwater in karst areas, use of channels draining into sinkholes, installation of drain pipes, or enlargement of sinkhole by excavation to increase drainage capacity are also considered. All these standard practices may cause several problems (Barner 1997): 1) runoff may exceed drainage capacity of sinkholes; 2) sediment and debris may plug sinkholes and associated karst conduits; 3) direct disposal of storm water into a karst aquifer may degrade water quality by introducing contaminants. All these problems derive from the lack of recognition of sinkholes as integrated parts of dynamic drainage systems, rather than as isolated entities. As repeatedly highlighted in the dedicated scientific literature, a real mitigation of the natural and anthropogenic risks in karst can be only reached after recognition of the peculiarities of this environment, and of its fragility as well (Cvijic 1918, Nicod 1972, White and White 1984, Mijatovic 1987, Parise and Gunn 2007). This is particularly true as regards the water management, and specifically flooding problems in karst areas, as outlined in this paper. References Arthur P (1999) Grubenhauser nella Puglia bizantina. A proposito di recenti scavi a Supersano (Le). Archeologia Medioevale 24: 171-178 Barner WL (1997) Comparison on stormwater management techniques in a karst terrane in Springfield, Missouri. In: BF Beck, JB Stephenson (eds) The engineering geology and hydrogeology of karst terranes. Proc. 6th Multidisc. Conf. Sinkholes, pp 253-258 Carrozzo MT, Delle Rose M, De Marco M, Federico A, Forte F, Margiotta S, Negri S, Pennetta L, Simeone V (2003) Pericolosità ambientale di allagamento nel Salento leccese. Quaderni di Geologia Applicata 2: 77-85 Ciaranfi N, Pieri P, Ricchetti G (1988) Note alla carta geologica delle Murge e del Salento (Puglia centro-meridionale). Memorie Società Geologica Italiana 41: 449-460 Colamonico C (1917) Le conche carsiche di Castellana in Terra di Bari. Boll. R. Soc. Geogr. It. 9-12:1-39 Cvijic J (1918) Hydrographie souterraine et évolution morphologique du karst. Rev. Trav. Inst. Géogr. Alpine 6:375-426 De Giorgi C (1884) Cenni di geografia fisica della provincia di Lecce. Lecce, 122 pp Delle Rose M (2007) Valutazioni dei rischi di allagamento per incremento degli eventi meteorici estremi in bacini endoreici della Puglia centro-meridionale. Proc. Conf. Climatic changes and geologic risk in Apulia, Sannicandro di Bari, 30/11/2007, pp 49-60 Delle Rose M, Marras V (2008) Attività di pretezione idrogeologica del GSN (anni 2004-2006). Proc. XI Regional Meeting of Speleology, Borgo Celano, 71-85 Fischer JA, Fischer JJ (1997) Wyndham Farms – A karst case history. In: Beck BF, Stephenson JB (eds) Proc. 6th Multidisc. Conf. on Sinkholes and the Environm. Impacts of Karst, Springfield, 6-9 April 1997, pp 287-292 Gams I (1978) The polje: the problems of definition. Zeit. für Geomorph. 22: 170-181 IPCC (2007) Climate Change 2007: Gli Impatti dei Cambiamenti Climatici, l’Adattamento e la Vulnerabilità. Sintesi per i decisori politici.

39

Lopez N, Spizzico V, Parise M (2009) Geomorphological, pedological, and hydrological characteristics of karst lakes at Conversano (Apulia, southern Italy), as a basis for environmental protection. Environmental Geology, DOI 10.1007/s00254-008-1601-9 Medagli P, Bianco P, Schirone B, D’Emerico S, Ruggiero L (1990) Il farnetto di Bosco Belvedere (Lecce). Annuario di Botanica 48: 77-82 Mijatovic BF (1987) Catastrophic flood in the polje of Cetinje in February 1986, a typical example of the environmental impact of karst. In: Beck BF, Wilson WL (eds) Proc. 2nd Multidisc. Conf. on Sinkholes and the Environm. Impacts of Karst, Orlando, 9-11 February 1987, pp 299-303 Masciopinto C (2006) Simulation of coastal groundwater remediation: the case of Nardò fractured aquifer in Southern Italy. Environmental Modelling and Software 21: 85-97 Nicod J (1972) Pays et paysages du calcaire. Presses Universitaires de France, Paris, 242 pp Orofino F (1990) Castellana-Grotte: le vicende storiche di Largo Porta Grande. Itinerari Speleologici 4:39-46 Parise M (2003) Flood history in the karst environment of Castellana-Grotte (Apulia, southern Italy). Natural Hazards and Earth System Sciences 3:593-604 Parise M (2009) Lakes in the Apulian karst (Southern Italy): geology, karst morphology, and their role in the local history. In: Miranda FR, Bernard LM (eds) Lake pollution research progress. Nova Science Publishers, Inc., New York, pp 63-80 Parise M, Pascali V (2003) Surface and subsurface environmental degradation in the karst of Apulia (southern Italy). Environmental Geology 44:247-256 Parise M, Gunn J (eds) (2007) Natural and anthropogenic hazards in karst areas: Recognition, Analysis and Mitigation. Geological Society, London, Special Publications, 279, 202 pp Parise M, Federico A, Delle Rose M, Sammarco M (2003) Karst terminology in Apulia (southern Italy). Acta Carsologica 32: 65-82 Provincia di Lecce (2006) Il piano comunale di Protezione Civile. 399 pp. Reina C (1972) Principi di difesa idraulica et idrogeologici dei bacini chiusi della regione pugliese. Atti Giornate di Studio 1° sez. CIRG, Firenze. Sgobba A (1896) Della inondazione avvenuta in Castellana il 9 novembre 1896. Stabilimento Tipografico N. Ghezzi, Monopoli, 15 pp Viterbo M (1913) Castellana e le alluvioni attraverso i secoli. Rass. Pugliese 28:1-23 White EL, White WB (1984) Flood hazards in karst terrains: lessons from the Hurricane Agnes storm. In: Burger A, Dubertret L (eds) Hydrogeology of Karst Terrains, 1, pp 261-264

40

Specificities of coastal karst aquifers with the hydrogeological characterisation of submarine springs – overview of various examples in the Mediterranean basin Nathalie DÖRFLIGER1, Perrine FLEURY1, Michel BAKALOWICZ2 , Hahmad EL HAJJ3, Abdoul AL CHARIDEH4, Mehmet EKMEKCI5 1

BRGM, Water Division, RMD Unit, 1039 rue de Pinville, 34000 Montpellier, France, e-mails: n.dorfliger@brgm; p.fleury@brgm 2 Hydrosciences, University of Monptellier II, c.c. MSE, 34095 Montpellier CEDEX 5, France, e-mail: [email protected] 3 CREEN ESIB, St-Joseph University, Beyrouth, Lebanon, e-mail: [email protected] 4 AECS, Damask, Syria, e-mail: [email protected] 5 HU-UKAM, University of Ankara, Turkey, e-mail: [email protected]

Abstract: Coastal karst aquifers are common around Mediterranean Sea. Belonging to aquifers with potentially important storage, they represent current or future groundwater resource. They discharge either at the coastal zone or directly into the sea at karst submarine spring (SMKS) level. The SMKS are often considered as non conventional resource which should be exploited. The occurrence of SMKS along the Mediterranean coast is discussed considering geological setting as well as the effect of the Messinian salinity crisis on karst development. Assessment of SMKS flow rates carried out mainly in the framework of the EU 6th FWP MEDITATE project, shows that the published values are often largely overestimated. Consequently, KSMS discharge may be insufficient to be captured. Moreover the frequent sea water intrusion makes the water unsuitable for consumption without any treatment. Techniques were developed for monitoring them in order to understand their functioning and assessing water resource. From studies of SMKS in France, Spain, Syria, Lebanon and Turkey, a classification of coastal karst aquifers is proposed, as a tool for managing coastal groundwater resources. Keywords: karst, submarine spring, groundwater resources

1 Introduction Coastal karst aquifers are known for frequently discharging at submarine or coastal brackish springs. The location of these springs is relatively well known and documented (Fleury 2005, Fleury et al. 2007b, Mijatovic 2007); at the beginning of the 20th century, Gruvel (1929, 1931) pinpointed already submarine springs along Syrian and Lebanese coasts. Thermography and infra red photography interpretation contributed to the location of these springs. Numerous coasts had been investigated, such as coasts of Languedoc (France), Lebanon, Greece, Sicily, Italy, former Yugoslavia, Turkey, Spain at the Mediterranean level. In addition, the occurrence of SMKS along the Mediterranean coasts, sometimes in a spectacular way with high discharge, makes them a potential unconventional resource which interests the decision makers in charge of regional development. Due to the hydrogeological setting, SMKS as karst springs are characterized by high variability of discharges but also concerning most of these springs by high variability of salinity. Their exploitation should thus

41

be considered as a function of these variations; due to the salinity variability consequently SMKS capture work should be coupled to a small unit of desalination in most case. The objectives of the paper are to propose, by means of some examples in Lebanon, Syria and Turkey, coastal karst aquifers studied in the framework of the 6th FWP MEDITATE European project, an overview of various types of SMKS linked to coastal karst aquifers and also give information about the potential use of SMKS for water management. 2 Origin of SMKS around the Mediterranean basin Limestone outcrops are widely spread around the Mediterranean basin, linked to PeriMediterranean Alps chains (Pyreneans, Alps, Dinarides, Taurus i.e.). Karst systems take place within these limestone outcrops; karst conduits networks are developed into depth due to the accommodation of the karst system to the variation of the base level. Theses variations are due either to tectonic movements or to Eustatic level changes. Given the numerous changes in the sea level (Haq et al. 1987), especially since Miocene (Rouchy and Saint-Martin 1992, Blanc 2002), sea level changes have strongly more influenced coastal karst evolution than tectonic movements which are localized even if in some place tectonics movement of high magnitude may play a major role (in Taurus i.e.). The evolution of karst systems and consequently their hydrogeological functioning is different if they developed under marine regression or transgression. During sea level falling, outlets of karst systems are located above the new base level corresponding to the sea level, in this case. The increase of hydraulic head within the karst system allows developing of new vertical conduits connected to previous channel conduits to the new base level (Ford and Williams 1989). At the opposite, a sea level rise leads to drown of springs and conduits. Depending on the hydraulic conditions, spring allows discharge permanently or seasonally, and/or intrusion of sea water.

Figure 1 Diagram giving the depth below sea level at which ad SMKS may outflow, in function of the hydraulic head in the conduit discharging at the spring (Fleury 2005). These hydraulic conditions are driven by water density difference between fresh water and sea water, and the water head in the conduits as well as the sea water level, such as described by Fleury (2005), considering the sea water fresh water relationships described by Ghyben and Herzberg (Stringfield and LeGrand 1971). The necessary hydraulic head in the conduit at the spring to get a fresh ground water submarine discharge at a determined depth is shown in Figure 1. At any site above the line (Figure 1) a karst conduit may discharge fresh water into the sea. At any site below the line, no fresh water discharges and the inflow of sea water into the 42

conduit is possible. For instance, for a 120 m deep conduit the fresh water head in the conduit must be at least 3 m above sea level for discharging ground water at a SMKS. These values are approximate, as the head losses in the conduits are neglected. The main consequence is the possible seasonal changes in the functioning of some springs, outlets of conduits subject to large water head variations. SMKS and submerged conduits present along the Mediterranean coast are in favour of bathymetric variations that occurred along the coast. During Quaternary the lowest sea level, i.e., 120 to 140 m below the present one, took place during the Würm glacial maximum, around 15 ka (Ford and Williams 1989). During the interglacial periods, the sea level was close to the present one. These variations have been recorded on all coasts worldwide. Otherwise deep karst conduit systems developed around Mediterranean Sea. Their depth is not in agreement with low glacial sea levels: (i) Port-Miou conduit (France) has been explored down to -172 m (Arfib et al. 2006a, Cavalera et al. 2006); (ii) in Fontestramar coastal spring (France), the phreatic karst has been dove up to a depth of -164 m (Brandt 1997; Aunay et al. 2003); (iii) in Chekka (Lebanon), the deepest spring is between -110 and -150 m below sea level (Moulard et al. 1965, Kareh 1967, El-Hajj et al. 2006, Bakalowicz et al. 2007b), (iv) in Almyros of Iraklio coastal karst (Crete, Greece), a sea water intrusion was identify at a depth of around 500 m below m bsl, involving a deep karstification (Arfib et al. 2002, Arfib and de Marsily 2004, Arfib and Ganoulis 2004, Arfib et al. 2006b) and (v) Karst conduit and cavities were hit at a depth around 600 m below the sea level during deep drilling for oil exploration in the Taurus mountains. Those depths are not compatible with the lowest known sea levels during the glaciations and in addition due to the fact that important uplifts took place along the Mediterranean coasts during Pliocene and Quaternary (Elias, et al. 2007). Consequently, the deep Mediterranean conduit systems must be due to a larger-scale event (Aunay et al. 2003, Bakalowicz et al. 2003a, Bakalowicz et al. 2004, Blavoux 2004), such as the Messinian salinity crisis (Clauzon 1982, Rouchy and Saint-Martin 1992, Rouchy 1999, Rouchy et al. 2006). Due to the closure of the connection of the Mediterranean Sea to the Atlantic Ocean and to a very high evaporation rate of the basin, the sea level dropped up to 1500m below the present day sea level, between -5.8 Ma and -5.32 Ma (Clauzon et al. 1996). This situation lasted around 500,000 years and the climate may have been close to that of the Red Sea today (Rouchy 1999). These conditions are in favour of the karst development of coastal carbonate aquifers and to sharp incision in limestone formations with canyons. Vertical shafts and galleries toward a greatly lowered base level took place (Blanc 1995). Depending on the local geological conditions, the deep conduits may be plugged with sediments or remain open to the sea. Then their role in the functioning of coastal aquifers varies, according to the depth of the open conduit and the fresh water hydraulic head within the aquifer. In consequence, the occurrence of SMKS of limestone coastal aquifers in Mediterranean is widely spread (Figure 2). 3 Assessment of the fresh water discharge from SMKS with some examples Since 1960’s, the fresh ground water discharge assessment from SMKS to the sea has been a real challenge, due to the strong interest by water management stakeholders, considering that the use of SMKS water could contribute to the regional economic development. Several methods for assessing their discharge were attempted (direct estimate by divers, flowmeter above the spring, mass balance of chloride or tritium in the water mass above the SMKS and thermal infrared surveys correlated to flow with analogy of results of river discharge assessment); the flow rate values obtained were unfortunately rarely looked at critically. The water budget is one of these methods to assess the unknown term, i.e. the discharge of the SMKS. Another method is the direct measurement, based on the monitoring 43

of flow rate as well as temperature, pressure and electrical conductivity directly at the SMKS; all these devices are implemented on a PVC pipe settle on the outlet and being at anchor on the basement with concrete blocks. These two methods were applied on three tests sites located in Lebanon (Chekka Bay) and in Syria (Bassieh-Banias); in Turkey (Gökova Bay) the water budget was assessed.

Figure 2 Distribution of karst areas (brick symbol) and coastal and SMKS (black dots) around the Mediterranean Chekka area is located in North Lebanon, the karst system is recharged by diffuse infiltration on 155 km² are of the Cenomanian-Turonian outcropping limestone and by point recharge into swallow holes in the two main river beds, El Jawz and El Asfour Rivers. It discharges at a complex outflow system with several springs, permanent and seasonal, connected to karst channels developed in depth. Karst features such as paleokarst cavities infilled with fine, detrital sediments, capped by carbonate deposits (stalagmite “pavement”) observed in Cretaceous limestone and the occurrence of SMKS at depth below -100 m bsl, are argues in favor of the impact of the Messinian crisis on the karst development of Chekka karst system. Occurring at depth below -100m, SMKS were located even deeper at the beginning of Quaternary, at least -400 m bsl, and even deeper at the beginning of Pliocene, according to the uplift estimated to 300 m during Quaternary in the Levantine (Sanlaville 1977). The water budget was assessed for the Chekka coastal karst system (CCKS), considering rainfall, river losses, on shore outflow and withdrawal. The total discharge of freshwater at the SMKS corresponds to a mean value between 1.7 and 2.7 m3/s, according hydrological years or cycles. These values are about ten times lower than assumed by previous studies, using flow velocity assessment i.e. (Kareh 1967). Moreover, this discrepancy was validated by direct monitoring of the flow rate of the main permanent spring; the mean flow rate monitored over a 3 months period is around 35 L/s at S2 spring. Considering some other small outlets, the total flow does not exceed 70 L/s, corresponding to 40 L/s of fresh water if the salinity is about 40%. The salinity variability of the S2 spring is important: during the recession, the salinity increases (500 PS/cm to 24 mS/cm); during low water stage, the value of the salinity is high (34 to 38 mS/cm), due to salt water intrusion. The fresh water ratio is then about 35%. During low water stage, the salinity as well as the discharge are linked to sea level variation (El Hajj 2008). 44

In Syria, in the area of Bassieh-Banyas (South Mediterranean Syrian coast, North to Lebanon-Syrian border), the confined upper Cretaceous aquifer discharges at the submarine springs, due to sufficient hydraulic head to enhance the vertical uprising of water through the confining layers. About thirty submarine springs were located along the coast of this area (IBG/DHV 2000), at various depths ranging between 5 to 35 m below sea level in average. One of the SMKS located at -5 m bsl was equipped with appropriate devices to monitor flow velocity; the discharge is few l/s -2- “ L/s in summer 2007 (MEDITATE, D#23 2007), without variation of Conductivity and discharge. The groundwater out letting is fresh water. The catchment area of this aquifer discharging in the Sea is assessed to be 855km². On the basis of a conceptual model, with some assumptions that have not been tested such as the no recharge of the aquifer by water losses in sinkholes, a rough assessment of the water budget was established. The total fresh water discharge from submarine springs in Bassieh bay is estimated to be 162 millions m3/y or 5.1 m3/s, considering 980 millions m3/y from rainfall, 197 millions m3/y for evapotranspiration, 180 millions m3 for surface water flow and 42 millions m3/y for withdrawal (MEDITATE, D#13 2007). Previous assessments using a combination of flow velocity and fresh water content gave discharge ranges from 3 m3/s at Bassieh Gulf to 3.5 m3/s at Tartous harbour (IBG/DHV 2000). Al Charideh (2006) using d18O/CL- relationship showed that all submarine springs water is a mixture of groundwater and seawater, with a large range of variation of percentage, between 12 to 97%. The total assessment of fresh water from permanent SMKS (four springs in total, 2 in Banyas Gulf and 2 in Tartous harbour) discharge considering flow velocity and isotopic and chemical data is 350 millions m3/year, i.e. 11 m3/s. (Al Charideh 2007). The assessment of discharge of this aquifer is function of high uncertainty on data, water catchment area. In the Gökova Bay, located on the Western Taurus Mountains in the Western Mediterranean region of Turkey, the coastal and submarine karst springs system shows well developed karst features. Plateau, peaks and coastal plains range between 800-1000 m above sea level (asl) to sea level. The whole area covers a drainage area of 1200 km² between Mugla and Ula Bodrum in the North of Gökova. The springs of Gökova are organized into 5 groups. They mainly discharge along fault lines within carbonate rock masses of Mesozoic age. The drainage area may be subdivided into 3 sub regions, Yatagan, Mugla and Ula regions, considering groundwater flow paths and drainage routes, as well as springs location. According to the water budget calculation, there is a surplus of about 640 millions m3/year (i.e. 20 m3/s), that should correspond to the unmeasured submarine springs. Considering an average amount of the precipitation over the area calculated by isohyets method, this amount reduces to about 200 millions m3/year (i.e. 7 m3/s). All these SMKS are located close to the coast, with diffuse discharge, making difficult to set up a monitoring surface as it was done in Syria and Lebanon. So it is more reliable to postulate that the submarine springs discharge comprises in the range of 200 to 640 million m3/year in average (MEDITATE 2007). 5 Conclusion These three examples of coastal karst aquifers with SMKS give an overview of various hydrogeological settings on one hand and on another hand, uncertainties around the assessment of SMKS discharge. However, results highlight the systematic discrepancy between the previous estimates of SMKS and the MEDITATE results, as previously shown in France (Fleury 2005). The total annual discharge is about ten times lower than the previous estimates, what makes these submarines springs globally not so interesting targets for developing new water resources. Direct flow measurement is not always possible; specifically when discharge is diffuse into the sea, such as it is the case partly in Syria and in Turkey. In this case, the aquifer functioning is not really karstic, as discharge is relatively low and varies little with season, as 45

it has been observed in Syria. Such type of coastal carbonate aquifer is characterized by narrow conduits and fractures, with high head water losses near the outlets, avoiding saltwater intrusion. For the case of Chekka, SMKS may have strong seasonal variability of discharge and physical parameters; some SMKS are temporary. The salinity of the water is often low during high flow but rises as the flow decrease and as there is withdraws by pumping on the system on the coast. This type of coastal karst aquifer describes karst aquifer with well developed karstic network, often on several levels. The karstification may be quite extensive under the present base level. This is the main type of coastal aquifers around Mediterranean Sea, in agreement with a regional geology driven by the Messinian salinity crisis. In addition, this type of karst is one of the most difficult to be developed for resources allocation, either with some on shore boreholes or directly at the submarine spring. Due to salinity occurrence and a seasonal discharge, it may be necessary to treat water at the level of a small unit of desalinization by inverse osmosis, for such capture work at the SMKS. A survey with an autonomous underwater vehicle (AUV) (MEDITATE 2007, Lapierre and Jouvencel 2008) could constitute possible alternative to calculate discharge. Such AUV is propelled and driven with control surfaces sensor; it is equipped of scientific sensors allowing CTD sampling and acoustic and video images acquisition (bathymetry and acoustic profiling). It will allow to survey above SMKS and to obtain a 3D physical parameter field. This one may be used in 3D oceanography modelling of the fresh plume of the SMKS, in order to determine discharge. In general the harnessing of SMKS cannot be appropriate answer to solve water shortage. However the monitoring of SMKS must be considered as a necessary survey of the appropriate management of coastal karst aquifers. This type of monitoring has to be integrated within an early warning system to prevent seawater intrusion. If harnessing SMKS is one of the solutions selected as alternative water resource, it may be connected to a small unit of desalination based on Reverse Osmosis depending on spring water salinity (feasible and interesting economically, MEDITATE 2007) and used either for irrigation, for industry or for water supply. Acknowledgments This work is the result of the 6th FWP MEDITATE (PL509112 – FP6-2002-INCOMPC1 of the European Commission), as well as of the pH D work of Fleury (2005) in Paris VI University. References Al Charideh AR (2007) Environmental isotopic and hydrochemical study of water in the karst aquifer and submarine springs of the Syrian coast. Hydrogeology Journal 15:351–364 Arfib B, Cavaliera T, Gilli E (2006a) Influence de l'hydrodynamique sur l'intrusion saline en aquifère karstique côtier. Comptes Rendus Geosciences 338 :757-767 Arfib B, Ganoulis J, de Marsily G (2006b) Locating the zone of saline intrusion in a coastal karst aquifer using springflow data. Ground Water 45:28-35 Arfib B, de Marsily G (2004) Modelling the salinity of an inland coastal brackish karstic spring with a conduit-matrix model. Water Resources Research 40 Arfib B, Ganoulis J (2004) Modélisation physique de l’intrusion d’eau de mer dans un aquifère karstique: cas de l’Almyros d’Héraklion (Crète). Comptes Rendus Geoscience 336, 999-1006 Arfib B, de Marsily G, Ganoulis J (2002) Les sources karstiques côtières en Méditerranée: étude des mécanismes de pollution saline de l'Almyros d'Héraklion (Crète), observations et modélisation. Bulletin Société géologique de France 173:245-253

46

Aunay B, Dörfliger N, Le Strat P, Ladouche B, Bakalowicz M (2003) Évolution géologique, mise en place de la karstification et thermalisme des aquifères karstiques périméditerranéens. Exemple du karst des Corbières d'Opoul. Colloque "Circulations hydrothermales en terrains calcaires", 10ème journée technique, Carcassonne, CFH-AIH Bakalowicz M, Aunay B, Le Strat P, Dörfliger N, Fleury P (2004) Karst development potential and base level changes in Mediterranean regions: a unique reference model. IGCP 448 Meeting, Bowling Green, KY, USA. Guangxi Normal University Press Bakalowicz M, El Hakim M, El-Hajj A (2007b) Karst groundwater resources in the countries of eastern Mediterranean. Environmental Geology DOI 10.1007/s00254-007-0854-z Blanc JJ (1995) Étapes et facteurs de la spéléogenèse dans le sud-est de la France. Karstologia 13-26 Blanc PL (2002) The opening of the Plio-Quaternary Gibraltar Strait: assessing the size of a cataclysm. Geodinamica Acta 15:303-317 Blavoux B, Gilli E, Rousset C (2004) Alimentation et origine de la salinité de la source sous-marine de Port-Miou (Marseille-Cassis). Principale émergence d'un réseau karstique hérité du Messinien. Comptes Rendus Géoscience 336:523-533 Brandt C (1997) Font Estramar, 11 ans de recherches en plongée dans un grand réseau noyé en bordure de la Méditerranée. Exploration et Spéléologie, Symposium 4, 12th International Congress of Speleology, La Chaux-de-Fonds, Switzerland Cavalera T, Arfib B, Gilli E (2006) Ressource karstique côtière en Méditerranée: les sources sous-marines de Port Miou et du Bestouan (Marseille - France). Proc. 8th Conference on Limestone Hydrogeology, Neuchâtel (Switzerland). Presses universitaires de FrancheComté, Besançon, France Clauzon G (1982) Le canyon messinien du Rhône: une preuve décisive du "dessicateddeep model", Bulletin de la Société Géologique de France 24:597-610 Clauzon G, Rubino JL, Suc JP (1996) Les Rias pliocènes du Var et de Ligurie: comblement sédimentaire et évolution géodynamique. Livret guide - Excursion commune du Groupe Français d'Etude du Néogène et du Groupe Français de Géomorphologie El-Hajj A, Bakalowicz M, Najem W (2006) Hydrogéologie des calcaires crétacés du Nord Liban. Conference on the ”Water Resources in the Mediterranean Basin” , WATMED 3, Tripoli (Lebanon) El-Hajj A (2008) L’aquifère carbonate karstique de Chekka (Liban) et ses exutoires sous-marins. Caractéristiques hydrogéologiques et fonctionnement, Thèse de doctorat de l’université Montpellier II (France) et de l’université Saint-Joseph (Beyrouth, Liban), 261 pp Elias A, Tapponnier P, Singh SC, King GCP, Briais A, Daëron M, Carton H, Sursock A, Jacques E, Jomaa R, Klinger Y (2007) Active thrusting offshore Mount Lebanon: Source of the tsunamigenic A.D. 551 Beirut-Tripoli earthquake. Geology 35:755-758 Fleury P (2005) Sources sous-marines et aquifères côtiers méditerranéens. Fonctionnement et caractérisation. PhD, Sciences de la Terre, Université Paris 6, France Fleury P, Bakalowicz M, de Marsily G (2007b) Submarine springs and coastal karst aquifers: a review. Journal of Hydrology 339:79-92 Fleury P, Bakalowicz M, de Marsily G, Cortes JM (2007c) Functioning of a coastal karstic system with a submarine outlet, in southern Spain. Hydrogeology Journal DOI 10.1007/s10040-007-0215-8 Ford DC, Williams PW (1989) Karst geomorphology and hydrology. London: Academic Division of Unwin Hyman Ltd Gruvel A (1929) De quelques particularités océanographiques observées sur les côtes de Syrie. Comptes rendus hebdomadaires des séances de l’Académie des sciences, Paris 189: 111-112

47

Gruvel A (1931) Les Etats de Syrie. Richesses maritimes et fluviatiles. Exploration actuelle – avenir. Société des éditions géographiques, maritimes et coloniales, Paris 453 pp Haq BU, Hardenbol J, Vail PR (1987) Chronology of fluctuating sea levels since the Triasic. Science 235:1156-1167 IBG/DHV (2000) Project of supplying part of the water demand of Damascus city and its countryside form the Syrian coastal area water surplus, Hydrogeological investigation, NR.1 Phase 1 – first stage, Damascus water supply and sewerage authority, Damascus, Syrian Arabi Republic Kareh R (1967) Les sources sous-marines de Chekka (Liban). PhD, Faculté des Sciences, Géologie, Université de Montpellier Lapierre L, Jouvencel B (2008) Robust non linear path following control of an AUV, Journal of the Oceanic Engineering 33:89-10 MEDITATE (2007) Final report and public deliverables available on the web : www.meditate-eu.org Mijatovic B (2007) The groundwater discharge in the Mediterranean karst coastal zones and freshwater tapping: set problems and adopted solutions. Case studies. Environ. Geol. 51:737-742 Rouchy JM, Saint-Martin JP (1992) Late Miocene events in the Mediterranean as recorded by carbonate-evaporite relations. Geology 20:629-632 Rouchy JM, Suc JP, Ferrandini J, Ferrandini M (2006) The Messinian Salinity Crisis revisited. Sedimentary Geology 188-189:1-8 Sanlaville P (1977) Etude géomorphologique de la région littorale du Liban. Publ. Univ. Lib. Section des etudes géogaphie, 2 tomes, 859 pp Stringfield VT, LeGrand HE (1971) Effects of karst features on circulation of water in carbonate rocks in coastal areas. Journal of Hydrology 14:139-157

48

PaPRIKa a multicriteria vulnerability method as a tool for sustainable management of karst aquifers Example of application on a test site in SW France Nathalie DÖRFLIGER1, Valérie PLAGNES2, Konstantina KAVOURI 2 1

BRGM, Water division, RMD Unity, 1039 rue de Pinville, 34000 Montpellier, France, e-mail: n.dorfliger@brgm 2 UMR Sisyphe, Pierre et Marie Curie University – Paris 6, France, e-mails: [email protected]; [email protected]

Abstract: In Europe, carbonate terrains occupy 35 % of the land surface, contributing up to 50% of drinking water and in some countries being the only available water resource. Karst groundwater is thus an important water resource, which is, however, particularly sensitive to contamination, due to structure and hydrological behaviour; it is thus considered to be particularly vulnerable to pollution. Karst aquifers require consequently, specific groundwater protection. Groundwater vulnerability assessment schemes have been developed during the last decade, specifically for karst aquifers, based on the early EPIK method (Dörfliger et al. 1999) that influenced later ones in the framework of the European COST Action 620 as well as at the national level. These methods are either resource or source vulnerability mapping. The proposed vulnerability mapping method PaPRIKa takes into consideration criteria for both structure and functioning of the aquifer. Based on EPIK and RISK resource methods, PaPRIKa method was developed as a resource and source vulnerability mapping method, allowing assessing vulnerability with 4 criteria: P for Protection (considering the most protective of a combination of Soil cover, Unsaturated zone and Epikarst aquifer behaviour criteria), R for Rock type, I for Infiltration and Ka for Karstification development. The development of this method took into consideration results of development of other methods such as PI (Goldscheider 2005) and COP source method (Andreo et al. 2009). Additional criteria are applied such as groundwater travel time and active conduit network, on the resulting vulnerability map. In this manner adequate source protection can be assessed; results may be used to outline protection areas of capture work in karst aquifers. The PaPRIKa method was applied in several test sites in France, in various geological and hydrogeological settings, particularly in Water agencies territories with particular high density of karst aquifers (Rhône-Mediterranean, Adour-Garonne and Seine-Normandy basins). In this paper an example is given as illustration of the method, the Ouysse karst system in the Causses area in South-West of France. The obtained results are discussed and the role of vulnerability map in the scheme of groundwater protection with protection zones of capture work as well as with priority of protection in catchment area is given. Keywords: karst, vulnerability, water management

1 Introduction Groundwater from karst aquifers is an important resource for drinking water supply of community around the world. Carbonate rock outcrops of which a large part is karstified cover up to 12% of the planet’s dry, ice-free land. About 25% of the global population is supplied largely by karst groundwater (Ford and Williams 2007). In Europe; carbonate terrains outcrops cover 35% of the land-surface; this percentage is similar for the French territory. These terrains contribute up to 50% to the water supply and even more in some regions where karst aquifers are the only fresh water resource (COST 65 1995). 49

However karst groundwater systems are generally considered to be particularly vulnerable to pollution, because of their unique structure. This structure is strongly heterogeneous. A network of conduits of high permeability is surrounded by large volumes of low permeability fissured rock. At the surface, heterogeneity is linked to the karst features morphology as well as the existence of various compartments from surface to vadose zone: recharge occurs by both dispersed and concentrated water entry (water losses of river, vertical shaft, dolines, dry valleys…). The recharge process is strongly influenced by these compartments in the verticality, namely the epikarst, defined as a perched saturated horizon close to the surface allowing temporary storage of water within dissolution features (Mangin 1975, Williams 1983). Special characters of karst aquifers, such as fast concentrated infiltration and rapid transport into the conduits network over long distances (several kilometers) make them particularly vulnerable to contamination due to various human and land-use activities. Consequently, karst groundwater requires specific and appropriate groundwater protection schemes in order to combine land use practices and sustainable water management. The concept of groundwater vulnerability mapping constitutes a common basis to set up protection zoning and land-use planning. The term of vulnerability was used in the sixties in France, introduced as a scientific term by Albinet and Margat (1970). The vulnerability was described by Vrba and Zaporozec (1994) as a relative, non measurable and dimensionless property, making a distinction between intrinsic and specific (for a particular contaminant) vulnerability. In the scope of the development of the PaPRIKa approach, the vulnerability definition is similar as the one defined by Dörfliger et al. (1999) and in the framework of the COST Action 620 (Zwahlen 2004): the intrinsic vulnerability of groundwater to contaminants represents the inherent hydrogeological and geological characteristics which determine the sensitivity of groundwater to contamination by human activities; it is independent of the nature of the contaminants and the contamination scenario. Several methods have been developed to characterize the intrinsic vulnerability of resource (water body reached by infiltration water with contaminants originated from landsurface the groundwater surface) and/or source (capture work: spring or well are the target after) (Goldscheider et al. 2000), during last decades. General methods applied for porous and fissured aquifers exist e.g. DRASTIC (Aller et al. 1987), SINTACS (Civita and De Maio1997), GLA (Hötling et al. 1995) and PI (Goldscheider et al. 2000). Specific methods were developed to take into consideration karst aquifers structure and hydrogeological behavior: EPIK method (Dörfliger et al. 1999), REKS Method (Malik and Svasta 1999), RISKE (Pételet et al. 2000) and the COP method (Vias et al. 2006, Andreo et al. 2009). The PaPRIKa method is an updated method derived from previous specific vulnerability mapping method for karst groundwater. The aim of this study is to present this new method for which guidelines are written to serve as a common basis in France for the karst groundwater protection, i.e. protection zones outlining for capture work and land-use planning. 2 PaPRIKa vulnerability mapping method 2.1 Overview of existing specific vulnerability methods for karst aquifers The first method considering the specific properties of karst groundwater is EPIK (Dörfliger and Zwahlen 1998, Dörfliger et al. 1999). This method considers 4 criteria, such as Epikarst (E), Protective cover (P), Infiltration conditions (I) and Karst network development (K). It is a multi-attribute weighting rating method (overlay and index method), point-count system based on the DRASTIC model. Several weakness of the method has been assessed later on after several tests carried out during the COST Action 620 in European countries and as well in France. The method RISK(E) based on EPIK method introduced modification into 50

the definition of the criteria, in order to avoid overlapping of criteria (e.g. in Infiltration and in Epikarst definition) as well in the weighting and rating system. More consistency was introduced, defining the total weight of criteria as 100% or 1. In the EPIK method Epikarst criteria was based on the mapping of karstic features such as karrenfields, dolines, dry valleys and in the same time concentrated infiltration considered as well water losses and dolines catchment. Epikarst is an ambiguous criterion, as due to its hydrogeological behavior it may play a role of temporary storage of infiltrated water and in contrary possibly allows concentrated infiltration through vertical shaft directly connected to channel network. In the RISKE method, Epikarst criterion was then considered through its hydrogeological functioning and not through its karstic features. Concentrated infiltration spots were considered in the Infiltration attribute. Theses methods were considered as resource vulnerability mapping methods. In the framework of COST Action 620, the PI method (Goldscheider et al. 2000) served as a basis for the European approach and for the COP method. The PI method considers the protective function of the layers above the saturated zone (P) and the infiltration conditions (P). The P factor is based on the assessment scheme proposed by Hölting et al. (1995).It includes the topsoil, the subsoil, the non karst rock and the unsaturated zone of the karst rock. Protectiveness is assessed on the basis of the effective field capacity of the soil, the grain size distribution of the subsoil, the lithology, fissuring and karstification of the non karst and karst rock, the thickness, the mean annual recharge and artesian pressure in the aquifer. The I factor takes account for the karst specific recharge and infiltration processes (Ravbar and Goldscheider 2009). It describes the infiltration conditions and in particular the degree to which the protective cover is bypassed as a result of lateral surface and subsurface flow that enters the karst aquifer. It depends of the slope, the soil properties and the vegetation and the position of a given point inside or outside the catchment of a sinking stream (Goldscheider 2005). The final protection factor is the product of P and I; five classes of vulnerability are distinguished. The COP method (Vias et al. 2006) designed for resource vulnerability mapping includes the overlaying layers factor (O), the concentration of flow factor (C) and the precipitation regime factor (P). For source vulnerability, water of wells or spring is the target; the horizontal flow in the saturated zone depicted as the karst network development is taken into consideration in the COP+K method. K factor is based on the transit time, the information on karst network, the degree of connection of it to the spring or well (Andreo et al. 2009). 2.2 PaPRIKA method The PaPRIKa method is a resource and source vulnerability mapping. Two major differences exist with PI and COP methods and derived methods from COST Action 620. The first one is the absence of considering the precipitation regime factor. Regime factor is not considered as an inherent and intrinsic attribute. The second one is a simple indexation in 5 classes of the vulnerability of each criterion as well as of the final index (1 for the most protective and 4 for the most vulnerable). The four attributes of the PaPRIKa method belong to two groups: P and R for attributes linked to the structure of the aquifer and I and Ka attributes linked to the hydrogeological behavior. The Protective cover assessment scheme is based on the consideration of the soil characteristics (texture, structure and thickness), the non saturated zone (thickness, lithology and fracture degree) and epikarst aquifer. The cross mapping of these criteria allows to keeping the most protective value. Additionally, within catchments of water losses the P criterion characterizes the non infiltration properties of soils and sub-soils. In this case, an elevated degree of vulnerability is attributed to impervious formations like clays, and a lower one to more permeable formations like sands and unconsolidated conglomerates. The Rock 51

factor considers the lithology and the degree of fracturing of the aquifer body. The Infiltration conditions factor is defined taking into consideration slope and the karstic features allowing direct infiltration such as water catchment of water losses, sinkholes, dolines etc… The Karstic degree Ka factor is based on the assessment of the karst degree considering discharge and chemical variability at the spring as well as velocities and restitution rates showed by artificial tracing tests. The weighting formula of these four criteria is the following: all weights are equal to 1. It is proposed that the total weight for the structure factors is equal to 0.4 and the total weight for the hydrogeological functioning is equal to 0.6. The obtained vulnerability map is considered as a resource vulnerability map. In order to obtain a source vulnerability map of the karst system or aquifer, the underground transit time is considered as an additional factor in the I map. We define isochrones, between the capture work and defined isochrones using velocity data from artificial tracing tests and as well on the karst conduits maps or supposed karst conduits. 2.3 Source vulnerability mapping in the pilot site of Ouysse The studied area is located in the department of Lot within the territory of the Natural Regional Park of the Causses of Quercy. The Quercy region is a typical karst terrain subdivided in four calcareous plateaus: the Causse of Gramat, the Causse of Martel, the Causse of Limognes and the Causse of St Chels. The Ouysse system constitutes the north part of the Causse of Gramat. The Causse of Gramat is constituted of a tabular carbonate sequence of middle to late Jurassic, slightly tilted towards the SW (Figure 1). Thanks to this structure the entire sedimentary sequence outcrops within the limits of the studied area. The Ouysse river basin constitutes a binary karst system of 590 km², from which 154 km² correspond to superficial water catchments whose runoff penetrates the aquifer through a group of swallow holes developed on the eastern edge of the calcareous series (Aalenian). Numerous speleological surveys allowed mapping three active conduits: the river of Vitarelles on East, the river of Lacarrière on West and the river of Viazac-Planagrèze on the South part of the basin. Artificial tracer tests show the direct connection of the karst conduits to springs at the NE of the basin. The territory is rich in karst geomorphology with numerous dolines and swallow-holes. Ouysse is captured at two springs (Cabouy and Fontbelle) and by a well located in the underground river of Vitarelles in the central part of the basin (Bèdes). The first part of this study concerns the assessment of the intrinsic vulnerability of the resource. This application took under consideration specificities of the Ouysse system such as the size of the catchment, the existence of a related surface catchment area and the spatial variability of the karst network development. As a result, the PaPRIKa map of Ouysse represents three classes of vulnerability, from moderate to extremely high (Figure 2). As less vulnerable (V2) is characterized the West part of the studied area which is covered by semipermeable layers (marls and silt limestone of late Kimmeridgian). Tectonic accidents, water losses’ catchments and mapped conduits are grouped in highest classes of vulnerability (V4) while the rest of the basin is characterized vulnerable mainly because of the high number of dolines (V3). At last, the vulnerability is generally accentuated within two areas considered to be connected with active karst draining axes, the first having a South to North direction and the second one an East to West. In the second part of this study an assessment of the intrinsic vulnerability (source vulnerability) for each capture work has been realized. An additional criterion of tracer’s travel velocity has been introduced in the Infiltration criterion, in order to assess “intervention time isochrones”. The principle of this approach is based on highlighting the zones where concentrated infiltration and rapid transfer towards the capture work is more probable. The intervention time constrains allow the representation of lateral transportation parameter in a 52

two dimensions plan. Several time limits have been suggested (12h, 24h, 36h and 48h) in order to provide to decision makers the possibility to choose among the most adapted to field’s particularities, opposed to a scenario of accidental pollution.

