COBISS 1.01 WATER RESOURCES ANALYSIS OF THE RJEČINA KARST SPRING AND RIVER (DINARIC KARST) ANALIZA VODNIH VIROV KRAŠKEGA IZVIRA IN REKE RJEČINA (DINARSKI KRAS) Ognjen BONACCI1*, Maja OŠTRIC2 & Tanja ROJE-BONACCI2 Abstract UDC 551.444.3:556.3(497.561) Ognjen Bonacci, Maja Ostric & Tanja Roje-Bonacci: Water resources analysis of the Rjecina karst spring and river (Di-naric karst) The paper deals with complex hydrological/hydrogeological behaviour in the Rjecina karst spring and river basin located in the north-western part of the deep and developed Croatian Dinaric karst. The Rjecina Spring is one of the major karst springs in Croatia, used for water supply of Rijeka City and its surrounding area. Beside the use of Rjecina spring for water supply, the development of the Rijeka hydroelectric power plant (HEPP) also changed hydrological and hydrogeological regime of the whole catchment. In order to analyse the anthropogenic influences in the system, hydrological analysis of the Rjecina river and spring discharge was done, as well as the analysis of the available data of groundwater measurements. The analysis showed that, due to the increase of water caption and decrease of precipitation, the average annual spring discharge decreased for approximately 25% in the 1980-2016 period. Detailed analysis of groundwater measurements indicated aquifer behaviour and the need for additional measurements and catchment delineation. Key words: hydrology, hydrogeology, karst, Rjecina Spring, Rjecina River, Dinaric karst. Izvleček UDK 551.444.3:556.3(497.561) Ognjen Bonacci, Maja Ostric & Tanja Roje-Bonacci: Analiza vodnih virov kraškega izvira in reke Rječina (Dinarski kras) Članek obravnava zapletene hidrološke/hidrogeološke razmere v zaledju kraškega izvira in reke Riečine v severozahodnem delu globokega in dobro razvitega hrvaškega Dinarskega krasa. Izvir Rječina je eden največjih kraških izvirov na Hrvaškem, zajet je za oskrbo s pitno vodo za mesto Reka in njeno okolico. Poleg rabe izvira Rječine za oskrbo z vodo je razlog za spremenjen hidrološki in hidrogeološki režim celotnega zaledja tudi gradnja hidroelektrarne Reka (HEPP). Za analizo antropogenih vplivov na sistem smo zato naredili hidrološko analizo pretokov reke in izvira Rječina pa tudi analizo razpoložljivih podatkov merjenja podzemne vode. Analiza je pokazala, da se je zaradi povečanja zajemanja vode in zmanjšanja količine padavin povprečen letni pretok izvira v obdobju 1980-2016 zmanjšal za približno 25 %. Podrobna analiza meritev podzemne vode je nakazala značilnosti delovanja vodonosnika ter pokazala potrebo po dodatnih meritvah in določitvi zaledja izvira. Ključne besede: hidrologija, hidrogeologija, kras, izvir Rječina, reka Rječina, Dinarski kras. 1 Split University, Faculty of Civil Engineering, Architecture and Geodesy, 21000 Split, Matice hrvatske 15, Croatia, e-mail: obonacci@gradst.hr 2 Hrvatske vode, 51000 Rijeka, Dure Šporera 3, Croatia, e-mail: maja.ostric@voda.hr 3 Split University, Faculty of Civil Engineering, Architecture and Geodesy, 21000 Split, Matice hrvatske 15, Croatia, e-mail: bonacci@gradst.hr, * Corresponding author Received/Prejeto: 31.08.2017 ACTA CARSOLOGICA 47/2-3, 123-137, POSTOJNA 2018 OGNJEN BONACCI, MAJA OSTRIC & TANJA ROJE-BONACCI INTRODUCTION Karst covers a vast area worldwide with various formations and evolution. Rich water resources as well as other valuable resources can be found (e.g., Flori-dan aquifer, Dinaric karst aquifer system) in planetary karst areas. Complex, dynamic and vulnerable karst system is a part of the Earth's system, sensitive to climate, environmental changes and human influence. Karst water resources provide drinking water for nearly 25% of the world population (Ford & Williams 2007), while in the USA up to 40% of water supply is from karst water and in some European countries 50 % (Goldscheider 2005). There are many human activities that, intentionally or not, produce severe impacts on karst, often with irreparable damage (e.g., Calo & Parise 2006; Parise & Gunn 2007). Although karst aquifers have a huge importance to water supply in many countries, due to the high discharge variations and vulnerability to pollution they are sometimes difficult to use (Goldscheider 2005; Brkic et al. 2018). The highest lack of water occurs during summer period due to the low amount of precipitation and high evapotranspiration (Zupan Hajna et al. 2010). The most usual sources of contamination are urban settlements, industry, agriculture and traffic (Petric et al. 2011; Kogovsek 2011). Beside the use of water for water supply, it is also used for irrigation and flood control, but the main impact on the karst has the use of waters in hydropower stations (Zupan Hajna et al. 2010) In karst terrains, groundwater and surface water constitute a single dynamic system. Determination of the catchment area and boundaries in karst environment presents an extremely difficult and complex task, which very often remains unsolved. The interdependency between surface water and groundwater in karst occasionally or permanently changes in time and space. These changes are caused by natural processes (intensive and abundant precipitations, collapse of surface and underground karst features, breaking out of clay clogs formed in karst conduits, etc.) as well as anthropogenic interventions, which have been very intensive in the last hundred years or so and are generally uncontrolled. This paper presents a hydrological and hydrogeo-logical analysis of a complex karst Rjecina River and Spring catchment, that are located in deep and developed Dinaric karst. The Rjecina River basin is located in the north-western part of the Croatian Dinaric karst. This karst system is of crucial importance for the town of Ri-jeka and its broader surrounding. Rijeka is the principal seaport and the third largest city in Croatia, located in the Kvarner Bay of the Adriatic Sea with a population of 128,624 inhabitants, according to the population census from 2011. The metropolitan area has a population of more than 240,000. Tourism causes high seasonal water needs. The main objective of this paper was to perform hydrological and hydrogeological analysis of a complex karst Rjecina River and Spring catchment in order to indicate the impact of the water use of Rjecina Spring in water supply and use of water of Rjecina River for hydropower plant (HEPP). STUDY AREA GEOLOGY, HYDROGEOLOGY AND CLIMATOLOGY The study area is located in the north-western part of the Dinaric karst in Croatia between 45°40' and 45°15'N and 14°20' and 14°30'E. Fig. 1 shows a location map of the study area indicating the Rje čina River and Spring, Zvir Spring, Zvir 2 Intake gallery, few other springs in the catchment, Rijeka meteorological station, Marčelji rain gauging station, Rijeka hydroelectric power plant (HEPP), Valici Reservoir, planned Kukuljani Reservoir, locations of nine deep piezometers (P1, P2, P2B, P3, P4, P5, B1, B2, B3) and seven hydrological gauging stations. Piezometers P2 and P2B are located within 10 m of distance and due to the scale of map it is not possible to show them separately, but instead two labels are placed on location of one piezometer. The main sources for water supply of Rijeka and a wider area are Rjecina Spring and Zvir. Rjecina Spring is used during most part of the year due to the hypsometric position that enables gravitational water supply of the city but it dries out in a summer period. In that period, permanent spring Zvir and intake gallery Zvir II are used for water supply. Maximal capacity of Rjecina Spring as well as the Zvir Spring (installed capacity) is 2000 l/s. Intake gallery Zvir II was constructed close to the spring Zvir in the 1982 as a tunnel with 6 wells of 12 m depth. Zvir II with capacity of 600 l/s is used only occasionally, in very dry seasons, when permanent spring Zvir drops its capacity. HEPP system Rijeka was built on Rjecina river in 1968 with Valici reservoir (V=0.6 *106 m3) and installed capacity of 36.8 MW. 124 ACTA CARSOLOGICA 47/2-3 - 2018 WATER RESOURCES ANALYSIS OF THE RJECINA KARST SPRING AND RIVER (DINARIC KARST) Fig. 1: Study area indicatingRjecina River and Spring, Zvir Spring, Zvir 2 Spring, few other springs in the catchment and neighbouring area, Rijeka meteorological station, Marcelji rain gauging station, Rijeka hydroelectric power plant (HEPP), Valici Reservoir, planned Kukuljani Reservoir. Fig. 2 shows results of tracer tests, and four hy-drogeological units. This is a well-developed karst system with well-developed groundwater flow, where also surface flow occurs on low permeable deposits (Fig. 2). Within the studied area, the aquifer is composed of: (1) well permeable group of deposits (dark green in Fig 2.), mostly Upper Cretaceous (K2) and Palaeogene limestone (E12), (2) medium permeable rocks, mostly Lower Cretaceous (K12) limestone and dolomites, (3) very low permeable Palaeogene flysch (E23) that forms a barrier to groundwater flow and (4) Quaternary sediments (Q) of different permeability. Quaternary deposits are very important for understanding of the surface and ground-water system in the area and have a