Figure 1 Simplified hydrogeological map and vertical section of the studied area (modified after Astruc et al. 1994)

Figure 2 Intrinsic vulnerability map of the Ouysse karst system. PaPRIKa method 53

The mapping’s target being the capture work, this application concerns only the well’s/ spring’s basic alimentation area. The example of the application for the well of Bèdes for an intervention time of 12 h and 48 h is given (figure 3a and 3b respectively). The catchment area of the capture work of Bèdes covers 199km² and corresponds approximately to 1/3 of the total catchment’s area. It has been delimited for the needs of this study by coupling topography with geological characteristics and hydrological information provided by tracer tests. Comparing these two maps with the PaPRIKa resource mapping (see also figure 2) one may note that the vulnerability value appears decreasing within the distant zones, located beyond the isochrone’s limits.

Figure 3 Intrinsic vulnerability map of the well of Bèdes for an intervention time of a) 48 h and b) 12 h 3 Conclusion The PaPRIKa method is the result of a major improvement of previous specific intrinsic vulnerability methods developed for karst aquifers, taking into consideration European research work on this topic. It was tested on numerous sites in France. It allows obtaining resource and source vulnerability mapping of karst aquifer, and constitutes a common methodology based on specific structure and functioning of this type of aquifer. The method is described in detail in common guidelines that are available for both French administrations and engineer’s enterprises in charge of carrying out vulnerability mapping in order to determine protection zoning. The vulnerability assessment needs to be reliable and in order to avoid subjectiveness, the validation of vulnerability maps should be done. Validation procedures should be applied considering on one hand integrating tracers (nitrates, pesticides such as developed in the framework of the EU FOOTPRINT project, i.e.) and on the other hand artificial tracer tests under variable hydrological conditions such as suggested by Ravbar and Goldscheider (2009). Acknowledgments This work is a contribution to the project co-funded by ONEMA as well as by BRGM in the framework of public activities for the Environment Ministry. The application to the test site was also supported by the Parc naturel Régional des Causses du Quercy (J. Trémoulet) and Adour-Garonne Water Agency. The authors thank the working group (P. Marchet (AEAG), L. Cadilhac (AERM&C), D. Humbert (AESN), Ph. Muet (Ginger Environment), Ph. Crochet (ANTEA) and P-Henri Mondain (CALLIGEE)) for their comments and suggestions in order to improve the methodology.

54

References Albinet M, Margat J (1970) Cartographie de la vulnérabilité à la pollution des nappes d'eau souterraine. Bulletin du BRGM (deuxième série) III(4):13-22 Aller L, Bennet T, Lehr JH, Petty RJ, Hackett G (1987) DRASTIC: a standardized system for evaluating ground water pollution potential using hydrogeological settings. US Environmental Protection Agency, EPA/600/2- 87-036, 455 pp Andreo B, Ravbar N, Vías JM (2009) Source vulnerability mapping in carbonate (karst) aquifers by extension of the COP method: application to pilot sites, Hydrogeology Journal 17:749–758 Astruc JG, Coustou JC, Cubaynes R, Galharague J, Lorblanchet M, Marcouly R, Pelissie T, Rey J (1994) Notice explicative, Carte géol. France (1/50 000ème), feuille Gramat (833). Orléans : BRGM, 69 p. Carte géologique par JG Astruc (1994) Cività M, De Maio M (1997) SINTACS: un sistema parametrico per la valutazione e la cartografia della vulnerabilita degli acquiferi all’inquinamento: metodologia e automatizzazione [SINTACS: a parametric system for the assessment and mapping of the groundwater vulnerability to contamination: methodology and automation]. Pitagora, Bologna, 208 pp COST 65 (1995) Hydrogeological aspects of groundwater protection in karstic areas, Final report (COST action 65). European Commission, Directorate-General XII Science, Research and Development, Report EUR 16547 EN, Brussels, 446 pp Dörfliger N, Jeannin PY, Zwahlen F (1999) Water vulnerability assessment in karst environments: a new method of defining protection areas using a multi-attribute approach and GIS tools (EPIK method). Environ Geol 39(2):165–176 Dörfliger N, Zwahlen F (1998) Practical guide: groundwater vulnerability mapping in karstic regions (EPIK). Swiss Agency for the Environment, Forests and Landscape, Bern, 56 pp Ford DC, Williams PW (2007) Karst hydrogeology and geomorphology. Chichester, UK, Wiley Goldscheider N, Klute M, Sturm S, Hotzl H (2000) The PI method: a GIS-based approach to mapping groundwater vulnerability with special consideration of karst aquifers. Z Angew Geol 463:157–166 Hölting B, Haertle T, Hohberger KH, Nachtigall KH, Villinger E, Weinzierl W, Wrobel JP (1995) Konzept zur Ermittlung der Schutzfunktion der Grundwasser berdeckung [Concept for the evaluation of the protective function of the layers overlying groundwater]. Geol Jahrb C63:5-24 Malik P, Svasta J (1999) REKS - an alternative method of karst groundwater vulnerability estimation. Hydrogeology and Land Use Management, Proceedings of the XXIX Congress of the International Association of Hydrogeologists, Bratislava, 79–85, September 1999 Mangin A (1975) Contribution à l’étude hydrodynamique des aquifères karstiques (Contributions to the hydrodynamic of karst aquifers). Thèse, Université de Dijon, 124 pp Pételet-Giraud E, Dörfliger N, Crochet Ph (2000) RISKE: Multicriteria assessment of karst aquifer vulnerability mapping. Application to the Fontanilles and Cent-Fonts karstic aquifers (Hérault, S. France) (in French), Hydrogéologie, 22 pp Ravbar N, Goldscheider N (2009) Comparative application of four methods of groundwater vulnerability mapping in a Slovene karst catchment, Hydrogeology Journal 17: 725–733

55

Vías JM, Andreo B, Perles JM, Carrasco F, Vadillo I (2006) Proposed method for groundwater vulnerability mapping in carbonate (karstic) aquifers: the COP method: application in two pilot sites in southern Spain. Hydrogeol J 14(6):912–925 Vrba J, Zaporozec A (eds) (1994) Guidebook on mapping groundwater vulnerability. International association of hydrogeologists. Heise, Hannover Williams PW (1983) The role of the subcutaneous zone in karst hydrology. In: Back W and LaMoreaux PE (Guest-Editors) V.T. Springfield symposium - processes in karst hydrology. J Hydrol 61:45–67 Zwahlen F (ed) (2004) Vulnerability and risk mapping for the protection of carbonate (karst) aquifers. Final report of COST Action 620. European Commission, DirectorateGeneral XII Science, Research and Development, Brussels

56

Investigation about recharge sources of Bistrica karst spring, the biggest spring of Albania, by means of environmental hydrochemical and isotope tracers Romeo EFTIMI ITA Consult, Rr. Rreshit Çollaku, pll. 10/3/18, Tirana, Albania, e-mail: [email protected]

Abstract: Bistrica karst spring, the biggest spring of Albania, issues at western side of Gjere Mountain karst massif; his mean discharge is 18.4 m3/s. In the eastern side the Gjere Mountain karst massif comes in contact with the abundant gravely basin of Drinos River. From the balance calculations results that only about 70 % of the groundwater resources of the massif could be recharged by the infiltration of the precipitation. Because Bistrica spring is located at elevation about 45 m lower than the mean elevation of gravely basin of Drinos River, it was supposed that the spring partially could be recharged by the gravely basin groundwater. A sampling program was applied during 1989-1990 with some additional sampling during 1996. There were analyzed the oxygen-18, deuterium as well as water chemical analyses were performed. The measurements were made at 6 springs in the karst area, at one borehole of Drinos River valley and at Drinos River. It was calculated that about 60-65 % of the water issuing from the Bistrica spring originate from the infiltrated precipitation in the karst massif and about 30-35 % represent the Drinos valley groundwater seepage into the massif. This result fit very well with the karst water balance study of the Gjere Mountain karst massif. Keywords: karst spring, environmental isotope tracers, environmental hydrochemical tracers

1 Introduction Gjere Mountain is located in south Albania and is one of the biggest karst massifs of the country. In the western side of the basin issues the largest Albania’s spring, Bistrica, with mean discharge of 18.4 m3/s. From water balance investigations it was concluded that the karst water resources of Gjere Mountain basin only partially could be replenished by precipitation. The goal of this research was the investigation of the mechanism of recharge of the Gjere Mountain karst massif feeding the Bistrica spring. Environmental isotope methods combined with hydrochemical investigation and groundwater level observation were used to elucidate the internal hydrology of this karst basin. 2 General characteristics of study area Gjere Mountain karst massif is located in the south-eastern part of Albania, on the border with Greece. The total surface of the karst massif is 440 km2, mostly located in Albanian territory (about 400 km2). The highest picks of Gjere Mountain are 1798 m and 1759 m a.s.l., while the mean altitude is about 900 m a.s.l. The crest of the Gjere Mountain is the natural surface water divide between the Drinos River basin located on the east, and Bistrica River basin located on the west. The geology of the study area, as well as the sampling locations are shown in Fig. 1. Gjere Mountain is an anticline dipping to the east with 25-30ι, while the structure is overthrown to the west. The western tectonic plane deeps to the east angle 40-45ι(Fig. 2)Ǥ

57

Figure 1 Hydrogeological map of Gjere mountain karst massif area 1. Quaternary fluvial deposits 2. Quaternary talus cones 3. Quaternary brachia 4. Palaeogene stratified limestone 5. Cretaceous stratified limestone 6. Jurassic limestone with siliceous rocks 7. Triassic dolomites 8. Paleogene and Neogene flysch 9. Permian-Triassic clay and gypsum 10. Spring, average discharge (a.d.) less than 10 l/s 11. Spring a.d. 10 to 100 l/s 12. Spring a.d.100 to 1000 l/s 13. Spring a.d. more than 1000 l/s 14. Temporary karst spring 15. Contours lines of fluvial gravely aquifer 16. Groundwater flow direction 17. Perennial stream 18. Part of a stream with intermittent runoff caused by the infiltration of surface waters in the river bed gravely deposits 19. Main surface water divides 20. Drilled water well 21. Geological limit 22. Hydrogeological section limit. Note: Numbers on the map show the location of sampling pints. The oldest rocks of the area, the Triassic gypsum-clayey formation, have only a small outcrop in the western side of the Gjere Mountain anticline. The carbonate sequence constitutes most of Gjere Mountain. 58

Figure 2 Hydrogeological cross-section of Gjere Mountain 1. Gravely fluvial deposits 2. Palaeogene-Neogene flysch deposits 3. Palaeogene limestone 4. Cretaceous limestone 5. Jurassic limestone 6. Permian-Triassic gypsum and clay 7. Groundwater level 8. Perennial spring 9. Temporary spring 10. Main groundwater flow direction. It consists of Upper Triassic dolomites, and different limestone rocks of Jurassic, Cretaceous and Palaeogene. The carbonate rocks are surrounded by Palaeogene and Neogene flysch formations (alternating beds of siltstone, claystone and sandstone). Only in the centraleastern side of the Gjere Mountain structure, from Jergucat village at the south to Dervician village at the north (6.5 km) the flysch formation is missing and carbonate rocks contact the Drinos valley gravely deposits. Using the methods described by Turc and Kessler the effective infiltration of the mean annual precipitation recharging the karst groundwater of Gjere Mountain massif is estimated equal to about 1175 mm/year (5.17*108 m3/year, or 16.4 m3/s). The evapotranspiration resulted to be about 573 mm/year (2.25*108 m3/year, or 8.0 m3/s). About 80 % of the karst water resources of the Gjere massif discharge in its western side. In this side of the massif issues Bistrica spring (nr 1-6) with a mean discharge of 18.4 m3/s, as well as some other springs, each with a mean discharge of less than 100 l/s. The largest spring of the eastern side of this massif is the temporary spring Viroi (nr 14), flowing only about 8 to 9 months per year, and having maximal discharge of 40 m3/s. Third important spring, Lista (nr 17) with a mean discharge of 1.7 m3/s issues in the southernmost edge of this massif, in Greek territory. Total discharge of all Gjere karst massif springs is estimated to be about 7.42*108 m3/year, (23.6 m3/s). As can be noticed, the total discharge of the springs of Gjere karst massif, corresponding to a quantity of about 2.26*108 m3/year (7.17 m3/s), is about 30 % larger than the calculated mean efficient precipitation infiltration in the massif,. 3 Environmental tracer approaches Environmental isotopes are most useful in problems related to the origin of water and the dynamics of water systems. Based upon the altitude effect, isotopes may be used for the identification of waters of the potential source areas of recharge (IAEA 1981, 1983, Payne 1978, Eftimi and Zoto 1997, Eftimi 2007). The environmental isotope techniques are being used here together with hydrochemical and to verify a partial replenishment of the karst water resources of Gjere Mountain massif by the gravely aquifer of Drinos valley. The formulation of this hypothesis takes into consideration the good hydraulic connection between the 59

aquifers, the gravely aquifer and the karst one, as well as, the natural groundwater slope to Bistrica spring, which the elevation is about 45 m lower than this of Drinos valley (Fig.2). The sample collection began in January 1988 and continued until December 1999, and some sporadic sampling mostly for investigation of the hydrochemical composition is done during 1996. The samples were analyzed for oxygen-18 and deuterium. Four springs were selected for sampling in the western side of Gjere karst massif; Bistrica on six outlets (points 1 to 6), Vrisi (point 7), Kardhikaq (point 11) and the small Lefterohor spring (point 9). On the eastern side of the karst massif were sampled two springs, Sopot (point 12) and Viroi (point 14) as well as a borehole near Jorgucat village tapping the Drinos valley gravely aquifer (point 20) and the Drinos River near Jergucat (point 21). The sampling locations are shown in Fig. 1. The results of isotope measurements performed on Isotope Hydrology Laboratory of IAEA are presented on Table 1. The analytical errors are 0.1 %o for G18O and 1.0 %o for GD. In Table 2 are presented the mean concentrations of some chemical components. 4 Discussions and interpretation of environmental tracer data 4.1 Isotope data The isotopic composition of all the six outlets of Bistrica spring (nr1-6) is very similar and homogenous, the standard deviation of mean values for each outlet vary within 0.02-0.08 for G18O and within 0.6-1.1 for GD values. This is an indication of a good mixing respectively a long residence time of the karst groundwater of Gjere Mountain massif. The presented on Table 1 G18O and GD values of Bistrica spring are the weighted values of six spring’s outlets. The correlation function between mean G18O and GD values of sampled points and shown in Fig. 3 results in two equations (1) and (2): (1) GD = 7.29 G18O + 13.18 (r = 0. 99) (2) GD = 2.28 G18O – 26.58 (r = 0. 99) Table 1 Mean isotopic composition of water samples of Gjere Mountain karst massif No

Sampling location

Elevation Mean discharge (m) (L/s) (1-6) Bistrica spring 152 18.400 7 Vrisi spring 177 70 9 Lefterohor spring 590 0.4 11 Kardhikaq spring 185 90 12 Sopot spring 1000 1.5 14 Viroi spring 196 0-40.000 15 Jergucat borehole 198 16 Drinos River, Bularat 197 -

Number of samples 34 8 6 8 5 5 3 3

G18O [‰] -7.58 r 0.16 -6.66 r 0.13 -6.95 r 0.18 -6.71r 0.16 -8.26 r 0.12 -7.99 r 0.19 -6.80 r 0.13 -6.80 r 0.13

Deuterium GD excess - [‰] [‰] -43.88 r 2.4 +16.76 -35.10 r 2.0 +18.80 -37.00 r 2.7 +18.60 -36.40 r 1.6 +17.28 +18.88 -47.20 r 0.6 +19.02 -44.90 r 2.9 +12.20 -42.20 r 1.8 +12.40 -42.00 r 1.7

Equation (1) could be considered a local meteoric water line and his slope is close to 8 of the world meteoric water line. The deuterium excess is +18‰ instead of +10‰ of the global meteoric water line (Dansgard 1964). Albania belongs to the east Mediterranean anomalous zone concerning the deuterium axis, determined by Gat and Carmi (1970). Equation (2) describes a mixing line; the mixing of infiltrated into karst massif precipitation and of the groundwater of Drinos valley gravely aquifer is responsible for the isotope composition of Bistrica spring. The infiltrated precipitation into karst massif is represented by the values of intercept of both lines in Figure 3, which practically coincides with the isotope composition of Viroi temporary spring (nr 14) G18O = - 7.99‰ and GD = 44.90‰. The recharged by the Drinos River the gravely aquifer groundwater is effected by the relative GD and G18O enrichment; it seems that the mean recharge area of the Drinos River 60

catchment has lower elevation that this of the Gjere Mountain karst massif. The isotope composition of gravely aquifer groundwater is represented by the Jergucat borehole G18O = 6.80‰ and GD = - 42.20‰.

Figure 3 Relationship between G18O and GD for waters of Gjere Mountain karst massif Based on simple two-component mixing analyses, the Drinos Valley gravely aquifer is estimated to contribute about 30 % to the replenishment of the Bistrica Spring. This is equal to 5.52 m3/s or 1.74*108 m3/year. Not enough springs at different elevations with well-defined recharge areas are known in the study area for the accurate definition of the altitude effect. Figure 4 shows the relation between G18O values and elevation of springs. Four big springs Bistrica, Vrisi, Kardhikaqi and Viroi have approximately the "same" elevation but quite different G18O values. Consequently their data can’t be used to study of the relation between G18O values and elevation of springs. It is considered that the mean recharge altitude (Mra) of the small local springs number 9 and 12 practically coincides with their corresponding elevations. The data of these springs are used to define the altitude effect (Fig. 4), and the prediction of the Mra from the G18O values of the springs can be based on the following equation (3): Mra (m) = -313 G18O – 1585 (3) The altitude effect as defined using equation (3) is -0.32 %o in G18O-values per 100 m. According to literature data the altitude effect for G18O %o - values vary from 0.15 %o to 0.5 %o per 100 m (Payne et al. 1978, IAEA 1981, Leontiadis at al. 1997). Using the relation presented in Figure 4 the Mra of the big springs of study area results as below: Bistrica - 800 m; Vrisi - 500 m; Kardhikaqi - 515 m and Viroi 915 m.

Figure 4 G18O versus altitude of springs 61

4.2 Hydrochemical data Referring to data shown on Table 2, from a hydro-chemical point of view the studied waters can be classified in three groups. To the first group belong the small local springs Lefterohor (nr 9), and Sopoti (nr 12), as well as the big ones Viroi (nr 14) and Vrisi (nr 7). Table 2 Mean of some chemical components of water samples of Gjere Mountain The chemical type is determined only by ions which concentration is more than 25 % me/L No

1-6 7 9 11 12 14 15 16 17

Sampling location

Number of Elevation samples (m) Bistrica spring 34 152 Vrisi spring 8 177 Lefterohor spring 6 590 Kardhikaq spring 8 185 Sopot spring 5 1000 Viroi spring 9 196 Jergucat borehole 3 198 Drinos River, Bularat 3 197 Lista (Greece) 10

T

pH

(°C) 12.4 14.5 12.4 14.3 8.4 11.2 12.8 13.8 12.4

7.64 7.62 7.50 7.53 7.89 7.72 7.56 8.05 8.00

Electrical conductivity (PS/cm) 585 430 337 567 189 371 786 950 270

Chemical type

SO4-(mg/l)

HCO3-SO4-Ca HCO3-Ca HCO3-Ca HCO3-SO4-Ca HCO3-Ca HCO3-Ca SO4-HCO3-Ca SO4-HCO3-Ca HCO3-Ca

135.0 32.2 9.9 121.4 18.1 47.0 258.0 385.0 11.0

Figure 6 G18O versus electrical conductivity for waters of Gjere Mountain karst massif

Figure 5 G18O versus sulphate concentration for waters of Gjere Mountain karst massif

The springs of the first group are of HCO3-Ca type and their sulphate concentration varies from 18.1 to 46.5 mg/L. To the second group belongs Bistrica and Kardhikaq spring. They are of HCO3-SO4-Ca type; sulphate concentration is relatively high, it varies about 121-135 mg/L. To the third group belong the groundwater of Drinos valley (point 20) and the Drinos River (point 21). The water type is SO4-HCO3-Ca and sulphate concentration of five wells of Drinos valley gravely aquifer varies from 175 to 308 mg/L. Very indicative for the relation of hydrochemical and isotope methods of study are the relations of SO4 concentrations and of conductivity values with the G18O concentration (Figures 5 and 6). On both graphs clearly appears that the Bistrica spring stay on the mixing line which mixing end-members are the infiltration of the precipitation in the karst massif and the groundwater of Drinos valley gravely aquifer. Bistrica spring (nr 1-6) and Kardhikaq spring (nr 11) issue in the contact with Upper Triassic clayey-gypsum deposits and are characterised by similar sulphate concentration. One could think that geological implications are responsible for the increased sulphate 62

concentration of both springs. In the same time both springs have significant differences of the concentrations of isotope, sulphate concentration, conductivity and temperature. A careful examination of graphs in Figures 3, 5 and 6 enables us to understand that the origin of sulphate ion at both springs seems to be quite different. While Bistrica spring at all three above mentioned graphs lay on the mixing line, Kardhikaq spring lay far from the mixing line. These facts help to conclude that the sulphate concentration of Bistrica spring results of the mixing of two different recharge sources, one of which has high sulphate concentration. Kardhikaq spring is recharged by the precipitation and the source of the sulphate ion is the gypsum of Triassic deposits. The result obtained with the environmental isotopes for the different recharge sources of Bistrica spring is supported by the hydrochemical observations. The proportion of the mixing of two different waters could be determined comparing the amount of the so-called conservative ions. A typical conservative ion is considered Cl ion but SO4 ion is used also (Eftimi 2006). For the estimation of the mixing proportions contributing to Bistrica spring sulphate ion is used as a neutral ion. Karst massif groundwater mixing component is represented by Viroi spring (point 14), sulphate ion concentration 47.0 mg/L. Because of lack of systematic monitoring, the sulphate concentration of the other mixing component, that of Drinos valley groundwater could not be considered well defined. This could be represented by the Jergucat borehole (nr 20), which mean sulphate concentration is 258 mg/L, but values of more than 300 mg/l are measured in other boreholes. Let remember also that the mean sulphate concentration of the Drinos River recharging the gravely aquifer vary about 350-400 mg/l, but values about 650-700 mg/l are measured also. The sulphate concentrations of two mixing components used for the calculations are 47.0 mg/l for Viroi spring and 300 mg/l for the Drinos valley groundwater. Based on simple two-component mixing analyses, is estimated that the contribution of the Drinos valley gravely aquifer consist about 35 % of mean discharge of Bistrica spring equal to 6.44 m3/s or 2.03*108 m3/s. As could be seen on Table 3, the results obtained by environmental tracer methods, isotope and hydrochemical are well comparable (Table 3). Table 3 Estimated recharge sources of Bistrica spring Method

Isotope Hydrochemical

Recharge sources of Bistrica spring Infiltrated precipitation in the karst basin Seepage of Drinos valley groundwater in the karst massif % m3/s % m3/s 70 12.88 30 5.52 65 11.96 35 6.44

According to the balance calculation of the Gjere Mountain karst massif the total discharge of the springs of the karst massif results about 7.17 m3/s larger than the calculated mean efficient precipitation infiltration in the karst massif. This number is well comparable particularly with the value of the seepage of Drinos valley groundwater in the karst massif obtained by the hydrochemical method. The comparison is justified if all the seeped groundwater of Drinos valley in the karst massif recharges only Bistrica spring, which seems to be a reasonable supposition. 5 Mechanism of groundwater circulation in the karst massif The mechanism of groundwater circulation in Gjere Mountain karst massive is responsible for the formation of main springs of this massif, as well as for the characterisation of the Drinos river valley surface and groundwater. These phenomenons have distinct seasonal character. 63

The period with most intensive infiltration is December-April. During this period the karst groundwater level successively increases and a temporary water divide is created inside the karst massif. The karst groundwater flows at both directions, to the west and to the east (Fig. 2). Beside large Viroi temporary spring (nr 14), many other short term springs appear at the eastern side of Gjere karst massif which duration varies some hours to some days. Yearly amplitude of the karst water level fluctuation of about 35-m is observed in Goranxi cave. Usually by the end of May up to October or November the karst groundwater level starts to decrease steadily. During this period the karst groundwater of Gjere Mountain massif flows mainly to the west, to Bistrica spring. In the eastern side of the karst massif all the springs dry up including also Viroi spring (usually starting from July). The Drinos River surface flow totally infiltrates in the river bed gravely deposits. During the concerned period the piezometric level of Drinos valley gravely aquifer suffers an unusual decrease, also. The yearly level fluctuation amplitude of the gravely aquifer is about 30 m; the biggest one observed in fluvial gravely aquifers of Albania. Water level contours suggest the seepage of gravely aquifer groundwater to the karst aquifer, mainly to Bistrica spring. As the chemical composition of Bistrica spring is very homogenous through the year round one could suppose about a good mixing of infiltrated precipitation into the karst massive and the seepage groundwater from Drinos valley. 6 Conclusions The application of the environmental isotope and hydrochemical methods in the study of the Gjere Mountain karst massif has allowed some important conclusions about the mechanism of groundwater flow in the study area. It is considered that about 60-65 % of groundwater resources of Bistrica spring originates from the infiltration of precipitation in the Gjere Mountain karst massif and about 30-35 % originates from the seepage of Drinos valley groundwater into the Gjere Mountain karst massif. Further systematic investigations are necessary to estimate better the mixing components of Bistrica spring. Acknowledgements The author is greatly indebted to Prof. Hans Zojer and Prof. Josef Zoetl respectively Director and former Director of Institute of Hydrogeology and Geothermie of Graz, and to T. Akiti, former Expert of Isotope Section of IAEA for the useful discussions during the starting phase of the investigation. The IAEA Isotope Hydrology Laboratory is also acknowledged for performing the isotope analyses. References Craig H (1961) Isotope variations in meteoric waters, Science 133:1702-1703 Dansgard W (1964) Stable isotopes in precipitation. Tallus 16:436 Eftimi R, Zoto J (1997) Isotope study of the connection of Ohrid and Prespa lakes, In: International Symposium, "Towards Integrated Conservation and Sustainable development of Transboundary Macro and Micro Prespa Lakes". Korcha, pp 32-38 Eftimi R (2006) Investigation about the recharge sources of Pocemi springs in Albania by means of environmental hydrochemical tracers. In: Sudar M, Ercegovac M, Grubi© (eds) Proceedings of XVIIIth Congress of the CBGA., September 3-6, 2006, Belgrade, pp123-126 Eftimi R (2007) Groundwater circulation in two transboundary carbonate aquifers of Albania; their vulnerability and protection. In: Witkowski AJ, Kowalczyk A, Vrba J (eds) Groundwater Vulnerability Assessment and Mapping, Internat. Conf. Ustron, Poland 2004, Taylor & Francis/Balkema, pp 199-211 Gat JR, Carmi J (1970) Evolution of isotopic composition of atmospheric waters in the Mediterranean Sea area. J. Geophys 75:30-39 64

Gat JR, Dansgaard W (1970) Stable isotope survey of the freshwater occurrences in Israel and the Jordan Rift Valley, J. Hydrol. 16:177-212 International Atomic Energy Agency - IAEA (1981) Stable isotope in hydrology, IAEA. Tech Rep. Ser., No 210, pp 339 International Atomic Energy Agency - IAEA (1983) Guidebook on Nuclear Techniques in Hydrology, IAEA. Tech Rep. Ser., No 91, pp 439 Leontiadis LL, Smyrniotis CH, Nikolaou V, Georgiadis P (1997) In: Günay G, Johnson I (eds) Karst Waters, Balkema, Rotterdam, pp 239-247 Payne BR, Leontiadis LL, Dimitrulas I, Dounas A, Kallergis G, Morfis A (1978) A study of the Kalamos Spring in Greece with environmental isotopes. Water Resources Research 14(4):653-658

65

Importance of transboundary karst aquifer resources in South Eastern Europe (SEE) Jacques GANOULIS1, Alice AURELI2 , Neno KUKURI3 1

UNESCO Chair /The International Network of Water-Environment Centres for the Balkans (INWEB) Aristotle University of Thessaloniki, 54124 Thessaloniki, Greece e-mail: [email protected] 2 UNESCO International Hydrological Programme (IHP) 1, rue Miollis, 75732 Paris, France 3 International Groundwater Resources Assessment Centre (IGRAC) P.O. Box 85467, 3508 AL Utrecht, The Netherlands

Abstract: Transboundary aquifer resources play a major role in SEE as sources of freshwater. Indeed, 65 Transboundary Aquifers (TA) were identified in the region in an inventory developed in 2007 by the UNESCO Chair/INWEB at the Aristotle University of Thessaloniki, in cooperation with UNESCO/IHP, IGRAC and the United Nations Economic Commission for Europe (UNECE). Two main types of TA were distinguished: (1) karst aquifers ranging from a few tens to hundreds of square kilometres, and which generate major karstic springs, and (2) alluvial aquifers with greater areal extent, up to some thousands of square kilometres. In some countries transboundary karstic groundwater covers between 60 and 100% of total water use (e.g. in countries located in the Dinaric region), while for transboundary alluvial groundwater the proportion varies from 15 to 70% (e.g. along the river Danube). TA in SEE, and especially those which are karstic, are highly vulnerable to pollution from different pressure factors (agriculture, industry, mining, sewage/waste disposal and tourism). In this paper, the WEB-based metadata inventory on transboundary karst aquifers in SEE is described. This inventory forms the first step towards implementing the UNESCO/ISARM (Internationally Shared Aquifer Resources Management) programme in the Balkans (SEE). This programme uses a multidisciplinary methodological approach and is based on an effective cooperation mechanism between countries, in order to reduce groundwater and ecosystem vulnerabilities and contribute to the sustainable management of transboundary karst groundwater resources in SEE. Together with the Global Environmental Facility (GEF) and other partners the cooperative project DiKTAS (Dinaric Karst Transboundary Aquifer System) was formulated specifically for the Dinaric area. Keywords: karst, transboundary, aquifer, inventory, South Eastern Europe

1 Introduction As water availability is constantly declining in many parts of the world, transboundary water resources, both surface and groundwater, are becoming increasingly important sources of freshwater. Data from different countries, and especially from those with an arid or semiarid climate (Revenga et al. 2000), indicate serious issues of water scarcity and water stress. Between the years 1970 and 2000 the average water volume per capita decreased by almost 30% and by the year 2025 is expected to have fallen even further to below 1700 m3/y/cap, which is the limit below which a situation of water stress exists. This is because of population growth and the overuse of water, mainly in agriculture and ambitious agricultural development projects. However, the figure above is a global average and does not reflect the uneven distribution of water availability between continents and particular regions, or show where countries are already under “water stress” and experiencing water scarcity. It was 67

estimated that in the year 2000 almost 40% of the world’s population, or about 2.3 billion people, were living in water basins with less than 1700 m3/y/cap, i.e. under water stress. This percentage is expected to have reached 50% by the year 2025 (Revenga et al. 2000). The figures above underline the importance of transboundary water resources, which for many countries are essential sources for drinking, agricultural irrigation and other purposes. Internationally shared river catchments affect 40% of the global population and cover about 45% of the total land on earth. In Europe this percentage is about 54% and in SEE about 90% (World Bank, 1987). Some countries receive almost all their surface water from outside their international borders (e.g. about 98% for Egypt). As can be seen from part of the inventory compiled by INWEB in 2008 (Figure 1), 17 Sub-Danubian transboundary river and lake basins have been identified in SEE.

Figure 1 Sub-Danubian transboundary surface water basins in SEE (INWEB 2008) Internal and external water resources availability is highly variable between countries in SEE, as shown in Figure 2.

68

Figure 2 Water resources availability in some SEE countries (World Bank 1987) Transboundary groundwater aquifer resources are also very important in many regions such as North Africa, Middle East and SEE. This paper focuses on transboundary groundwater resources in SEE and mainly on internationally shared karst aquifers in the region. 2 The UNESCO/ISARM/Balkans Programme 2.1 Programme structure In order to promote studies on transboundary aquifers, the UNESCO General Assembly decided in 2002, that as part of the UNESCO IHP, it would launch an intergovernmental initiative called the “Internationally Shared Aquifer Resources Management” programme (ISARM) (UNESCO/ISARM 2001). The programme has identified the following five key areas for the sound management of transboundary aquifer water resources: x scientific-hydrogeological approaches, x legal aspects, x socio-economic issues, x institutional considerations, x environmental protection. Each of these areas not only has different aims but is also managed by different groups or bodies as shown in Table 1. The UNESCO ISARM programme has been implemented in different parts of the world. The first phase was initiated in Africa in 2002, the second phase in the American continent in cooperation with OAS (Organisation of American States) in 2003, and the third phase was launched in the Balkans by the UNESCO Chair INWEB in cooperation with UNESCO IHP and the International Association of Hydrogeologists/Transboundary Aquifer Resource Management Commission (IAH/TARM) in 2004 . INWEB held a workshop in Thessaloniki in October of that year to present and assess its first results (see www.inweb.gr).

69

Table 1 Type of scope, aims and target groups concerned with managing shared aquifer resources Type of scope

Aims

ScientificTo support the development of Hydrogeological/ national and regional management Technical/Technological policies and strategies To prevent groundwater pollution, Environmental environmental degradation and loss of biodiversity To ensure endorsement by Legal/Political governments and international partners and minimise/prevent conflicts (national and regional) Institutional and SocioTo ensure endorsement, appropriate economic implementation and sustainability of actions

Target groups Regional scientific and research institutions, researchers, policy makers Environmental scientists, researchers, policy makers Governments, users and international partners

Policy makers, the public and international partners

2.2 Methodology The main objective of the programme was to extend the inventory of the Balkan’s transboundary aquifer resources. This refers to: x geographic, hydrogeological, environmental and socio-economic data x information on water policies (international agreements, national institution setting, projects and critical problems to be addressed). The output took the form of a situation analysis for transboundary groundwater shared by two or more of the following countries: Slovenia, Croatia, Romania, Serbia, Bosnia and Herzegovina, Montenegro, the Former Yugoslav Republic of Macedonia, Albania, Bulgaria, Greece and Turkey. The UNESCO Chair INWEB also cooperated closely with UNECE: Working Group on Monitoring & Assessment, Switzerland, to follow up the European study they had compiled in 2000, as well as with the UN Economic and Social Commission for Western Asia (ESCWA) and the Observatoire du Sahara et du Sahel (OSS) for the Mediterranean inventory. The inventory of transboundary aquifer resources in SEE is available in a revised form on INWEB’s web site (http://www.inweb.gr/) (INWEB 2008). In order to collect the data a questionnaire was prepared and sent to UNESCO/IHP national committees. At the same time, the UNECE-Working Group on Monitoring & Assessment, Switzerland followed up their European assessment of the status of transboundary groundwater in SEE. This revised inventory aimed to set out the scale and scope of transboundary groundwater in the region. to assess their importance in satisfying demand for water, to examine the pressure factors exerted on them, to provide information on status, trends and impacts in relation to both water quantity and quality, and to describe the management measures being taken, planned or needed to prevent, control or reduce negative impacts on transboundary groundwater in the region. With the support of UNESCO IHP a workshop to review the draft of the assessment was held in Thessaloniki in April 2007 (see www.inweb.gr).

70

2.3 Results Many transboundary groundwater bodies in the region had been identified years ago, and had been listed in the earlier UNECE and INWEB inventories. However, because SEE has seen major conflict and political change in the last fifteen years, aquifers and groundwater that had previously been located within a single country are now shared between different, newly formed countries. Thus while the previous UNECE inventory had recorded 23 transboundary aquifers (TA) in the region and the draft INWEB inventory had reported 47, this latest assessment identified 65Two main types of TA were distinguished: (1) karst aquifers ranging from a few tens to hundreds of square kilometres, and which generate major karstic springs, and (2) alluvial aquifers with greater areal extent, up to some thousands of square kilometres. The locations of these aquifers are shown in the overview map in Figure 3.

Figure 3 Overview map of transboundary aquifers in SEE (INWEB 2008) Furthermore, the Balkans programme aimed at exploring transboundary karst and porous aquifers in the region by country, and presenting data and information for comparative purposes. The importance of karst transboundary aquifers by country is given in Figure 4 and Figure 5.

71

Figure 4 Total number of transboundary and karst transboundary aquifers in the SEE countries

Figure 5 Proportion of karst transboundary aquifers (as % of the total) in the SEE countries Groundwater use Transboundary groundwater resources play a significant role in SEE. The physical attributes of the region i.e. the geology, topography and major catchments, is such as to promote the occurrence of productive aquifers. These aquifers are of two distinctive main types, the limestones of the karstic type area of the Dinaric coast and its mountainous hinterland, and the thick alluvial sedimentary sequences of the Danube basin. In some 72

locations the alluvial sediments overlie and are in hydraulic contact with the limestones or comprise relatively thin aquifers in river or lake sediments overlying ancient metamorphic rocks. The geographical distinction between the two main aquifer types, and the fact that much of the national borders of several of the countries of the region are traversed by transboundary groundwater, can be clearly seen in the map in Figure 3. Transboundary karstic groundwater aquifers were reported to provide 60 to 80% of total water usage in their respective areas, and some of the Dinaric karstic aquifers of Bosnia, Serbia, Croatia, Montenegro and Albania provide as much as 90 or even 100%. In terms of numbers, the importance of transboundary karstic aquifers is shown in Figures 4 and 5. Compared to surface waters, the range of use of alluvial aquifers is much greater, varying from only 15 to up to 70% for the important Banat, Baka and Srem alluvial aquifers along the River Danube in Serbia, Croatia and Hungary. Pressure factors The majority of transboundary aquifers, except for those located in remote, sparsely populated areas, are very vulnerable to anthropogenic pollutants emitted from both point and non-point sources. Karstic aquifers, with their lack of soil cover and rapid flow paths leaving little time for attenuation, are almost invariably classified as highly vulnerable. Alluvial aquifers are also likely to be vulnerable, unless they contain a high proportion of clay-rich material to reduce their permeability, are overlain by a protective confining layer of clays and/or the water table is relatively deep. The transboundary groundwater of the SEE region are likely, therefore, to be highly vulnerable to pollution if the pressure factors caused by agriculture, industry and tourism described below produce significant loadings of mobile and persistent pollutants. In general, both alluvial and karstic aquifers have reported groundwater quality problems. Of the questionnaires received, only a few specifically reported that there were no groundwater quality issues at all. Agricultural activities put major pressure on freshwater systems in SEE in terms of both quantity and quality. Some 70% of overall water use is for agriculture and severe problems can result when this heavy usage depends on groundwater abstractions. Moreover, whether or not the land is irrigated, intensive cultivation invariably means the heavy use of fertilisers and pesticides. Intensive cultivation and animal production can produce increased levels of nutrients and pesticides in groundwater from infiltrating surface run-off from agricultural land, leaching from the soil through the unsaturated zone and sometimes from return waters from irrigation channels. In contrast to the amount of pressure caused by agricultural activities, the pressure factors on transboundary groundwater in the region caused by industry appear overall to be quite limited. However, in the summer months tourism creates a huge demand for drinking water and recreational activities can also constitute a pressure. 3 Conclusion The implementation by the UNESCO Chair/INWEB of the first phase of the UNESCO/ISARM programme in SEE during the period 2003-2008 resulted in the development of a regional inventory indicating the main hydrogeological and socio-economic characteristics of transboundary aquifer resources in the region. A Web-based interactive database combining the Google map and Google earth technologies is available on INWEB’s WEB site (www.inweb.gr). Two main types of transboundary aquifers were identified: (1) transboundary karst aquifers and (2) alluvial porous aquifers. Transboundary karst aquifer resources in SEE are important sources of freshwater for different purposes and mainly for drinking water supply. In some countries, like Croatia, Bosnia and Montenegro, transboundary karstic groundwater 73

provides as much as 90% of water needs. In many countries, groundwater drains into rivers. In Slovenia, Croatia, Bosnia-Herzegovina, Serbia, Montenegro and Albania about half of the surface water drains underground into karst aquifers and springs and flows into the Adriatic Sea. This is the Dinaric karst aquifer system, which is very important for the ecological biodiversity and the socio-economic development of the region. At the Thessaloniki UNESCO/ISARM workshop in 2004 it was decided that the Dinaric karst aquifer system should be given first priority as a characteristic case study in SEE. The DiKTAS (Dinaric Karst Aquifer System) project on the “Protection and Sustainable Use of the Dinaric Karst Aquifer System” was formulated as a potential GEF (Global Environmental Facility) project during 2006-2007 and after approval by the GEF Council a preparatory phase was executed by UNESCO and UNDP in 2008-2009. The full size project involving Albania, Bosnia-Herzegovina, Croatia and Montenegro is expected to be implemented between 2010 and 2014. The results obtained during the first phase 2000-2008 of the UNESCO/ISARM programme, the lessons learnt and the key requirements that resulted from different case studies around the world will form the base for the second phase of the programme. This will be implemented in the coming decade and will further promote scientific knowledge, capacity building and aide to policy formulation in order to increase effective and sustainable use and protection of the world’s precious transboundary aquifer resources. References Ganoulis J (2007) Integrated Management of Transboundary Aquifers in Southeastern Europe. (A report within GEF IW: LEARN Activity D2). GWP-Med, UNESCO Chair and Network INWEB, Thessaloniki INWEB (2008) Inventories of Transboundary Groundwater Aquifers in the Balkans, UNESCO Chair and Network INWEB, Thessaloniki, Greece http://www.inweb.gr/ Revenga C, Brunner J, Henniger N, Kassem K, Payne R (2000) Pilot Analysis of Global Ecosystems: Freshwater Systems. Washington, DC: World Resources Institute UNESCO/ISARM (2001) A Framework Document. Paris, UNESCO, Non Serial Documents in Hydrology WORLD BANK (1987) Water Resources Management in South Eastern Europe, Volume I, Issues and Directions

74

The resources, environment and development in Fengshan Geopark karst area Jiang Guanghui GUO FANG Karst Dynamics Laboratory, Ministry of Land and Resources, Institute of Karst Geology, CAGS, No.50, Qixing Road, Guilin 541004, China, e-mail: [email protected]

Abstract: A remote place named Fengshan in karst of south China is changing quickly. Practices which improve life of local people and environment can be referenced by other karst regions. The policy includes much work such as traffic, education, health, energy, environment, tourism, and drinking water, which are all necessary and eager to be resolved. The character of karst is research in order that the karst resources and environment are well managed by the government. The environment suffers from soil erosion, flooding, water shortage and desertification. Three types of soil erosion are related to karst: allogenic soil erosion, autogenic soil erosion, and conduit soil erosion. Flooding in karst becomes more serious for sediment in conduit, however tributary withstanding may also lead to flooding. Epikarst flow is stored for drinking by a small domestic water cellar. Biogas is popularized to stop deforestation. Splendid karst landscape is protected in Geopark. It also attracts tourists. Keyword: peak cluster; epikarst; karst engineering

1 Introduction Karst environment has not been definite strictly. It has wide area and plenty of study objects. In common sense karst environment is made of soil, karst landscape, atmosphere, karst water, and biosphere (Yuan et al. 1988). The five parts connect each other, and the combination of them is different in special. Many environmental problems have been discussed in academe (Yuan 1988, De Waele 2008). Deforestation, soil erosion (Turnage et al. 1997), desertification (Hu et al. 2004, Wang et al. 2004), flooding and drought (Andriani 2008), collapse, water pollution (Kacaroglu 1999) occurs frequently in karst area. Although these problems are related to fragile of karst environment, impact of human activity is important. This paper gives reasons and answers of soil erosion, flooding, water shortage, poverty and deforestation in a subtropical karst region. 1.1 Study area Fengshan lies in west of Guangxi province, China (Fig. 1). It has been a national geopark in 2005 and is being a world geo-park for its beautiful, typical, and huge karst landscape including karst springs, collapse depressions, caves, natural bridges, karst windows. However poverty stands here for many years and series of environmental problems affect local people. Fengshan karst area is within subtropical climate, with annual precipitation of 1550mm, annual temperature of 19 degree. The elevation changes from 700m to 1100m. It is an anticline in geological structure. Triassic sandstone closes Permian and carboniferous limestone. Plenty of allogenic water coming from sandstone areas is important for karst formation. The area of peak cluster is 400 km2. Karst conduits are about 90km long, and half of them are measured. Karst groundwater is drainaged from Poxin spring with base flow discharge of 4.2 m3/s. Economy of Fengshan County is undeveloped. The establishment of Fengshan Geopark is believed to bring hopes for the County; however karst environmental management is important for sustainable development. 75

Figure 1 Location of Fengshan in Guangxi 1.2 Life in karst area There are few good land resources and little drinking water in karst mountain area. But many people choose living here. Most of land in Fangshan is covered by shrub and forest. Cultivated field distributes in bottom and slopes of depressions. Small paddy field lies in a few poljes where soil and allogenic water are sufficient. People live on cultivated field. Their agriculture activity induces environmental problems. 2 Results 2.1 Environmental problems Soil erosion Slope cultivation and deforestation induce soil erosion. Soil erosion is accelerated by steep sloops of peak cluster and monsoon climate. It leads to series of problems such as destroying farmland, blocking conduits, reducing soil quality and forming rocky desertification. This is mostly an irreversible process. Eroded slope is covered by little soil and much stone teeth .It can’t feed any plant but grass. Three types of soil erosion are related to karst: allogenic soil erosion, autogenic soil erosion, and conduit soil erosion. Allogenic soil erosion is caused by allogenic rivers, autogenic soil erosion is made by autogenic flow, and conduit soil erosion is induced by conduit flow (Fig. 2). Poljes or depressions near non-karst area usually suffer from allogenic soil erosion. Allogenic soil erosion leads to debris flow which buries farmland. The suspended particles are made of clay, sand and scree. The debris flow enters into swallow hole and suspended particles deposit somewhere which makes the conduits blocked. Once it happened flooding will occur in the poljes. Cropland in several poljes has been destroyed by iterative flooding. Autogenic soil erosion happens after heavy rain. Overland flow, epikarst flow erodes soil in slope and bottom of depression. Intensity of erosion is related to rainfall, vegetation, slope degree. Cultivation on slope will make erosion quicker. Suspended particles related to autogenic soil erosion are mainly clay and some block of rock. They stop moving at the bottom or near sinkhole. Autogenic soil erosion induces rocky desertification in south of China. A depression above conduit may suffer from conduit soil erosion. Overflow in rainy season from sinkhole or karst windows at bottom of the depression become an intermittent river. The intermittent river takes away soil and forms a deep valley at the bottom of 76

depression. The elevation of the depression has decreased about 5m in the past 20 years. Land has been destroyed gradually (Photo 1).

Figure 2 Soil erosion in karst aquifer, three types of soil erosion is related to karst, the first is allogenic river related, the second is autogenic flow related, and the third is conduit flow related Flooding Flooding in depressions and poljes makes huge damage. This hydrological phenomenon is related with conduits of karst aquifer. Allogenic water enters into karst aquifer by swallow hole in poljes. However it stops to form a lake around the hole when the conduit is too small to permit such large flow to pass after intensity rainfall. Conduits which enlarged by karstification can be blocked by sediment. Blocking of conduit is a really reason for flood. Then finding the location of block is all-important for opening conduits. Roughness of conduits impacts the capacity of conduit discharge. Wall of conduit should be smooth with wash of water, but stalagmite and sediment of mash, sand or gravel in the floor increase the roughness. On the other hand gravel moving together with water makes the flow slower. Moreover sediments usually reduce the hydraulic radius of the conduit. So except for blocking conduit at some points, sediments reduce flow by increasing roughness or decreasing hydraulic radius. Shima polje is the largest poljes which suffers from flooding. There was ten million tons of capacity in 1976 and 500 acres of field couldn’t be used. The flood comes from rivers in sandstone area with the area of 44 km2. The conduit can’t be examined for finding block pots because it is always filled. A tunnel had been built for drainage, however flooding can’t be removed completely. Moreover flooding time becomes longer and longer from 1970s to now. Shima is becoming a real lake. Someone thinks that it is better for Shima to being a lake or reservoir than returning to fields.

77

Flood in Shima polje had been studied in 1970s. A tunnel was built to increase the discharge of conduits. However the conduits were still be blocked somewhere unknown downstream. So the flood hasn’t been solved completely now. There are six connected swallow holes in Shima polje. Water level in these holes was monitored for finding the blocks of conduits which connected the swallow holes. It was found that there was 3 m different in elevation between Ping cave and others when only Ping cave accepted allogenic rivers, but no difference when all swallow holes accepted rivers. So it could be concluded that there was block between Ping Cave and others, however some block unknown still existed downstream. The tunnel was built connecting Ping cave and others. It discharged small flood but remained large flood unsolved (Fig. 3, Fig. 4).

Photo 1 Soil erosion in karst area, A: paddy field buried by allogenic soil erosion; B: big blocks in allogenic soil erosion; C: slope eroded by autogenic soil erosion, only stone teeth remained; D: depression eroded by conduit soil erosion

Figure 3 Swallow holes in Shima polje, a tunnel was built to discharge flood 78

Figure 4 Water level in swallow holes, which indicates location of blocking in conduit Collapse in conduit reduces flood. A tunnel was built in collapsed detritus for discharge flood in Dongni depression in 1960s. The collapsed detritus was formed below a cliff. It is the erosion of intermittent river in the depression which makes collapse of conduit. A new sinkhole formed after collapsing, however which can’t discharge the flood by time. The tunnel in collapsed detritus was very effective, and flooding in depressions upstream was even diminished (Photo 2). Tributary of conduit may also reduce flood. It is found that flooding occurs along the tributary in north, while there is little flooding in the south tributary. The south tributary is recharged by larger allogenic rivers. Moreover it has shorter length and bigger hydraulic grads. So the flood in south tributary comes earlier than the north, which may enhance the head in the main conduit and prevent flow in the north conduit (Fig. 5). Water shortage Annual discharge of Poxin spring is 6 m3/s. Its area is 803 km2 with about half of carbonate rock and half of non-carbonate rock. Allogenic rivers from non-carbonate rock recharge karst aquifer by swallow holes. Rainfall recharges karst aquifer through sinkhole and fractures. There are little rivers in karst area, while karst groundwater is plentiful. However it is difficult to use karst groundwater, because of its heterogeneous. Finding groundwater is mostly impossible in most area except some karst windows and sinkholes. Local people far away from groundwater have to build small reservoir for harvesting rainfall.

Figure 5 structure of conduit in Poxin, tributary conduit is recharge by allogenic rivers 79

Photo 2 flooding in karst area and engineering for discharge, A: tunnel in Dongni for discharge flooding in the depression; B: flooding impacts paddy field in polje Thousands of small reservoirs have been built in the past ten years. Almost every family own one reservoir, and some of them build one big for several families. Water shortage for life is resolved by this way, moreover people build reservoir near their field so as to irrigation. By these means corn field is changed into paddy field. They eat ice produced by them, which is so imagined and exciting for them. They even want to feed fish in reservoir, and it is really a magic idea. Water quality in the reservoir is possibly polluted by bacteria and turbidity under high temperature and heavy rainfall. Treatment is necessary before water entering, and the environment surrounding of the reservoir is also important. The sources of reservoir water are overland flow, epikarst flow and rainfall. Enough catchment area of reservoir is important. Epikarst flow has better quality and longer life, so it is best source. Plant above reservoir is important too. Forest can clean water and cut flooding peak (Photo 3).

Photo 3 series-wound water cellars for harvesting rainfall, epikarst flow and overland flow in karst area, a piece of paddy field among corn is irrigated by the cellars

80

2.2 Resources Epikarst water The perched saturated zone near surface in peak and depression is called epikarst. Epikarst is characterised by intensified dissolution and corresponding high porosity because of high CO2 concentration in soil and low saturation of rainfall. Somewhere small springs are formed in the process of karst evolution. These springs flood soon after rainfall, while base flow continued several months. But their discharge usually attenuates to zero in dry season, so a small reservoir is needed which storages flood for dry. Plant Many kinds of plant in karst area are use for herb. Some of them are only adapt to karst environment. Wild herb is precious and expensive, so it has been over picked. Government organizes people to crop herb for business. This work not only brings money for local people but also does well to wild vegetation protection. Tourism Fengshan becomes famous because of the establishment of national geopark. The geopark includes huge and long caves, big karst windows, splendid stalagmite, and long life old. There is the longest life old in the world. The secrecy of long life is its special geological environment. Several viewpoints have come into being, and some of them have been commercial parks, such as Sanmenhai and Shuijinggong. Energy Although soil is thin, rocky slope can still be covered by shrubs and bushes if there is no cultivation. Production of vegetation is much because sun light and rainfall is enough. These grass or shrub feed pigs, goats horses and cows. Government organizes technician to help people to build biogas digester. It is very effective for environmental protection. Biogas will replace firewood as family energy. Poverty elimination Fengshen is poverty-stricken county, where per capita income of people is below 150 EUR. Government spends much money in traffic, education, health. Tourism has been developing since the beautiful karst landscape is found. Orchard brings hope for local people, and return of the large investment will gradually come into being in the future. Practices for poverty alleviation should be environment-friendly. Best management practices are needed for environmental protection during developing. Soil erosion should be controlled according to its character. More strict vegetation protection practices are necessary for non-karst area, where main sediments come. A buffer zone near swallow hole can intercept big sediment from allogenic rivers. Conversion of cropland to forest in slope may reduce rocky desertification, which also control autogenic soil erosion. Conduit flow soil erosion can be prevented by building concrete bank. It is difficult to open once conduit is blocked. Maybe erosion of turbulence flow can naturally remove sediment if there is small new sediment from surface. Removing stem from conduit by engineering technique is effective sometime, but which is impossible when the conduit is long and filled by water. Opening a new tunnel may be useful but only when proper location is determined. Storage of epikarst flow is a cheap way to solving water shortage. A small protection zone for reservoir is necessary.

81

3 Conclusions Poverty elimination work in karst area of South China should be done with proper ways. Traditional business can not be encouraged because it is inefficient and unsustainable. New business for example biogas, reversion cropland to forest, and tourism not only does well to environmental protection but also enhance earning for local people. Karst environment is characterised by thin soil, little water resources on surface, and long underground conduit net. It is very fragile. Cultivation and deforestation intensify soil erosion; soil erosion induces flooding and water shortage, and then induces poverty; poverty induces cultivation and deforestation. This is a vicious circle, and should be replaced by sustainable way. Reference Andriani GF, Walsh N (2008) An example of the effects of anthropogenic changes on natural environment in the Apulian karst (southern Italy). Environmental Geology, DOI 10.1007/s00254-008-1604-6 De Waele J (2008) Evaluating disturbance on Mediterranean karst areas: the example of Sardinia (Italy). Environmental Geology, DOI 10.1007/s00254-008-1600-x Hu Baoqing et al. (2004) Design and application of dynamic monitoring and visualization management information system of karst land rocky desertification. Chinese Geographical Science 14(2):122-128 Kacaroglu F (1999) Review of groundwater pollution and protection in karst areas. Water, Air and Soil Pollution 113:337-356 Turnage KM et al. (1997) Comparison of soil erosion and deposition rates using radiocesium, RUSLE, and buried soils in dolines in East Tennessee. Environmental Geology 29(1/2):1-10 Wang Lachun et al. (2004) Karst environment and eco-poverty in South-Western China: a case study of Guizhou Province. Chinese Geographical Science 14(1):21-27 Yuan Daoxian (1988) Environmental and engineering problems of karst geology in China. Environmental Geology 12(2):79-87 Yuan Daoxian, Cai Guihong (1988) The science of karst environment. Chongqing: Chongqing publishing house, pp 23

82

Linking generic models to site-related models of conduit evolution Bernhard HUBINGER, Christoph REHRL, Steffen BIRK Institute for Earth Sciences, Karl-Franzens-Universität Graz, Heinrichstr. 26, 8010 Graz, Austria, e-mails: [email protected]; [email protected]; [email protected]

Abstract: The hydrogeological characterization of karst aquifers can be supported by the analysis of the processes involved in the evolution of solution conduits. To this end, a numerical model coupling flow and dissolution processes is employed for simulating conduit development. Both generic models representing hypothetical carbonate environments and site-related models referring to the gypsum karst settings of the Western Ukraine are considered. The generic models comprise networks of interconnected protoconduits with spatially varying initial diameters of about one millimetre and less. To identify the influence of hydrological conditions on conduit evolution different hydraulic boundary conditions are considered in the simulations. If the maximum flow rate in the karst aquifer is not strongly limited a stable bimodal aperture distribution is found to develop; only a limited number of conduits continue to grow while the other apertures stay small. The number of large-sized conduits tends to decrease with decreasing maximum flow rate. However, if flow rates are strongly limited uniform aperture distributions are obtained. The principles revealed by the above-described generic simulations are confirmed by site-related models representing the multi-storey artesian settings of the gypsum karst terrain of the Western Ukraine. In this type of setting, soluble units are supplied with chemically aggressive water from beneath. Due to the development of solution conduits a hydraulic connection between the lowly soluble aquifers underlying and overlying the soluble unit is established. At this late stage of conduit development, the maximum flow rate is controlled by the regional boundary conditions and the permeability of the aquifers. If the permeability is sufficiently high conduit development is found to be competitive, thus leading to bimodal aperture distributions. In a low-permeability formation, however, conduit development is found to be less selective due to the limitation of flow through the conduit system. Thus, multiple pathways develop and the frequency distribution of conduit apertures appears to be unimodal rather than bimodal. Keywords: karstification, speleogenesis, aperture distribution, numerical model

1 Introduction Karst waters are highly vulnerable to contamination due to the rapid spreading of pollutants in solution conduits. Effective strategies for management and protection of water resources in karst terrains thus must be based on reliable information about the properties of the karst conduit system. Investigations of processes involved in the evolution of solution conduits represent one approach to support the hydrogeological characterization of karst conduit systems. In this work, a numerical model coupling flow and dissolution processes is employed for simulating conduit evolution in both generic and site-related model settings. The purpose of the modeling is to provide insight into the interrelation between the properties of the evolving conduit systems and environmental factors, such as the hydraulic boundary conditions and the permeability of the rock formation. 2 Modelling approach Conduit evolution is simulated using the discrete tube network implemented in the modeling tool CAVE (Clemens et al. 1996, Liedl et al. 2003). Flow rates in the individual 83

tubes of the network are calculated using the Hagen-Poiseuille or Darcy-Weisbach equation for laminar and turbulent flow, respectively. At the nodes of the network water is mixed and routed downgradient. After each time step the tube diameters are increased to account for the dissolution of rock. The dissolution rate F is calculated by (Svensson and Dreybrodt 1992, Dreybrodt et al. 1996, Liu and Dreybrodt 1997, Eisenlohr et al. 1999, Jeschke et al. 2001) F = k( 1 

c n ) c eq

where ceq is the equilibrium concentration with respect to the dissolved mineral. The rate constant k is generally dependent on factors such as type of rock, flow conditions, etc. The exponent n equals unity if the dissolution process is diffusion controlled, i.e. if the ratelimiting step is the diffusion of the dissolved species from the conduit wall into the mobile conduit water. If the dissolution rate is limited by the surface reaction, n is a positive number that has to be determined experimentally. Both in limestone and gypsum a switch from n | 1 to n > 1 is observed at high relative saturation states. In the following simulations, we use typical values provided by the aforementioned references. 3 Model application and results 3.1 Generic settings The generic models comprise tube networks representing interconnected protoconduits in limestone with spatially varying initial diameters. The networks are realized by regular grids of 200 m length and width consisting of 20 x 20 nodes connected by 722 water-filled tubes. The initial apertures are spatially uncorrelated and log normally distributed with a mean of 0.5 mm and a standard deviation of 0.1 mm. The hydraulic head is given at two opposite sides with a difference of 5 m, and no-flow boundaries at the other sides. The parameters used for calculating the dissolution rates are shown in Table 1. Table 1 Model parameters used for generic simulations of limestone dissolution Symbol ceq cs

Description calcium equilibrium concentration switch concentration

Unit

Value

mol m-3

2

-3

0.9 · ceq = 1.8

-2 -1

mol m

k1

surface reaction rate constant (n = n1)

mol m s

4 · 10-7

k2

surface reaction rate constant (n = n2)

mol m-2 s-1

4 · 10-4

n1

dimensionless exponent for c < cs

-

1

n2

dimensionless exponent for c  cs

-

4

All generic simulations have been carried out for 30 realizations of the assumed initial aperture distribution. It is generally found that different runs lead to similar results. Simulation results of typical scenarios are shown below. To identify the influence of hydrological conditions on conduit evolution different hydraulic boundary conditions are considered in the simulations. To account for the limited availability of flow inherent in any type of hydrogeological environment the initial hydraulic gradient is reduced with ongoing conduit development such that a predefined maximum flow rate is not exceeded at the nodes on the inflow side. Thus, the fixed head at an inflow node is 84

replaced by a fixed flow rate if the fixed head results in too high flow rates. For this investigation the total maximum inflow rate was varied in steps of one order of magnitude from 102 m3 s-1 to 10-11 m3 s-1. In a first step, the results of a weakly limited total maximum inflow rate of 0.1 m3 s-1 are discussed. This value is sufficiently high that breakthrough can occur in the system. This means that at least one pathway develops where the solute concentration is below the switch concentration along the entire length from the inflow to the outflow boundary and thus dissolution rates are governed by the fast first-order kinetics. This leads to competitive conduit development: The conduits along this pathway continue to grow rapidly while others stay small or become connected to the already developed preferential pathway. The latter only happens if the amount of inflowing water is still sufficient to allow further conduit development. Figure 1 depicts this situation when two separate preferential flow paths with larger conduits evolved and join together. The cumulative frequency distributions of conduit apertures reveal that a stable bimodal aperture distribution evolves after quite a short time (Figure 2, left panel). Afterwards only the already strongly widened conduits continue to grow further.

Figure 1 Typical pattern of generic settings with weakly limited inflow rate after breakthrough. The main flow direction is from left to right. The circles represent nodes, the arrows conduits. The grey shadings indicate the dissolution kinetics at the outlet. Each arrow points towards the node with the lower head showing the flow direction in the tube. Arrow thicknesses are scaled to the conduit apertures and tildes denote turbulent flow regimes.

85

A comparison of scenarios with different maximum flow rates shows that the relative number of large conduits evolving in a given system decreases with decreasing maximum flow rate, i.e. with decreasing availability of recharge. Yet the generic systems show a very different behavior if the maximum flow rate is very strongly limited below a certain threshold disabling breakthrough. Here this is realized by setting the total maximum inflow rate to a value of e.g. 10-8 m3 s-1, i.e. every tube at the inflow side receives a maximum of 5 · 10-10 m3 s-1. In this case, the evolution of large conduits is significantly slowed down; breakthrough is disabled, and conduit development is non-competitive. Some tubes are significantly widened but they do not collect all the available water. Thus, other tubes are still able to grow with time, too. Hence, the evolving apertures cover some orders of magnitude without any gap in the distribution (Figure 2, right panel). The resulting aperture distributions thus are unimodal rather than bimodal and no distinctive modes can be distinguished anymore. It is further found that longer simulation times lead to more uniform aperture distributions.

Figure 2 Cumulative frequency distributions of apertures at different times for generic models with weakly limited (left) and strongly limited (right) maximum flow rates. The first breakthrough occurs after approximately 4100 years if high maximum flow rates are assumed (left) and is disabled when the available water is strongly limited (right). 3.2 Site-related setting The principles revealed by the above-described generic simulations are confirmed by site-related models representing the multi-storey artesian settings of the gypsum karst terrain of the Western Ukraine (Rehrl et al. 2008). The development of maze caves found in this region is probably initiated in artesian settings prior to an uplift of the rock formation. Klimchouk (1997, 2000a) developed a corresponding conceptual model, where a gypsum layer is supplied with chemically aggressive water from a confined aquifer beneath.

86

Figure 3 Conceptual model of a typical artesian basin in the Western Ukraine (top) and its translation into a simplified numerical profile model (bottom) (from Rehrl et al. 2008). The model domain represents a vertical slice consisting of a gypsum layer placed between two insoluble aquifers. Dark solid lines within the model domain represent the tube network at the beginning of the simulation, dashed lines represent equipotential lines. Rehrl et al. (2008, 2009) translated a simplified conceptual model into two-dimensional numerical profile models to investigate the influence of various hydrogeological parameters on conduit development (Figure 3). To be able to represent the unkarstified fissured porous rock and the regional boundary conditions the tube network is hydraulically coupled to the continuum model MODFLOW-96 (Harbaugh and McDonald 1996) using a linear exchange term (Barenblatt et al. 1960).Two confined aquifers consisting of insoluble material are separated by an initially less permeable gypsum layer. The gypsum layer consists of three storeys, each enclosing a fissure network with distinct pattern and density (Klimchouk 2000b, 87

2007). The pronounced vertical heterogeneity caused by the discordance of fissure networks between the different storeys is represented by a low number of vertical protoconduits in the mid level of the numerical model setting. The recharge area is represented by a fixed hydraulic head at the right. The left hand side and the bottom of the model domain are noflow boundaries. Discharge is through the upper boundary, which is represented by a headdependent boundary condition. More details are provided by Rehrl et al. (2008). To account for the uncertainty of the hydraulic conductivity of the rock formation two different scenarios were investigated: A high-conductivity setting adapts the hydraulic conductivities from Birk et al. (2003, 2005); in a second scenario the hydraulic conductivities of all units were reduced by a factor of one hundred (low-conductivity setting). At the beginning of the simulation the hydraulic head of the upper aquifer is lower than that of the lower one (Figure 3). This causes upward flow of chemically aggressive water and thus the enlargement of conduits following the upward directed hydraulic gradient. At the early stage of conduit development the flow rates in the tube network are limited by the narrow outlets at the top of the gypsum. Because of the low flow rates the switch concentration is approached soon and thus dissolution rates are low, following higher-order dissolution kinetics. With ongoing simulation period, however, the solutionally widening of conduit apertures leads to increasing flow rates. Thus, aggressive water propagates farther upward in the tube network. In the high-conductivity scenario, flow rates increase until water with a solute concentration below the switch concentration emerges at an outlet at the top of the gypsum (breakthrough). The more effective first-order dissolution kinetics, which is active then, enhances the growth of a highly conductive pathway connecting lower and upper aquifer. Yet the maximum flow rate is limited by the regional boundary conditions and the permeability of the rock formation. Similar to the generic scenarios considered above conduit development is found to be competitive and leads to a stable bimodal aperture distribution with a less developed region of conduits between a few millimetres up to one decimetre (Figure 4, left).

Figure 4 Cumulative frequency distributions of apertures at different times for the highconductivity (left) and low-conductivity (right) scenario (modified after Rehrl et al. 2008). In a low-permeability formation (Fig. 4, right), however, breakthrough events are very rare and temporary because of the suppressed flow. Thus, conduit development is less competitive, multiple pathways develop, and the frequency distribution of conduit apertures is unimodal with a smooth transition from nearly undeveloped protoconduits (< 1 mm) to welldeveloped conduits (> 1m). More details on these and other scenarios are provided by Rehrl et al. (2008, 2009). 88

4 Conclusions Generic as well as site-related model scenarios show that a stable bimodal aperture distribution evolves if the maximum flow rate in the karst system is not strongly limited; only a limited number of conduits continue to grow while the others stay small. The number of large-sized conduits tends to decrease with decreasing maximum flow rate. However, if flow rates are strongly limited unimodal aperture distributions are obtained. In the latter case, conduit development proceeds very slowly in a limestone environment such as that considered in the generic model scenarios. However, dissolution rates are much higher in gypsum than in limestone. Thus, the site-related scenarios referring to the gypsum karst settings of the Western Ukraine reveal that conduits may significantly be enlarged within reasonable times under suppressed flow conditions. In this case, conduit development is noncompetitive and aperture distributions show a smooth transition from nearly undeveloped protoconduits to well-developed conduits. Acknowledgments This work was funded by the Austrian Science Fund (FWF) under grant no. P20014N10 and the German Research Foundation (DFG) under grants nos. BI 809/2-1 and 436 UKR 113/86/0-1. References Barenblatt GI, Zheltov IP, Kochina IN (1960) Basic concepts in the theory of seepage of homogeneous liquids in fissured rocks. J. Appl. Math. Mech. 24: 1286-1303 Birk S, Liedl R, Sauter M, Teutsch G (2003) Hydraulic boundary conditions as a controlling factor in karst genesis: A numerical modelling study on artesian conduit development in gypsum. Water Resources Research 39(1), 1004,doi:10.1029/2002WR001308 Birk S, Liedl R, Sauter M, Teutsch G (2005) Simulation of the development of gypsum maze caves. Environmental Geology 48(3): 296-306 Clemens T, Hückinghaus D, Sauter M, Liedl R, Teutsch G (1996) A combined continuum and discrete network reactive transport model fort he simulation of karst development. In: Calibration and Reliability in Groundwater Modelling, IAHS Publ. 237, pp 309-318 Dreybrodt W, Lauckner J, Zaihua L, Svensson U, Buhmann D (1996) The kinetics of the reaction CO 2  H 2 O o H   HCO 3 as one of the rate limiting steps for the dissolution of calcite in the system H 2 O  CO 2  CaCO 3 . Geochimica & Cosmochimica Acta 60: 3375-3381 Eisenlohr L, Meteva K, Gabrovek F, Dreybrodt W (1999) The inhibiting action of intrinsic impurities in natural calcium carbonate minerals to their dissolution kinetics in aqueous H2O-CO2 solutions. Geochimica & Cosmochimica Acta 63: 989-1002 Harbaugh AW, McDonald MG (1996) Programmers documentation for MODFLOW-96 – an update to the US Geological Survey modular finite-difference groundwater model. USGS Open-File Report 96-486 Jeschke AA, Vosbeck K, Dreybrodt W (2001) Surface controlled dissolution rates of gypsum in aqueous solutions exhibit nonlinear dissolution kinetics. Geochimica & Cosmochimica Acta 65(1): 27– 34 Klimchouk AB (1997) Artesian speleogenetic setting. In: Proceedings of the 12th Int. Congress of Speleology, La Chaux-de-Fonds, Switzerland Vol. 1, pp 157-160 Klimchouk AB (2000a) Speleogenesis under deep-seated and confined settings. In: AB Klimchouk, D Ford, N Palmer, W Dreybrodt (eds) Speleogenesis - Evolution of Karst Aquifers. National Speleological Society, Huntsville, Ala., pp. 244-260 89

Klimchouk AB (2000b) Speleogenesis of great gypsum mazes in the Western Ukraine. In: AB Klimchouk, D Ford, N Palmer, W Dreybrodt (eds) Speleogenesis - Evolution of Karst Aquifers. National Speleological Society, Huntsville, Ala., pp 261-273 Klimchouk AB (2007) Hypogene Speleogenesis: Hydrogeological and Morphogenetic Perspective. NCKRI Special Paper no. 1, National Cave and Karst Research Institute, Carlsbad, NM., 102 pp Liedl R, Sauter M, Hückinghaus D, Clemens T, Teutsch G (2003) Simulations of the development of karst aquifers using a coupled continuum pipe flow model. Water Resources Research 39(3), 1057, doi:10.1029/2001WR001206 Liu Z, Dreybrodt W (1997) Dissolution kinetics of calcium carbonate minerals in H2OCO2 solutions in turbulent flow: The role of the diffusion boundary layer and the slow reaction H 2O + CO 2 œ H   HCO-3 . Geochimica et Cosmochimica Acta 61: 2879-2889 Rehrl C, Birk S, Klimchouk AB (2008) Conduit evolution in deep-seated settings: Conceptual and numerical models based on field observations. Water Resources Research 44, W11425, doi:10.1029/2008WR006905 Rehrl C, Birk S, Klimchouk AB (2009) Influence of initial aperture variability on conduit development in hypogene settings. Zeitschrift für Geomorphologie, accepted for publication Svensson U, Dreybrodt W (1992) Dissolution kinetics of natural calcite minerals in CO2-water systems approaching calcite equilibrium. Chemical Geology 100: 129-145

90

Sustainability in a karst - the Bungonia Caves, New South Wales, Australia Julia M. JAMES1, Andy SPATE2 1

School of Chemistry, F11, University of Sydney, NSW 2006, Australia e-mail: [email protected] 2 Optimal Karst Management, 2/10 Victoria Street, Hall, ACT 2618, Australia e-mail: [email protected]

Abstract: The Bungonia Caves are located on a substantial block of commercial quality limestone and since their discovery in1830 it has always been under threat from mining. The Bungonia Caves Reserve was first designated for the protection of caves and then more the karst was incorporated into the Bungonia State Recreation Area then underwent a name change to the Bungonia State Conservation Area. These last designations have the same objectives as a national park that is to protect natural and cultural heritage values and to provide recreational opportunities but with mineral exploration and mining permitted. Following its early gazettal the Bungonia karst was managed in a variety of ways until 1992 when the National Parks and Wildlife took over protection it bringing all the expertise and resources available to a state organization. Commercial cave tourism at Bungonia was brief lasting only from 1889 to 1909. The Minister of Lands officially closed the caves in 1931 to all, because they contained foul air. The first Bungonia Caves Reserve Trust appointed in 1932, took on a karst that had been neglected during the Depression. Major consequences of the neglect were the karst was stripped of native vegetation and there was erosion caused by overgrazing by sheep and cattle. After 1949, numerous speleologists worked at Bungonia exploring deep into the foul air regions of the caves and carrying out research into all aspects of the caves and karst. Despite this the reserve continued to be further degraded by the rubbish dumping, wood gathering, uncontrolled camping, fire and grazing and uncontrolled caving activities such as digging both on the surface and the caves. In 1970, conservationists and speleologists united to fight a threat from a limestone quarry that wanted to extend its mining leases to include the north wall of the Bungonia Gorge. The case was successful and the publicity it generated enabled the Bungonia karst to start a slow road to recovery and a future of sustainability. The major impacts on the karst are now due to the presence of recreational cavers who are requested to follow the codes and guidelines of the Australian Speleological Federation notably the Code of Ethics, the Minimal Impact Caving Code and Safe Caving Guidelines. In addition, a number of measures have been implemented to protect the cave fauna and cavers. The future sustainability of the Bungonia karst will be further ensured if the New South Wales state conservation area’s review recommendation that the karst and caves area is designated as a new national park called the Bungonia National Park. Keywords: karst, sustainability, Bungonia Caves, Australia

1 Introduction Bungonia Caves are located in the Southern Tablelands of New South Wales approximately 28 km east of Goulburn (Figure 1). The karst containing the caves is on the limestone members of the Bungonia Group (Silurian to Early Devonian) (Bauer 1998). The caves, some of the deepest on the Australian mainland, are found on a plateau that is dissected by one of the finest limestone gorges in Australia. In 1902, an area of 565 ha around the caves was gazetted a reserve for the preservation of caves and became known as

91

the Bungonia Caves Reserve (Figure 1). The sustainability1 of the Bungonia karst will be discussed in the following paper.

Figure 1 The location of Bungonia Caves 2 The early years The Bungonia Caves were discovered in 1830 (Nurse 1972). In 1872, the same year as the World’s first national park, Yellowstone in the USA, was created the caves and karst area was gazetted a Reserve for Public Recreation and Water Supply. At that time exploration of the Bungonia karst was by graziers familiarizing themselves with the local lands and surrounding properties; surveyors from the Department of Lands and the Department of Mines extending and mapping resources or those concerned with the development of tourist activities. In 1889, a caretaker was appointed by the Department of Mines resulting in three caves being developed as tourist caves. In 1897, the caves were protected by a regulation “No person be allowed to enter the caves unless accompanied by a caretaker or authorized guide”. 3 The dark ages The 1902 legislation for the preservation of the caves as the Bungonia Caves Reserve lasted until it was revoked partially in 1921 and finally in 1936. During the intervening years much conservation legislation was enacted but all rapidly became extant or was revoked (Middleton 1972). By 1909, commercial cave tourism at Bungonia had effectively finished; there was no longer a caretaker and the caves were regarded as being closed to the public. The Minister of Lands officially closed the caves in 1931 to all, because they contained foul air (high levels of carbon dioxide (up to 7%) and low levels of oxygen (down to 15%) (James et

1

Sustainability, in a broad sense, is the capacity to endure. In ecology the word describes how biological systems remain diverse and productive over time. For humans it is the potential for long-term improvements in wellbeing, which in turn depend on the wellbeing of the natural world and the responsible use of natural resources. Wikipedia. 92

al. 1975) Foul air is encountered throughout the caves and severely limited early cave exploration. The first Bungonia Caves Reserve Trust was appointed in 1932 in order to administer the karst area that had been neglected in the Depression. Major consequences of the neglect were that the surface was stripped of native vegetation and there was erosion caused by overgrazing by sheep and cattle. In 1949, with the advent of caving groups, numerous speleologists worked at Bungonia establishing the significance of the caves, exploring deep into the foul air regions and carrying out research into all aspects of the caves and karst. Many publications resulted in both national and international journals. Nevertheless, the reserve continued to be further degraded by the rubbish dumping, wood gathering, uncontrolled camping, fire and grazing. Unwise actions by cavers resulted in irreversible damage to parts of the karst. For example, it was believed that if the water in the phreatic regions of the caves was drained and vadose flow established fresh air would remove the foul air as suggested by JC Wiburd in 1925 (cited in Bonwick 1972). The foul air caves all drain to one spring that had been blocked by breakdown of the cliff above it. This site was excavated from 1923 until 1967 with the excavation of a 30 m long trench that lowered the outflow by some 15 m (Bonwick 1972). The excavation left a massive scar on the side of the gorge and caused the death of an ancient white cedar tree and other rainforest remnants by diverting their water. The water level in the final chamber of the closest cave was lowered at the same time and banded sediments were exposed (James 1973). Once exposed the sediment bands were oxidized destroying their paleoclimatic value (Contos 1998). A cave at the spring was revealed by the excavation but its passage slumped after 20 m. In an unsuccessful attempt to access the flooded passage at the spring thousands of litres of water were siphoned from the phreas (Bonwick 1972). This operation could have had an enormous impact on any stygofauna in the cave system. The greatest threat to the Bungonia Caves Reserve materialized in 1970 when its exemption from the leasing provisions of the 1906 Mining Act was revoked. Immediately a proposal to extend its limestone mining leases along the north wall of the gorge came from company that owned the Marulan Quarry. The quarry was to the north of the reserved karst and across the 300 m deep Bungonia Gorge (Figures 1&2). Conservationists and speleologists united to fight this proposal in the Mining Warden’s Court. To support the case a book, Bungonia Caves (Ellis et al. 1972) was published documenting many aspects of the karst and limestone quarrying in the Bungonia area. A publicity campaign was run with the slogan “Keep Bungonia Gorgeous”. The case was successful and the dark ages had ended. 4 The age of enlightenment The publicity that the keep “Bungonia Gorgeous” generated enabled the Bungonia Caves Reserve to start the slow road to recovery and a future of sustainability. A Caves Reserve Trust managed the reserve under the Crown Lands Consolidation Act 1913. In 1974 saw the Bungonia Caves Reserve become part of a much larger (4007 ha) Bungonia State Recreation Area (Figure 2). This category of reserve protects natural and cultural heritage values and provides recreational opportunities. However, unlike national parks and nature reserves, the designation allows other uses including mineral exploration and mining and petroleum exploration and production. The economic minerals in the Bungonia State Recreation Area were significant deposits of orogenic and alluvial gold, polymetallic and silica deposits as well as abundant limestone. Mining of these has been carried out at a number of locations within Bungonia State Recreation Area and adjacent to it. The Marulan Quarry is the largest limestone quarry in New South Wales.

93

Figure 2 Reserves at Bungonia In 1974, the Department of Lands appointed a ranger for the Bungonia State Recreation Area and during his tenure the karst was fenced against sheep and cattle. A Bungonia cleanup weekend was organized by the speleologists and tonnes of rubbish were removed from the karst. A well-appointed campground was built away from the karst and out of its catchment. Two caves were gated and a series of regulations were promulgated to protect the caves and surface. These relatively simple regulations to enhance ethical use of the caves and karst have since evolved into complex, and, perhaps less sensible, regulations under the National Parks and Wildlife Act 1974. In 1992 the National Parks and Wildlife Service took over the management of the Bungonia State Recreation Area bringing to it all the expertise and resources available to a state organization. The old designation as a State Recreation Area was honoured and free access to most caves was allowed after consultation with recreational cavers and karst scientists. Under National Parks and Wildlife Service management, the campground was greatly upgraded, as were the water supply, roads, car parks and lookouts. The publication Bungonia Caves had been prepared at a time when the integrity of Bungonia Gorge was under threat from limestone quarrying. In recent years, the popularity of the Bungonia caves for recreation has increased to a point where there are visible impacts on the karst and within some caves. The linked impacts on the fragile ecosystems that exist within the caves cannot be quantified. Education of recreational cavers was recognized as being critical to the preservation of caves for future generations. In 1998 a second publication, Under Bungonia (Bauer and Bauer 1998) was published in order to educate the recreational caver and supply them with clear and concise conservation information. All wildlife is protected in areas set aside under the National Parks and Wildlife Act 1974. On the Bungonia karst remedial or preventative measures to protect the cave fauna have been applied. These have included revegetation and stabilisation of soil erosion at cave entrances and the avoidance of channelling runoff from car parks and road surfaces into sinkholes and caves. The greatest threat to cave fauna at Bungonia is the uneducated caver. Eberhard (1998) made recommendations that would enable cavers to avoid their activities impacting on the cave fauna if universally adopted. The species most likely to be affected at 94

Bungonia by recreational caving is the endemic silverfish Trinemura anemone. One of the caves in which it is found has been gated to ensure its viability. At Bungonia there are two cave bats on the vulnerable list of the Threatened Species Conservation Act 1995. At Bungonia these bats are protected at the times of their life cycle when they are most at risk and caves are closed at these times. Because seasonal conditions change, the times of closures are not fixed and are subject to change. The seasonal closures fall into the following pattern: over wintering caves closed from 1 May until 30 September, staging caves are closed from 1 December until 31 December; and maternity caves are closed from 1 November until 31 March. Cavers are advised to avoid of the disturbance of individual bats, small clusters and populations is also sensible at other times of the year. Further details of the sustainable conservation of bats at Bungonia can be found in Spate (1998). A Bungonia Recreational Activities Group advises on caving and other activities within the State Conservation Area. The Group includes scientists, recreational cavers and representatives of the Department of Defence, the Scout Association, commercial operators and so on. It meets three to four times a year and is chaired by a specialist karst scientist. Activities in the Conservation Area can only be carried out with the consent of the DirectorGeneral of the National Parks and Wildlife Service. In order to facilitate this, consent is obtained by completion of an Activity Register at the park office. The register is available 24 hours a day. Cavers are requested to follow the codes and guidelines of the Australian Speleological Federation notably the Code of Ethics, the Minimal Impact Caving Code and Safe Caving Guidelines. Leaflets on the dangers of entering the “foul air” regions of the caves and on safe responsible practices are also available. At times caves may be closed because of exceptionally hazardous levels of foul air or after flood events. The cave closures for the protection of fauna and visitors that occur from time to time have been respected. Digging activities, fixing markers, placing bolts for vertical caving, entry to gated caves and research activities all require the explicit consent of the DirectorGeneral and must to comply with the Environmental Planning and Assessment Act 1979 and the National Parks and Wildlife Act 1974 and regulations under those Acts. Removal of flora, fauna, soil, archaeological materials, speleothems and fossils is prohibited unless collection of such materials has been approved by the Director-General. In 2002, the Bungonia State Recreation Area was gazetted as a State Conservation Area with no change to its status with respect to management or mineral exploration and mining. At this time future sustainability of the Bungonia karst appeared to be the best available and was extremely promising and practicable. 5 The final act Since the keep “Bungonia Gorgeous” campaign in 1970 conservationists and speleologists have continuously stated that the best protection for the Bungonia karst is that it should become part of the adjacent Morton National Park. The same suggestion was made for the larger Bungonia State Conservation Area (Figure 2). To some speleologists that have had a long working relationship with the Bungonia karst it does not seem right that it should lose its identity by being subsumed into the much larger national park. It could also lose its focus on caves and karst. There was some discussion that the Bungonia karst be better designated as a Karst Conservation Reserve with the specific objectives of the conservation of the karst environment and its biodiversity. The karst conservation reserve concept is relatively new, and of the more than 780 parks in New South Wales, only four have the karst conservation reserve designation. Such reserves were specifically designed to increase the conservation status of reserved lands on karst whilst allowing for recreational caving and scientific research.