role in groundwater retention and flow. Grobnik Polje is a depression composed of Quaternary deposits, with numerous intermittent karst springs as well as ponors, occurring on its outskirts. Ponors mostly occur in the southern part of Grobnik Polje, while intermittent springs occur in its northern part (Fig. 2). The wider area of the Rjecina River valley is part of a dominant morphostructural unit, which strikes in the direction of the Rjecina river valley - Bakar Bay - Vinodol Valley (Velic & Vlahovic 2009). This geologic structure could be considered as a Palaeogene flysch syn- ACTA CARSOLOGICA 47/2-3 - 2018 125 OGNJEN BONACCI, MAJA OSTRIC & TANJA ROJE-BONACCI Fig. 2: Study area indicating results of tracer tests and four hydro-geological units: (1) well permeable carbonate rocks, (2) medium permeable carbonate rocks, (3) very low permeable flysch deposits, (4) Quaternary, low permeable deposits. (Simplified hydrogeological map and tracer tests results from data base of Croatian waters). cline limited by faults (Benac et al. 2011). The syncline indicated in the Fig. 2 is an important structure that has a huge impact on hydrogeology of the study area. The syncline is striking from Slovenia to the coastline of Novi Vinodolski (more to SE, outside of the Fig. 2), with the submerged part in the Bakar Bay. The flysch sediments generally have very low permeability and depending on their structural position, they represent a hydrogeologi-cal barrier to groundwater flow in karst areas. The spatial distribution of flysch deposits affects the development of hydrologic network as well as the occurrence of the springs in the area. A complex undiscovered underground system composed of conduits, cavities and caves is an inherent characteristic of the analysed environment. Groundwater and surface water are hydraulically connected through numerous karst forms which facilitate the exchange of water between the surface and subsurface as well as between neighbouring karst aquifers and catchments. In karst terrains, one of the basic data for characterization of hydrogeological characteristics of the aquifer, and determination of groundwater flow direction and recharge area is obtained by groundwater tracer tests. The results of numerous tracer tests, given in Fig. 2, point to 124 ACTA CARSOLOGICA 47/2-3 - 2018 WATER RESOURCES ANALYSIS OF THE RJECINA KARST SPRING AND RIVER (DINARIC KARST) Fig. 3: Mean annual rainfall measured at the Marcelji rain gauging station divided into two time series for periods: (1) 1961-1979 and (2) 1980-2016. Solid lines indicate trend: blue line for the first subpe-riod, red line for the second. extremely complex hydrogeological setting in this karst system. The largest number of tracer tests was carried out from intermittent or permanent ponors and some from the boreholes. In Fig. 2 it is possible to see the results of tracings from transboundary Croatian-Slovenian aquifers, too. As tracer tests indicate, a part of groundwater flow from the Grobnik Polje is directed to springs in Ri-jeka City (Zvir Spring and Zvir 2 Intake gallery) and the other part to the Bakar bay springs (Perilo, Dobra, Dobri-ca in Fig. 2). In the north-eastern coast of the Bakar Bay there are numerous coastal and submarine permanent and intermittent karst springs. This location represents the lowest placed contact between large karst aquifer and flysch lithogenetic complex that forms hydrogeological barrier (Benac et al. 2003). Karst system of the Rjecina River and Spring is characterized by many endorheic depressions. An endorheic basin (terminal basin or internal drainage system) is a closed drainage basin that retains water and prevents outflow to other external bodies of water (Fiorillo & Pag-nozzi 2014). Recharge of the analysed karst aquifer from these endorheic depressions occurs when rainfall and snowmelt infiltration passes the soil surface and percolates through the vadose zone. The infiltration can occur in both, concentrated and diffuse forms. The first systematic analysis of the stable isotope composition of the complex karst hydrological systems in the broader area of the Rjecina River catchment revealed that: (1) stable isotope composition of the spring water suggests the recharge is dominated by winter precipitation, (2) seasonal variations were not observed in the stable isotope composition of the precipitation, (3) the dual-porosity system is dominated by baseflow (a fissure-porous aquifer), (4) the hinterlands of the indi- vidual springs have different degrees of karstification, (5) stable isotope analyses