95

The status quo as part of the Bungonia State Conservation Area had been reluctantly accepted the best option. There was a review of all the state conservation areas in New South Wales in 2008 (DECC 2008). For the Bungonia State Conservation Area the reviewers noted that in south-western part of area exploration and mining titles still apply which means it cannot be reserved as a national park or nature reserve under the National Parks and Wildlife Act (1974), and must remain a State Conservation Area to allow for exploration or mining, subject to environmental assessment. However, in the north-eastern part of the area, there were no exploration or mining titles and limited geological evidence of valuable mineral resources and the dual-purpose State Conservation Area category is no longer required. The reviewers propose that about 770 ha become a national park. This is a greater area than the 563 ha of the 1902 Bungonia Caves Reserve as it includes more of the karst on the north side of the gorge. The proposal is shown by the dark blue line in Figure 2 and would be known as Bungonia National Park. We believe that this designation, or better still as Bungonia Karst Conservation Reserve, is desirable as it would keep the historical Bungonia identity for the karst, retain the excellent management by the National Parks and Wildlife Service supported by the Bungonia Recreational Activities Group and further guarantee the sustainability of the karst by protection from mineral exploration and mining. The change would require gazettal under the National Parks and Wildlife Act, 1974. References Bauer J (1998) Geology of the Bungonia Group. In: Under Bungonia Bauer and Bauer (eds), JB Books Oakflats, NSW, Australia, pp 2-19 Bauer J, Bauer P (1998) Under Bungonia, JB Books, Oak Flats, Australia Bonwick BS (1972) The Efflux, B67. In: Bungonia Caves, Ellis et al. (eds.) Sydney Speleological Society, Sydney, pp 55-62 Ellis R, Hawkins L, Hawkins R, James JM, Middleton G, Nurse B, Wellings G (1972) Bungonia Caves, Sydney Speleological Society, Sydney Contos AK (1998) Biomineralisation in Caves. PhD Thesis, University of Sydney, NSW, Australia DECC (2008) Bungonia SCA. In Review of State Conservation Areas, Department of Environment and Climate Change, Sydney, pp 212-213 Eberhard S (1998) Cave Invertebrates. In: Under Bungonia (eds.) Bauer and Bauer, JB Books, Oakflats, NSW, Australia pp 74-83 James JM (1973) Sediments in a Bungonia cave (B24). International Speleology 1973, Proc.6th Int.Speleol.Cong.,Vol.3, pp 449-456 James JM, Pavey AJ and Rogers AF (1975) Foul air and resulting hazards to cavers. Transactions of the British Cave Research Assoc. 2:79-88 Middleton GJ (1972) Conservation and Mining of the Bungonia Limestone Belt in Bungonia Caves Ellis et al. (eds.) Sydney Speleological Society, Sydney, pp 179-199 Nurse BS (1972) Summary of the History of Bungonia Caves and Area in Bungonia Caves Ellis et al. (eds.) Sydney Speleological Society, Sydney, pp 13-26 Spate A (1998) The Cave Bats of Bungonia in Under Bungonia Bauer and Bauer (eds), JB Books, Oakflats, NSW, Australia, pp 67-73

96

Hydrogeological conditions and water quality of the karstified formations of Louros basin, Epirus, Greece Konstantina KATSANOU1, Euaggelos NIKOLAOU2, George SIAVALAS3, Eleni ZAGANA1, Nikolaos LAMBRAKIS1 1

Laboratory of Hydrogeology, Section of Applied Geology and Geophysics, Department of Geology, University of Patras, Greece, e-mail: [email protected] 2 Institute for Geology and Mineral Exploration, Preveza, Greece 3 Section of Earth materials Department of Geology, University of Patras, Greece

Abstract: In the frame of the present study the hydrogeological and hydrochemical conditions that prevail in Louros River drainage basin are analyzed and interpreted. The karstified basin occupies 2 an area of 926 km and is located in the Epirus region at the northwestern part of Greece. It covers the drinking water needs of four towns constituting one of the most important karstic hydrogeological systems of this region. Geologically the basin is hosted in the formations of the Ionian geotectonic zone, which is the western part of the External Hellenides, which in turn constitute the extension of the Dinarides towards the South. The older formations of the study area consist of evaporitic layers of Triassic age, which outcrop at the western part of the Louros drainage basin. The Triassic evaporites are overlain by a thick sequence of carbonate and clastic rocks reflecting a continuous sedimentation from Late Triassic to Upper Eocene. Oligocene flysch outcrops at the margins of Louros basin, whereas at the lower part these formations are overlain by Neogene and Pleistocene sediments as well as Holocene fluvial deposits. The basin is delimited by large north–south trending faults, which affect the land topography and play a principal role in the function of the karst springs. The karst system is developed in the upper part of the carbonate sequence (Middle Jurassic-Upper Eocene), mostly consisting of thick-bedded and hydraulically interconnected formations, due to tectonic activity. The chemical composition of the groundwater showed the prevalence of Ca-HCO3 and Ca-HCO3-SO4 water types. The presence of SO4 is attributed to the evaporites. Schoeller and Piper diagrams also show that the chemical composition of groundwater is affected by the Triassic gypsum, which is expected to expand underground, throughout the largest part of the area. In the southern part, the reduction of sulphates, contributes to the rise of hydrogen sulphide-rich waters. R-mode factor analysis reveals that the most important process that controls groundwater chemistry is the dissolution of minerals along with redox environment. Reducing conditions are predominant in the periphery of the basin, mainly to the western and southern parts. Nitrates, which derive from agricultural activities, affect groundwater quality. Keywords: karst, Ionian zone, hydrochemistry, statistic analysis, hydrogeology

1 Introduction The study area is located in the Epirus region at the northwestern part of Greece, including the drainage basin of Louros River that occupies an area of 926 km2. It supplies the drinking water needs of four towns constituting one of the most important karstic hydrogeological systems of this region. Louros River has a length of about 75 km and an average flow of about 10.6 m3/s. It discharges into Amvrakikos gulf to the south and its water irrigates about 120 km2 of cultivated area and supplies a small hydroelectric dam with 10 MW installed capacity. Louros karstified aquifer is discharged by 17 main springs, with yields ranging between 2.5 m3/sec and 6.7 m3/sec. At the estuaries of Louros there are lagoons such as Tsoukalio, Rodia, Logarou, and Tsopeli, which are wetlands designated under the Ramsar Convention and the European 97

Communities Legislation. About 2.5% of Louros catchment surface is cultivated and a number of large and small agricultural industries and fish-farms operate in the same region. Karst aquifers, which supply drinking water to an estimated 25% of the global population (Pulido-Bosch 1999), are very vulnerable to contamination because of their hydrogeologic characteristics (Escolero et al. 2002) and display properties that are directly related to the investigation of microbiological contamination. Mahler et al. (2000) have reported four reasons that justify the above considerations for karst aquifers: (1) the direct and rapid connection between the surface water and groundwater systems, (2) the sediment mobility, (3) the karst heterogeneity and (4) the possibility of sudden water quality variations (episodic contamination). The study emphasizes on karstic aquifers, hosted in the carbonate formations of Louros drainage basin, aiming to elucidate the major hydrochemical processes, which control the chemical composition of the groundwater, but also to estimate potential anthropogenic impacts on groundwater quality. 2 Climate and morphology Evaporation and humidity are relatively high throughout the year. Climate can be characterized as mild temperate, to continental, changing by the influence of geographical position and relief (Boltsis 1986). The hydrological regime of northern Epirus is characterized by uneven seasonal and regional distribution of precipitation. An average precipitation for Louros drainage basin is 1150-1800 mm. Out of them 50% is evapotranspiration, 35% is recharge and 15% is discharge (Fig. 1; Nikolaou 2005).

Figure 1 The distribution of precipitation in Louros drainage basin Morphologically, Louros karstic system is characterized of high elongated mountain ranges and narrow valleys, due to the tectonic (anticline structures) and geologic conditions of the region and the lithologic alternation between limestones and flysch. Louros basin is also characterized by numerous exokarstic and endokarstic features such as caves, sinkholes, karst lakes and springs that make the system vulnerable to groundwater pollutants. 3.1 Geology Geologically, the basin is hosted in the formations of the Ionian geotectonic zone, being the western part of the External Hellenides, which constitute the extension of the Dinarides towards the South. The Ionian zone was first described by Philippson (1896), who also included in it the Pre-Apulian and Gavrovo zones. However, the main stratigraphic features and the geotectonic position were presented by Renz (1955), who named it "Adriatsche- Ionische Zone". Aubouin (1959) discussed the stratigraphic and geotectonic evolution of the basin, subdividing it into Internal, Central and External. The subsequent investigations made by IGRS-IFP (1966) in Epirus, Corfu and Lefkas, and by British Petroleum (BP) in Central Greece and the neighbouring Ionian Islands, provided further information on the geology of the zone. 98

The older formations of the study area consist of evaporitic layers of Early to Middle Triassic age, which outcrop at the western and southern part of Ziros Lake (Fig. 2). The Triassic evaporites were deposited in a shallow supersaline basin and they are locally overlain by breccias derived from calcification processes (Karakitsios and Pomoni-Papaioannou 1998). These formations are overlain by a thick sequence of carbonate and clastic rocks reflecting a continuous sedimentation from Late Triassic to Upper Eocene. Until the Middle Liassic the Ionian Zone formed part of the Apulian Platform where thick, neritic limestones were deposited. From bottom to top this sequence includes Foustapidima formation that consists of black, hypolithographic limestones and dolomites of Carnian age. They are overlain by massive Early to Middle Liassic neritic limestones, the socalled Pantokrator Limestones (Karakitsios and Tsaila- Monopolis 1988). In Louros drainage basin their thickness is estimated to range from 1000 to 1500 m (Leontiadis & Smyrniotis 1986). A transition to pelagic sedimentation prevailed within the basin during the late Liassic giving rise to the following formations: The Siniais Limestone composed of thin-bedded limestones with chert nodules and cherty interbeds occurring locally on top of the Pantokrator Limestone (IGRS-IFP, 1966; Skourtsis-Coroneou et al., 1995), the Lower Posidonia Shales or their lateral equivalent, the marly Ammonitico Rosso, both of Toarcian age, Middle Jurassic limestones with filaments, and the Upper Posidonia Shales of Callovian-Kimmeridgian age, that consist of very thin-bedded, brown, green and black cherts alternating with thin beds of shales and marls (Fig. 2). Vigla Limestone, which was deposited during the latest Kimmeridgian-Santonian (Skourtsis-Coroneou and Manacos 1995) over the Upper Posidonia Shales, is the first pure pelagic phase in the zone’s sequence. This consists of light-coloured, thin-bedded limestones with abundant calpionellids, foraminifera and radiolaria, and cherty lenses and interbeds, which become more common in the upper parts, locally forming the ‘upper chert horizon’. ‘Clastic’ limestones, overlying the Vigla Limestone, were deposited during the late Senonian–Eocene interval (Aubouin 1959). They include microbrecciated limestones and breccias with clastic materials that invaded the basin from its marginal areas and from the neighbouring zones, as well as numerous intercalations of pelagic, micritic limestones towards the top (Skourtsis-Coroneou et al. 1995). A Flysch unit was deposited during the Oligocene on top of the ‘Clastic’ limestones. The transition is characterized by marly limestones and calcareous marls (Skourtsis-Coroneou et al. 1995, 1999). Unconsolidated deposits outcrop at the lowlands. These consist mostly of lacustrine deposits of Pliocene age, glacial drift, siliceous deposits, riparian terrace deposits, talus cones, and alluvial sediments. 3.2 Tectonics Within the Ionian Zone several major thrusts have been described as well as other important tectonic features like great E-W trending strike-slip fault systems (IGRS-IFP 1966). Most of the thrusts show a "normal" westward dip (Louros, Paramithia, Margariti and Parga Thrusts) and similarly, most of the folds are asymmetric with east-dipping axial planes. Nevertheless important tectonic structures with an opposite dip also exist (Xerovouni and Tomaros back-thrusts). Since all the Alpide edifice of the Hellenides was created by westdipping thrusting, we refer here to the east- dipping thrusts as back-thrusts. Moreover, the large E-W fault zones, like the Souli strike-slip fault system, also known in the literature as Petoussi Fault, play an important role in the Alpide tectonic evolution of the area. Indeed, on the two sides of these faults the amount of displacement and the style of the NNW-SSE trending shortening structures often differ (IGRS-IFP 1966).

99

225000

230000

4

! A

235000

LP40 ! A

! A

LP31

! !A A

Melia

4370000

LP10

LP44LP45

"

! A ! A

240000

LP2

"

4370000

220000

LP30

Pesta ! A

LP43

LP12 ! A R1 ! A

"

"

LP38

Legend:

Terobon

! LP39 A ! A " Derbiziana

! A

LP41

! A

LG59 ! A

LP14

LP42 " Assos

! A

4360000

4360000

! A

! A

Panagia "

LP46

! A

LG26 ! A

! A

! A

LP32

LG107

LG29

! A

LG47

LG37

! A LP6 ! A

! A

LG102

LG104 R2

! A

! A

LP15 LP35 ! A ! A

! A

LP18 LP17 !

LG3

LP3 "

"

Agios Georgios

km

Ammotopos

4350000

4350000

! A

! A

A

" Thesprotiko

! A

LP27

! A

! A

LP1

Sampling sites Louros River Normal faults Thrusts Holocene deposits Quaternary deposits Neocene deposits Flysch Terra Rossa Clastic Limestones Senonian limestones Vigla limestones Limestones with Fillaments Ammonitico Rosso Shales with Poseidonies Pantokrator & Sinies limestones Dolomites & Dolomitic limestones Triassic Breccias Triassic gypsum

LG64 ! A

! A

LP8

0

1

2

4

STRATIGRAPHIC COLUMN OF THE IONIAN ZONE

LG9

LP9 al Ziros Lake ! A

Upper EoceneLower Mocene

Flysch

LP33LP4 "

A !! A Kampi

! A

! A

LP26

! A

! A

! A

Eocene

LP23

4340000

LP7

4340000

! A

LP11

R3

LP5

Chanopoulo

"

Limestone

Paleocene

Limestone

Senonian

Limestone

Up. Albian-Turonian

Chert

Malmian -Albian Dogger Up. Liassic (Toarcian)

Vigla Limestone Limestone with Filaments Chert, shale with Posidonia Ammonitico Rosso

Thin-bedded limestone

Pantokrator limestone

Dolomite

4330000

4330000

Upper TriassicMiddle Lias sic

Breccias

Triassic ! A

sea 1

200 m

Evaporite

100

Amvrakikos Gulf

0 220000

225000

230000

235000

240000

Figure 2 Geographic position, geological map and lithostratigraphical column of Louros drainage basin 100

4 Hydrogeological Conditions The study area is structured from geological formations of different hydrogeological properties. The geological conditions, the lithostratigraphical diversity, and the complicated tectonics in combination with the geomorphological conditions have resulted in the formation of independent or semi-independent hydrogeological units. The geological setting of Epirus illustrates characteristic alternant limestone anticlines and flysch syncline structures with general axis direction of NW-SE. Groundwater circulates following the same directions resulting in the formation of extensive hydrolithological units hosted in the karstified calcareous anticlines separated by the impermeable flysch synclines (Leontiadis and Smyrniotis 1986) Carbonate formations are the major units constituting the highlands of the study area. These units include dolomites, the Pantokrator Limestone, the Vigla Limestone, Late Senonian limestones, Paleocene–Eocene limestones, and Posidonian chert. The most important aquifers in the broader area have been developed in carbonate rocks (mainly Pantokrator and Upper Senonian limestones). These formations show high permeability due to their intense karstification and fracture porosity and their occurrence is extended especially in the upper parts of Louros draining basin. The flow rates of the major karstic springs in the region (Fig. 3) are higher than those expected by the hydraulic properties of individual units. This suggests the existence of hydraulic interconnections among carbonate formations of various ages, as well as replenishment of groundwater from rivers and lakes.

Figure 3 Discharge (m3/s) of the major springs of the study area (Nikolaou 1991) According to Leontiadis and Smyrniotis (1986) Louros hydrogeological basin extends beyond the surface of the hydrological one. From the hydrogeological point of view the basin can be separated into five different hydrogeological systems. The first consists of the subbasin of Thesprotika mountains, which outcrop in the western part of Louros basin, and covers an area of 120 km2. The second is the Priala-Rizovouni system, the third is Chanopoulo springs system being in the eastern part of the study area and the other two are the upstream and downstream part of Louros River and cover 195 and 53 km2, respectively. 5 Groundwater chemistry 5.1 Sampling and analytical procedures In order to define the hydrochemical composition of the karstic aquifer in the frames of the present study a total number of 108 samples were collected during October 2008 and May

101

2009 from springs and boreholes along Louros drainage basin according to US EPA (1976) procedures. Two polyethylene bottles of 1 and 0.1 L volume, respectively, were collected for each sampling site. The first bottle contained bulk water sample proper for the anion concentrations analyses. The second bottle contained water filtered through a Whatman 0.45 m cellulose membrane and acidified with 0.5 mL ultra pure HNO3. The unstable physicochemical parameters including temperature, pH, dissolved oxygen, electric conductivity and redox potential were measured in situ using Hana® HI 9828 portable equipment. Additionally, alkalinity was also determined on the field using Hach® Digital Titrator. All chemical analyses were performed in the Laboratory of Hydrogeology, University of Patras. Anion (NO3-, NO2-, PO43-, SO42-, and F-), NH4+ and SiO2 concentrations were measured in a Hach® DR 4000 spectrophotometer. For the determination of Cl- content titration techniques were applied by using AgNO3 0.1 N. Major cation (Ca2+, K+, Mg2+, Na+) concentrations were determined in a GBC® Avanta flame atomic absorption spectrophotometer. Trace element concentrations were measured using inductively coupled plasma-mass spectrometry (ICP-MS) in an ELAN 6100 Perkin-Elmer. It should be noted that the percentage error of the chemical analyses results calculated according to ion mass balance does not exceed 5%. 5.2 Descriptive statistics and groundwater classification The samples are classified into two water types according to the Piper classification (Fig. 4). In the first water type the samples are characterized by the Ca2+ and HCO3predominance, whereas in the second, the anion site is shared between HCO3- and SO42-. Samples of the second type display higher salt content, especially sulphates, compared to the first. This could be attributed to the presence of a sulphate-rich source, like evaporite minerals that outcrop on the south-western part of the study area. Concentration (meq/l) 1000.

80

80

Legend:

100.

Legend: 60

WS SW RW R3 LP5 LP11

10.

60

40

20

1.

WS SW RW R3 LP5 LP11

40

20

Mg

SO4

80

80

0.1

60

60

40

0.01

40

20

0.001 Mg

Ca

Na+K

Cl

SO4

HCO3

20

80 Ca

60

40

20

20 Na

HCO3

40

60

80 Cl

Figure 4 Schoeller and Piper diagrams of the analyzed samples. WS: groundwater sample, SW: seawater, RW: rainwater The descriptive statistics of the two main water types of the study area are shown in Table 1.

102

The Scholler diagram (Fig. 4) demonstrates that rCl/rNa and rNa/rCa ratios reveal values similar to those of rainwater. On the contrary, rCl/rSO4 ratio is diverse, probably due to the interaction of water with the evaporites. Samples R3, LP5 and LP11 represent hydrogen sulphide-rich waters, occurring at the southern parts of the study area. Table 1 Descriptive statistics for the analyzed water samples. N: number of samples

Twa (oC) Eh (mv) pH Cond Alk (mg/L) HCO3 K (mg/L) Na (mg/L) Mg (mg/L) Ca (mg/L) NH4 (mg/L) NO3 (mg/L) NO2 (mg/L) SO4 (mg/L) F (mg/L) Cl (mg/L) B (mg/L) Mn (mg/L) Fe (mg/L) Sr (mg/L)

N 26 27 27 27 33 33 33 33 33 33 33 33 33 33 33 33 33 33 33 33

Min 7.72 -20.00 7.00 200.00 95.00 115.90 0.15 1.64 0.72 40.00 0.00 0.00 0.00 0.00 0.00 0.90 0.00 0.00 0.01 0.03

Ca-HCO3 Max Mean 17.97 14.59 277.00 104.54 8.21 7.53 604.00 349.41 276.00 161.41 336.72 196.92 8.84 1.23 19.12 5.71 10.11 3.43 117.50 69.08 1.47 0.08 21.00 7.15 0.47 0.03 50.90 17.19 1.48 0.15 30.00 8.84 0.03 0.01 0.19 0.01 0.12 0.03 0.24 0.11

Std.Dev. 2.50 82.52 0.34 93.97 41.85 51.06 1.60 3.23 2.45 17.14 0.26 4.97 0.09 15.42 0.27 6.60 0.01 0.03 0.03 0.06

N 22 25 25 24 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25

Min 13.63 -290.00 6.99 338.00 92.00 112.24 0.38 3.36 3.66 52.50 0.00 0.00 0.00 42.50 0.08 3.40 0.01 0.00 0.00 0.11

Ca-HCO3-SO4 Max Mean 18.70 15.54 170.00 84.45 8.85 7.47 3920.00 753.96 246.00 164.32 300.12 200.47 44.00 2.93 733.00 59.55 80.80 12.55 414.00 106.80 0.52 0.04 26.00 7.08 0.34 0.03 1120.00 172.00 1.38 0.34 830.00 69.33 0.41 0.04 0.07 0.01 2.00 0.12 3.84 0.55

Std. Dev. 1.47 96.70 0.39 847.54 26.83 32.73 8.94 167.69 17.02 68.02 0.10 6.55 0.07 238.09 0.28 189.34 0.10 0.01 0.41 0.73

5.3 Factor analysis R-mode factor analysis was applied, according to the steps described by Davis (1987), to study the interrelations among 19 selected variables measured in the Louros basin water samples. The aim of this analysis is to reduce a large number of variables in the original data to a significantly smaller number of ‘factors’, each of which is a linear function of the original variables (Ashley and Lloyd, 1978; Adams et al, 2001). The selection of the four factors in Table 2, was based on the criterion that eigenvalues must be higher than 1. The chosen four factor model explains more than 77% of the total variance. All variables display very high communalities (1.000), indicating that the 4-Factor model describes them very well. In a next step, the contribution of each factor at every site (factor scores) was calculated (Fig. 5). The first factor explains 45.7% of the total variance, and shows that most of the covariance in the properties of the system may be represented by variances of Na+, Mg2+, Ca2+, SO42-, F-, Cl-, B, Fe, and Sr2+. This indicates the principal role of these elements in the chemical composition of the groundwater. Figure 5 shows that samples LP11 and LP6, which are chemically affected by the presence of underlain evaporites, have a major contribution to the formation of factor 1. Samples LG3, LG47 and LG48 from the east-central part of Louros basin, with high positive scores on this factor indicate a possible extension of the same geological conditions in this area. The first factor also highlights the effect of redox potential in the groundwater chemical composition. The second factor accounts for 13.84% of the total variance and shows a negative relationship between two redox sensitive constituents of groundwater (NH4+, Mn). Samples LP23, LP27, LP29, LP5 and LP9 showing high positive scores on this factor (Fig. 5) delimit an area in the western part of Louros basin. This area is characterized by outcrops of Triassic breccias.

103

Table 2 Factor loadings of the selected variables Factor 1 T(wa) Eh pH HCO3 K Na Mg Ca NH4 NO3 NO2 SO4 F Cl B Mn Fe Sr Total % of Variance Cumulative %

Factor 2

Factor 3

Factor 4

-.689 -.679 .730 .969 .812 .883 .963 -.788 .968 .713 .968 .849 .924 .943 .937

Communalities 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000

Rotation Sums of Squared Loadings 8.238 2.484 1.958 1.353 45.768 13.800 10.878 7.519 45.768 59.568 70.445 77.964

Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization.

Figure 5 Factor 1 and 2 scores of the analyzed samples The third factor, which accounts for 10.9% of the total variance, is characterized by the well-known relationship between pH and HCO3 ion, whereas the fourth factor accounting for 7.5% of the total variance is related to NO3- and indicates a probable anthropogenic impact.

104

6 Conclusions The chemical composition of the groundwater in Louros drainage basin showed the prevalence of Ca-HCO3 and Ca-HCO3-SO4 water types. The presence of SO4 is attributed to the Triassic evaporites and the overlying breccias, which also affect the chemical composition of groundwater in general, as it is suggested by Schoeller and Piper diagrams. In the southern part, the reduction of sulphates, contributes to the rise of hydrogen sulphide-rich waters. R-mode factor analysis reveals that the most important process that controls groundwater chemistry is the dissolution of minerals along with redox environment. Reducing conditions are predominant in the periphery of the basin, mainly to the western and southern parts. Nitrates, which derive from agricultural activities, affect groundwater quality. References Adams S, Titus R, Pietersen K, Tredoux G, Harris C (2001) Hydrochemical characteristics of aquifers near Sutherland in the Western Karoo, South Africa, Journal of Hydrology 241:91-103 Ashley RP, Lloyd JW (1978) An example of the use of factor analysis and cluster analysis in groundwater chemistry interpretation. Journal of Hydrology 39: 355-364 Aubouin J (1959) Contribution a l’etude geologique de la Grece septentrionale: Les confins de l’Epire et de la Thessalie. Ann.Geol. d. Pays Hell. X:1-525 BP (British Petroleum Company Limited) (1971) The geological results of the petroleum exploration in western Greece. Institute of Geology & Subsurface Research 10:173 Boltsis T (1986) Contribution to the study of the water equivalent of precipitation in Epirus region. PhD Study, University of Athens, Greece Davis JC (1987) Statistics and Data Analysis in Geology. John Wiley & Sons Inc. Singapore, 646 pp Escolero OA, Marin LE, Steinich B, Pacheco AJ, Cabrera SA, Alcocer J (2002) Development of a Protection Strategy of karstic Limestone aquifers: The Mérida Yucatán, Mexico Case Study. Water Resources Management 16:351-367 IGRS-IFP (Institut de Geologie et de Recherches du Sous-sol et Institut Francaise du Petrole) (1966) Etude Geologique de l’Epire (Grece nord-occidentale), Technip, Paris, 306 pp Karakitsios V (1995) The influence of preexisting structure and halokinesis on organic matter preservation and thrust system evolution in the Ionian Basin, Northwest Greece. The American Association of Petroleum Geologists Bulletin 7: 960-980 Karakitsios V, Tsaila-Monopolis S (1988) Donnees nouvelles sur les niveux superieurs (Lias inferieur-moyen) des calcaires de Pantokrator (zone ionienne moyenne, Epire, Grece continental): description des calcaires de Louros. Revue de Micropaleontologie 31: 49-55 Karakitsios V, Pomoni-Papaioannou F (1998) Sedimentological Study of the Triassic Solution-collapse Breccias of the Ionian zone (NW Greece) Carbonates and Evaporites 13(2): 207-218 Katsikatsos G (1992) The Geology of Greece. University of Patras, Greece, 451 pp Leontiadis IL, Nikolaou E, Dotsika E (2006) Environmental isotopes in determining groundwater flow systems, Epirus, Greece. Bulletin of the Geological Society of Greece, 14(2):47-70 Leontiadis I, Smyrniotis Ch (1986) Isotope hydrology study of the Louros Riverplain area. Proceedings of 5th International Symposium on underground water tracing, Athens, pp 75-90 Macleod DA, Vita-Finzi C (1982) Environment and provenance in the development of recent alluvial deposits in Epirus, NW Greece. Earth surface processes & landforms 7:29-43

105

Mahler BJ, Personné JC, Lods GF, Drogue C (2000) Transport of free and particulateassociated bacteria in karst. Journal of Hydrology 238:179-193 Nikolaou (1991) Study of diet of the groundwater systems of Epirus. Report, Institute for Geology and Mineral Exploration, Preveza, Greece, 200 pp Nikolaou (2005) Quantitative and qualitative features of groundwater supplies of Epirus-management suggestions. Report, Institute for Geology and Mineral Exploration, Preveza, Greece, 9 pp Philippson A (1896) Reisen und Forschungen in Nordgriechenland. Zeitschrift der Gesellschaft fuer Erdkunde 1:193-249 Pulido Bosch A (1999) Karst water exploitation. In: Karst Hydrogeology and Human Activities. Impacts, Consequences and Implications, D. Drew y Heinz Hötzl (eds) International Contributions to Hydrogeology, Balkema, Rotterdam, 20:235-240 Renz C (1955) Stratigraphie Griechenlands. Institute of Geology and Subsurface Research, Athens, 637 pp Rigakis N, Karakitsios V (1998) The source rock horizons of the Ionian Basin (NW Greece). Marine Petroleum Geology 15:593-617 Skourtsis-Coroneou V, Manacos C (1995) New micropaleontological data on the age of the onset of the deposition of the Vigla Limestones Formation. Special Publication of the Geological Society of Greece 4:269-274 Skourtsis-Coroneou V, Solakius N (1999) Calpionellid zonation at the Jurassic/ Cretaceous boundary within the Vigla Limestone Formation (Ionian Zone, Western Greece) and carbon isotope analysis. Cretaceous Research 20:583-595 Skourtsis-Coroneou V, Solakius N, Constantinidis I (1995) Cretaceous stratigraphy of the Ionian Zone, Hellenides, Western Greece. Cretaceous Research 16:539-558 U.S. Environmental Protection Agency (1976) Quality criteria for water. Washington, DC, 501 pp

106

Karstology and motorway construction Martin KNEZ, Tadej SLABE Karst Research Institute, Scientific Research Centre of the Slovenian Academy of Sciences and Arts, P.O.Box 59, Postojna, Slovenia, e-mails: [email protected]; [email protected]

Abstract: One of the major ongoing projects in Slovenia is to link the country with modern expressways. Almost half of Slovenia is karst and more than half of its supply of water comes from karst aquifers. Slovenia is the home of the Classical Karst region, which gave its name to numerous world languages for the type of landscape that develops on carbonate rock and where the science of karstology began to develop. Comprising an important part of our natural and cultural heritage, the sensitive karst landscape demands from us good knowledge and serious effort for its preservation. Since 1994, Slovene karstologists have cooperated closely in the planning and construction of expressways in karst regions. With the consideration of the integrity of the karst landscape in the foreground, we have recommended avoiding more important areas of karst phenomena (sinkholes, poljes, collapse dolines, karst walls, etc.) and already known caves in the selection of routes for expressways and railway lines. We have devoted special attention to the impact on karst waters of building and using the expressways. Expressway should be impermeable. Water from the road surface is first collected in oil separators and then released clean into the karst. We have also studied the pollutants in the water that flows off the expressways everyday. Construction work has provided a series of important discoveries about the formation of karst and its development on various bedrock, in different conditions and through various processes. We have studied the karst along expressways between towns Razdrto, Kastelec, and Fernetii (southwest Slovenia), the central part of the Dolenjska karst region (south Slovenia), and the young karst in the Vipava Valley (southwest Slovenia). This selection of sites includes the most important areas of our karst regions. We have acquired a great deal of information about surface karst phenomena and the epikarst, and where excavation work has cut deeper in the surface and in tunnels about the vadose zone and the paleokarst. Everywhere the development of the karst left important traces, above all in the numerous old caves. More than 350 new caves have been opened. The regular research of karst features revealed during the construction of expressways has enriched our knowledge of the natural and cultural heritage and deepened karstological knowledge. The research results are also a starting point for spatial planning in karst areas and for protecting the karst landscape. Keywords: motorway construction, karst, karst cave, classical karst (kras), Slovenia

1 Introduction Over the last fifteen years the construction of modern expressways in Slovenia has been one of the major construction projects aimed at connecting important parts of the country and opening them to Europe. Almost half of Slovenia is karst and more than half of the water for the supply of the population comes from karst aquifers. Slovenia is home to the classical karst region of Kras that gave its name for this unique carbonate rock landscape to numerous world languages and is also the cradle of karstology. We need to better understand this fragile karst landscape and do everything to preserve it since it is an important part of our natural and cultural heritage. We focused on examples from the Classical karst, the low karst of the Dolenjska region, and the karst breccia in the Vipava Valley. Special attention is devoted to Kras, a karst plateau rising above the north westernmost part of the Adriatic Sea that is bordered on the southwest by a vast flysch area with elevations 107

exceeding 600 meters. Lying between 200 and 500 meters above sea level, the plateau covers 440 square kilometres and in a broad sense belongs to the Outer Dinaric Alps. From the viewpoint of the theory of tectonic plates, the plateau lies at the northern deformed edge of the Adriatic plate and is the result of tectonic overlapping. Only Cretaceous and Paleogenic rocks are found here. They are characterized by exceptionally varied limestone that mostly formed in relatively shallow sedimentation basins with lush fauna and flora. On the Kras plateau there are no remains of the surface waters used in the past to explain the development of the plateau. Originally, the plateau was surrounded and covered with flysch and therefore flooded. Vertical percolation was minimal. The water table later dropped several hundred meters into the karst. At the contact between the carbonate rock and flysch, surface waters created characteristic and extensive contact karst. Today, all Kras rivers sink where they flow from flysch onto limestone bedrock and flow underground toward the springs of the Timava River in Italy. The largest stream is the Reka River, which sinks in the Škocjan Caves, while 65% percent of the water sinks from the surface in a dispersed fashion. From the ecological standpoint, Kras has one of the most vulnerable natural systems in Slovenia. The low karst of the Dolenjska region is mostly covered with a variety of alluvia under which a unique karst surface formed with stone forests as one of its most distinctive features (Knez et al. 2003). The water table is often just below the surface, and the valley systems are occasionally flooded. Karst areas also developed in the breccias that formed from the scree on the slopes of Mount Nanos. They lie on more or less permeable flysch, and water flowing at the contact carved the largest caves in this area. For a number of years, karstologists have cooperated in the planning and construction of expressways in the Kras region (Kogovšek 1993, 1995, Knez et al. 1994, Knez and Šebela 1994, Šebela and Mihevc 1995, Slabe 1996, 1997a, 1997b, 1998, Mihevc and Zupan Hajna 1996, Mihevc 1996, 1999, Kogovšek et al. 1997, Mihevc et al. 1998, Šebela et al. 1999, Knez et al. 2003, 2004a, 2004b, Bosák et al. 2000, Knez and Slabe 1999, 2000, 2001, 2002a, 2002b, 2004a, 2004b, 2005, 2006). In the selection of expressway and railway routes, the main consideration is the integrity of the karst landscape and therefore the routes chosen avoid the more important surface karst features (dolines, poljes, collapse dolines, karst walls) and already known caves. Special attention is devoted to the impact of the construction and use of expressways on karst waters. Expressways should therefore be impermeable so that runoff water from the road is first gathered in oil collectors and then released clean onto the karst surface. We studied the impact of traffic routes on karst waters. Kogovšek (1993, 1995) determined the contents of polluted water flowing daily from the expressways. Small quantities of stagnant water found in caves along the expressways contained traces of mineral oils (Knez et al. 1994). During the construction of expressways we also perform karstological monitoring. We study newly revealed karst phenomena as an important part of our natural heritage and advise on how to preserve them if the construction work allows it. At the same time our new findings are of great help to the construction companies. We have acquired a number of new findings on the formation and development of the karst surface, epikarst, and the perforation of the aquifer. 2 Exploring the karst surface and new caves during expressway construction The removal of soil and vegetation from the karst surface and of course major earthworks such as the excavation of cuts and tunnels reveal surface, epikarst, and subsoil karst features. Our task is to study these features as part of the natural heritage, advise on how

108

to preserve them, and of course share our new findings with the builders. These findings are used to overcome construction obstacles. The karst surface is dissected by dolines and unroofed caves. Dolines are a sign of the current shaping of the surface by precipitation water that percolates vertically through it and passes through the vadose part of the aquifer to the underground water. Some dolines are more distinctly filled with soil than others. There are shafts and fissures at their bottoms through which water flows. The soil must be removed from the dolines and their bottoms reinforced with rocks arranged in a vault-like pattern; the mouths of shafts are often smaller than the chambers beneath them. The dolines are then filled with layers of rubble. Unroofed caves have a similar form or are more oblong. These are old caves that appear on the surface due to the lowering of the karst surface and no longer have the upper part of their circumference. The fine-grained fill, in this case old cave alluvia, must be removed and replaced with rocks and rubble. Otherwise, water could gradually carry the alluvia away and cause subsidence on the surface. The epikarst is crisscrossed with fissures that are more distinctive in Cretaceous limestone and less so in Paleogenic limestone, and many of them open at the bottoms and slopes of dolines. In most cases they are filled with soil and their walls are dissected with subsoil rock relief forms. Due to the lowering of the karst surface, many shafts are now located just below the surface. More than 350 caves were opened on the 70-kilometer section of expressway built in Kras in the last few years (Figure 1). Relative to the development of the aquifer, we distinguish between old caves through which watercourses flowed when the karst aquifer was surrounded and covered by flysch and shafts through which water vertically percolates from the permeable karst surface to the underground water. The deepest shaft found measured 110 meters. Some old caves are empty, almost two thirds of them are filled with alluvia, and one third are unroofed caves.

Figure 1 Discovered caves and closing and preservation of them 109

Caves are opened when vegetation and soil is removed from the surface, and a large number of caves were opened during the excavation of cuts. Blasting caused their roofs to collapse, and cross sections of passages were preserved in embankments. The most shafts were opened at the bottoms of dolines when the soil and alluvia were removed. We studied all the caves, drew their plans, determined their shape, examined the rock relief, collected samples of alluvia for paleomagnetic and pollen analyses, and sampled flowstone for mineralogical analyses and age determination. We extrapolated the further extent of the caves on the basis of their shapes and the geological conditions, which is especially useful for road builders. 3 Studies that accompanied construction produced new findings on karst development The unroofed cave is a special and frequent karst form. Today, this significant karst surface feature is a familiar phenomenon, but it had not been thoroughly studied before the construction of the expressway across Kras. Great attention has been devoted to unroofed caves since the occurrence of this phenomena turned out to be considerably higher than previously expected, and numerous articles on unroofed caves and the construction of new expressways are now available (Knez and Šebela 1994, Šebela and Mihevc 1995, Slabe 1996, 1997a, 1997b, 1998, Mihevc and Zupan Hajna 1996, Mihevc 1996, Kogovšek et al. 1997, Mihevc et al. 1998, Šebela et al. 1999, Knez and Slabe 2000, 2001, 2002a, 2002b, 2004a, 2004b, 2005, 2006). The shape of an unroofed cave is the consequence of the type and shape of the cave and the development of the karst aquifer and its surface in various geological, geomorphological, climate, and hydrological conditions. The distinctiveness of the surface shape of an unroofed cave is dictated by the speed at which the alluvium was washed out of the cave relative to the lowering of the surrounding surface. If the speed was low, we can often see soil and vegetation or areas of alluvium and flowstone on the surface; where it was faster, unroofed caves on the karst surface resemble dolines, a string of dolines, or oblong depressions. Frequently they are an interweaving of various old forms such as caves and recent shaping of the karst with dolines and shafts. A large proportion of the caves were filled with alluvia, in most cases fine-grained flysch alluvia with intervening layers of gravel. We took alluvia samples for paleomagnetic research from caves at Kozina and Divaa and determined they originated in the Olduvai period. We therefore concluded that the caves were filled after the Messinian crisis approximately 5.2 million years ago (Bosák et al. 2000). In short, unroofed caves are an increasingly recognized feature of the karst surface, an important part of the epikarst, and an exceptional trace of the development of the karst aquifer. Determining the age of the alluvia helps us understand the oldest periods of karstification and has proven that the oldest caves in Kras are much older than earlier karstologists thought. 4 Planning road construction Before construction starts we verify the accuracy of known data about caves in the field and add possible new measurements and explanations of their development. To throw light on the situation we present existing data on perforation of the aquifer and elaborate prognosis subsurface maps with special emphasis on the anticipated lithological and tectonic changes in rock composition and structure. Before the start of construction work we try to present the perforation of the karst as accurately as possible. We can determine the position of underground caves by drilling, and along with measurement indicators we also determine the type of potential fill material (flowstone, alluvia). To a certain extent we can anticipate the 110

shape, type, and occurrence of caves in the vicinity using our knowledge of the known surface and underground features. Perforation puts a special stamp on the construction of expressways in the Kras region. In addition to its varied development, Slovenia’s karst is marked by tectonic and lithostratigraphic diversity and it is therefore difficult to determine in advance where caves will occur. As a rule, caves occur more frequently along the contacts of flysch with limestone. The perforation of the karst aquifer is therefore determined primarily on the basis of good and comprehensive knowledge of the karst and continuous intensive work in planning and constructing the expressways. When planning expressways, the link between surface and underground karst features requires the karstological evaluation of the karst surface as well as the karst underground, the hydrological situation, and the presented variables. On all the expressway construction sites in Kras we encountered numerous karst phenomena including dolines, filled and empty caves, and sections of old and current drainage systems through the karst. The lowering of the karst surface exposed many karst caves that are now visible in the Kras region. In recent years, we have focused on unroofed caves “discovered” during the construction of expressways. We are certain that a quality karstological study of the area where a road is planned enables the better selection of a route and is one of the basic starting points for planning expressway construction in this unique and vulnerable landscape. We begin by assembling published literature, archives, and various unpublished studies to learn about the surface karst features, and thus identify dolines, collapse dolines, and other morphological features in particular. Through a field survey we establish the starting points for mapping the areas of the selected route. In the field, we evaluate different types of rock from the karstological aspect. On theme maps we present the known entrances to underground caves and supplement them with potential new entrances. We anticipate the branching of underground cave systems on the basis of surface mapping and explanations of the development of morphologically identified unroofed caves visible in the relief. On the basis of surface mapping we also consider possibilities for dumping waste material if necessary. We know from experience that during construction every route crossing Kras will sooner or later encounter underground caves or parts of cave systems. To a certain degree we can predict the shape and type of caves using our knowledge of surface and underground phenomena. We trace the caves in the wider area of the traffic route, determine their type, position, and role in the aquifer, their shape, rock relief, the alluvia and flowstone found in them, and present them on suitable maps. To make the maps easier to read, we present the previous data on the perforation of the aquifer and elaborate predictions with special emphasis on anticipated lithological and tectonic changes in the rock. Due to the special characteristics of carbonate rock, karst rivers and stream that sink in the studied area easily find direct routes into the underground (karst aquifer); it only takes them an hour to percolate through 100-meter thick rock beds. Although the flysch rock beds in Kras found in permanent direct contact with carbonate rock are often presented as exclusively impermeable beds, it must be emphasized that the flysch (often in thin beds) is only an isolated lens lying on permeable carbonate rock. Furthermore, it must also be observed that a smaller number of underground conductive channels do occur in flysch and that precipitation water collecting on flysch runs off onto carbonate rock. We therefore undertake hydrological mapping in the field. For this purpose we delineate and define the basic characteristics of hydrogeological units in the wider route area, identify hydrological objects (captured and uncaptured springs, surface streams, water caves, boreholes, measuring stations, etc.), and establish the physical and chemical properties of springs. When necessary we perform tracing experiments during low and high waters, primarily to determine the direction and velocity of underground flows in the wider area of the traffic route. With the 111

results of field mapping and tracking experiments, we elaborate and upgrade the existing hydrogeological maps, build a database on the state of the environment, and assess the impact of the construction on karst waters. The basic guidelines for planning traffic routes include: 1. the selection of a route shall be based on a comprehensive assessment of the karst with emphasis on local features; 2. the selected traffic route shall avoid specific exceptional karst features; 3. the conservation of karst aquifers shall be one of the priority goals of planning. 5 Preserving as many karst caves as possible The shafts were easiest to preserve and concrete plates were used to close the smaller entrances (Figure 1). It was similarly possible to preserve old caves with solid circumferences but caves located in fractured rock or opened during blasting had to be filled. Rock walls were used to close caves crossed by road cuts with entrances on embankments. Their circumferences were fractured to such an extent that they were unsuitable for visiting, and water could wash clay from caves filled with alluvia and deposit it on the roads. One well preserved cave was left open for travellers crossing the border with Italy to visit. The most interesting and best preserved caves were completely secured and made accessible for visiting even though they were located under the expressway or even wound around a tunnel as with the Kastelec tunnel. They are accessible via concrete culverts closed at the roadside and in the tunnel with a door. We also studied the consequences of blasting in caves, which will help us in road construction and the preservation of karst features in future. 6 Protecting the karst during expressway construction and use Experience acquired tracing waters and accidental spillages of various substances on the karst surface drew attention to the great perforation of the karst aquifer, which the number of caves newly discovered during construction confirmed. Low permeability occurs only on individual relatively small patches such as the bottoms of dolines heavily covered with soil washed usually from a larger surrounding area, dolines on the Paleogene limestone of the Divaa valley system that are transformed into ponds during rainy periods, and smaller patches of clay that filled old caves. Maximum precautions must therefore be employed during both the construction and use of roads. Daily traffic leaves numerous environmentally harmful substances on road surfaces (Kogovšek 1993), and mineral oils were found in stagnant waters in caves located near traffic routes (Knez et al. 1994). Due to these findings and the persistence of karstologists, expressways are made to be impermeable. Pipes and gutters (Figure 2) along the roads lead to wastewater collectors. Untreated water should never reach the permeable karst surface and the specifications for drainage systems must meet this requirement. The existing wastewater collectors are often too small and abundant precipitation can easily wash the sediments from them. 7 Conclusion It is clear that the cooperation of karstologists in the construction of expressways in the Kras region has brought positive results. It is important that karstologists participate in the planning and construction of expressways and later that they monitor the impact of the expressways on the environment, that is, throughout the entire process of encroachment on the vulnerable karst landscape. This logical cooperation helps preserve natural heritage and increase our basic knowledge about the creation and development of karst and about the construction of expressways in this unique environment. There are many types of karst and each requires a unique approach, which calls for permanent and continuous cooperation 112

between road builders and karstologists. Over the last ten years, the cooperation between the planners and builders of expressways and karstologists has resulted in knowledge used in the planning and implementation of other encroachments in karst areas.

Figure 2 Drainage gutters at the edge of the road

References Bosák P, Pruner P, Mihevc A, Zupan Hajna N (2000) Magnetostratigraphy and unconformities in cave sediments: case study from the Classical Karst, SW Slovenia. Geologos 5:13–30 Knez M, Kranjc A, Otoniar B, Slabe T, Svetlii S (1994) Posledice izlitja nafte pri Kozini. Ujma 9:74–80 Knez M, Otoniar B, Slabe T (2003) Subcutaneous stone forest (Trebnje, Central Slovenia). Acta Carsologica 32(1):29–38 Knez M, Slabe T (1999) Unroofed caves and recognising them in karst relief (Discovered during motorway construction at Kozina, South Slovenia). Acta Carsologica 28(2):103–112 Knez M, Slabe T (2000) Jame brez stropa so pomembna oblika na kraškem površju: s krasoslovnega nadzora gradnje avtocest na krasu. In: A Gostinar (ed) 5. slovenski kongres o cestah in prometu. Družba za raziskave v cestni in prometni stroki Slovenije, Zbornik povzetkov referatov, pp 29 Knez M, Slabe T (2001) Karstology and expressway construction. Proceedings of 14th IRF Road World Congress, CD Knez M, Slabe T (2002a) Unroofed caves are an important feature of karst surfaces: examples from the classical karst. Z. Geomorphologie 46(2):181–191

113

Knez M, Slabe T (2002b) Lithological and morphological properties and rock relief of the Lunan stone forests. In: F. Gabrovšek (ed) Evolution of karst: from prekarst to cessation, pp 259–266 Knez M, Slabe T (2004a) Karstology and the opening of caves during motorway construction in the karst region of Slovenia. International Journal of Speleology 31:159–168 Knez M, Slabe T (2004b) Highways on karst. In: Gunn J (ed) Encyclopedia of caves and karst science, Fitzroy Dearborn, New York, London, pp 419–420 Knez M, Slabe T (2005) Caves and sinkholes in motorway construction, Slovenia. In: T Waltham, F Bell, M Culshaw (eds) Sinkholes and Subsidence. Karst and Cavernous Rocks in Engineering and Construction, Springer, Praxis, Chichester, pp 283–288 Knez M, Slabe T (2006) Krasoslovne raziskave pri gradnji avtocest preko slovenskega krasa. Annales 16(2):259-266 Knez M, Slabe T, Šebela S (2004a) Karstification of the aquifer discovered during the construction of the expressway between Klanec and rni Kal, Classical Karst. Acta Carsologica 33(1):205–217 Knez M, Slabe T, Šebela S (2004b) Karst uncovered during Bi–Korenitka motorway construction (Dolenjska, Slovenija). Acta Carsologica 33(2):75–89 Knez M, Šebela S (1994) Novo odkriti kraški pojavi na trasi avtomobilske ceste pri Divai. Naše Jame 36:102 Kogovšek J (1993) Kakšna je sestava voda, ki odtekajo z naših cest? Ujma 7:67–69 Kogovšek J (1995) The surface above Postojnska jama and its relation with the cave. The case of Kristalni rov. Proc. of Symposium International Show Caves and Environmental Monitoring, pp 29–39 Kogovšek J, Slabe T, Šebela S (1997) Motorways in Karst (Slovenia). Proceedings & Fieldtrip excursion guide, 48th highway geology symposium, pp 49–55 Mihevc A (1996) Brezstropa jama pri Povirju. Naše Jame 38:65–75 Mihevc A (1999) The caves and the karst surface – case study from Kras, Slovenia. Etudes de géographie physique, suppl. XXVIII, Colloque européen – Karst 99, pp 141–144 Mihevc A, Slabe T, Šebela S (1998) Denuded caves. Acta Carsologica 27(1):165–174 Mihevc A, Zupan Hajna N (1996) Clastic sediments from dolines and caves found during the construction of the motorway near Divaa, on the Classical Karst. Acta Carsologica 25:169–191 Slabe T (1996) Karst features in the motorway section between ebulovica and Dane. Acta Carsologica 25:221–240 Slabe T (1997a) Karst features discovered during motorway construction in Slovenia. Environmental Geology 32(3):186–190 Slabe T (1997b) The caves in the motorway Dane–Fernetii. Acta Carsologica 26(2):361–372 Slabe T (1998) Karst features discovered during motorway construction between Divaa and Kozina. Acta Carsologica 27(2):105–113 Šebela S, Mihevc A (1995) The problems of construction on karst - the examples from Slovenia. In: BF Beck, FM Pearson (eds) Karst geohazards, engineering and environmental problems in karst terrain. Proceedings of the Fifth Multidisciplinary Conference on Dolines and Engineering and Environmental Impacts on Karst, pp 475–479 Šebela S, Mihevc A, Slabe T (1999) The vulnerability map of karst along highways in Slovenia. In: BF Beck, AJ Pettit, JG Herring (eds) Hydrogeology and engineering geology of dolines and karst. Proceedings of the Seventh Multidisciplinary Conference on Dolines and the Engineering and Environmental Impacts on Karst, pp 419–422

114

Shilin - lithological characteristics, form and rock relief of the Lunan Stone Forests (South China karst) Martin KNEZ, Tadej SLABE Karst Research Institute, Scientific Research Centre of the Slovenian Academy of Sciences and Arts, P.O.Box 59, Postojna, Slovenia, e-mails: [email protected]; [email protected]

Abstract: Stone forests are unique karst surface landforms. The Lunan stone forests developed from underground karren, and where this type of surface is highly developed in China, it is defined as a “shilin” landscape. Numerous examples of stone forests (shilin) that developed in almost identical conditions show that the diverse shape of the pillars is primarily a consequence of the properties of the rock, the distribution and density of joints and fissures in the rock, and its diverse stratification and composition. The Lunan stone forests formed in early Permian carbonates of the Qixia and Maokou formations. Characteristic of these are frequent alternations of very pure limestone, dolomitized limestone, and dolomite, the alternation of thin and thick layers, and in some places distinctive late diagenetic dolomitization and secondary porosity. The layers are mostly horizontal or inclined by five to ten degrees. Due to vigorous tectonic action, they are fractured by numerous vertical and subvertical joints and fissures. The diverse fracturing, stratification, and rock composition are reflected in the shapes of the stone forests and their stone pillars. In the same stone forest, which developed on diversely composed rock, pillars may be of various but typical shapes, the consequence of their development on different levels of a diverse rock column. The shape of stone pillars occurring on thicker and uniformly composed rock strata reflects primarily the development from subsoil karren into a stone forest, and the traces of subsoil factors are gradually reshaped by rainwater. Cross sections of stone pillars occurring on thin rock strata are often jagged, and their tops (even of thinner pillars), which as a rule are pointed, are often flat, the consequence of the rapid disintegration of thin strata. Porous rock strata are most often perforated below the ground and disintegrate faster on the surface; the pillars are therefore narrower and the tops on such rock have no characteristic shapes. More resistant rock strata protrude from the cross section. The tops of the narrower pillars are sharp, sharpened as much by subsoil factors as by rainwater. The broader tops, however, are dissected by points and funnel-like cups. Because of the exceptional characteristics of this karst phenomenon in China, we propose that the term “shilin” be used for this type of stone forest in the professional literature. Keywords: Shilin, karst surface development, rock relief, Yunnan, China

1 Introduction The Lunan stone forests (Figures 1, 2, 3) are a unique form of karst karren (Chen et al. 1998, Knez 1998, Slabe 1998, Knez and Slabe 2001a, 2001b, 2002). The karren, crisscrossed by fissures along cracks, are composed of rock pillars (Song 1986, Habi 1980, 109) or rock teeth (Song 1986, Song and Liu 1992, Yuan 1991). Rock teeth (Figure 4) are smaller protrusions not exceeding five meters; high teeth tower more than three meters in height while lower teeth do not exceed one meter (Song and Liu 1992). According to their shape, they are divided into conical, angular, and oblong rock teeth (Song and Liu 1992). The pillars reach from five to fifty meters high and display a variety of shapes. Large stone forests are a characteristic form of subtropical climate conditions (Song 1986, 3, 5). According to their position, Song (1986, 6, 7, 8) distinguishes three types of stone forests. Tall forests with intervening sinkholes and collapse sinkholes occur in valley systems 115

and valleys. Underground waters flow underneath them; they are periodically flooded or water flows through them. Stone forests located on the tops of hills appear on the tops of hills

Figure 1 Naigu stone forest, Lunan appear to be lower (10–30 m), the pillars grow from a uniform base, and the sediment cover above them is thin. Forests located on the slopes of hills are an intermediate form between these two types.

Figure 2 Pu Chao Chun stone forest, Lunan

Figure 3 Lao hei Gin stone forest, Lunan

116

The Lunan stone forests are often described as a form of covered karst (Chen et al. 1983, Maire et al. 1991, Sweeting 1995, 125). The carbonate rock in which the karren formed was covered by thick layers of sediment that had a major influence on the formation and shape of the stone forests. Relative to the sediments, stone forests can be bare, covered, or buried. Hantoon (1997, 311) describes stone forests as an epikarst form. Mangin (1997, 106) believes that the epikarst of stone forests reaches depths up to 100 meters. The Lunan stone forests formed mainly when the water below the soil, which contains biogenic CO2 and sediments, dissolved the rock under the ground. The water widens the cracks and separates rocks. Under acidic soil, wide and deep fissures formed between pillars and deep channels developed on their walls (Yuan 1997). Exposed carbonate rocks are further shaped by rainwater. Teeth are first to form and over time develop into forests (Song 1986, 13). At first, limestone that was already undergoing karstification (Yu and Yang 1997, Song and Wang 1997, 433) was covered by Permian basalt and tufa that influenced its formation and the rock, which metamorphosed in places (Song and Li 1997, Ford et al. 1996, 34). Water percolated through the basalt and tufa leading to the formation of underground karst. The Mesozoic saw the denudation of parts of the limestone (Song and Wang 1997, 433). In the Oligocene and Miocene, blocks of rock rose and lowered, and the lower parts of the karst relief were reshaped by erosion (Yu and Yang 1997). In the Eocene, the Lunan ridge collapsed and thick layers of lake sediments were deposited (Chen et al. 1983, Zhang Geng et al. 1997; Song and Wang 1997, 433). In the tropical climate, thick layers of laterite soils formed on these sediments (Sweeting 1995, 124, Ford et al. 1997, 112). The Pliocene saw the beginning of the development of the current stone forest (Yu and Yang 1997, 66). In the Quaternary, a large part of the sediment was removed although some was preserved in the cracks.

Figure 4 Rock teeth, Lunan

117

The level of the underground water played an important role in the development of the forest (Ford et al. 1997, 114). The development of underground water courses initiated the transformation of teeth into pillars (Zhang, Geng et al. 1997). The oscillating underground water widened the cracks (Yuan 1991). Today, an integrated and diverse network of water courses had developed below the stone forests (Zhang, Geng et al. 1997, 5). The fragmenting of rock due to tectonic action caused the lowering of the underground water table, swept the sediments from the surface, and accelerated the growth of the forests. The central part of the stone forest spreads over 80 hectares, and larger and smaller stone forests are distributed over some 350 square kilometres (Chen et al. 1983, Zhang Geng et al. 1997). The central part is situated 1,625 to 1,875 meters above seal level in a valley system. After abundant precipitation, the underground water, which is close to the surface, rises by ten meters. Most (70–80%) of the annual rainfall of up to 936.5 mm falls between June and October (Chen et al. 1983). The average temperature is 16.3º C, and temperatures range from -2º C to 39º C. The tallest pillars are located in the central part of the valley system where surface waters drain into the underground, and more sediments are deposited at the edges of the stone forest (Sweeting 1995, 125). Habi (1980, 110) calls it “shallow karst” because water also flows on the surface in the lower part of the stone forest. The tourist areas of the Lunan stone forests are visited by more than two million visitors each year. The forests are a unique and integral natural and cultural landscape where tourism is now an important part of the Sani minority’s economy. The shape of the pillars in the forest and their height are characteristic of individual types of rock and their topographical position (Zhang, Wang et al. 1997, 73). Numerous examples of stone forests that developed in almost identical conditions show that the diverse shape of the pillars is primarily a consequence of the properties of the rock, from the distribution and density of joints and cracks in the rock to its stratification and composition. However, we must also consider the significance of the effect on their shaping of subsoil factors and rainwater, that is, the course of their development in various periods. 2 The formation of stone forests 2.1 Lithological characteristics of rock The stone forest area consists of Lower Permian carbonates of the Qixia and Maokou formations. These are two of the most important basal formations from which numerous stone forests emerged in the southern Yunnan province of Lunan. Typical for Qixia formations are micrite limestone with intercalary dolomite and dolomitized limestone with intervening layers of schist. In the lower part of Maokou formations, limestone alternates with dolomite and dolomitized limestone. In the upper part we find a sequence of limestone layers that in some places are thin and in others several meters thick as well as solid limestone that contains several-decimetre large nodules of chart in individual horizons. The main lithological properties of the Maokou formations are roughly similar to those of the Qixia formations, except that in Maokou carbonates we do not find a major influence of late diagenetic dolomitization and in some places a considerable secondary porosity. However, both show a strong diagenetic alteration of the basic rock, which is undoubtedly also a consequence of intensive volcanic (basalt lava) activity during the transition from the Palaeozoic to the Mesozoic. The rock contains an extremely high percentage of carbonate. In the area studied we find considerable variations in thickness, porosity, and degree and type of dolomitization, in the components of inclusions, and in the colour of individual layers that are reflected in the formation of the stone forests.

118

2.2 Rock fissuring Morphological characteristics reflect various factors, among which geological factors are the most important. The fracturing of the rock is without a doubt one of the basic factors affecting the shape of the forest and the size of the rock pillars. The distribution of the pillars (stone forest ground plan) corresponds to the fracturing of the rock. The pillars can be linked in rows between distinct fault areas or close together, and the stone forest or parts of it can consist of individual wide or narrow pillars. As a rule, pillars with smaller cross sections occur with a dense network of cracks, and larger rock masses with broader tops with a sparser network. 2.3 Rock stratification Rock stratification is a very important factor affecting the shape of rock pillars. It has little or no effect on the pillars that developed on thick layers and layers with uniform rock composition. Longitudinal sections of pillars on thin rock strata are often jagged since they are dissected by notches occurring along the bedding planes and their shapes reflect the uneven resistance of the different rock strata to the factors of their formation. 2.4 Rock composition The composition of the rock is another important geological factor. The rock’s composition, particularly if it is diverse, can decisively influence the shape of the rock pillars, as much their longitudinal sections as the size of the cross sections. Porous strata are often perforated and disintegrate more rapidly, while rock strata with less soluble components usually protrude from the walls. 2.5 Subsoil shaping of rock To understand both the regional and local development of the rock pillars we must also consider the influence of the subsoil rock shaping that fostered the characteristic formation of the pillars on various rocks and the unique rock relief. Rainwater sharpens the tops of the pillars and reshapes the traces of the original subsoil shaping. 2.6 Rock relief The unique development of the stone forests is also reflected in their subsoil rock relief. The most distinct and particularly the largest rock forms are subsoil and composite rock forms. Subsoil rock forms include scallops, large channels (Figure 5), notches and half-bells (Figure 6), and subsoil channels and cups on broader tops. Composite rock forms include the channels that lead from subsoil channels or cups and dissect pillar walls. Many pillars are undercut below the ground, while their tops have been reshaped by secondary subsoil rock forms and forms carved by rainwater. The rock relief of larger rock pillars is unique as well, particularly those with broader tops, either on thick rock strata where secondary subsoil forms occur or on tops that developed due to the disintegration of thin rock strata when subsoil tubes occurring along bedding planes developed into subsoil forms or large channels that were reshaped by rainwater. Both forms indirectly influence the shape of the pillar walls due to water flowing from them and carving channels. Rainwater gradually transforms the subsoil rock relief. As a rule, smaller rock forms do not occur on dolomite rock, on very porous rock, or on rock with larger intraclasts. 3 Discussion and conclusions Numerous examples of stone forests that developed in almost identical conditions show that the diverse shape of the pillars is primarily a consequence of the properties of the rock, from the distribution and density of joints and cracks that crisscross it to its stratification and 119

composition. However, we must also consider the significance of the effect on their shaping by subsoil factors and transformation by rainwater, that is, the course of their development in various periods. The Lunan stone forests formed in Lower Permian carbonates of Qixia and Maokou formations. They are characterized by a varied alternation of very pure limestone, dolomitized limestone, and dolomite, the alternating of thin and thick strata, distinct late diagenetic dolomitization in places, and secondary porosity. The strata are mostly horizontal or inclined by five to ten degrees. Due to vigorous tectonic action, they are fractured by numerous vertical and subvertical joints and cracks. The diverse fracturing, stratification, and rock composition are reflected in the shapes of the stone forests and their rock pillars. In the same stone forest, which developed on diversely composed rock, pillars may be of various but typical shapes, the consequence of their development on different levels of a diverse rock column.

Figure 5 Large channels, Lunan

Figure 6 Half-bell, Lunan 120

The shape of rock pillars occurring on thicker and uniformly composed rock strata reflects primarily the development from subsoil karren into a stone forest, and the traces of subsoil factors are gradually reshaped by rainwater. Cross sections of rock pillars occurring on thin rock strata are often jagged, and the tops, even of thinner pillars that as a rule are pointed, are often flat, the consequence of the rapid disintegration of thin strata. Porous rock strata are most often perforated below the ground and disintegrate faster on the surface; the pillars are therefore narrower and the tops on such rock have no characteristic shapes. More resistant rock strata protrude from the cross section. The tops of the narrower pillars are sharp, sharpened as much by subsoil factors as by rainwater. Broader tops, however, are dissected by points and funnel-like cups. The unique rock relief occurs when the rock is shaped under the ground. The denuded rock is reshaped by rainwater that carves flutes, channels (Figure 7), and solution pans.

Figure 7 Rock, reshaped by rainwater, Lunan The development of stone forests and their rate of growth in a particular period are also influenced by the position and development of karst caves below them, that is, by the manner the water—and the sediment and soil with it—flows from the karst surface. Various development periods can be determined from the karst caves. Because of the exceptional characteristics of this karst phenomenon in China, we propose the term “shilin” be used for this type of stone forest in the international professional literature. References Chen Z, Song, L, Sweeting, MM (1983) The pinnacle karst of the stone forest, Lunan, Yunnan, China: an example of sub-jacent karst. In: K. Paterson, MM Sweeting (eds) New Directions in Karst, Proceedings of the Anglo-French Karst symposium, pp 88-124 Chen X, Gabrovšek F, Huang C, Jin Y, Knez M, Kogovšek J, Liu H, Petri M, Mihevc A, Otoniar B, Shi M, Slabe T, Šebela S, Wu W, Zhang S, Zupan Hajna N (1998) South China Karst I. ZRC Publishing, Ljubljana, 247 pp 121

Ford D, Salomon JN, Williams P (1996) Les »Forets de Pierre« ou »Stone forests« de Lunan. Karstologia 28(2):25-40 Ford D, Salomon JN, Williams P (1997) The Lunan stone forest as a potential world heritage site. Stone forest, a treasure of natural heritage. China environmental science press, pp 107-123 Habi P (1980) S poti po kitajskem krasu. Geografski Vestnik 52:107-122 Hantoon PW (1997) Definition and characteristics of stone forest epikarst aquifers in South China. Proceedings of 12th International Congress of Speleology 1(8):311-314 Knez M (1998) Lithological properties of the three Lunan stone forests (Shilin, Naigu and Lao Hei Gin). In: X Chen et al. (eds) South China Karst I, pp 30-43 Knez M, Slabe T (2001a) Shape and rock relief of pillars in Naigu Stone Forest (SW China) Acta Carsologica 30(1):13-24 Knez M, Slabe T (2001b) The lithology, shape and rock relief of the pillars in the Pu Chao Chun stone forest (Lunan stone forests, NW China). Acta Carsologica 30(2):129-139 Knez M, Slabe T (2002) Lithologic and morphological properties and rock relief of the Lunan stone forests. In: F Gabrovšek (ed) Evolution of karst: From prekarst to cessation, pp 259-266 Maire R, Zhang S, Song S (1991) Genèse des karsts subtropicaux de Chine du Sud (Guizhou, Sichuan, Hubei). Karstologia mémoires 4:162-186 Mangin A (1997) Some features of the Stone forest of Lunan, Yunnan, China. Stone forest, a treasure of natural heritage. China environmental science press, pp 105-106 Slabe T (1998) Rock relief of pillars in the Lunan Stone Forest. In: X Chen et al. (eds) South China Karst I, pp 51-67 Song L (1986) Origination of stone forest in China. International Journal of Speleology 15:3-33 Song L, Liu H (1992) Control of geological structures over development of cockpit karst in south Yunnan, China. Tübingen Geographische Studien 109:57-70 Song L, Li Y (1997) Definition of Stone forest and its evolution in Lunan County, Yunnan, China. Stone forest, a treasure of natural heritage. China environmental science press, pp 37-45 Song L, Wang F (1997) Lunan Shilin Landscape in China. Proceedings of 12th International Congress of Speleology 1(8):433-435 Sweeting MM (1995) Karst in China. Its Geomorphology and Environment. Springer, Berlin, Heidelberg, New York, 265 pp Yuan D (1991) Karst of China. Geological Publishing House, Beijing, 224 pp Yuan D (1997) A global perspective of Lunan Stone forest. Stone forest, a treasure of natural heritage. China environmental science press, pp 68-70 Yu Y, Yang B (1997) Paleoenvironment during formation of Lunan Stone Forest. Stone forest, a treasure of natural heritage. China environmental science press, pp 63-67 Zhang F, Wang F, Wang H (1997) Lunan Stone forest landscape and its protection and conservation. Stone forest, a treasure of natural heritage. China environmental science press, pp 71-77 Zhang F, Geng H, Li Y, Liang Y, Yang Y, Ren J, Wang F, Tao H, Li Z (1997) Study on the Lunan stone forest karst, China. Yunnan Science and Technology Press, Kunming, 155 pp

122

Contribution of time series analysis to the study of the Malenšica karst spring, Slovenia Gregor KOVAI1, Metka PETRI2 1

University of Primorska, Faculty of Humanities Koper, SI-6000 Koper, Titov trg 5, Slovenia 2 Karst research institute SRC SAZU, SI-6230 Postojna, Titov trg 2, Slovenia

Abstract: The paper deals with the application of the time series analysis in the hydrological background of the Malenšica karst spring in the southwestern Slovenia. The results of the analysis contribute to the better understanding of the functioning of the complex Malenšica karst spring system, which extends over 700 km2. Correlation and spectral analysis between daily and hourly hydrologic time series between the input function; precipitation and secondary infiltration via ponors; and output function (springs) to characterize the hydrologic functioning of the Malenšica karst spring catchment was performed. The results of the autocorrelation analysis (daily values) show that the storage capacity of the Malenšica karst spring system is moderate, but the highest among all of the measuring points in its hydrological background and vicinity. The cross-correlation function between the daily precipitation in the background and daily discharges of the spring show the 2 day delay. The autocorrelation and cross-correlation analysis for the hourly precipitation, discharges, water levels, electric conductivity (EC) and temperature (T) values from the measuring points in the near catchment of the Malenšica karst spring and its capture for the hydrological year 2008 was performed as well. The storage capacity of the Malenšica karst spring is rather low. Autocorrelation shows a high memory effect for the T and some less for the EC parameter. From the methodological point of view, the univariate and bivariate time series analysis performed for the daily values in the hydrological year 1975 and hourly values in the hydrological year 2008 showed different results. This indicates that the selection of the investigated hydrological year and time lag are important factors in time series analysis. Keywords: time series analysis, correlation, spectral analysis, karst aquifer, the Malenš ica karst spring, Slovenia

1 Introduction The purpose of this study is to demonstrate that the time series analysis can contribute to the hydrogeological study of a karst aquifer on a regional scale. The Malenšica karst spring is one of the most important karst springs in Slovenia. Because of its favourable hydrological characteristics, since 1971 it has been captured for the drinking water supply. Hydrogeologic and hydrodynamic characteristics of the Malenšica karst spring and its catchment are well investigated, but still we do not have all the answers considering the quantities and directions of recharge towards the spring in different hydrological conditions (low or high waters). In this regard the correlation and spectral analysis between daily and hourly hydrologic time series of the input function (precipitation and direct infiltration through ponors) and output function (springs) to characterize the hydrogeologic processes of the Malenšica karst spring catchment was performed. 2 General characteristics of the Malenšica karst spring The mean annual discharge of the Malenšica karst spring in the reference period 19611990 is 6.704 m3/s. The ratio between Qmin, Qmean and Qmax is 1 : 6.1 : 9, Qmin is always above 1.1 m3/s. In the average hydrological year two maximums, primary in April (snowmelt) and secondary in December (precipitation), and two minimums, primary in August and secondary 123

in February (snow retention) appear. The catchment area of the spring extends over karst poljes of Notranjska region and in the Snežnik and Javorniki high Dinaric karst plateaus on 726 km2. In the Malenšica karst spring catchment rocks with karst-fissure porosity prevail, covering over 73 % of the area, 13 % of the area is characterized by fissured porosity (dolomites), in some parts of the catchment area also surface runoff is present. The Malenšica karst spring is recharged both by autogenic and allogenic recharge from sinking rivers. The mean annual precipitation in the catchment area of the spring (period 1961-90) is estimated to 1776 mm, the mean annual evapotranspiration is 720 mm (period 1971-2000). The mean annual runoff is estimated to 1055 mm, the runoff coefficient is around 60 % (Kovai 2009). The mean annual discharge of the Malenšica karst spring is 6.92 m3/s in the hydrological year 1975 and 5.83 m3/s in the hydrological year 2008. The amount of annual precipitation in the catchment of the investigated spring in hydrological year 2008 is around 1600 mm.

Figure 1 The Malenšica karst spring at high waters (Photo: G. Kovai, April 4th 2008) 3 Methodology Time series analysis of hydrological data comprises the mathematical analysis of karst systems’ response to recharge (precipitation, concentrated infiltration via ponors) and indirectly obtains information regarding the functioning of karst hydrological systems (Box and Jenkins 1970, Box et al. 1994). The univariate analysis characterizes the structure of individual time series, either in time (autocorrelation) or frequency domain (spectral density). Autocorrelation function quantifies the memory effect of the system and gives indirect information about the storage capacity of the system. Spectral density function quantifies the regulation time, which defines the duration of the influence of the input signal and it gives an indication of the length of the impulse response of the system (Larocque et al. 1998). Crosscorrelation and cross spectral density functions imply the transformation of the input signal (precipitation, direct infiltration via ponors) to the output signal (karst spring, wells). The transformation of the signal within the karst system depends on its structure. For more detailed theoretical basis of time series analysis and explanations of the application of the methodology to different test sites the reader is referred to Angelini (1997), Larocque et al. (1998), Mathevet et al. (2004), Padilla and Pulido-Bosch (1995), Panagopoulos and Lambrakis (2006), Rahnemaei et al. (2005) and Samani (2001).

124

Firstly we analyzed the daily precipitation values from 13 precipitation stations and daily discharge (water level) values from 22 gauging stations in the hydrological year 1975, when the most number of gauging stations in the Malenšica karst spring in the observation time was operational. The length of the time series for the hydrological year 1975 is 411 days, from August 28th 1974 to October 12th 1975. These data were received by the Environmental Agency of the Republic of Slovenia. Since we wanted to get more precise information on hydrologic behaviour of the Malenšica karst aquifer, in year 2007 we installed seven measuring instruments on the watercourses in the background of the spring, in its water capture and in its vicinity. In this way we gathered hourly data on water levels, discharges, EC and T for the hydrological year 2008 from September 9th 2007 to October 17th 2008 (9840 hours, 411 days). Hourly precipitation data were collected from 3 automatic rain gauge devices installed in the different locations within the catchment. Univariate and bivariate time series analyses between gathered hydrologic time series were performed. The hydrological time series in the hydrological year 1975 were investigated in time lag of 1 day, in the hydrological year 2008 in time lag of 1 hour. 4 Results of the time series analysis Results of autocorrelation function for the daily values in the hydrological year 1975 show, that the Malenšica karst spring has the largest storage capacity among all of the investigated karst springs and other measuring locations on watercourses in its background and vicinity. The autocorrelation coefficient reaches the rk = 0,2 value after 32 days, revealing that the memory effect of its system is according to Mangin’s (1994) classification moderate. The spectral density function of the spring has a regulation time of 32 days, which indicates that the aquifer’s impulse response is moderate as well. Low storage capacity of the spring indicates high karstification degree of the system. The delays between precipitation events and discharges of the springs in the catchment area of the Malenšica karst spring and itself are mostly within the interval of one day, though noticeable distinction between different pairs of analyzed series can be observed, also in the value of maximum correlation coefficients (from 0.28 to 0.73). The cross-correlation function between the daily precipitation from the most representative station Jurše and daily discharges of the spring show the 2 day delay and a correlation coefficient of 0.28. The results of cross-correlation analysis show that within the period of one day the precipitation events in the background of the Malenšica karst spring occur almost simultaneously, in many cases also the sums of daily precipitation at different stations are roughly equivalent. The autocorrelation and cross-correlation analysis for the hourly precipitation, discharges and water levels from the measuring spots in the near catchment of the Malenšica karst spring and its capture in the hydrological year 2008 show rather different results. The autocorrelation function shows, that the storage capacity of the Malenšica karst spring is rather lower, reaching the rk = 0.2 value after 557 hours (23 days) and the rk = 0 value after 973 hours (40.5 days) (Figure 2a). The spectral density function has a regulation time of 159 hours (6.6 days), which indicates that the aquifer’s impulse response is even shorter as it is in the year 1975, although the discharge of the spring diminishes very slowly over long period. The shape of autocorrelation function of hourly precipitation at the station Postojna diminishes very rapidly and reaches the rk = 0.2 value immediately (3 hours) and the 0 value after 36 hours. Autocorrelation analysis proves a high memory effect for the T and some less for the EC parameter (Figure 2b).

125

1.0

1.0 Malenšica - Q Unica - Q Kotlii - h Postojna - P

0.8 0.6

Malenšica - T Unica - T

0.8

rk

rk

0.6 0.4

Malenšica - EC

Kotlii - T

400

800

Kotlii - EC 0.4

0.2 0.0

Unica - EC

0.2

-0.2 0.0 0

200

400

600

800

1000

0

200

Lag (k)

600

1000

Lag (k)

Figure 2 Autocorrelation functions for the hourly time series in the hydrological year 2008: a) discharges (water levels) and precipitation time series; and b) temperatures and electric conductivity time series 1.0 0.8

Discharge Precipitation - discharge T EC

0.6

r xy (k)

0.4 0.2 0.0 -0.2 -0.4 -1000 -800

-600

-400

-200

0

200

400

600

800

1000

Lag (k)

Figure 3 Cross-correlation functions between precipitation values (station Postojna) and Kotlii karst spring (discharge, T and EC) as input and Malenšica karst spring as output A cross-correlation function for the hydrological year 2008 between the Kotlii karst spring as input and the Malenšica karst spring as output shows no delay and immediate response, meaning that both springs react simultaneously on precipitation events in their background. The distance between the two springs is 4.6 km. The delay between the precipitation events and response of the spring is 37 hours, but the correlation is rather weaker as that for the hydrological year 1975 (daily values), which is quite understandable. The cross-correlation is stronger for the parameters T and EC. T values show long impulsive response between the springs, the correlation is strong, rxy(0) = 0,96. The maximum rxy(k) value for the parameter EC between springs is attained 49 hours after the signal in the Kotlii karst spring (rxy = 0.84), the response time is rather shorter as it is for the parameter T.

126

5 Conclusions The results of the time series analysis as a complementary (indirect) method for the investigation of hydrogeological characteristics and functioning of karst systems show that the complex recharge area of the Malenšica karst spring is functioning rather homogeneously. The Malenšica karst springs shows longer memory effect in comparison to the other karst springs and watercourses in its background and vicinity, but in comparison to the examples of karst springs from above cited literature, its storage capacity is characterized only as moderate. In hydrogeological sense this is the consequence of well developed underground water circulation, which supplies large amounts of water towards the spring, where the limited outflow capacity of small fissures reduces the maximums discharges. That is why during the high waters the Malenšica karst spring contributes only 10 % to all waters of the Planina Polje. The delays between the reactions from ponors to springs are within the interval of one hour, the cross-correlation coefficients of different parameters (discharge, water level, T and EC) diminish very slowly and reach the rxy = 0 value in far lags. Springs and watercourses react instantly and simultaneously to rather homogeneous precipitation events in the background of the Malenšica karst spring, but with different intensity.

Figure 4 Cross-correlation coefficients for the period of hydrological years 1975 and 2008 The univariate and bivariate time series analysis performed for the daily values in the hydrological year 1975 and hourly values in the hydrological year 2008 showed different results. This indicates that the selection of the hydrological year (structure of time series) and the selection of a time lag are important factors in time series analysis, which influence on results and corresponding hydrogeological interpretation of them. In this way it is important to consider these facts while interpreting the results of time series analysis and describing the hydrogeological characteristics of a specific karst aquifer in absolute sense, since the data interpretation can be biased. It is authors’ opinion that the time series analysis is a valuable tool for the hydrogeological investigations of karst springs, since the time series data is 127

relatively easy to collect, but has also its limits and should be used together with other methods used in karst hydrology as a complementary method for the investigation of hydrogeological characteristics and functioning of karst systems. References Angelini P (1997) Correlation and spectral analysis of two hydrogeological systems in Central Italy. Hydrological Science - Journal-des Sciences Hydrologiques 42(3):425-438 Box GEP, Jenkins GM (1970) Time series analysis: Forecasting and control. Holden Day, San Francisco, 575 pp Box GEP, Jenkins GM, Reinsel C (1994) Time series analysis: forecasting and control. 3rd ed., Prentice Hall, New Jersey, 598 pp Larocque M, Mangin A, Razack M, Banton O (1998) Contribution of correlation and spectral analyses to the regional study of a large karst aquifer (Charente, France). Journal of Hydrology 205:217-231 Kovai G (2009) Hidrologija kraškega izvira Malenšica in njegovega hidrografskega zaledja, Doctoral thesis, University of Primorska Koper, 354 pp. Mangin A (1994) Karst Hydrogeology. In: J Gilbert, DL Danielopol, JA Stanford (eds) Groundwater Ecology, Academic Press, San Diego, pp 43-67 Mathevet T, Lepiller M, Mangin A (2004) Application of time-series analyses to the hydrological functioning of an Alpine karstic system: the case of Bange-L'Eau-Morte. Hydrology and Earth System Sciences 8(6):1051-1064 Padilla A, Pulido-Bosch A (1995) Study of hydrographs of karstic aquifers by means of correlation and cross-spectral analysis (France, Spain). Journal of Hydrology 168:73-89 Panagopoulos G, Lambrakis N (2006) The contribution of time series analysis to the study of the hydrodynamic characteristics of the karst systems: Application on two typical karst aquifers of Greece (Trifilia, Almyros Crete). Journal of Hydrology 329:368-376 Rahnemaei M, Zare M, Nematollahi AR, Sedhi H (2005) Application of spectral analysis of daily water level and spring discharge hydrographs data for comparing physical characteristics of karst aquifers. Journal of Hydrology 311:106-116 Samani N (2001) Response of karst aquifers to rainfall and evaporation, Maharlu basin, Iran. Journal of Cave and Karst Studies 63(1):33-40

128

Capacity for managing local development in karst areas Sanja MALEKOVI, Sanja TIŠMA, Anamarija FARKAŠ Institute for International Relations, Zagreb, Lj. F. Vukotinovia 2, Croatia, e-mails: [email protected]; [email protected]; [email protected]

Abstract: The paper will analyze the existing capacity and skills on the local and county level necessary for efficient management of local development and environmental protection in karst areas. The authors will reflect upon the key development problems of karst areas and will propose possible guidelines for further improvement of overall socio-economic development of these areas from point of view of local development initiatives and more efficient management of local development. On the basis of the analysis of current development problems and needs of karst areas, and in line with the guidelines and objectives of the current Croatian regional policy, the authors will present their views in regard to strengthening of the existing capacity of local/county development “actors, i.e. stakeholders, through further raising of levels of knowledge and skills on the local level - particularly from point of view of more effective preparation and realization of development programs. From point of view of specific characteristics of karts areas, proposals will also be given in regard to public participation in planning and managing of sustainable local development. Keywords: karst areas, local and regional development, management of local development

1 Introduction The paper provides an assessment of the current situation in the Croatian karst areas, with particular focus on the existing capacity and skills on the local and county level necessary for efficient management of local development. It provides an outline of the key development problems and needs of karst areas, as well as some guidelines from point of view of more efficient management of local development, which can be fostered through successful preparation and implementation of local development projects, including infrastructure projects and large environmental investments. The paper reflects the main results of research studies elaborated in IMO in the past 3 years, the main study being “The Assessment of Capacities, Competences and Needs of the County Development Agencies in Decision Making in Croatia” elaborated in the Ministry for Regional Development, Forestry and Water Management (MRRFWM). Its basic aim is to point out the current situation in the karst areas in regard to existing and necessary capacity for managing local development. Since local and regional development agencies are the main support institutions and main instruments for promoting overall local development, the paper is focused on the existing capacity in the county development agencies (CDAs). It also analyses the capacity related to physical planning and environmental policy in the departments in towns and county offices (local and regional self government) in the karst areas, since we are aware that the situation in this regard is unfavourable, i.e. related to efficient management of local development, due to the existing bottlenecks. 2 Current situation in the karts areas One of the characteristics of the Croatian karst areas are substantial differences in socioeconomic development among the different areas. This is particularly stressed if we analyse the main socio-economic development indicators in the north of the Adriatic (NUTS II 129

region) karst areas, priory in the Istria and Primorje Goranska county, in comparison to those in the southern Adriatic counties, but also the Karlovac county. The karst areas, with the exception of the major towns and local units along the coast, as well as the mentioned northern Adriatic areas, are to a great extent considered as disadvantaged regions in Croatia, in which around one fifth of local units lag behind in development more then 50% in comparison to the most developed areas in North West Croatia (Puljiz and Malekovi 2007). In such circumstances, this was particularly stressed on the local level, which, till very recently, forced to act on its own, basically followed the practice in relation to lack of development planning. Some karts areas, particularly those in Lika, the hinterland of Dalmatia, but also the majority of the islands, have for a long time suffered from unfavourable geographic conditions, reflected primarily in poor transport communications with major urban areas. Geographic isolation coupled with other unfavourable economic and social factors has resulted with intense and long term migration which has severely deteriorated the human resource base and seriously endangered their long-term development perspectives (Puljiz and Malekovi 2007). The socio-economic capacity for managing development in the karst areas is also reflected in the poor fiscal capacity, resulting with a lack of the basic grounds for development. Further, the poor horizontal and vertical coordination of local development initiatives and activities, which is still present in most of the karst areas, particularly the non urban, and those located in the southern Adriatic counties, as well as the Karlovac county, are a major obstacle to the efficient management of local development. However, the active involvement of the county development agencies is gradually changing this situation, and, from point of view of the recent activities in the framework of the drafted National Strategy for Regional Development (NSRD 2008), it is expected that this negative circumstance will undergo a radical change in the forthcoming year. This process is particularly being promoted from the part of the related Ministry - MRRFWM. Related to environmental pollution, it occurs more frequently in karst areas, particularly in the coastal cities due to the ongoing urbanisation, accumulation of waste, traffic, noise and damage to human health. However, in this field also, significant changes have been made in the last two years, particularly in the field of waste and waste waters. 3 Existing capacity for managing local development in the karst areas By the term “capacity” we refer here to the existing knowledge, experience and specific skills necessary for development planning and programming (such as preparation of project fiches, project cycle management (PCM), monitoring and evaluation skills, elaboration of environmental impact analysis (EIA), knowledge regarding Practical Guide for External Actions (PRAG), elaboration of cost benefit analysis (CBA), feasibility studies, tender dossiers and other) necessary for the efficient preparation of project applications for projects funded from international sources, in line with EU standards. Namely, the efficient and effective management of local development asks for very specific expertise and capacity of local actors and county administration. This expertise is currently mostly present in the established CDAs due to the fact that they were among the first institutions on the local level/regional level involved in the implementation of development projects, mostly funded by international sources. It was particularly these international financial sources (mainly EU pre-accession technical assistance support), which, as in other pre-accession countries, aimed priory at raising the capacity of government administration as well as relevant actors on all government levels for efficient preparation, management and implementation of development projects – enabling thus good governance.