of groundwater and precipitation suggests a meteoric origin of the groundwater, (6) the isotopic compositions of the baseflow and the rapid-flow components of springs within the Rjecina River catchment (especially Rjecina Spring) originate at higher elevations than the other springs located outside of the catchment (Mance et al. 2014). The terrain configuration, with mountains rising steeply just a few kilometres inland from the shores of the Adriatic Sea, provides some striking climatic and landscape contrasts within a small geographic area. The climate of the study area varies between the North Mediterranean (near the Adriatic Sea coast) and mountain (in the upper part, which reaches the altitude of 1000 m a.s.l.). Two stations: Marcelji rain gauging station and Ri-jeka meteorological station were chosen for the rainfall analysis due to the availability of data. Rainfall for the period 1961-2016 and 1948-2016 were available respectively. The annual rainfall measured at the Marcelji rain gauging station during the 1961-2016 period (Croatian Hydrological and Meteorological Service - DHMZ) varies in a very large range between 1198 mm (2015) and 2889 mm (2010). Fig. 3 shows mean annual rainfall measured at the Marcelji rain gauging station divided into two time series: (1) 1961-1979 and (2) 1980-2016. The average annual rainfall for the whole period (1961-2016) is 2090 mm. Two previously mentioned subseries were defined using rescaled adjusted partial sums (RAPS) method (Garbrecht & Fernandez 1994). A time series analysis can detect and quantify trends and fluctuations in records. The rescaled adjusted partial sums (RAPS) ACTA CARSOLOGICA 47/2-3 - 2018 125 OGNJEN BONACCI, MAJA OSTRIC & TANJA ROJE-BONACCI t (year) Fig. 4: Time series of the mean annual temperature for the Rijeka meteorological station in the 19482016 period. Solid lines indicate temperature trend: violet line linear trend and red line second-order polynomial trend. approach highlights small yet systematic changes over time that are often hidden in a standard time series plot by the comparatively large magnitude and variability of data itself (Garbrecht & Fernandez 1994). In the first subperiod average annual rainfall was 2262 mm, while in the second it dropped 260 mm (12%) to 2002 mm, which is statistically significant (t-test p<0.01). Areal rainfall distribution follows a well-developed topography. The average annual rainfall measured at the Rijeka meteorological station, which is located near the Adriatic Sea coast (Fig. 1), in the 1948-2015 period (DHMZ) was 1547 mm, varying between 1021 mm (2003) and 2105 mm (2010). The rainfall in Rjecina Spring basin area is abundant with a very high intensity that can reach the value of 100 mm h-1 (Gajic-Capka et al. 2014). Fig. 4 depicts a time series of mean annual tempera- ture for the Rijeka meteorological station in the 19482016 period (DHMZ). It includes linear and second-order polynomial trend lines. The average mean annual air temperature is 14.1°C. The minimum mean annual air temperature of 12.7°C was recorded in 1980 and the maximum of 15.8°C in 2014. The linear trend is statistically significant (p<0.01) with a linear correlation coefficient, r = 0.428. The second-order polynomial trend with a coefficient of nonlinear correlation, r = 0.747, is much higher. It provides a better fit for the analysed data than the linear one. Fig. 4 shows a strong increasing trend of mean annual air temperatures over approximately last 35 years (started in early 1980s), which could have a strong influence on the Rjecina Spring and River hydrological regime, especially in combination with annual rainfall decreases, which started in 1980 (Fig. 3). HYDROLOGY OF THE RJECINA SPRING AND RIVER THE RJECINA KARST SPRING HYDROLOGY Fig. 5 shows a photography of the Rjecina Spring. It should be noted that small concrete dam has been constructed at the spring exit. The top of this dam is at the altitude of 325.24 m a.s.l. Knezevic (1999) cited that the recharge area of Rjecina River covers 163.9 km2. This number represents only a rough estimation. Exact boundaries and catchment area are not yet defined. More recent studies estimate the catchment area to be 2-3 times larger (app. 500 km2) than previously presumed (Munda et al. 2009; Kuhta et al. 2014). In order to differentiate the catchment area of coastal springs in Rijeka City from the coastal springs in the Bakar Bay, several tracer tests from Grobnik Polje were performed in the last decade. Rjecina Spring catchment should be defined on the basis of detailed interdisciplinary