130

This was an extremely valuable process since the provided training and transfer of expertise from EU member states provided the basis for implementing projects according to best existing practice, i.e. on the basic principles of EU regional policy in all EU member states. 3.1 Human resource potential Not all CDAs in the karst areas currently meet the minimum requirements for development planning and programming and implementation of county development policy. Most of CDAs have approximately 6 permanent employees, while those covering karst areas having an average in the range of 8-10. Among those already quite experienced, involved in project management are IDA in Istria, PORIN in the Primorje Gorski Kotar count, DUNEA in the Dubrovnik Neretva county and the Regional development agency of Šibenik Knin (ŠBK). As to the necessary elementary skills in the process of planning and programming – the skills for preparing project applications, looking from an overall perspective of established CDAs; it is visible that both in CDAs and counties the current capacity for the preparation of project applications is substantial. Only 3.3% see their capacity as weak, as visible from the following graph (Malekovi and Puljiz 2007): Current capacity for project application preparation 60% 50% 40% 30% 20% 10% 0% Total

CDAs Very good

Average

Counties Weak

As to the existing specific skills for the preparation and implementation of projects financed from international sources the carried out survey confirms that the CDAs are most skilled in the preparation and implementation of internationally funded projects, project cycle management and delivery of information and organization of training events. On the other hand, by far the weakest skills area relate to the preparation of tender dossiers, as visible from the following graph:

131

Skills for preparation and implementation of intern. funded projects in CDAs Project Fiche PCM Delivery of information and training events Grant Schemes Monitoring and Evaluation EIA CBA and Feasibility Studies PRAG Tender Dossiers

0% Good

20% Average

40%

60%

80%

100%

Weak

Tasks and responsibilities of the established CDAs in the karst areas differ considerably, as well as their competencies, expertise and overall capacity for managing local development. Not all of them, particularly not those recently established, have an acknowledged role in promoting the overall development in the respective county. Their current role is usually related to preparing and monitoring the development strategy of the county, supporting local municipalities’ and entrepreneurs’ socio-economic development programmes and projects, including their fund mobilization as well as support in their implementation (Armstrong 2006, Malekovi and Puljiz 2007). 3.2 Experience on internationally funded projects The first technical support provided through internationally funded projects was the main means for acquiring the first skills and experience necessary for preparing, managing and implementing local development projects. This gained knowledge and experience varies significantly between the agencies. The most experienced ones are surely IDA, PORIN with practical experience on 18 projects) who often played a role as project applicant and main manager, and lately DUNEA, while the least experienced are RERA, ZADRA and LIRA Participation in cross border cooperation (CBC) projects so far was mainly initiated by Italian institutions, often CDAs not being able to provide inputs in preparation phase, again - due to inexperienced staff. 3.3 Human resource potential for local development in local and regional self government units in karst areas A similar situation related to low level of knowledge and skills is visible with public servants in local and regional self government units, particularly related to strategic planning, project preparation, project finances, procurement and project management. They have good overall analytical skills, but need to improve specific knowledge related to e.g. project prioritization and long term investment planning. Servants working in local regional self government are not used to project management tools, they have knowledge gaps in using criteria for project appraisal, project record keeping and progress reporting and they lack experience with stakeholder analysis. According to their self assessment, respondents have the poorest knowledge in writing a Quality Manual for a project, implementing cost-effective 132

monitoring and evaluation systems, preparation of international contracts, and solving claims during the implementation of the project (Tišma et al. 2008). Financial management skills need to develop from an overall understanding of the budget process to be able to set realistic performance target, prepare financial proposals for funding form external sources, manage information system to make decisions and to deal with budget constraints. Insufficient experience in conducting cost effectiveness and cost benefit analysis are potential obstacles in work performance as well. Being aware of its importance, the respondents have identified the project preparation activities as their least proficient area. They have expressed the desire to have training in many of the topics that deal with project preparation such as preparation, controlling and managing financial feasibility study for national institutions, international agencies or EU grants as well as writing a project fiche or application form for structural funds (Tišma et al. 2008). Table 1 Knowledge and skills on the level of local and regional self-government Topic Level of skills Project management (PCM, implementation, monitoring)

High

Strategic management

Medium

Communication and reporting

Medium

Legal procedures

Medium

Legislation learning

Medium

Legislation compliance, enforcement and EU approximation

Medium

Administrative tasks

Medium

Human resources management, training and education

Low

Financial management

Low

Procedures, methodologies, strategic document elaboration

Low

Document checking, approving and issuing

Low

Data collection and reporting

Low

Procurement

Low

Communication with clients

Low

Inspection

Low

Institutional evaluation and development

Low

The public servants evidently see communication information technology skills, expert and legislation knowledge and tendency to team work as the most relevant ones. Highly important for gaining success are also foreign language skills and organisational and analytical skills. Among those ranked as relevant were also Project Cycle Management and Stress management (Tišma et al. 2008).

3.3 Cooperation with other development institutions/stakeholders In all counties in the NUTS II Adriatic region, covering the karst areas, CDAs have developed good cooperation with all of the relevant development institutions and stakeholders – employment services, chambers of commerce and crafts, entrepreneurship centres and 133

incubators, technology centres, etc.- This was particularly the case with IDA, PORIN, ŠBK, LIRA and DUNEA. The cooperation between the CDAs in the karst areas is particularly developed through regular cooperation within RAJ (Development Agencies of Adriatic), which is still not institutionalised in any formal way. From point of view of joint activities of CDAs in the karts areas, relevant for addressing relevant common development problems and priorities, joint lobbying aiming at winning tenders and implementing strategic development projects, as well as their more effective joint implementation, they are however still poorly developed. 4 Main problems and needs related to existing and necessary capacity for managing local development The survey identified a number of problems that CDAs as well as the local and regional self government units responsible for large environmental investments related to improving the capacity of local and regional actors to ensure a more efficient and effective process of strategic planning. Among the key issues identified were the following: x More then 70% of respondents stated that their major problem in this area was the inadequate staff number and their low capacity for preparing and implementing project proposals, which related to the already mentioned small number of so far implemented projects and in only several counties. x A lack of motivation is perceived as a further obstacle by 8% of respondents, including lack of ideas related to the realization of project proposals on the basis of an analysis of development needs. CDAs in this regard stress that all stakeholders have requests and needs, but this does not result in well prepared project proposals. x Lack of knowledge related to monitoring and implementation of projects; x Lack of knowledge of foreign languages; x Inadequate internal info flow (within county institutions); x Lack of interest and awareness as to the relevance and use of EU programs from the part of counties; x Lack of information related to concrete programs for which applicants are applying; x Inadequate knowledge regarding project management, related to implementation procedures x Lack of experience with cooperation with a large number of partners. 4.1 Basic needs related to existing capacity in the karst areas On the basis of the mentioned research results in the framework of the MRDFWM the basic needs related to more efficient project preparation and implementation, leading thus to more successful management of local development, were in the following segments: - Awareness raising (SF awareness; IPA and SF OPs elaboration and management; awareness raising of municipalities and other local actors on funding opportunities); - Strategic planning: Methodology for drafting and preparing the County Development Strategy; (Strategic planning in specific areas; design and development of GIS in order to support the local development and management planning process; - Partnership: Management of partnership, role of CDA in Local Action Groups, other; - Project pipeline: How to create and manage a project pipeline, SF requirements for bigger regional projects), Preparation and management of (internationally funded) contracts in the context of big regional projects, Technical skills for infrastructural projects, How to set-up feasibilities studies, cost-benefit analyses, Environmental Impact Analyses, Funding opportunities for projects; 134

-

Business planning: Business planning procedures, Financing models for CDAs, Organizational development; Instruments: Methods to support SMEs, Guarantee fund management, Business zones management (coordination), Investments promotion, PPP facilitation, cluster management; Skills development: Lobbying and direct contacts with EC services, Analytical skills, English proficiency in relation to EU-projects terminology, Selection of indicators (especially in HR, environment, energy).

5 Main recommendations and conclusions It is a fact that EU funds provide vast possibilities as financial sources for local development, on the basis of which the karst areas could benefit. However, in order to mobilize these funds for development projects it is necessary to develop very specific skills and capacity, on the central, as well as regional and local level. Along with the relevance of the necessary administrative capacity on all levels, is the vertical coordination of activities of the local level in the karst areas with the county governments and respective Ministry for regional development (MRDFWM). This will be achieved through the implementation of the county development strategies, for which CDAs will be in charge. However, the process of strategic planning needs to be further developed, and particularly on the local level, and in the less developed karst areas, located in the southern part of the Adriatic (NUTS II) region. Related to cooperation and coordination with county/regional and central government institutions, it is necessary to: enable better info-flow and coordination of local development programmes among involved local/county bodies and institutions: organization of regular meetings with representatives of central government institutions related to relevant development programs and strategic documents elaborated on the central government level, with possible impact on the karst areas; to organize joint working groups on particular development problems of interest to more local units; to establish an internet forum of karst areas - enabling more efficient realization of tasks, information flow, sharing of experience and knowledge flow, as well as a means for fostering joint local actions and to establish data bases for local development projects in the karst areas,. Further, the joint cooperation of CDAs and other key local actors and stakeholders in the karst areas needs very concrete support – from joint discussions and consensus building related to development priorities and tasks to be undertaken; joint lobbying and fund mobilisation; joint cooperation on projects defined as development priority in the overall area, to sharing of experience based on implemented projects in each of the CDAs. Also, as mentioned previously, pubic awareness rising related to importance and benefits of local and regional development, as well as related to planning and managing of sustainable local development will have to be one of the key activities to be undertaken in the forthcoming period. The culture of partnership and interregional cooperation is thus fostered, and this is relevant from point of view of initiating and preparing large projects, based on joint cooperation of a number of localities and neighbouring regions. This shift from small infrastructure projects to larger projects with major development impact is very relevant for more efficient local development and optimal use of available resources for development needs in karts areas. The drafted Strategy for Regional Development particularly focuses the importance of the culture of partnership, stressing the needs for achieving consensus of development strategies, projects and activities on the county and NUTS II level. The results of this approach for local development are very evident. We would also like to emphasize that efficient local management of development in karst areas will ask for continuous and effective evaluation of the previously mentioned relevant development projects. Only evaluation (ex- ante and ex-post particularly) of 135

development projects results and impacts, will provide the basic learning tool for future projects as well as local actors and stakeholders which will be implementing them, enabling their continues improvement, and raising of efficiency and effectiveness for socio-economic development of karst areas. Finally, even though capacity for management of local development has been substantially raised in all agencies in the karst areas, as well as all respective county governments, it is a fact that regional competitiveness and substantive involvement in support to technologically propulsive innovative firms is very rare in these areas, particularly as we move towards the southern karst areas. This evidently remains as an issue asking for particular support and initiatives – not only from the part of mentioned institutions and government levels, but from all relevant key stakeholders and actors on the county level in the karst regions. References Armstrong S (2006) the use of the Nuts 2 level in the future management of National and EU Regional Policy in Croatia, Ministry of Sea, Transport, Tourism and Development, Zagreb Puljiz J, Malekovi S (2007) Current Situation and Future Perspectives of Regional Policy in Croatia, 6th International Conference: Economic Integrations, Competition and Cooperation, Opatija, pp 23 NSRD (2008) National Strategy for Regional Development, draft version, Ministry for Regional Development, Forestry and Water Management Tišma S, Gracin P, Boromisa A, Pavlus N (2008) Environmental Management in Urban Areas: Croatian Challenges towards European Union Association. Croatian International Relations Review 14(50/51):25-33

136

Hydrogeological functioning of the karst aquifer drained by Yedra Spring (Southern Spain) from hydrochemical components and organic natural tracers Matías MUDARRA, Bartolomé ANDREO Centre of Hydrogeology of University of Malaga (CEHIUMA), Málaga, Spain, e-mails: [email protected]; [email protected]

Abstract: Hydrogeochemical natural tracers are commonly used to characterize infiltration processes, the travel time of recharge waters and flow conditions in karst aquifers. These tracers are the principal chemical components and environmental isotopes, as well as organic components from soil origin. Among the latter, the Total Organic Carbon (TOC) is the parameter most commonly used as a natural tracer. In addition to this, the natural spectrofluorescence of water due to chemical compounds like humic and fulvic acids and protein organic matter can be measured. From April 2008 to January 2009, the hydrochemical and hydrodynamic responses of the Yedra spring (Malaga province, Southern Spain) were monitored to characterize its hydrogeological functioning. In this study, the major chemical parameters, together with the TOC and the natural fluorescence were analysed. The hydrochemical data were analysed by means of principal component analyses (PCA), temporal evolution, frequency curves and dispersion plots. The Yedra spring water shows important decreases in electrical conductivity and in major chemical components due to dilutions provoked by recharge periods. The frequency curve of the electrical conductivity of the water is of a multimodal type, with 100 μS/cm variation range. A very obvious inverse relation exists between tracers from the soil and infiltration +2 through the unsaturated zone (TOC and NO3 ) against the Mg content, which is a natural tracer of the residence time of the groundwater within the aquifer, mainly in its saturated zone. Furthermore, there is a strong, direct relation between TOC and the natural fluorescence associated with humic and fulvic acids. Both parameters present a very similar response to rainfall events, with significant increases in the recharge situation and gradual falls in depletion periods. This relation means that TOC comes mainly from the aforementioned organic acids. The results demonstrate the existence of rapid infiltration processes, usually with a lag of less than one day to rainfall periods, which is typical of a karst aquifer with conduit flow behaviour, rapid drainage and low capacity of natural regulation. The combined use of conventional hydrochemical parameters and organic natural tracers enables us to characterize the degree of functional karstification in the aquifer drained by the Yedra spring, this being especially developed in the unsaturated zone. This information can be used, moreover, to validate the vulnerability to contamination of the aquifer. Keywords: Yedra Spring, karst aquifer, hydrochemical response, TOC, spectrofluorescence

1 Introduction Studies of karst systems have traditionally focused on analysing the natural responses at their springs (Goldscheider and Drew 2007), whether physical (Andrieux 1978, Genthon et al. 2005), chemical (Bakalowicz 1979, Mudry 1987, Hess and White 1993) or hydrodynamic (Mangin 1975, Bonacci 1993). The individual or combined use of these responses makes it possible to determine the hydrogeologic characteristics and, in general, the functioning of karst aquifers. In this respect, the joint use of natural hydrogeochemical tracers (TOC, NO3-, natural fluorescence), in conjunction with some of the major hydrochemical components of the water (such as the Mg2+ content), is useful for characterizing infiltration processes, the transit time of water from the surface to the spring, flow conditions and, above all, the degree of participation of the saturated and unsaturated zones in the functioning of karst systems. Some of these tracers originate in the organic activity taking place in the soil covering the epikarst. Of these, Total Organic Carbon (TOC) is the one most commonly used as a natural tracer, because when there is no human-originated contamination or flows from the surface, it is an important indicator of rapid infiltration (Batiot et al. 2003a, 2003b Emblanch et al. 1998, 2003). In addition, it has been shown that the natural fluorescence of the organic matter dissolved in the water is related to 137

certain chemical compounds present in the soil (humic and fulvic acids, and protein organic matter) which, in turn, form part of TOC, and consequently could be interesting indicators of rapid infiltration (Baker et al. 1997, 1999, Blondel 2008, Cruz Jr et al. 2005). Another aspect is that the Mg2+ content, which is a major chemical component of the water, informs on the residence time of the water in the aquifer, mainly in the saturated zone (Batiot et al. 2003a Emblanch et al. 1998, 2003). From April 2008 to January 2009, the hydrochemical and hydrodynamic responses of the Yedra spring (Malaga province, Southern Spain) (Fig. 1) were monitored in order to determine the relative importance of the saturated and the unsaturated zones in the hydrogeologic functioning, the water resources and the contamination vulnerability of the aquifer that drains this spring. The aim of the present study is to contribute to characterizing infiltration processes and to determine the hydrogeologic functioning of the carbonate aquifer examined, by means of natural hydrochemical tracers and the major chemical components in the spring water. 2 General characteristics of the study area The Yedra spring is the main drainage point for the carbonate aquifer of Sierra de Las Cabras. This aquifer has a surface area of approximately 10.5 km2 and it is located in the province of Malaga, in southern Spain. It forms part of a larger water body known as Sierra de las Cabras – Camarolos – San Jorge (70 km2). The landscape is very rugged, with altitudes ranging from 800 to 1300 m a.s.l. The prevailing climate in the area is of temperate Mediterranean type, with a mean annual temperature of 16ºC. Precipitations, with a mean annual value of 700 mm, are characterized by severe annual and interannual variations.

Figure 1 Geographic location and geological-hydrogeological sketch of the Sierra de las Cabras experimental area, showing the situation of the Yedra spring From a geological standpoint, Sierra de Las Cabras is located in the Betic Cordillera, and it is made up of oolitic limestones and Jurassic dolostones, which are bounded, at the base, by Upper Triassic clays mixed with gypsums; to the north and south, it is bordered by Flysch-type clays (Fig. 1). The geologic structure is constituted of an anticlinal fold oriented approximately E-W. The karst features in Sierra de Las Cabras aquifer is relatively well developed at the surface of the carbonate outcrops, mainly in the Jurassic oolitic limestones that are found at the highest altitudes, where karrenfield may be observed. Recharge takes place by the direct infiltration of rainfall, and the aquifer discharges through springs at the northern border, the most important of these being the Yedra spring, which was selected for the present study. 138

Some previous researches have been made of this area, and the hydrogeologic characteristics of the aquifer have been described (Pulido Bosch and Cerón 1991, Marín et al. 2007, Mudarra et al. 2007, 2008). 3 Methodology From April 2008 to January 2009, sampling was performed twice weekly, daily during some periods of high water conditions, and twice monthly during periods of depletion. Concurrently with the sampling programme, the temperature and electrical conductivity of the water were measured, using portable equipment (to a precision of ±0.1º C and ±1 μS/cm respectively). The hydrochemical parameters were analyzed at the Laboratory of Hydrogeology of the University of Málaga. Total Alkalinity Content (TAC) was determined by volumetry with H2SO4 0.02 N to pH 4.45. The chemical analyses of the major components (Ca+2, Mg+2, Na+, Cl-, SO4-2, NO3-,) were performed by applying ionic chromatography. TOC was measured using a Carbon analyzer (made by Shimadzu, model V-TOC). Natural fluorescence was measured with a spectrofluorimeter (Perkin-Elmer, LS-55). WinLab software was used to obtain the excitation-emission matrices (EEM) corresponding to each of the samples. In every case, the excitation (ex) began with a wavelength of 200 nm, and ended with one of 350 nm with a 5 nm slit, while the emission (em) ranged from wavelengths of 250 to 550 nm. The fluorescence values recorded for EEM refer to Fluorescence Intensity Units (FIU) measured in Uf.nm2 (Baker et al. 1997, Blondel 2008). We only took into account the highest values for fluorescence intensity visible in the EEM, of which only two maximum values appear in every sample, corresponding to the organic chemical substances considered in this study (upper and lower peaks). The upper fluorescence peak appears between 300-340 nm excitation and 400-440 emission, while the lower one was identified between 225-245 nm excitation and 400-450 nm emission. The hydrochemical data were analyzed using different procedures. The frequency curves for the electrical conductivity values of the spring water inform on the variability of the mineralization and on the chemical composition of the water (Bakalowicz 1979). If the curve presents a unimodal shape, with a small range of variation, this indicates a low degree of development of functional karstification. If, on the other hand, there is no clearly defined mode (i.e. we are considering a multimodal situation), then there exists a high degree of functional karstification in the aquifer drained by the spring (Bakalowicz 1979, Mudry 1987). The temporal evolutions of the chemical components of the water informs on the lag between rainfall (the input signal) and the hydrochemical response (the output signal). The magnitude of the dilutions recorded permits to estimate the degree of functional karstification of the aquifer. Of all the components analyzed, particular account was taken of the contents of NO3-, Mg+2 and TOC, together with the natural florescence of the water. The two-dimensional diagrams of these chemical components give information on the degree and type of relation of dependence that may exist between them, depending on the characteristics of the groundwater flow. Principal Component Analysis (PCA) is a multivariate statistical technique that is used for grouping variables and water types that may be associated with the same hydrodynamic conditions or with a common origin. PCA has been applied in numerous studies for the quantitative interpretation of hydrochemical data, for grouping or classifying the same data, and for identifying hydrogeologic processes. 3 Results Table 1 presents certain statistical parameters concerning the chemical composition of the water drained by the Yedra spring. This water has a low level of mineralization (mean values of electrical conductivity of 280-369 μS/cm) and is of a calcium bicarbonate type.

139

Table 1 Statistical parameters of the water samples analyzed from the Yedra spring examined in this study during the period from April 2008 to January 2009. (n) number of samples, () standard deviation, (v) coefficient of variation. Yedra spring E.C. (μS/cm) n 41 mean 321  27.4 V (%) 8.54% Min 280 Max 369 Range 89 n mean  V (%) Min Max Range

Mg (mg/L) 41 10.48 3.83 36.62% 6.22 16.87 10.64

T (ºC) 41 14.0 1.0 7.18% 12.8 16.7 3.9

Q (L/s) 41 20.17 20.53 101.78% 1.00 100 99

T.A.C. (mg/L) 41 186.79 18.6 9.96% 152.5 219.6 67.1

SO4 (mg/L) 41 10.62 4.09 38.56% 5.68 18.22 12.54

Cl (mg/L) 41 6.43 0.34 5.28% 5.81 7.30 1.49

Na (mg/L) 41 3.27 0.30 9.06% 2.81 3.78 0.98

Ca (mg/L) 41 59.59 2.6 4.37% 53.59 64.47 10.91

Up.P.fluor. (Uf.nm2 ) T.O.C. (mg/L) 41 41 121.00 0.39 41.66 0.16 34.43% 40.22% 67.67 0.20 226.05 0.80 158.38 0.61

NO3 (mg/L) Lo.P.fluor. (Uf.nm2 ) 41 41 11.59 63.26 2.38 22.78 20.52% 36.00% 7.53 33.51 17.99 122.54 10.46 89.03

The frequency curve for the electrical conductivity of the water (Fig. 2) does not have a clearlydefined mode; it is, in fact, clearly multimodal with the presence of 4 distinct peaks (P1, P2, P3, P4) which represent up to 14% of the total accumulated frequency (P2). The range of variation is from less than 300 μS/cm, during periods of dilution (high water conditions), to 370 μS/cm during periods of low water, using electrical conductivity classes of 5 μS/cm.

Figure 2 Frequency curve of the electrical conductivity of the water obtained from periodic measurements in Yedra spring. Figure 3 shows the temporal evolution of the parameters measured for the spring water, versus the precipitations that occurred between April 2008 and January 2009. The hydrograph (Fig. 3) reveals sharp rises in water flow rates in response to recharge events (April, November and December 2008 and January 2009). The hydrochemical parameters show a general seasonal variation in electrical conductivity, temperature and majority chemical components (TAC, SO4-2, Mg+2, Ca+2, Na+, Cl-), in which there can be appreciated two different behaviour patterns: one for periods of low water (MayOctober 2008), when the values for all these parameters are higher, and another for periods of high water (April, November and December 2008 and January 2009), when the values of the same components are lower.

140

Figure 3 Temporal evolution of discharge rate, water temperature, major chemical components and natural fluorescence peaks At each recharge event, the spring water underwent sharp decreases in electrical conductivity, accompanied by decrease in almost all the chemical components. In general, these major compounds evolved in a way similar to that of the electrical conductivity, except as regards the content of Ca2+ and Cl-, which during the important flood that occurred in November 2008 increased rather than decreasing, especially the Cl- ion. The temperature of the spring water presented a clearly cyclical pattern, with higher values in the spring and summer, and lower ones in the autumn and winter. However, some natural hydrogeochemical tracers reacted in the opposite way to that described above. Neither the contents of TOC and of NO3- nor those of the two peaks of natural fluorescence followed the general pattern for electrical conductivity. Their values, in contrast, increased during each recharge event, with maxima at the beginning of the hydrogeologic year (in November 2008), when the first important recharge event occurred. All the water flow peaks coincided with increases in the contents of NO3-, TOC and natural fluorescence, except for the increases of December 2008 and January 2009, during which the NO3- decreased. 141

Figure 4 shows the Mg+2 content versus the concentrations of NO3-, TOC and the peaks of natural fluorescence of the spring water, and the relation between the peaks of natural fluorescence and the TOC (D). The TOC - Mg+2 diagram (Fig. 4A) shows, in general, an inverse exponential relation between the two components for the whole set of sample analyzed. The spring water presents a wide range of variation in Mg+2 and TOC content. In the NO3- - Mg+2 diagram and that for natural fluorescence - Mg+2 (Fig. 4C), it may be also observed the same general inverse relations between the components, although this is less evident in the case of NO3- - Mg+2. In other words, the contents of NO3-, TOC and natural fluorescence produce similar results when represented versus the Mg+2 content. This fact is more evident for the case of the natural fluorescence peaks of the water and TOC, as can be seen in Fig. 4D, which shows a positive linear relation between these two substances.

Figure 4 Diagrams of the Mg2+ content versus TOC (A), NO3- (B), and natural fluorescence peaks (C). The last one shows relation between lower and upper fluorescence peak. Plot D shows relation between TOC content and fluorescence peaks. A Principal Component Analysis (PCA) was carried out to determine the relation between the different variables and the samples measured (Figs. 5A and 5B). The two main axes of the PCA account for 85% of the sampling variation. Axis 1 (71.31%) is determined by electrical conductivity, temperature, TAC, Mg+2, SO4-2, Na+, TOC, the two natural fluorescence peaks and the discharge (Fig. 5A), while Axis II (13.65%) mainly addresses the Cl- ion. The Ca+2 variable was not assigned to either of these two principal factors, but rather was associated with Axis III, which is not described in the present paper. In factorial plane I-II of the statistical units (Fig. 5B), three water groups can be distinguished: Group I is constituted entirely of water samples located in the positive part of Axis I, and presents a wide range of variation, both negative and positive, in relation to Axis II. The cases making up Group 3 are located in negative positions on Axis I and close to zero on Axis II, and at the same time are closely grouped. The samples included in Group 2 present a dispersion and location that are intermediate between those of the other two groups, being centred on the null values of the two axes.

142

4 Discussion The Yedra spring, in response to precipitation, undergoes important increases in water flow volumes for short periods of time (of less than one day). This suggests there are karst conduits in the aquifer drained by the spring. Furthermore, the values of the coefficient of variation of all the hydrochemical parameters (Table 1) and the frequency histogram for the electrical conductivity values measured in the spring water (Fig. 2) show that the water has a high level of hydrochemical variability. According to Bakalowicz (1977), this type of frequency histogram is accounted for by the existence of karst conduits within the aquifer drained by the spring.

Figure 5 Principal Components Analysis on hydrochemical data, including natural fluorescence peaks, plus discharge. Plot of the variables (A) and plot of the statistical units (B). The temporal evolution of the water temperature, electrical conductivity and of almost all the majority hydrochemical components analyzed for the spring water (Fig. 3) reveals sharp, rapid decreases in their values following each important pluviometric event, followed by an increase in the same, if no further recharge takes place. This fact has been observed in previous studies (Pulido Bosch & Cerón, 1991; Mudarra et al., 2008). On the other hand, the increase in the content of the Cl- ion during the important pluviometric event of early November 2008 indicates that this anion proceeds, in part, from the concentration of rainfall in the epikarst and its subsequent mobilization with the first rainfall of the autumn. The contents of NO3-, TOC and of natural fluorescence evolved in a way inverse to that of the major components, with increases in high water situations and decreases in periods of low water. All these substances proceed from the soil and are natural tracers of infiltration (Batiot et al., 2003a; 2003b; Baker et al., 1997; 1999). The highest concentrations detected at the beginning of the hydrogeologic year (in November 2008) were caused by the leaching of the soil that had been subjected to biological activity, in the absence of rainfall, since late spring. For this reason, the infiltration water coming from the first heavy rainfall in the autumn usually has a higher content of these components, which decreases progressively as the summer advances. Nevertheless, and except for the decrease in NO3- content during the high water periods of December 2008 and January 2009, in all the recharge events there were relative increases in these parameters, as a consequence of the arrival at the spring of water that had infiltrated through the soil. The time lag observed between the peaks of the NO3- ion and these high water periods was due to the different biogeochemical kinetics produced in the soil by its nitrogen content, with respect to other organic compounds. The information provided jointly by the majority chemical components and by the natural organic tracers of the water reveals that, for each recharge event, there occurs a mixing of water from the saturated zone with water from the unsaturated zone. The water that has recently infiltrated trough the latter zone flows rapidly from the surface of the epikarst to the spring, via karst conduits. 143

The diagrams in Fig. 4 describe the characteristics of the flows through the aquifer. Thus, and in accordance with Batiot et al. (2003a), the graphs of NO3- and TOC content, which are indicators of the rapid infiltration through the soil and the unsaturated zone, in relation to the Mg2+ content (Figs. 4A and 4B), a natural tracer of the residence time of the water in the aquifer (mainly in its saturated zone), reflect the high variability of these natural tracers of the water in the Yedra spring. The graph in Fig. 4 shows that the two natural fluorescence peaks of the organic matter dissolved in the water also present a high degree of variability with respect to its Mg2+ content. The direct relation observed between the natural fluorescence peaks and the TOC content (Fig, 4D) suggest that this latter parameter is associated with the organic acids that are detected under spectrofluorometry. Accordingly, organic substances can be used, in the same way, as natural tracers of rapid infiltration. The variations in the content of the natural tracers of soil and of the Mg2+ ion are due to the fact that both the unsaturated and the saturated zones participate in the functioning of the system, the first of these doing so mainly during periods of aquifer recharge. Principal Components Analysis (PCA) provides quantitative assistance in determining the main factors responsible for the total variance. In factorial plane I-II of the variables (Fig. 5A), there are seen to be, in general, two groups of variables located at the positive and negative extremes of Axis I, and with very little representation on Axis II. The group of the negative part (E.C., T, TAC, Mg+2, SO4-2, Na+) comprises the variables considered to be tracers of the mineralization of the karst, or indicators of greater residence time of the water within the aquifer, while the group situated in the positive part is made up of the natural tracers of the infiltration (TOC, natural fluorescence peaks and, to a lesser degree, NO3-), plus the discharge. In contrast, Axis II only includes the variable Cl-, which is a marker of the epikarst, in its positive part. In factorial plane I-II of the statistical units, three sets of water samples (Fig. 5B) were distinguished. Group 1 was made up of the samples with a low level of mineralization, and which were obtained during periods of high water. In general, these presented lower concentrations of major chemical components and higher discharge rates, as well as TOC, natural fluorescence and NO3contents. The water samples in Group 3 were highly mineralized and corresponded to low water periods. They had higher contents of almost all the principal chemical compounds, and lower ones of the infiltration tracers. The cases included in Group 2 were in an intermediate situation between Groups 1 and 3, and represented the water samples undergoing a process of mineralization, following the end of a recharge event. The cases situated on the positive part of both axes represent the water samples taken following the important pluviometric event of November 2008. These samples presented considerable increases in contents of organic tracers and Cl- (Fig. 3). 5 Conclusion The joint use of conventional hydrogeochemical parameters and of the natural organic tracers present in the water drained by the Yedra spring (in southern Spain) permit to characterize the degree of functional karstification present in the hydrogeologic system of Sierra de las Cabras. The analysis and processing of these data revealed that the aquifer has a highly developed degree of functional karstification, which is typical of conduit flow systems, with a rapid drainage and a low capacity of natural regulation. This karstification is developed both within the unsaturated zone and in the saturated zone, as evidenced by the hydrodynamic and hydrochemical response to precipitation events. Karst drainage favours the rapid infiltration and flow of rainwater towards the spring, which leads to increases in outflow rates and decreasing in almost all the hydrochemical parameters, except those components leached from the soil by the infiltration water. These characteristics mean that the volume of water stored in this aquifer is not very considerable and that the karst spring associated with it is highly vulnerable to contamination, an aspect that should be taken into account for its adequate protection and management. Finally, it was observed that the Total Organic Carbon (TOC) measured in the water from the Yedra spring is associated with organic acids detected by spectrofluorescence. The relation observed 144

between these substances and the TOC is linear and direct, which shows that the characterization of the organic matter dissolved in the water of the spring, using excitation-emission fluorescence matrices, can be used as another indicator of processes of leaching and rapid infiltration of water from the soil/epikarst towards the spring. Natural fluorescence may be another factor that could contribute to evaluate and validate the contamination vulnerability of karst aquifers. Acknowledgments This work is a contribution to projects CGL2005-05427 and CGL2008-06158 BTE of DGICYT, P06-RNM 2161 of Junta de Andalucía and IGCP 513 of UNESCO and to Research Group RNM-308 of Junta de Andalucía. References Andrieux C (1978) Les enseignements apportés para la thermique dans le karst. Colloque de Tarbes, Le karst: son originalité physique, son importance économique. Association des Géologues du SudOuest (AGSO), France, pp 48-63 Bakalowicz M (1979) Contribution de la géochimie des eaux à la connaissance de l'aquifère karstique et de la karstification. Thèse Doct. Sci. Nat., Univ. P. et M. Curie, París-VI, Géol. Dyn., 269 pp Baker A, Barnes WL, Smart PL (1997) Variations in the discharge and organic matter content of stalagmite drip waters in Lower Cave, Bristol. Hydrological Processes 11:541-555 Baker A, Genty D, (1999) Fluorescence wavelength and intensity variations of cave waters. Journal of Hydrology 217:19–34 Batiot C, Emblanch C, Blavoux B, (2003a) Carbone organique total (COT) et Magnésium (Mg2+): Deux traceur complémentaires du temps de séjours dans l´aquifére karstique. C R Geoscience 335:205-214. Batiot C, Liñán C, Andreo B, Emblanch C, Carrasco F, Blavoux B (2003b) Use of TOC as tracer of diffuse infiltration in a dolomitic karst system: the Nerja Cave (Andalusia, southern Spain). Geophysical Research Letters 30(22):2179 Blondel T (2008) Traçage spatial et temporel des eaux souterraines dans les hydrosystèmes karstiques par les matières organiques dissoutes. Thèse doctorel. Académie d’Aix-Marseille. Université d’Avignon et des Pays de Vaucluse. Bonacci O (1993) Karst spring hydrographs as indicators of karst aquifers. Journal of Hydrological Sciences 38:51-62 Cruz Jr FW, Karmann I, Magdaleno GB, Coichev N, Viana Jr O (2005) Influence of hydrological and climatic parameters on spatial-temporal variability of fluorescence intensity and DOC of karst percolation waters in the Santana Cave System, Southeastern Brazil. Journal of Hydrology 302:1–12 Emblanch, C, Blavoux B, Puig JM, Mudry J (1998) Dissolved organic carbon of infiltration within the autogenic karst hydrosystem. Geophysical Research Letters 25:1459-1462 Emblanch C, Zuppi GM, Mudry J, Blavoux B, Batiot C (2003) Carbon 13 of TDIC to quantify the role of the unsaturated zone: the example of the Vaucluse karst systems (Southeastern France). Journal of Hydrology 279(1-4):262-274 Genthon P, Bataille A, Fromant A, D'Hulst D, Bourges F (2005) Temperature as a marker for karst waters hydrodynamics. Inferences from 1 year recording at La Peyrére cave (Ariège, France). Journal of Hydrology 311:157-171 Goldscheider N, Drew D, (2007) Methods in Karst Hydrogeology. Taylor & Francis, 262 pp Hess JW, White WB (1993) Groundwater geochemistry of the carbonate karst aquifer, southcentral Kentucky, USA. Applied Geochemistry 8:189-204 Mangin A (1975) Contribution à l’étude hydrodynamique des aquifères karstiques. Ann. Spéléol., 29(3):283-332, (4):495-601, 30(1):21-124 145

Marín AI, Andreo B, Mudarra M (2007) In: XXXV IAH Congress International Association of Hydrogeologists, Groundwater and Ecosystems. Lisbon, Portugal Mudarra M, Andreo B (2007) In: Atlas Hidrogeológico de la Provincia de Málaga 2:113-118. Diputación provincial de Málaga-IGME-UMA Mudarra M, Andreo B, Marín AI (2008) Consideraciones sobre la importancia de la zona saturada y no saturada en el funcionamiento hidrogeológico del acuífero carbonatado de la Alta Cadena (provincia de Málaga). Geo-Temas 10 (VII Congreso Geológico de España, Canarias 2008), pp 579-582 Mudry J (1987) Apport du tracage physico-chimique naturel à la connaissance hydrocinématique des aquifères carbonatés. Thèse Sciences Naturelles, Université de Franche-Comté, Besançon, 378 pp Pulido Bosch A, Cerón JC (1991) Variaciones hidrogeoquímicas de periodicidad anual en surgencias kársticas. Ejemplo del manantial de la Yedra (Málaga). Revista de la Sociedad Geológica de España 4(1-2):51-59

146

The hydrology of turloughs as groundwater dependent terrestrial ecosystems Owen NAUGHTON, Paul JOHNSTON, Laurence GILL Department of Civil, Structural and Environmental Engineering, Trinity College Dublin, Dublin 2, Ireland, e-mail: [email protected]

Abstract: Turloughs are one of the characteristic features of the Irish karst landscape. They are transient lakes resulting from a combination of high rainfall and accordingly high groundwater levels in topographic depressions in the karst. A turlough is effectively a hydrogeological feature defined as “A topographic depression in karst which is intermittently inundated on an annual basis, mainly from groundwater, and which has a substrate and/or ecological communities characteristic of wetlands” (Tynan et al. 2005). The hydrological regime in a turlough results in a characteristic ecology associated with the pattern of groundwater inundation. The behaviour of a turlough as a wetland is fundamentally driven by its hydrology. Discharges through turloughs are typically difficult to assess because of the often uncertain location and nature of the inflow and outflow points. The only realistic hydrological measures are based on water level. This study quantifies the hydrological regime of a set of turloughs and suggests a conceptual model to explain turlough operation. Semi-permanent water level monitoring stations were established in twenty-two selected turloughs using pressure transducers at or near the lowest point of each turlough. Time series datasets of water levels were then developed using hourly intervals over a 30 month period from November 2006 to May 2009. Stage–Volume relations were derived using digital terrain models generated from detailed topographic surveys. These were then combined with water level data to give volume time series. A range of turlough response and recession characteristics were observed with some having multiple flood events in the course of a year whereas others show a single event with a slow recession. Maximum flow capacities were derived through analysis of the volume recessions. The observed hydrological behaviour has been used to develop a conceptual hydrological model for the functioning of turloughs. Together with ecological and land-use data, this model will aid in the evaluation of turlough conservation status as groundwater dependent terrestrial ecosystems as defined in the Water Framework Directive (2000/60/EC). For the first time the role turloughs occupy within a karst groundwater system have been defined; risks posed to these protected ecosystems may now be evaluated and quantified. Whether through abstraction/drainage or through hydrochemical pressures on trophic status, these potential risks are assessed in terms of appropriate hydrological indicators relevant to the characteristic ecology of a turlough. Keywords: turloughs, hydrology

1 Introduction 1.1 Definition Turloughs are one of the characteristic features of the Irish karst landscape. They are transient lakes resulting from a combination of high rainfall and accordingly high groundwater levels in topographic depressions in the karst. A turlough is effectively a hydrogeological feature defined as “A topographic depression in karst which is intermittently inundated on an annual basis, mainly from groundwater, and which has a substrate and/or ecological communities characteristic of wetlands” (Tynan et al. 2005). 1.2 Ecological Importance By their nature, turloughs support many characteristic flora and fauna species (Reynolds 147

1996). Under the Water Framework Directive (2000/60/EC), turloughs are classified as Groundwater Dependent Terrestrial Ecosystems (GWDTEs) and as a Priority Habitat in Annex 1 of the EU Habitats Directive (92/43/EEC). Consequently, under national legislation, many have been designated Special Areas of Conservation (SAC), that is, areas of ecological importance which are afforded the highest level of protection as ‘natural’ sites. Both EU directives necessitate the monitoring and management of these habitats to ensure favourable conservation and groundwater status is achieved. In particular, the Water Framework Directive requires a good understanding of the hydrological linkage between the turlough wetland, its ecological functioning and the connected groundwater body. Development of a conceptual model for the hydro-ecology and hydrochemistry of turloughs is currently the subject of a major research project being carried out by Trinity College Dublin (Republic of Ireland) on behalf of the National Parks and Wildlife Service (NPWS) of the Department of the Environment, Heritage and Local Government in aid of a management strategy for these wetlands. 1.3 Hydrology Turloughs are at the interface between groundwater and surface water. They fill mainly by rising groundwater levels through estavelles and springs together with surface runoff; they ultimately empty through estavelles and swallow holes (Coxon 1986). Filling normally occurs in late autumn due to periods of intense or prolonged rainfall; with emptying typically occurring from April onwards. The karst flow system, of which a turlough is a surface expression, possesses a flow capacity which is defined by the size and connectivity of the flow paths present within the rock (Drew and Daly 1993). Rainfall of insufficient duration or intensity can be accommodated by subsurface flow paths; hence no surface flooding is visible in the turlough basin during these dry periods. However once the required combination of rainfall intensity and duration occurs the storage of the system is exceeded and flooding begins. A range of turlough response and recession characteristics exist with some having multiple flood events in the course of a year whereas others show a single event with a slow recession as shown in Figure 1. The level, duration and extent of flooding vary greatly among turloughs with maximum flood depths of 3-14 metres and flooded areas of over 60 hectares recorded during the monitoring period. 8 7

Water Depth (m)

6 5

Lough Aleenaun

Termon

4 3 2 1 0 Oct-06

Nov-06

Jan-07

Mar-07

Apr-07

Jun-07

Aug-07

Sep-07

Date

Figure 1 Contrasting turlough hydrological regimes of Lough Aleenaun and Termon 148

1.4 Geology Turloughs are found in areas of thin, relatively permeable subsoil on well-bedded, pure grey calcarenite. They occur predominantly on Dinantian pure bedded limestone due to its purity and well developed bedding (Coxon 1987). Turloughs were originally considered hollows in glacial drift with underlying karst drainage systems (Williams 1964). However, Drew (1976) asserted that turloughs invariably lie in bedrock hollows and were solutional features requiring a far longer period to develop than has passed since the last glaciation. Coxon and Coxon (1997) later suggested that turloughs are polygenetic with both processes playing a part in their formation. The presence of lacustrine marl in the basin of many turloughs also provides evidence that the flood regime has altered over time (Coxon 1994). 2 Methodology 2.1 Hydrology The behaviour of a turlough as a wetland is fundamentally driven by its hydrology, essentially groundwater but with some surface water interaction that includes direct rainfall. Discharges through turloughs are typically difficult to assess due to the often uncertain location and nature of the inflow and outflow points. The only realistic hydrological measures are based on water level. Semi-permanent water level monitoring stations were established in twenty-two selected turloughs (Figure 2) using proprietary pressure transducerbased instruments (“Diver”). Time series of water levels were collected using Divers installed at or near the lowest point of each turlough and recorded on an hourly basis. Sites were instrumented in early November 2006 and monitoring is still ongoing at selected sites.

Figure 2 Turlough monitoring site locations To supplement long term rainfall records obtained from synoptic stations operated by the Irish Meteorological Service (Met Eireann), three rain gauges were installed in Kilchreest and Francis Gap in Co. Galway and Ballintober in Co. Roscommon. Rainfall was measured using an ARG100 tipping bucket rain gauge which recorded cumulative rainfall at 15 minute intervals. 149

2.2 Topography In the absence of direct measurement of flows in or out of a turlough, the approach was taken to estimate net flows by determining the volume of the turlough and deriving flows from a combination of time changes in stage related to the relevant volume at that stage. Thus a depth-volume relationship for each turlough was essential in determining flow. A detailed topographic survey was carried out on each turlough using Trimble differential GPS equipment. An average of over 900 points was taken per turlough with a mean horizontal point spacing of 12 m. Using this extensive topographic dataset, digital terrain models (DTM) were generated (Fig. 3). Statistical methods from relevant literature (Yanalak 2003) were used to assess this effect on turlough topographic models by calculating and comparing the standard deviation of derived surfaces using various gridding methods for a range of grid spacing. Kriging and multi-quadratic radial basis function were found to produce the most accurate results, with radial basis consistently showing lower  values. However at high grid resolution radial basis generated unrealistic physical features and so kriging with a 2m grid resolution was selected as the preferred method for all DTM work.

Figure 3 Digital terrain model for Termon Turlough, Co. Galway (Maximum extent of flooding highlighted in bold)

To determine turlough volume and net flow rates the relationship between stage and volume in each turlough basin was established. This was achieved by calculating the volume between the lower surface of the turlough and an upper horizontal surface representing a specific water level at 20mm intervals over the recorded range. Applying this relation allowed time series of volumes and associated flow rates to be obtained from recorded water levels for each turlough. 150

3 Discussion 3.1 Regime Quantification The behaviour of a turlough as a wetland is fundamentally driven by its hydrology. Discharges through turloughs are typically difficult to assess because of the often uncertain location and nature of the inflow and outflow points. The only realistic hydrological measures are based on water level. x x

Flood depths of 3-14 metres were measured during the monitoring period with a mean flood depth of 5 metres determined. Maximum flood volumes of 400,000-4,000,000 m3 were found during the monitoring period with a median maximum volume of 750,000 m3 determined.

3.2 Drainage Capacity Upon applying stage – volume relationships to recorded hydrographs a common characteristic was observed across the set of turloughs. During the emptying phase the net outflow did not vary with decreasing head. Instead net outflow remained constant for a large part of the recession, the outflow only reducing at low water levels. This implied that there exists a maximum rate at which each turlough can drain, a rate which is independent of water level within the turlough. By applying linear regression to the recession curves the maximum drainage capacity was derived. The constant flow rate during recession is only clearly defined when there is little or no rainfall during the recession period. Recharge from rainfall events causes water levels in the catchment to rise thus slowing the rate of emptying of the turlough. It also enters the turlough directly via surface flow and direct rainfall. In an attempt to limit this effect, regression analysis was carried out on data from mid March to early April 2007, a period where little or no rainfall fell. Sites found to empty before this period or having multiple recessions had regression analysis carried out on all available recessions and the highest recorded rate taken as the maximum flow rate (Figure 4). 3.3 Hydro-ecological Indicators From a turlough’s water level record and DTM appropriate measures for hydroecological risk are being established. Frequency-duration curves for different stages in the turlough are uniquely related to characteristic ecological communities (Tynan et al. 2005), while the period of inundation or hydroperiod has been shown to influence mean abundance and taxon richness of macroinvertebrates (Porst and Irvine 2009). As such depth – duration – frequency curves have been generated over the monitoring period. Disturbance has an important effect on macroinvertebrate community structure in turloughs, with high disturbance generally supporting lower faunal diversity (Porst et al. 2009). The areal rate of change, defined as the average rate of change of area between the time of maximum areal inundation and the emptying of a turlough, has been used to represent turlough disturbance (Porst et al. 2009).The higher the areal reduction rate the more rapid the changes in water levels and, thus, disturbance. The rate at which soil nutrients within the turlough basin are released to the water column could potentially be dependent on the areal rate of change.

151

Straight Line Recession: Turloughmore 450000

Q = 0.42m3/s

400000

R2 = 0.9992

Q = 0.37m3/s Q= R2 = 0.9972

0.40m3/s

350000 R2 = 0.9985 Volume (m 3)

300000 250000 200000 150000 100000 50000 0 19-Nov

29-Nov

09-Dec

19-Dec

29-Dec

08-Jan

18-Jan

28-Jan

07-Feb

Date/Tim e

Figure 4 Recession analysis for Turloughmore Turlough, Co. Clare

3.4 Conceptual Model The recession behaviour observed from 2 years of hydrological data led to the development of a conceptual model for the operation of turloughs. As the turloughs are typically empty during summer months despite rainfall events, there exists a flow system capable of carrying this recharge without surface flooding. When the rainfall on the catchment exceeds the capacity of this system, it becomes surcharged and flooding occurs. This process is represented by the conduit flow system shown in Figure 5.

Figure 5 Schematic for conceptual turlough model The single conduit represents an actual conduit or a system of interconnected fractures or conduits. The turlough is represented by a pond with depth – volume characteristics 152

derived from the digital terrain model (DTM) fitted to the topographic survey data. Two catchments are defined in the model. The first is the greater catchment area which drains via the conduit system beneath the turlough. The second is a smaller local catchment which supplies water to the turlough via direct rainfall, surface runoff and shallow groundwater flow. Rainfall on the greater catchment enters the turlough via the conduit flow system; the capacity of which is controlled by the restriction. During recession periods flow through the conduit system does not enter the turlough. Instead it controls the rate of release of water from the turlough by varying the pressure in the conduit. The hydrological data collected to date also confirmed the basic operation of a turlough as a ‘surge tank’ in a hydraulic sense and that most turloughs operate through a limited number of entry points rather than behaving as ‘flow through’ devices where the surrounding boundaries are all permeable. Once a turlough fills, it has very little mixing association with the underlying groundwater. This model has implications for interpretation of potentially polluting pressures and on the management of associated risks. 4 Conclusion The behaviour of a turlough as a wetland is fundamentally driven by its hydrology. Discharges through turloughs are typically difficult to assess because of the often uncertain location and nature of the inflow and outflow points. The only realistic hydrological measures are based on water level. This study quantifies the hydrological regime of a set of turloughs and suggests a conceptual model to explain turlough operation, thus allowing hydro-ecological indicators to be defined. For the first time the role turloughs occupy within a karst groundwater system have been defined; risks posed to these protected ecosystems may now be evaluated and quantified. Whether through abstraction/drainage or through hydrochemical pressures on trophic status, these potential risks are assessed in terms of appropriate hydrological indicators relevant to the characteristic ecology of a turlough. Acknowledgements This research is funded by a grant from the National Parks and Wildlife Service (Department of the Environment, Heritage and Local Government). This research forms a part of the large interdisciplinary project on Assessing the Conservation Status of Turloughs, being carried out in Trinity College Dublin in 2006-2010, funded by the National Parks and Wildlife Service of the Department of the Environment. We are also grateful to Patrick Veale for his field assistance. References Coxon C (1994) Carbonate Deposition in Turloughs (Seasonal Lakes) on the Western Limestone Lowlands of Ireland. Irish Geography 27(1):14-27 Coxon CE (1986) A study of the hydrology and geomorphology of turloughs, Trinity College Dublin. PhD Coxon CE (1987) The Spatial Distribution of Turloughs. Irish Geography 20:11-23 Drew D, Daly D (1993) Groundwater and Karstification in Mid-Galway, South Mayo and North Clare. Geological Survey of Ireland. Dublin, Republic of Ireland Porst G, Irvine K (2009) Distinctiveness of macroinvertebrate communities in turloughs (temporary ponds) and their response to environmental variables. Aquatic Conservation: Marine and Freshwater Systems 19(4):456 - 465 Porst G, Naughton O et al. (2009) The importance of disturbance for seasonal and interannual succession of macroinvertebrates in turloughs

153

Reynolds JD (1996) Turloughs, their significance and possibilities for conservation. The Conservation of Aquatic Systems. Royal Irish Academy, Dublin, pp 38-46 Tynan S, Gill M et al. (2005) Development of a methodology for the characterisation of a karstic groundwater body with particular emphasis on the linkage with associated ecosystems such as turlough ecosystems, Environmental Protection Agency Williams P W (1964) Aspects of the Limestone Physiography of Parts of Counties Clare and Galway, Western Ireland. Unpublished PhD Thesis, University of Cambridge Yanalak M (2003) Effect of Gridding Method on Digital Terrain Model Profile Data Based on Scattered Data. Journal of Computing in Civil Engineering 17(1):51-67

154

Hazards in karst Mario PARISE National Research Council, IRPI, Bari, via Amendola 122-I, Bari, Italy, e-mail: [email protected]

Abstract: Karst is an extremely fragile natural environment. The geological, morphological, hydrological, and hydrogeological features of karst determine an overall high vulnerability to a number of potentially dangerous events. The delicate equilibrium of karst ecosystems can be therefore changed very easily, sometimes dramatically and irreversibly, up to its destruction. This may occur as a consequence of both natural and anthropogenic impacts. The present paper examines the main peculiarity of karst, and discusses the hazards affecting karst settings, by subdividing them into the two main categories above mentioned: natural vs. anthropogenic hazards. Sinkholes, mass movements, floods, loss of karst landscape will be treated, together with other hazards more directly related to man’s activities, such as marine intrusion, groundwater vulnerability, pollution, loss of the valuable archives contained in karst caves, quarrying, etc. Eventually, a particular focus will be done on the necessity to protect karst, an environment that needs specific regulations to be properly safeguarded, and where, more than in any other geological and natural setting, a sustainable approach for the actions implemented by man has to be searched for. In this sense, some insights about the application of a recently developed method (Karst Disturbance Index) for the evaluation of the degree of disturbance done by man to the natural karst will be also provided.

Keywords: hazard, karst, sinkhole, flood, water management

1 Introduction Hazard is defined as the probability of occurrence of a potentially damaging event in a given area and in a given time frame (Varnes 1984). The need to constrain the temporal frame, in addition to the spatial one, distinguishes hazard from susceptibility. Hazards can be subdivided into two main categories, whether the hazard is natural or anthropogenic in origin. In the first case, we deal with any event, neither provoked or dependant by man and his activities, that causes a more or less sudden change in the natural environment: to provide some examples of events belonging to this category, earthquakes, floods, tsunamis, mass movements may be mentioned. When the event is directly caused by man, or can be linked to human activities, the hazard is defined as anthropogenic. In this latter case, the situation may be even worst, and the produced changes often become irreversible, which means that recovering the pristine state of the natural environment may be high costly or impossible to achieve. Karst is one of the most fragile and vulnerable environments on Earth. This is due to the main hydrologic, geomorphologic and hydrogeological features of karst (White 1988, Ford and Williams 2007, Parise and Gunn 2007). In karst, generally we do not see water flowing at the surface (Cvijic 1918): it finds rapidly its way underground through the complex network of conduits, fissures and caves, enlarging the voids in the rock mass by solution and mechanical erosion, and producing the richly decorated underground environment of karst caves. Scarcity of water at the surface, therefore, is the first element; but sometimes there are exceptions to the rule. Prolonged periods of rainfall, or short but intense rainstorms may cause flooding (White and White 1984, Mijatovic 1987, Parise 2003), and/or the formation of temporary lakes, which duration depends on the capacity of water to open its way through the clogged sites, and infiltrate underground. Another very important feature in karst is represented by the hydrological boundaries of the catchments (Palmer 2007). In other environments it is possible to delimit the catchment 155

looking at the surface morphology, following ridges and watersheds, dividing a certain area in several basins, computing the hydrological balance for each of them. Knowing the amount of rainfall that drops in a specific watershed, its size and geological characters, it is therefore possible to estimate how much water will be collected in the discharge area. Karst does not work like this. Identification of the surface watershed does not necessarily correspond to what is really happening underground. Water can infiltrate at a certain site, and, due to complex underground systems, be transported in another, nearby watershed. The only way to be sure of the course of water in karst is following it underground, when possible, or using dye tracers (Goldscheider and Drew 2007) to follow its path whenever man is not able to enter narrow or flooded passages. 2 Natural and anthropogenic hazards in karst The main natural and anthropogenic hazards affecting karst areas are briefly described in this section; due to the limited space and for sake of brevity, the following pages cannot be considered exhaustive, of course, but hopefully they will provide the reader with the main and updated references on the topic. 2.1 Sinkholes Sinkholes are definitely one of the most typical hazards in karst areas: originally defined as a circular depression in karst environments, with underground drainage (Bates and Jackson 1987), the term sinkhole has become later on more popular, and is nowadays used even to designate collapses in non karst areas, or in urban centres. In these latter cases, sinkholes are generally related to presence of man-made underground cavities that are at the origin of collapses of the ground with consequent damage to the built-up area above. Sinkholes may occur through different processes, in some cases being related to the presence of underground voids (collapse and cover collapse sinkholes in the classification by Waltham et al. 2005), whilst in others they are simply produced by dissolution of soluble rocks (solution sinkholes), or can be differentiated on the basis of the presence of cohesive (dropout sinkhole) or non-cohesive soils (suffosion sinkhole) in the overburden above a soluble bedrock. What is more important, in terms of the hazard posed to man, is the velocity of the process, especially as regards the catastrophic phase, that is the collapse. In this term, the most dangerous typologies are those characterized by high velocity, namely collapse and cover-collapse sinkholes (Tharp 1995). Lack of premonitory signs, and the rapid evolution in the catastrophic phase of the event, are elements worth to be further investigated by scientists, in the attempt to work toward mitigation of the risk deriving from sinkholes. As previously mentioned, however, man and his activities may also play a crucial role in the development of sinkholes: underground cavities, in some ways “forgotten” or abandoned by man, and later on become part of the built-up areas, are the main problem. Knowledge of where these voids are, what characteristics they do have (depth, width, length, etc.), and their stability conditions as well, is fundamental to avoid the possibility of encountering problems, or causing collapses in the above roads or buildings. Any other human action, that may determine changes in the hydrological regime, has to be carefully planned in karst areas, and should take into consideration the possibility to induce or trigger sinkhole development. An interesting case study in this sense is that at Marina di Lesina, in Apulia (southern Italy) where some decades ago the flanks of an artificial channel connecting the nearby lake to the Adriatic Sea have been object of works to replace the previously existing concrete structures with gabions, considered to have a lower environmental impact. Being the channel opened in Triassic evaporite rocks, interested by presence of many cavities filled by sandy-silty materials, the gabions bounding the channel facilitated flow of water toward the channel itself, and draining out of the sands and silts from the cavities. The new 156

voids so produced determined collapses at several sites (Fig. 1), with the process still going on since many years, and progressive shifting of the area affected by sinkholes away from the channel, up to the built-up areas. Several buildings show at present cracks and slight deformation in the foundations. The lesson learned from this event is that carrying out engineering works without taking into the due consideration the peculiar aspects of karst can be extremely dangerous, and cause risk to the man-made infrastructures and buildings.

Figure 1 Sinkhole in the Triassic evaporite at Marina di Lesina (Apulia, Italy). In the background, the buildings that are at present being involved in sinkhole development. 2.2 Mass movements In a broader sense, sinkholes are a form of instability, and can be considered as a category of mass movements, even though they generally affect very low-gradient, if not subhorizontal, areas. Slope movements s.s. too are frequent in karst, in those areas where relief energy is high, rock mass conditions are poor, and (which is peculiar of karst) caves are present (Waltham 2002, Santo et al. 2007, Parise 2008). Many different typologies of mass movements can be observed in karst, the most significant being those involving rock mass such as falls, or toppling failures from steep to vertical valley sides and canyons. It has to be mentioned that in many cases breakdown processes, mostly occurring through progressive failures from the vault, represent the main type of evolution of caves, once they have been left by water (White and White 1969). Analysis of breakdown deposits within cave systems, in relation to geology, and to the above existing man-made structures at the surface, are therefore of extreme importance for hazard assessment, by adding a further view (that from the inside) to the common analysis from the surface (Klimchouk and Andrejchuk 2002). To evaluate the sectors within cave systems mostly prone to further failures, it is also very important to consider the weathering processes in the rock mass (Fookes and Hawkins 1988, Zupan Hajna 2003), which often causes strong reductions in the mechanical properties of the rocks, thus contributing to its overall decrease in strength, and facilitating its proneness to failures. In some cases, depending upon depth of the underground voids and local 157

geological conditions, upward propagation of such failures may lead to development of sinkholes at the surface (Culshaw and Waltham 1987, Waltham and Lu 2007). 2.3 Floods Water generally infiltrates underground in karst through the complex network of karst conduits, caves and fissures, until it is transferred to the outflow zones. There is therefore a very limited amount of surface runoff; when the swallow holes become, for some reasons, partially or totally clogged, water is not able to enters the ground, and thus it accumulates at the surface, creating floods and inundating areas (Bonacci et al. 2006), especially in the case of very flat, polje-like, valleys. Floods in karst may be frequent and destructive (Fig. 2). Many examples worldwide show that, unfortunately, man tends to loose the memory of past events, and keep planning and building in karst areas without taking into account that sometimes in the past those areas had been affected by flooding and inundation (Parise 2003). As a consequence of such behavior, we are still facing great damage on the occasion of floods, and probably it is easy to forecast further, increasing damage in the next future.

Figure 2 Damage produced by flooding in karst environment: the effects of the September 2003 flood event in the Taranto province (Apulia, Italy) 2.4 Quarrying One of the most destructive activities by man for karst areas is quarrying (Gunn 1993, 2004): quarries may cause partial or total destruction of caves, degradation of the landscape, changes in the natural hydrography. Once the activity stops, they often become sites of illegal landfills, inevitably causing pollution to karst ecosystems, particularly the aquifers. Beside surface quarries, quarrying may also be developed underground, with a long and difficult work to realize subterranean galleries, and bring at the surface the excavated material. In southern Italy, remarkable examples of this type of quarries are at Cutrofiano, where the local calcarenite have been extensively quarried at depths ranging from about 10 to 45 meters, with galleries extending for total lengths of some tens of km. Underground quarrying activity may also result in development of sinkholes at the surface, because of instability problems underground (Fig. 3) that then progressively move upward.

158

Figure 3 Failures from the ceiling and the walls in an underground quarry at Cutrofiano (Apulia, Italy). The quarry is part of an extensive network of subterranean quarries that in overall counts over 30 km of underground galleries. 2.5 Loss of karst landscape Land use changes in karst may be related to efforts in gaining new pieces of land to agriculture, by terracing slopes, filling the depressions, and clearing the fields from the rocky stones. Especially when the karst landscape is very flat, it is extremely easy to cancel some of the landforms. In ancient times stone clearing was performed by hand, and the collected stones were used to build the dry-stone walls, a typical element in the rural architecture of many Mediterranean karst areas (Nicod 1972); in the last decades, thanks to the extensive use of machinery to remove and destroy the rocks, stone clearing has completely changed the natural setting, destroying the epikarst, diverting the surface runoff, causing erosion even on the occasion of small amount of rainfall, and determining in many areas a strong tendency toward desertification (Parise 2009). As a further consequence, many rocks taken out from the field are dumped into caves or accumulated at their entrances, in both cases determining a serious danger to cavers (Parise and Pascali 2003). 2.6 Pollution Pollution events in karst are often related to other hazards, as before said: abandoned quarries, or other types of anthropogenic cavities, may become sites where solid and liquid wastes are discharged, with severe impact on the quality of the karst groundwater resources (Zwahlen 2004). In particular social situations, such as in post-conflict scenarios, the situation becomes even worst and severe (Calò and Parise 2009). Natural caves, too, are not immune by pollution, since in those territories where no specific protection law exists, or, if it exists, it is not actually reinforced, dumping of different types of wastes (even toxic and highly dangerous) is carried out. Another type of pollution is typical of the coastal areas, where uncontrolled withdrawal of water determines upwelling of the interface between fresh and salt water, and the abstraction of brackish water. In most of the cases, this occurs because of the presence of tens, if not hundreds, of illegal wells along the coastlines. 3 Conclusions Man has definitely become one of the most powerful factors that can cause changes in the karst environment, produce direct damage, or predispose the territory to threatening events 159

(Milanovic 2002). A possible way to determine the impact deriving from human activities to the karst environment has been recently proposed in the form of the Karst Disturbance Index (van Beynen and Townsend 2005, North et al. 2009). Taking into account a number of indicators, subdivided into 5 different categories (Geomorphology, Atmosphere, Hydrology, Biota, Cultural) the disturbance to the karst environment deriving by anthropogenic actions can be determined (Table 1), to provide some insights about how the karst landscape has negatively been modified by man. Table 1 Karst Disturbance Index level categories (after North et al. 2009) Karst Disturbance Index Pristine Little disturbance Disturbed Highly disturbed Severely disturbed

Value 0.0-0.19 0.2-0.39 0.4-0.59 0.6-0.79 0.8-1-0

Given the high fragility of karst, what should be done to protect such environment, and safeguard the natural resources it contains? Regulations, laws, strict enforcement of existing rules are important and necessary things. However, in many cases they are not enough: a law can be written in an excellent way, but too often it remains valid only on paper, when no control is then actually implemented. Repression does not work very well, either. What needs to be done, which can be much more powerful than prohibition, is let people understand the reasons why they should not carry out certain actions. Explain to them the consequences of their activities. Let them understand that the damage they are producing does not only affect the environment, but also their own life, that of their children, the next generations, the water they drink, the food they crop and eat. Education, and direct involvement of the local inhabitants living in karst, is fundamental. When people really realize that each action carried out at the surface is transferred underground with a negative impact, the number of potential polluting events reduce. Scientists and cavers should produce stronger efforts in this direction: educating people, talking about karst to the people that do live in karst lands, carrying their knowledge outside the too often much closed world of karst science and speleology. It has not to take for granted that everybody knows what karst is, and what problems may affect it.

References Bates RL, Jackson JA (1987) Glossary of geology. American Geol. Institute, 3rd ed Bonacci O, Ljubenkov I, Roje-Bonacci T (2006) Karst flash floods: an example from the Dinaric karst (Croatia). Natural Hazards and Earth System Sciences 6:195-203 Calò F, Parise M (2009) Waste management and problems of groundwater pollution in karst environments in the context of a post-conflict scenario: the case of Mostar (Bosnia Herzegovina). Habitat International 33:63-72 Culshaw MG, Waltham AC (1987) Natural and artificial cavities as ground engineering hazards. Quart. J. Eng. Geol. 20:139–150 Cvijic J (1918) Hydrographie souterraine et évolution morphologique du karst. Rev. Trav. Inst. Géogr. Alpine 6:375-426 Fookes PG, Hawkins AB (1988) Limestone weathering: its engineering significance and a proposed classification scheme. Quart. J. Eng. Geol. 21:7–31

160

Ford D, Williams P (2007) Karst hydrogeology and geomorphology. John Wiley & Sons, 562 pp Goldscheider N, Drew D (eds) (2007) Methods in karst hydrogeology. International Contributions to Hydrogeology 26, Int. Ass. Hydrogeologists, 264 pp Gunn J (1993) The geomorphological impacts of limestone quarrying. Catena 25:187198 Gunn J (2004) Quarrying of limestones. In: Gunn J (ed) Encyclopedia of cave and karst science. Routledge, London, pp 608-611 Klimchouk A, Andrejchuk V (2002) Karst breakdown mechanisms from observations in the gypsum caves of the Western Ukraine: implications for subsidence hazard assessment. Int. J. Speleol. 31 (1/4):55–88 Mijatovic BF (1987) Catastrophic flood in the polje of Cetinje in February 1986, a typical example of the environmental impact of karst. In: Beck BF, Wilson WL (eds) Proc. 2nd Multidisc. Conf. on Sinkholes and the Environm. Impacts of Karst, Orlando, 9-11 February 1987, pp 299-303 Milanovic P (2002) The environmental impacts of human activities and engineering constructions in karst regions. Episodes 25:13–21 Nicod J (1972) Pays et paysages du calcaire. Presses Univ. de France, Paris, 242 pp North LA, van Beynen PE, Parise M (2009) Interregional comparison of karst disturbance: West-central Florida and southeast Italy. J. Environ. Management 90: 1770-1781 Palmer AN (2007) Cave geology. Cave Books, 454 pp Parise M (2003) Flood history in the karst environment of Castellana-Grotte (Apulia, southern Italy). Natural Hazards and Earth System Sciences 3(6):593-604 Parise M (2008) Rock failures in karst. In: Cheng Z, Zhang J, Li Z, Wu F, Ho K (eds) Landslides and Engineered Slopes. Proc. 10th Int. Symp. on Landslides, Xi’an (China), June 30 – July 4, 2008, 1, pp 275-280 Parise M, (2009) Land use changes in the karst landscape of Apulia, south-eastern Italy: the negative effects of stone clearing. Proc. Int. Conf. Geography, Gjirokaster (Albania), 2021 November 2009 Parise M, Pascali V (2003) Surface and subsurface environmental degradation in the karst of Apulia (southern Italy). Environ Geol 44:247-256 Parise M, Gunn J (eds) (2007) Natural and anthropogenic hazards in karst areas: Recognition, Analysis and Mitigation. Geological Society, London, sp. publ. 279, 202 pp Santo A, Del Prete S, Di Crescenzo G, Rotella M (2007) Karst processes and slope instability: some investigations in the carbonate Apennine of Campania (southern Italy). In: Parise M, Gunn J (eds) Natural and Anthropogenic Hazards in Karst Areas: Recognition, Analysis and Mitigation. Geol. Soc. London, sp. publ. 279: pp 59–72 Tharp TM (1995) Mechanics of upward propagation of cover-collapse sinkholes. Engineering Geology 52:23-33 van Beynen PE, Townsend K (2005) A disturbance index for karst environments. Environ Management 36:101-116 Varnes DJ (1984) Landslide hazard zonation: a review of principles and practice. Unesco, Paris, 63 pp Waltham AC (2002) The engineering classification of karst with respect to the role and influence of caves. Int. J. Speleol. 31(1/4):19–35 Waltham T, Lu Z (2007) Natural and anthropogenic rock collapse over open caves. In: Parise M, Gunn J (eds) Natural and Anthropogenic Hazards in Karst Areas: Recognition, Analysis and Mitigation. Geol. Soc. London, sp. publ. 279: pp 13–21 Waltham T, Bell F, Culshaw M (2005) Sinkholes and subsidence: karst and cavernous rocks in engineering and construction. Springer, Berlin. 161

White WB (1988) Geomorphology and hydrology of karst terrains. Oxford Univ. Press, 464 pp White E, White W (1969) Processes of cavern breakdown. Bull. Natl. Speleol. Soc. 31 (4):83–96 White EL, White WB (1984) Flood hazards in karst terrains: lessons from the Hurricane Agnes storm. In: Burger A, Dubertret L (eds) Hydrogeology of Karst Terrains, 1, pp 261-264 Zupan Hajna N (2003) Incomplete solution: weathering of cave walls and the production, transport and deposition of carbonate fines. Carsologica, Postojna-Ljubljana, 167 pp Zwahlen F (2004) Vulnerability and Risk Mapping for the Protection of Carbonate (Karstic) Aquifers. Final report COST action 620. European Commission, Brüssel

162

Modelling of rillenkarren formation Matija PERNE Karst Research Institute SRC SASA, Titov trg 2, SI-6230 Postojna, Slovenia e-mail: [email protected]

Abstract: The purpose of our work is development of mathematical model of rillenkarren formation. Presumptions of our model are based on experimental evidence. Rillenkarren are known to form from initially flat surface of soluble rock under constant uniform rain. Because of that we omit beginnings and ends of rain showers and only model constant rain. Under these circumstances all the rock surface is wet and rainwater is flowing over it in a thin film. The situation stays the same for minutes and hours, except for rock retreat which is much slower than water flow and can be safely neglected in flow calculations. The rock is mostly carried away in dissolved form. Rills form because the dissolution rates are different on different sites on the rock surface. This can only result from differences in chemical properties of the solution. These are influenced only by the history of the water since the arrival of the raindrop onwards because raindrops are chemically all the same. Dissolved rock has almost no influence on hydrodynamical properties of the solution. Water flow is therefore evaluated independently from dissolution and calculated in advance for all types of rock at the same time. While the presented model does not predict rillenkarren formation, it is known from nature and experiments that they do form under the modelled circumstances. So we can infer that some of the process which was not accounted for correctly is crucial for rillenkarren formation. Some of the approximations, simplifications and inaccuracies of our model are: 1) Instead of Nervier-Stokes equation lubrication approximation is used. The approximation works well on average but on the upper edge of the rock it does not. 2) Steady-state approximation of the water flow is used. 3) The phenomena at raindrop impacts are not accounted for. Fresh water is added only at the water surface, eventual penetration of drops into the film is neglected.

Keywords: rillenkarren, numerical modelling, dissolution

1 Introduction Rillenkarren are parallel shallow channels, 20-30 mm apart and separated by sharp ridges that form on inclined bare surfaces of soluble rocks due to dissolution by rainwater (Lowe and Waltham 1995). Their theory of origin is unknown (Field 2002) and the only detailed explanation in (Glew and Ford 1980) is incomplete and weakly supported. Rillenkarren can be used in determining history of particular rock surfaces, for example in (Gams 1989). From the stage of rillenkarren development it is inferred for how long the rock is directly exposed to influences of the atmosphere. Insight of the mechanism of rillenkarren formation would make the method more reliable. Rillenkarren form on quite different rocks (Ginés in preparation) so details of dissolution chemistry cannot be important for their formation. Live organisms are not always present on rillenkarren surface so they are also not crucial.

2 Numerical model 2.1 Protorill Because rillenkarren form on initially flat surface, they inevitably have to go through a stage in which the surface is only gently undulated, covered with shallow rills which are 163

becoming deeper. A satisfactory model of rillenkarren formation thus has to predict that a rill which is not as deep as mature rills will deepen. All the models are tested on the same rock surface form, dubbed protorill. It has parabolic shape because every gentle curve is parabolic in the first approximation. Its slope is 40° and it is 2 cm wide, which are typical values for rillenkarren, while it is only 0.5 mm deep, much less than real rillenkarren. The upper 16 cm of such a rill were studied. The density of water is taken to be 103 kg/m3 and its viscosity 10-3 Pas. 2.2 Model of the water flow Fluid flow is described by Nervier-Stokes equation which is difficult to solve. Because of that some reasonable approximations are used in order to simplify it. For the beginning, compressibility of water is neglected. Nervier-Stokes equation for incompressible fluid is ª wv º (1)  « + v ˜ ’ v » = f b  ’p + ’ 2 v (Kušer and Kodre 1994) ¬ wt ¼ where  is density, v is velocity, t is time, f b are body forces per unit volume, p is pressure and  is viscosity. A group of approximations known as lubrication approximation is then used and surface tension is also ignored. The lubrication approximation is based on neglecting inertial forces that is the whole left side of equation 1 and second derivatives of velocity in direction parallel to the surface.

Figure 1 Coordinate systems. A local Cartesian coordinate system in which the rock surface lies in xy plane introduced (see Figure 1). Here m is water depth, h is thickness of the water film, z 0 elevation of rock surface above a reference level, and  is the surface slope. For density water flow j it turns out after introducing the mentioned approximations that it proportional to the water film thickness to the third power and to the surface slope. That is g h 3 (2) j=  ’ z 0 + m ,  3 164

is is of is

where g is acceleration of gravity and ’ stands for derivation in x and y directions only. Conservation of water gives another equation: wh , (3) ’ ˜ j = v~r  wt where v~r is rain intensity in the z direction. The equations 2 and 3 can be solved numerically for arbitrary surface shape. The method of time propagation is efficient enough to find an approximate steady-state solution. That means an initial approximation for m is taken, j is calculated from equation 2 and the new m after a short time step is calculated from equation 3. The procedure is then repeated with the new m as initial approximation until m converges to a steady state and does not change anymore.

Figure 2 The left graph shows the rock surface and the right one the steady-state water surface when the rock is exposed to rain. Figure 2 shows a test result of the algorithm on a slope with a depression. A pool of water fills the depression the same way as in reality. The method performs as good as expected, or better. It should be noted that in this case an assumption of lubrication approximation that water and rock surfaces are nearly parallel is not fulfilled but the result is realistic anyway. Instead of depending on numerical solution, which is inherently only approximate, the water flow over the protorill can be calculated analytically. The steady-state form of the equation 3 is ’ ˜ j = v~d , (4) ~ where vd is time average of rain intensity. The flow density j is parallel to the water surface slope. The equation is nonlinear because j depends both on the third power of water film thickness, or m , and on gradient of m itself. But if the local water depth is much smaller than typical height differences between points on the rock surface, the rock surface slope is approximately equal to the water surface slope and can be used in its place. Thus the equation 4 becomes linear and easy to solve analytically. It can be solved using method of characteristics.

165

Figure 3 Both graphs show calculated water depth as a function of position on the protorill. On the left graph there are the results of analytical calculation and on the right graph of numerical one. On the left rear side is the top of the rill which extends to the right front side. The results are for rain intensity of 10 mm/h. All coordinates are in mm. Figure 3 shows the differences between numerical and analytical solution. Results are almost the same for upper part of the rill. On its lower part the numerical solution becomes smoother while the analytical one stays sharp. The difference could result either from numerical diffusion, which would mean that the analytical solution is better, or the additional approximation used only in analytical calculation, in which case the numerical solution is more exact. It turns out that the difference between both solutions becomes important only when the water is deep in comparison to the rill and the presumptions of the analytical method are no longer fulfilled. Thus both solutions agree with each other wherever the analytical one is correct. That means the difference results from the additional approximation in analytical method, so the numerical solution may be the better one and was used for calculating dissolution rates. 2.3 Dissolution kinetics The dissolution rates of calcite are determined by three rate-controlling processes (Kaufmann and Dreybrodt 2007): 1. the kinetics of dissolution at the mineral surface, which depends on the chemical composition of the solution at the mineral surface; 2. mass transport of the dissolved material away from the boundary by diffusion; 3. conversion of CO2 into H+ and HCO3. Which one of the processes is the slowest depends on the circumstances. In the case of a few tenths of a millimetre thick water film on a rock surface, the dissolution rate is limited only by the reaction on the surface (Dreybrodt et al. 2005). In such geometry diffusion is much faster and thus unimportant for dissolution rate. Diffusion is described by diffusion equation Dc (5) = D’ 2 c , Dt n where D is diffusion coefficient and c is concentration, c = . Diffusion has no influence V on dissolution when D o f . In this case the dissolution rate is limited by other processes and c does not depend on distance from the rock surface. The equation describing the surface reaction is 166

w S (6) =  ceq  c , (Dreybrodt et al. 2005) wt where  is a constant,  S is surface density of dissolved matter, and ceq is equilibrium

concentration. It is unclear whether the simple form of the equation 6 is only an approximation or if it has a deeper meaning. It is also questionable if diffusion is really unimportant for dissolution kinetics. In absence of a more suitable model the relation 6 and the limit D o f are used anyway for modelling rillenkarren on limestone. In the cases of gypsum or salt, on which rillenkarren also form, the surface reaction is so fast that dissolution under the same circumstances is limited only by diffusion (Jeschke et al. 2001). The situation is also less complex because CO2 does not enter the reaction and its conversion has no influence on dissolution rates. So dissolution can be modelled by equation wS wc = D~ , (7) wt ~wz z=0 where z is the coordinate normal to the rock surface. 2.4 Models of rillenkarren formation Both modes of dissolution, for limestone and for gypsum and salt, are first applied on flat surface and then on the protorill. For limestone, the relation 6 and limit D o f are used. The concentration of solute is thus independent of z , it is constant along a vertical profile through the water layer. m2 which is close For gypsum and salt, equation 7 is used with the value of D = 10 9 s to real diffusion constants for these substances at normal temperatures. 2.5 Limestone, flat surface The surface is oriented so that its upper edge is horizontal. Rainwater then flows in the direction of the slope. It turns out that the dissolution rate is the same everywhere on the surface. Steeper surface retreat slower but lowers faster than gentler ones if the rain is held constant. 2.6 Limestone, protorill Water flow is always parallel to the local water surface slope direction. The shape of the water surface is calculated as described in subsection 3.2 and can be used to calculate paths along which water flows downward. From here on these paths are called flow lines. The protorill is uniformly covered with such flow lines and dissolution rates are calculated along every one of them. Every flow line can be dealt with independently from others because water in the film neither enters nor leaves it, while lateral gradients of concentration are presumed to be very small so diffusion of solutes into or out of the flow line can be safely neglected. Dissolution rates on the points of rock surface that do not lie directly on a calculated flow line are obtained by interpolating. The results are shown on figure 4.

167

Figure 4 Speed of lowering of the surface of a protorill made of limestone. Higher points mean faster lowering. The top of the rill is on the top left side. The graph shows values of ceq  c for  = 10 7 m / s . ceq cos 2.7 Gypsum or salt, flat surface On these rocks, the concentration of dissolved rock is dependent on z while at the rock surface it is assumed to be at ceq , so mass transport in both x and z direction is taken into account. From the lubrication approximation water velocity field is calculated and advection in both x and z directions is accounted for. Diffusion in x direction is neglected because of small concentration gradients while in z it is the main force driving the mass transport and so has to be taken into account. A suitable coordinate transformation made it possible to solve the diffusion-advection problem with finite difference scheme. Dissolution rates directly follow from the resulting concentration field. In this case, dissolution rates on different points on the flat surface are different. 2.8 Gypsum or salt, protorill The model for the flat surface has to be only slightly modified to handle dissolution on a flow line along a curved surface. The same flow lines as for the limestone protorill are used, dissolution rates all over the rill are calculated and are presented on figure 5.

168

Figure 5 Speed of lowering of the surface of a protorill made of gypsum or salt. Higher points mean faster lowering. The top of the rill is on the top left side. The graph shows values of c 2  c1 in mm-1. Values are calculated with resolution of 51 nodes along z direction in ceq z 2  z1 finite difference scheme. 3 Conclusion The presented model does not predict rillenkarren formation. On the other hand, it is known from nature and experiments that they do form under the modelled circumstances. So we can infer that some of the process which were not accounted for correctly are crucial for rillenkarren formation. Some of the approximations, simplifications and inaccuracies common to all models are: x Instead of Navier-Stokes equation lubrication approximation is used. The approximation works well on average but on the upper edge of the rock it does not. x Steady-state approximation of the water flow is used. x The phenomena at raindrop impacts are not accounted for. Fresh water is added only at the water surface, eventual penetration of drops into the film is neglected. Conditions during rillenkarren formation are certainly not steady-state. The steady-state shape of water film on 2 cm by 16 cm rill is calculated using time propagation. It turns out that after three seconds of simulated water flow the film shape is very near the steady state, even if initial state is far from the steady one. We assume that state in reality is nearly steady if a lot of drops fall on the rill in less than 3 s. If we take 1 mm3 as an average raindrop

169

volume, in the simulated rainfall rate of 10 mm/h only 27 drops impact the rill in three seconds. The raindrop does not stay at the surface of the water film, as presumed for the calculations, but pushes off some of the old film. It is also possible that it does not stay at the site of impact, maybe it bounces toward the centre of the rill, effectively increasing rainfall rate at the centre. This effect is not taken into account either. So it would make sense to include raindrop impact and non-steady state situation into future models of rillenkarren formation.

References Bögli A (1960) Kalklösung und Karrenbildung. Zeitschrift für Geomorphologie, Supplementband 2: Internationale Beiträge zur Karstmorphologie, pp 4–21 Dreybrodt W (1988) Processes in Karst Systems: Physics, Chemistry, and Geology. Springer-Verlag, Berlin Heidelberg, 288 pp Dreybrodt W, Gabrovšek F, Romanov D (2005) Processes of speleogenesis: a modeling approach. Založba ZRC, Ljubljana, 375 pp Field MS (2002) A Lexicon of Cave and Karst Terminology with Special Reference to Environmental Karst Hydrology. U.S. Environmental Protection Agency, Washington, DC, 214 pp Gams I (1989) Dežni žlebii kot pokazatelji starosti deforestacije. Geografija in Aktualna Vprašanja Prostorskega Razvoja:127–138 Ginés À (in preparation) Rillenkarren. In: W Dreybrodt, A Ginés, M Knez, T Slabe (eds) Karst rock features and rock relief Glew JR, Ford DC (1980) A simulation study of the development of rillenkarren. Earth Surface Processes 5:25–36 Jeschke AA, Vosbeck K, Dreybrodt W (2001) Surface controlled dissolution rates of gypsum in aqueous solutions exhibit nonlinear dissolution kinetics. Geochimica et Cosmochimica Acta 65(1):27–34 Kaufmann G, Dreybrodt W (2007) Calcite dissolution kinetics in the system CaCO3– H2O–CO2 at high undersaturation. Geochimica et Cosmochimica Acta 71: 1398–1410 Kušer I, Kodre A (1994) Matematika v fiziki in tehniki. Društvo matematikov, fizikov in astronomov Slovenije, Ljubljana, 394 pp Lowe D, Waltham T (1995) A dictionary of karst and caves: a brief guide to the terminology and concepts of cave and karst science. BCRA, London, 40 pp Sweeting MM (1972) Karst landforms. Macmillan, Basingstoke, 362 pp

170

Salinization of the Vrana Lake in Dalmatia within the context of anthropogenic influences and climate changes (situation in 2008) Josip RUBINIC1, Ana KATALINIC2, Mirjana SVONJA3, Ivana GABRIC4, Gordana BUSELIC5, Maja CUZE6, Bojana HORVAT7 1

Faculty of Civil Engineering, University of Rijeka, V.C.Emina 5, 51000 Rijeka, Croatia, e-mail: [email protected] 2 Public Institution Nature Park Vransko jezero, Kralja P. Svacica 2, 23210 Biograd, Croatia, e-mail: [email protected] 3 Croatian Waters, Vukovarska 35, 21000 Split, Croatia, e-mail: [email protected] 4 Croatian Waters, Vukovarska 35, 21000 Split, Croatia, e-mail: [email protected] 5 State Meteorological and Hydrological Service, Gric 3, 10000 Zagreb, Croatia, e-mail: [email protected] 6 Public Institution Nature Park Vransko jezero, Kralja P. Svacica 2. 23210 Biograd, Croatia, e-mail: [email protected] 7 Croatian Waters, Ulica grada Vukovara 220, 10000 Zagreb, e-mail: [email protected]

Abstract: Vrana Lake is a coastal lake in Dalmatia (Croatia), protected as Nature park. Although it is by its surface of about 30 km2 the largest lake in Croatia, its volume comprises only about 82.5 mil m3. The lake is a cryptodepression (with maximum level depth -3m) separated from the sea by a 0.8 – 2.5 km wide limestone ridge within which the coastal karst aquifer interacts with both the sea and the lake. Therefore this lake is a very sensitive karst coastal system with a close interaction with the sea - a direct interaction through the 800 m long Prosika canal and the indirect interaction through the karst aquifer. The period of a few recent years (particularly 2007-2008) has been characterized by pronounced draughty hydrological conditions. Evidence of the lake critical condition is the total volume of water flowing out into the sea through the Prosika canal of only about 0,19 mil. m3 in the year 2008 which is minor as compared to the average of about 31,8 mil.m3 for the analyzed period 2000-2008. Consequently, during the summer of 2008, the water level of Vrana lake fell under the sill level of Prosika canal and in specific conditions of daily sea level fluctuations even below the sea level. That caused the sea to flow into the lake directly through Prosika canal and also through numerous springs on the lake shore. The result was an enormous increase in lake water salinity. Chloride concentration values in southern part of the lake where Prosika canal and a large number of salty inshore springs is situated increased up to 6500 mg/l in 2008, compared to the average of 590 mgl-1 for the previously analyzed period from 2000 to 2006.. Within the context, the sill level of the constructed Prosika canal, which was once sufficient for the stabilization of Vrana lake, can be a problem which generates an increase in lake salinization. Keywords: water salinization, karst, Vrana lake, Dalmatia

1 Introduction Coastal aquifer management is one of the largest contemporary water management challenges in the Mediterranean. Negative anthropogenic influences with increasing pressure on water resources, combined with the general trend of sea level increase (Lambeck et al. 2004, Lambeck and Purcell 2005, Orlic 1995, Pirazzoli 2000, 2005) and air temperature increase which occurs simultaneously with the precipitation volume decrease and catchment area outflow balance trends (Bolle 2003, Gajic-Capka and Zaninovic 2006, Svensson et al. 171

2005, Svonja et al. 2003) make coastal karst aquifers particularly sensitive. Salty water intrusions into deeper parts of coastal karst aquifers, mostly caused by an excessive exploitation of coastal water resources, occur in a large part of the Mediterranean coastal area (Benblidia et al. 1996, Custodio and Bruggeman 1982, Custodio 2002, Margat 2004).

A)

B)

Figure 1 Position of Vrana Lake in Dalmatia: A) Catchment area overview – supplemented according o Fritz (1984) Key: (1) – hydrogeological catchment area border, (2) regulation canal, (3) permanent natural water flow, (4) periodical natural water flow, (5) underground hydrogeological link, (6) source intercepted for water supply, (7) non-intercepted more important source, (8) brackish water source, (9) submarine spring, (10) estavelle, (11) sink hole, B) broader situation of the analyzed area

Vrana Lake in Dalmatia (Figure 1) is one of the most sensitive water resources of large biodiversity which is pointedly affected by all the previously mentioned influences. In 1999 the lake with its adjacent onshore territory was declared a Nature park covering an area of 57 km2. The area in question is an extremely valuable coastal lake site covering an area of about 30.8 km2 and comprising a volume of about 82,5 mil. m3. Vrana Lake is a cryptodepression, up to 5 m deep, with its bottom situated at about 3 m below sea level. The lake runs along the sea coast in its full length of about 13.6 km. It is separated from the sea by a 0.8 – 2.5 m wide limestone ridge within which the coastal karst aquifer dynamically interacts with both the sea and the lake and through which the lake also interacts with the sea. Already in 1770 the lake was directly connected to the sea by a 800 m long Prosika canal which has later been widened and deepened on many occasions in order to improve the drainage of a constructed hydrotechnical melioration system. The canal bottom level is at only 0.43 m above sea level (m a. s. l.) while the water level of the lake varies from 0.03 m a. s. l. (in 1990 and 2008) to 2.24 m a. s. l. (in 1974 and 1994), mean value being 0.81 m a. s. l. The lake-sea interaction is performed in two ways. Direct interaction occurs through Prosika canal with water mostly flowing out, except during long-term dry periods when the sea uplevels the lake, allowing sea water to flow directly into Lake Vrana. Indirect interaction occurs through the karst aquifer where salinization of the lake takes place through several spring groups. Therefore the lake is characterized by great variations in and chloride concentration (from about a 100 mgl-1 up to over 6.000 mgl-1). In the recent years, 2007 and especially in 2008, Vrana Lake suffered extremely unfavourable hydrological conditions of increased salt-water intrusion into the lake (Rubinic and Cuze 2009). Should the started negative processes of salinization continue, lowering of 172

biodiversity together with trophy grade increase and rapid lake degradation is to be expected. In order to develop measure to prevent such negative scenario, an analysis of the existing hydrological conditions is a prerequisite condition. This paper therefore analyzes some of the basic hydrological interrelations of the lake catchment area, its karst aquifer and the surrounding sea.

2 Vrana Lake characteristics and its hydrological dynamics Vrana Lake gets watered through numerous canals of the constructed hydromelioration system which, apart from the surface waters, also partly gather overflow waters from several karst springs as well as waters from their immediate karst aquifer (Figure 1). The complete potential catchment area surface is about 470 km2 (Fritz 1984). However, hydrological monitoring in the catchment area is not organized in a way that appropriate lake inflow control is applicable. Only in the past 7 – 8 years there are partial data about flow capacities at some of the more important lake inflow sites with a total of 1.8 m3s-1 of the lake inflow balance.

Figure 2 Distribution of annual flow capacities Since proper inflow balance could not be calculated based on the available, relatively short and hydrologically insufficiently relevant data on flow capacities, such estimation for a referential period from 1961-1990 was conducted based on the available climatological parameters. The methodology, based on applying the Langbein method in GIS environment is developed in the paper by Horvat and Rubinic (2006). Figure 2 shows the results of analyzing the specific flow capacity spatial distribution at which the spatial raster of the analyzed data is 1 x 1 km2. By using the latest data (Geotehnicki Fakultet Varazdin 2009) on the partially corrected surface of a potentially gravitating catchment area covering about 411 km2, the lake surface included, the balance of complete mean yearly lake inflows of 3.9 m3s-1 was calculated. The lake also has considerable losses both in form of surface outflow from the lake through Prosika canal and due to evaporation of the lake itself, usage of water in the catchment area and sinking into subterranean passages. According to the recorded data about the flows from the Prosika profile from 1996 to 2008, the average outflow from the lake on the Prosika canal profile entrance was about 1.01 m3s-1. Since the period in question was a relatively dry one with about 20% less precipitation and higher air temperatures, the perennial average amount of canal outflow should be 20 – 50% larger, up to 1.50 m3s-1. The loss of 173

water through evaporation is significant due to high temperatures and shallowness of the lake. Based on the analogy of the registered water evaporations from the salt-pans on the island of Pag from the paper by Berakovic (1983), it has been assessed that the losses are about 1.66 m/year. Similarly, according to the conducted empirical assessments of Meyer method (Hrvatska Vodoprivreda 1994), the average annual evaporation was assessed to 1.403 m/year. Using the mean value of the both data the balance contribution of about 1.50 m3s-1 is achieved regarding the mean lake surface. The annual usage of about 0.107 m3s-1 (Geotehnicki Fakultet Varazdin 2009) was recorded for the water supply needs from Vrana Lake catchment area. However, due to the unregistered usage of water for irrigation and local water supply, this amount is surely almost doubled – cca 0.200 m3s-1. The losses caused by sinking into subterrain passages are practically unknown and within the context of previous balance assessment closures they must be about 0.70 m3s-1. As a comparison, during the solely known water measurement of water loss from Prosika canal which was conducted on July 2nd 2009, the flow of 0.502 m3s-1 was measured at the canal entrance and up to the Prosika canal mouth the flow was just 0.055 m3s-1. This gives the loss from the canal of about 0.45 m3s-1 when the denivelation between the water level of the lake and the sea was about 0.66 m. It is quite certain that water losses at the lake perimeter depend not only on the interrelation of the lake water level and sea water level but also on the level of underground water in the karst aquifer. The obtained results of average Vrana lake inflow balance assessment are similar to those of the balance assessment conducted by Svonja (2003) which showed 4.2 m3s-1 for the same 30-year period. Somewhat lesser values were obtained by Pavic in the study of Vodotok (2008), 3.14 m3s-1 during the relatively draughtier period from 2000 – 2005. The balance assessments obtained for the period 1963 – 1980 must also be pointed out (Berakovic 1983). They relate to the 360 km2 of Vrana lake catchment area up to the formerly planned barrier profile in the very lake. The obtained mean annual flow was 2.48 m3s-1, but without counting the no quantified underground sinking losses from her paper. Even lesser mean annual Vrana lake inflows were obtained in the study of Hrvatska Vodoprivreda (1994) for the period from 1963 – 1992 in the amount of 1.96 m3s-1, which is, within the study itself, contrary to the offered assessments of lake water outflows through Prosika canal (1.5 m3s-1) and the mentioned evaporation assessments. Regarding the relatively questionable data about flows into the Vrana lake system, the paper in question used data from other hydrological parameters at disposal. Figure 3 shows the mean annual lake water level modular value (at Prosika site), the sea level (data from the water level recorder Prosika was partly supplemented with Split-Marjan tide-gauge measurements data, based on the conducted regression analysis) and data about annual precipitation volume from Biograd and Stankovci stations (due to their position in relation to the lake the average values were used) during the last 30-year period. It is clear that despite the observed precipitation decrease trend (0.64%, 5.5 mm/year), there is a mild lake water level increase trend (0.31%, 0.34 cm/year) caused by the present sea water level increase trend (1.08%, 0.26 cm/year) (Katalinic et al. 2007).

3 Analysis of the Vrana Lake salinization during 2008 Vrana lake and its aquifer are in a dynamic balance with the sea. Dry periods with low lake water levels are followed by extreme rises in lake salinity due to inflow of sea water both through Prosika canal, or, even to a larger extent, through salty springs in the southern and northwest part of the lake area. Water from these springs flows immediately into the lake or gets collected by constructed drainage canals on the north-western part of the lake area (such as Kotarka canal). This correlation between the lake-sea water level dynamics and extreme salinization is best seen on Figure 4. It shows the mean annual water level differences of lake and sea water levels as well as the maximum annual recorded salinity values at I – Kotarka 174

canal mouth and II – Crkvine stations (both situated in the northwestern part of the lake) starting from 1982 (with interruption during the war years). During the observed period lake salinity exceeded 2000 mg l-1 in three different occasions, all of which happened during and after extremely low lake water level periods. The first of those was recorded in 1989 – 1990 (Romic et al 1997), the second one in 2003-2004 and the most prominent one in the recent 2007 and 2008 in particular. What the extents of the unfavourable hydrological circumstances were during the year 2008 is best illustrated with the mean annual lake and sea water level difference. While the average lake water level elevation in relation to the sea water level was 51 cm during the observation period, during the year 2008 the reverse situation was noted for the first time which meant that the mean annual lake water level was lower than the mean annual sea water level for 7 cm.

1,60

y Sea = 0.0108x - 20.541

y Lake = 0.0031x - 5.2357

1,40

1,20

Sea Lake

1,00

K

Precipitation

0,80

0,60

yPrecipitation = -0.0064x + 13.843 0,40 1978

1983

1988

1993

1998

2003

Year

2008

Figure 3 Modular value of lake water level, precipitation and sea level variations (1979 2008)

Level diference (lake - sea) (m)

1,0

5000

Salinity

(mgl-1)

4000

0,6 3000

2000

0,2 1000

-0,2 1980

1985

1990

1995

2000

2005

0 2010

Level diference (lake - sea) Chlorides - I Chlorides - II

Figure 4 Annual lake water level and sea water level differences and maximum salinity recorded at the stations on Kotarka mouth (I) and Crkvina (II) in Lake Vrana (1982-2008) Figure 5 shows hourly lake and sea water levels, and Prosika canal outflow volume from Vrana lake for the recent critical year 2008. It is clear that the lake water level was under the Prosika canal bottom level for larger part of the year. During this period not only was there no lake water spillway, but in situations when the sea water was higher than the lake water (during high tide almost on daily basis) the sea slowed down the spillway from the lake and salted the water in the canal in return. During overdrying of the lake outflow the sea flew 175

immediately into the lake through the canal but also fed the karst aquifer. The critical conditions of the year 2008 can be best illustrated by the fact that the total annual outflow volume (from the lake into the sea) through Prosika canal was only about 0.19 mil.m3, which is neglectable in comparison to the average amount of about 31.8 mil.m3 for the observation period 2000 – 2008.

Figure 5 Observed Vrana lake water level and sea water level and outflow from the lake during 2008 Extreme salinization of the lake which was going on in 2007 and especially 2008 is shown in Figure 6. Recorded chloride concentration values at the measuring station Kotarka canal mouth in the northwest part of the lake varied from 424 mgl-1 to 4000 mgl-1 for 20072008 with the average being 2007 mgl-1 in relation to the average of 590 mgl-1 for the previously analyzed period. On the nearby measuring site Crkvine the recorded values during 2007 and 2008 varied from 360 mgl-1 to 4100 mgl-1, the average being 1669 mgl-1 in relation to the average of 650 mgl-1 for the previously analyzed period. In the south-eastern part of the lake near Prosika canal there is also a large number of coastal springs through which the salty water flows into the lake. There, even higher values of salinity were recorded: in 2007-2008 chlorides varied between 395 mgl-1 and 6500 mg/l, the average being 2331 mgl-1, in relation to the average of 590 mgl-1 for the previously analyzed period from 2000 to 2006. During the analyzed period of the so far highest recorded Vrana lake salinity (2007–2008), the mean lake water level was only 0.41 m a. s. l., while the mean value for the period 2000-2006 was 1.04 m a. s. l. During that critical period the sea water level was 0.38 m a. s. l which indicates an increase of 2 cm in relation to the average value of 0.36 m a. s. l for the previous period. The average precipitation volume recorded at Stankovci station during 2007-2008 was 757.2 mm, which is ca 17% less than the measured average value for the previous 7-year period. It can be stated that such an intensive salinization in the past two years was primarily caused by a coincidence of several unfavourable hydrological conditions also helped by anthropogenic influence, namely, existence of Prosika drainage canal, cut off at a level much too low for the present lake-sea water level interrelations. The observed and extremely significant problem of sea water flow into Vrana lake during the analyzed year of 2008 brought the need for constructing a lock gate on Prosika canal into focus. This lock gate construction proposal was accentuated five years ago because the lock gate could be used to control the lake water level and ensure the elevation of the lake 176

water level and lake system low water elevation increase during draught periods. Should that not be the case and should the extremely draughty hydrological conditions and sea water level increase continue, the consequences of once performed structural intervention by which the cultivable area around Vrana lake was enlarged would be incalculable in terms of the lake ecosystem. 0,9

7000

0,8

6000

Chloride concentration (mg/l) Precipitation ( * 10 mm)

Lake level (masl)

0,7 5000

0,6 4000

0,5

3000

0,4

Precipitation

Chlorides - I 0,3

2000

0,2 1000

Chlorides - II

0,1 Chlorides - III 0,0

0

7 7 07 07 -07 07 07 -07 -07 07 -07 -07 -08 -08 -08 08 -08 08 08 -08 -08 08 -08 -08 r ttrry y c v c v n n-0 b-0 rb n- Jul- ug ep n- Jul- ug ep Oc No De Ja Fe Ma Ap Ma Ju Ja Fe Ma Ap Ma Ju Oc No De A S S A

Lake level

Figure 6 Overview of monthly precipitation volume, water level and chloride content 4 Conclusion The paper has determined that the global processes in the coastal area which are characteristic for the aquatic appearances in the whole Mediterranean have been recently intensively manifested in the area of Lake Vrana and its karst aquifer. Global climate change / climate variations in the form of unfavourable precipitation trend and rise in sea water level, as well as anthropogenic interventions in Vrana Lake catchment area, influenced the lake water level variations and outflows from the lake, causing the occurrence of extremely salty water inflows into the lake system during 2007 and 2008. It is established that salinization occurs both immediately, through the Prosika drainage canal as well as through two sets of springs located in the south-eastern part of the lake round Prosika canal and in Vrana field. The risk of lake salinization is totally doubtless under given conditions. Therefore, the recommendation is to introduce a lake outflow regulation by lock gate construction which would ensure the rise in minimum lake water levels, in order to stabilize the lake system and preserve the existing ecosystem of the Vrana Lake Nature park. Acknowledgments The authors would like to thank the Nature Park Vrana Lake for their support and cooperation. References Benblidia M, Margat J, Vallee D, Glass B (1996) Water in the Mediterranean Region. Blue Plan for the Mediterranean. Regional Activity Centre, Sophia-Antipolis, France, 91 pp Berakovic M (1983) Proucavanje rezima voda Vranskog jezera. In: Proc. Jugoslavenski simpozij o inzinjerskoj hidrologiji, Vol I, Split, Croatia, November 1983, Gradevinski institut, Zagreb, pp 272-284 Bolle HJ (2003) Mediteran Climate – Variability and Trends. Springer Verlag, Berlin Custodio E (2002) Aquifer Overexploitation: What Does lt Mean? Hydrogeology Journal 10: 254-277 Custodio E, Bruggeman GE (1982) Groundwater Problems in Coastal Areas. Studies and Reports in Hydrology 45. UNESCO, Paris, 650 pp 177

Fritz F (1984) Postanak i starost Vranskog jezera kod Biograda na moru. Geoloski Vjesnik 37:231-243 Gajic-Capka M, Zaninovic K (2006) Long-Terms Trends in Temperature, Precipitation and Runoff of the Croatian Eastern Adriatic Coast. Proc. Balwois, Ohrid, 23.-26.5.2006 Geotehnicki Fakultet Varazdin (2009) Ocjena stanja i rizika cjelina podzemnih voda na krskom podrucju u Republici Hrvatskoj, Varazdin, unpublished Horvat B, Rubinic J (2006) Annual runoff estimate - an example of karstic aquifers in the transboundary region of Croatia and Slovenia. Hydrological Sciences Journal 51(2):314324 Hrvatska Vodoprivreda (1994) Vransko jezero - prethodna studija utjecaja na okolis akumuliranja vode Vranskog jezera, Zagreb, unpublished Katalinic A, Rubinic J, Buselic G (2007) Hydrology of two coastal karst cryptodepressions in Croatia: Vrana lake vs Vrana lake. Proc. of the 12th World Lake Conference Taal 2007, Jaipur, Ministry of Environment & Forests Government of India, pp 732-743 Lambeck K, Antonioli F, Purcell A, Silenzi S (2004) Sea-level change along Italian coast for the past 10 000 yr. Quaternary Science Reviews 23:1567-1598 Lambeck K, Purcell A (2005) Sea-level change in the Mediterranean Sea since the LGM: model predictions for tectonically stable areas. Quaternary Science Reviews 24:19691988 Margat J (2004) Mediteranean Basin Water Atlas. UNESCO, Paris, pp 46 Orlic M (1995) Vodostaj Jadranskog mora i klima. Proc: Zbornik radova 1. Hrvatske konferencije o vodama, Dubrovnik 24.-27.05.1995, Hrvatske Vode, Zagreb, pp 553-559 Pintur G (2003) Prijedlog projekta za unaprjedenje zastite i ocuvanja bioraznolikosti na sirem podrucju Parka prirode Vransko jezero. Proc. Round table Hydrological stabilization and conservation of biodiversity of the Vransko jezero Nature Park catchment area. Public Institution Nature Park Vransko jezero, Biograd, Croatia, pp 3-7 Pirazzoli PA (2000) Sea level changes – the last 20 000 years. John Wiley & Sons, Chichester, pp 211 Pirazzoli PA (2005) A review of possible eustatic, isostatic and tectonic contribution in eight late-Holocene relative sea level histories from the Mediteran area. Quaternary Science Reviews 24:1989-2001 Romic D, Tomic F (1997) Znacajke voda Vranskog jezera u Dalmaciji kao kriterija procjene pogodnosti za natapanje. Prirucnik za hidrotehnicke melioracije II/6, Gradevinski fakultet Rijeka, Rijeka, pp 243-258 Rubinic J, Cuze M (2009) Problem zaslanjenja sustava Vranskog jezera u Dalmaciji – stanje 2008.g. i ocjena mogucnosti osiguranja inicijalnog rjesenja zastite. Public Institution Nature Park Vransko jezero, unpublished Svensson C, Kundzewich ZW, Maurer T (2004) Trends in Flood ond Low Water Hydrological Time Series. WCASP 66. UNESCO, Paris & WMO, Geneve Svonja M (2003) Hidrologija Vranskog jezera. Proc. Round table Hydrological stabilization and conservation of biodiversity of the Vransko jezero Nature Park catchment area. Public Institution Nature Park Vransko jezero, Biograd, Croatia, pp 14-17 Svonja M, Pavic I, Rubinic J (2003) Analiza kolebanja karakteristicnih prosjecnih protoka vodotoka Jadranskog sliva u Hrvatskoj. In: 3. Hrvatska konferencija o vodama, Hrvatske vode, Zagreb, 123-130 Vodotok (2008) Prostorna sistematizacija hidrografskih podataka za vodno podrucje Dalmatinskih slivova, Zagreb, unpublished

178

Protection of the Miljacka karst spring: an underground connection between the rivers Zrmanja and Krka Josip TERZI, Ante PAVII, Tamara MARKOVI, Jasmina LUKA REBERSKI Croatian Geological Survey, Sachsova 2, 10000 Zagreb, Croatia, e-mails: [email protected]; [email protected]; [email protected]; [email protected]

Abstract: Miljacka is a karst spring situated in the Krka River canyon, in Dalmatia, Croatia. Its catchment is mostly built of karstified carbonate rocks and it is a part of well known Dinaric karst. During the last few years a bulk research program was performed in the Miljacka spring catchment area. Achieved results comprise all previous conclusions, together with some new aspects. The main reason for this research was establishment of sanitary protection for this strategically important karst spring. Sanitary protection zones were proposed, but because of some new understandings of this complex karst system, they will have to be extended in the future with parts of the Krka River catchment area. Although the Miljacka spring is situated in the Krka River canyon, majority of its water flows from quite distant Zrmanja River, which was proven by dye tracer tests in the past. During the recent research, hydrochemical investigation of the Miljacka proven that it is only partially true, and some portion of the spring water originate from the Krka River as well. Therefore, significant parts of the terrain will have to be added to the Miljacka spring catchment area and to sanitary zones as well. The most of the investigated area is built of karstified carbonate rocks, of Eocene-Oligocene (Promina rock mass), Cretaceous, Jurassic and Triassic. Complexity of this karst system is especially presented with the main characteristics of the Zrmanja River, which usually dries out in the swallow-hole zones during the summer dry season. These swallow-hole zones are active throughout the year and water that infiltrates the karst underground in that area discharges mostly at the Miljacka spring. This “hanging” part of the Zrmanja River was proven with boreholes in the past years, but that phenomenon should be investigated in detail. The main problem for any hydro-researcher of the Miljacka spring is shortage of the measured discharge data. This can be avoided in the future either by construction of the direct measuring place, or by measurements on the Krka River profile, but very close to the Miljacka (up and down stream), because the possibility of existence of swallow holes in river bed is quite high. Keywords: karst spring, hydrogeology, sanitary protection zones, Dalmatia, Croatia

1 Introduction The investigated area is situated in northern Dalmatia. It is part of the Dinaric karst region, known for its karstified carbonate rocks, deep and irregular karstification, and the preferential groundwater-flow paths. Miljacka is a typical karst spring situated in the canyon of the Krka River, only a few meters away from the river itself. The canyon is 150 m deep and very steeply cut, present in a relatively undisturbed plateau consisting of Eocene-Oligocene “Promina” rock mass (Fig. 1). The rock mass consists of almost horizontal and relatively thick beds of conglomerates, limestones, and marly carbonates. Although these beds differ in their hydrogeological properties, the rock mass can be considered relatively permeable because it is fractured and karstified. Although the Miljacka spring is in the Krka River’s canyon, it gets the major portion of its water from another river, Zrmanja. This has been proven with a tracing test (Fritz et al. 1986). In the past, the Miljacka spring catchment area has been researched many times, and the results are very different. Such diverse data indicate extremely complex relations in this karst system. Initially, the catchment was determined in 1967 by Komatina 179

(290 km2). A significant improvement was established after two tracer tests (Fritz et al. 1986, Kapelj and Fritz 1987), when Fritz and Pavii (1982, 1987, 1990) added huge parts of the Zrmanja River catchment, until the Miljacka spring catchment reached an area of 639 km2. In the last few years, detailed hydrogeological mapping and hydrogeochemical research were carried out (Terzi et al. 2008) with the purpose of establishment of the sanitary protection zones. As a result, this catchment was limited to an area of 516 km2, along with the remark that there is a partial influence from the Krka River too, which cannot be omitted.

Figure 1 Krka River canyon; at the right river-bank Miljacka spring and extraction site; at the left river-bank Miljacka power plant. Visible morphology – plateau, deep canyon, and carbonate rock mass with thick subhorizontal beds. 2 Regional hydrogeological setting The researched area belongs to the Dinaric karst region and it is mostly composed of fractured and karstified carbonate rocks. Because of the intensive tectonical dynamic of the area these rocks are karstified to a high extent (Herak et al. 1969, Pami et al. 1998). In the Miljacka spring catchment area there are clastic Triassic and Tertiary rocks, and the carbonates or Jurassic and Cretaceous prevail (Babi and Zupani 2007). In the catchment area there are practically no real barriers for the karst groundwater. Subhorizontal orientation of Eocene-Oligocene “Promina” rocks has a positive role considering Miljacka spring sanitary protection, because the water infiltrated through many diverse beds (over 100m thick all together) in close hinterland would not have as high impact on the spring water. Tectonical features in close hinterland make main preferential paths for the groundwater flow. There is synclinal structure through which groundwater preferentially flows toward the springing site. Quaternary sediments cover only some karst poljes and do not have significant influence on the karst groundwater flow. All of these karstified rocks can be taken as permeable, and the content of limestones and dolomites indicates if the rock mass is more or less permeable in a particular area. Due to tectonics (fracturing) and karstification (dissolution), preferential groundwater flow zones are developed and most of the groundwater flow happens in these

180

paths. Still, whole rock mass is fractured and karstified and flow through the rocks out of preferential zones should not be neglected as well.

Figure 2 Map of the researched area with sanitary protection zones (with position map in detail) (Terzi et al. 2009). 1-groundwater divide; 2-sanitary protection zone boundary; 3Miljacka spring; I, II, III, IV – proposed sanitary protection zones; A, B – possible parts of catchment, further research needed; C – direction of Krka River and confluents impact, has to be added to sanitary protection in the future. Zrmanja River near the Prevjes area flows into karstified carbonate rock mass, where lots of swallow holes were documented in the past. During the dry seasons, discharge of the river is lower than the swallow holes capacity, and the river dries out. The “hanging” character of the Zrmanja River was additionally proven by the observations in P-4 borehole during the 1980s (Fritz et al. 1986; Fig. 3), where throughout the year karst groundwater level never reached level of the Zrmanja River. Dye tracer tests in the past determined main groundwater flow from these swallow holes in the Zrmanja River bed to the Miljacka spring (Fritz et al. 1986). Noted groundwater velocity during this text was 1.31 cm/s, and that puts whole that area in the 3rd sanitary protection zone according to Croatian legislation (Fig. 2). After this test it was taken as a fact that water from the Miljacka spring originates from 181

Zrmanja. In the last researches (Terzi et al. 2008), hydrochemical study that was performed had proven that it is only partially true, and there is a significant proportion of the spring water that originates from the nearby Krka River. Such complexity of this karst system emphasizes the need for more detailed research of this mixing phenomenon.

Figure 3 Situation and cross section of the Zrmanja River near location Prevjes presenting “hanging” character of the river (Fritz et al. 1986).

Figure 4 Schematized hydrogeological profile of the researched area. 1-“terra rossa” quaternary soil; 2-carbonate rocks, mostly limestones; 3-limestones, dolomitic breccias and conglomerates exchange; 4-carbonates, dolomites prevailing; 5-normal geological boundary; 6-transgressive geological boundary; 7-fault with relative movement; 8-fold axis; 9schematized groundwater level

182

During the recent research (Terzi et al. 2008) the studied catchment was determined on an area of 516 km2. Still, the influence of the Krka River was proven. In the next research phases that will have to follow, the entire catchment of the Krka River upstream from the Brljan retention should be added to the protected zone, together with Krka’s confluents Butišnica, Kri, and Kosovica. This part of the catchment was never before included in the Miljacka spring catchment area, so it was not embraced in the research program. The chemical composition of the water from the Brljan retention (Krka River) showed very high sulphate concentration. Sulphate occurs in the Kosovica River (the Krka’s confluent), which partially flows through an area consisting of sulphate rocks (gypsum and anhydrite). The water from the Zrmanja near the Mokro polje contains very small concentrations of sulphate. Mass balance and equilibrium calculations were carried out, and the results indicate that 66– 84 % of the Miljacka water originates from the Zrmanja River, and the rest of the water is derived from Krka River. Currently, water from the Miljacka spring is being used for the public water supply for nearby settlements. The total amount extracted is 130 L/s; however, there are plans for further extension and supply of water to a larger area (consumption about 400 L/s). This exposes one of the main problems faced by past, recent and future researchers: lack of reliable data on Miljacka discharge. In numerous unpublished technical reports, very different minimal discharges were noted, ranging from 200 L/s to a few m3/s. The value of a few hundred L/s could be taken as an order of magnitude. The configuration of the spring in the canyon near the Krka River does not allow direct measurement. Measurements in the Krka, upstream and downstream from the Miljacka, are also not completely reliable because typical karst conditions prevail. Possible springs or ponors (swallow holes) can occur in riverbeds. Circumstances are additionally complicated by human influence because near the Miljacka spring, at the other riverbank, there is a hydroelectric power plant that uses water from the upstream Brljan retention (Fig. 1).

3 Hydrochemical research Two sampling campaigns were undertaken in the study area in different hydrological conditions: in May and September 2008. Water samples were collected from three springs (Miljacka, Zrmanja and Krka), and from rivers Zrmanja in Mokro polje and Krka at the Brljan retention. On site were measured EC, T, pH and oxygen content of sampled waters. Collected samples were analyzed for major cations (Ca2+, Mg2+, Na+, K+) and anions (HCO3-, SO42-, Cl-) in order to distinguished the influences of rivers Zrmanja and Krka. 5

Ca2+(mmol/l) Mg2+(mmol/l) Na+(mmol/l) K+(mmol/l) HCO3-(mmol/l) Cl-(mmol/l) SO42-(mmol/l)

4 3 2

Zrmanja Mokro Polje-09-08

Zrmanja Mokro Polje-05-08

Krka Brljan-09-08

Krka Brljan-05-08

Zrmanja spring-09-08

Zrmanja spring-05-08

Krka spring-09-08

Krka spring-05-08

Miljacka spring-09-08

0

Miljacka spring-05-08

1

Figure 5 Chemical composition of sampled waters in mmol/l 183

According to the chemical composition the waters from springs Krka, Zrmanja, Miljacka and river Zrmanja in Mokro polje belong to the Ca-HCO3 hydrochemical type, but the sampled waters from Krka at the Brljan retention belong to Ca-HCO3SO42- hydrochemical type (Fig. 5). The major ion chemistry of the springs systems are mainly controlled by weathering of carbonate sediments (limestone) with minor contribution of CO2 from atmospheric and soil source. Except the waters from Krka at the Brljan retention, where ion chemistry of the river water is partially controlled by the Kosovica River (the Krka’s confluent), which flows through an area consisting of sulphate rocks (weathering of gypsum and anhydrite) and also, weathering of carbonate sediments (limestone) with minor contribution of CO2 from atmospheric and soil source. The pH of the analyzed waters varies from 7.28 (Miljacka) to 8.41 (Krka at the Brljan retention) (Fig. 6). In general, pH of sampled waters is slightly alkaline to alkaline, and depends on amount of dissolved CO2 which is controlled by decomposition of organic matter in water and temperature (Stumm and Morgan 1996). The temperature of sampled waters varies from 9.7 to 18.6 oC (Fig. 6). The highest temperature values are measured in September in river waters of Zrmanja and Krka (Fig. 6). The temperature of spring water is annual average temperature of the air of the recharge area. The EC values vary from 304 (Zrmanja in Mokro polje) to 673 PS/cm (Krka at the Brljan retention) (Fig. 7). Spring and river waters are saturated with oxygen. NETPATH (Plummer et al. 1994) an interactive computer program, was used to interpret net geochemical mass-balance reactions between an initial (Zrmanja in Mokro polje and Krka at the Brljan retention) and final (Miljacka spring) waters along a hydrologic flow path. Also to compute the mixing proportions of two initial waters that can account for the observed composition of final water. Mass balance, equilibrium and mixing calculations gave as results that from 66 % (in September) to 84 % (in May) of the Miljacka water originates from the Zrmanja, and the rest of the water is derived from Krka (Fig. 8). This was calculated using STATISTICA software. Further studies are required in future to examine this “mixing” phenomenon in a large number of groundwater samples. 20

8,6

18

8,4 8,2

16

8,0 14 7,8 12

7,6

Zrmanja Mokro Polje-09-08

Zrmanja Mokro Polje-05-08

Krka Brljan-09-08

Krka Brljan-05-08

Zrmanja spring-09-08

Zrmanja spring-05-08

Krka spring-09-08

7,2

Krka spring-05-08

8

Miljacka spring-09-08

7,4 Miljacka spring-05-08

10

Figure 6 Temperature and pH distribution in sampled waters

184

T (oC)(L) pH(R)

700 650 600 550 500 450 400 350

Zrmanja Mokro Polje-09-08

Zrmanja Mokro Polje-05-08

Krka Brljan-09-08

Krka Brljan-05-08

Zrmanja spring-09-08

Zrmanja spring-05-08

Krka spring-09-08

Krka spring-05-08

Miljacka spring-05-08

250

Miljacka spring-09-08

300

110 105 100 95 90 85 80 75 70 65 60

EC (PS/cm)(L) O2 (%)(R)

Figure 7 EC and oxygen content of sampled waters 3,0 2,8

Krka Brljan

2,6 Miljacka spring

Ca2+(mmol/l)

2,4 2,2 2,0 1,8

Zrmanja in Mokro polje and other springs

1,6 1,4 -0,2

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

1,6

1,8

2,0

2,2

SO42-(mmol/l)

Figure 8 Relationship between waters sampled in Miljacka spring and Zrmanja in Mokro polje and Krka at the Brljan. 4 Sanitary protection zones On the basis of previous and recently achieved knowledge, and according to current Croatian legislation, groundwater sanitary protection zones were proposed (Fig. 2). Numerous open questions point to a need for further research. Depending on the results thereof, sanitary protection of this strategically important karst spring will have to be adjusted. Nevertheless, the main conclusion of the presented data is the fact that groundwater that springs out at Miljacka is a mixture of Zrmanja and Krka River waters, with higher ratio of Zrmanja. This fact only corroborates known fact that water supply spring in karst areas should be protected simultaneously, and not one by one. Sanitary zones of the Miljacka spring in the north are also sanitary zones of water extraction sites in the Zrmanja River downstream. There are numerous similar examples and researches for the sanitary protection zones of all these water supply sites (springs, wells, rivers; even in the presented case study) should be done at the same time within the same research program. Otherwise there will be high level of mix-up,

185

and of conflict of opinions. All these circumstances could only cause that the decision makers will be forced to neglect every effort for the groundwater preservation.

Acknowledgements The authors wish to acknowledge their gratitude to colleagues from Šibensko-Kninska county, investors of the research presented, and especially to dr. Živana Lambaša Belak. We are also grateful to all researchers and authors of unpublished data and technical reports connected with the studied area. The presented paper was done within the “Basic hydrogeological map of Croatia” project, founded by the Croatian Ministry of science, education and sports (project number 181-1811096-3165). References Babi Lj, Zupani J (2007) Major events and stages in the sedimentary evolution of the paleogene Promina basin (Dinarides, Croatia). Natura Croatica 16(4):215–232 Fritz F, Pavii A (1982) Hydrogeologically “hanging” parts of the Zrmanja and Krka rivers. Proc. “VII jugoslavenskog simpozija o hidrogeologiji i inženjerskoj geologiji” Novi Sad, 115-121, Fritz F, Pavii A (1987) The Miljacka karst spring catchment area in Krka River valley. Proc. “IX jugoslavenskog simpozija o hidrogeologiji i inženjerskoj geologiji” Priština, 97-101 Fritz F, Pavii A (1990) The Miljacka spring. Part of the hydrogeological investigations for the sanitary protection zones determination. Technical report 11/90. Croatian Geological Survey, Zagreb, unpublished Fritz F, Reni A, Pavii A (1986) Groundwater tracing test in Zrmanja River swallow hole near Mokro polje. Technical report 23/86. Croatian Geological Survey, Zagreb, unpublished Herak M, Bahun S, Magdaleni A (1969) Pozitivni i negativni utjecaji na razvoj krša u Hrvatskoj (Positive and negative influences on the development of the Karst in Croatia). Krš Jugoslavije 6:45-78 Kapelj J, Fritz F (1987) Groundwater tracing test near the Zrmanja River in the Ervenik karst polje. Technical report 16/87. Croatian Geological Survey, Zagreb, unpublished Komatina M (1967) Hydrogeological properties of parts of Dalmatia and western Bosnia and Hercegovina. Technical report, Zavod geol. geof. istraž., Beograd, unpublished Pami J, Guši I, Jelaska V (1998) Geodynamic evolution of the Central Dinarides. Tectonophysics 297(1-4):251-268 Plummer LN, Preston EC, Parhurst DL (1994) An interactive code (NETPATH) for modelling net geochemical reactions along flow path. Version 2.0 USGS Water-Resources Investigation Report, 94-4169, Reston, Virginia Stumm W, Morgan JJ (1996) Aquatic chemistry. Chemical Equlibria and Rates in Natural Waters. John Wiley & Sons, New York Terzi J, Pavii A, Frangen T, Markovi T, Luka Reberski J, Doli M (2009) Groundwater resources sanitary protection zones. Two case studies from Croatian karst. 6th EUREGEO - European Congress on Regional Geoscientific Cartography and Information Systems - Proceedings / Bayerisches Landesamt fur Umwelt, Munich, pp 225-228 Terzi J, Pavii A, Markovi T (2008) Groundwater researches with the purpose of sanitary protection of the Miljacka spring. Technical report 56/08. Croatian Geological Survey, Zagreb, unpublished

186

Contact information

INTERNATIONAL HYDROLOGICAL PROGRAMME (IHP) UNESCO/Division of Water Sciences (SC/HYD) 1 rue Miollis 75732 Paris Cedex 15 France Tel: (+33) 1 45 68 40 01 Fax: (+33) 1 45 68 58 11 Email: [email protected] http://www.unesco.org/water/ihp