investigations using continuous measurements of different climatologic, geologic, hydrogeologic and hydrologic parameters. At this moment, there is no enough data to fulfil this complex task, extremely important in order to 124 ACTA CARSOLOGICA 47/2-3 - 2018 WATER RESOURCES ANALYSIS OF THE RJECINA KARST SPRING AND RIVER (DINARIC KARST) protect these valuable karst water resources. It seems that the Rjecina River catchment changes in time and space depending on temporal hydrological conditions. Fig. 6 shows mean annual discharges measured at the Rjecina Spring hydrological station during the 1948-2016 period (missing 1960-1965, 2001) (DHMZ). The average annual discharge for the complete analysed period is 6.87 m3 s-1, ranging between 3.53 m3 s-1 (2011) and 11.26 m3 s-1 (1951). Using the RAPS method, the following two subseries with statistically significant average annual discharges are defined: (1) 1948-1979 (missing 1960-1965) and (2) 1980-2016 (missing 2001). The first subperiod shows 1.71 m3 s-1 (about 25%) higher average annual discharge than the second subperiod. Fig. 7 shows two linear regressions between mean annual discharges measured at the Rjecina Spring and annual precipitation measured at the Marcelji gauging station during two subperiods: (1) 1966-1979 and (2) 1980-2016 (2001 missing) (DHMZ). High statistically significant values of the coefficients of linear correlation (r1 = 0.765 and r2 = 0.848) indicate that the mean annual discharges substantially depend on annual rainfalls in both subperiods. In the first subperiod (1966-1979) the same annual precipitation causes higher Rjecina Spring annual discharges than during the second subperiod (1980-2016). For the lower precipitation, these differences are larger than for the higher precipitation. This is shown in Fig. 7: (1) for P = 1500 mm, AQ1 = 1.96 m3 s-1; (2) for P = 2500 mm, AQ2 = 0.964 m3 s-1. Occurrence of two relationships between P and Q, in two subperiods is caused by natural and anthropogenic factors. Natural factors are increase of air temperature (Fig. 4) and drop in annual precipitations (Fig. 3) during the second subperiod. Anthropogenic influence is caused by the construction of the new pipeline (in 1980) and increased water caption from the Rjecina Spring. The main problem is the lack of data that would enable quantification of the impact of both, natural climatic cause as well as human activities on the process. During the 1948-2016 period (missing 1960-1965 and 2001) (DHMZ) values of the maximum annual discharges range between 31.7 m3 s-1 (2007) and 62.9 m3 s-1 (2016) with average of 43.8 m3 s-1. Although intensive short-time precipitations are frequent in the analysed region maximum discharges emerging from the spring are not extremely high. This karst spring belongs to the springs with limited outflow capacity. There are many factors that limit maximum discharges of karst springs, like the size of the karst conduit, intercatchment overflow, pressure flow and occurrence of intermittent springs in the same catchment (Bonacci 2001; Barbera & Andreo 2015). According to the available data (2003-2016, Croatian waters monitoring data base) the minimum measured water temperature at the Rjecina spring was 6.6°C, Fig. 5: The dam constructed at the exit of the Rijecina spring cave (Photo: VIKRijeka). a maximum of 10°C with an average value of 7.8°C. These values clearly indicate that water retains within a deep karst aquifer relatively long and the air temperature in the spring catchment area, that varies in a wide range, has no significant effect on the temperature. It should be noted that spring can dry up practically any day during the year. Minimum discharges occur during the hot summer period when the demands for the water are the highest. The average value of the days in a year when Rjecina Spring dried up in the 1948-2016 period (missing 1960-1965 and 2001) was 44.6 days, and it ranges between 0 (1948, 1968, 1977, 2014) and 157 days (1949) (Bonacci et al. 2017). When the Rijecina spring dries out, water from the Zvir Spring is used for the water supply. Besides the Rjecina Spring, that covers approx. 75% of water needs of the Rijeka City and the surrounding area, there are several other springs and wells used for water supply. During the 1997-2015 period, between 13.4 x 106 m3 (2007) and 20.9 x 106 m3 (2014) of water of high quality was used yearly from the Rjecina Spring, with an average of 17.7 x 106 m3, that is app. 560 l/s (Bonacci et al. 2017). RJECINA RIVER HYDROLOGY There are seven hydrological stations at the relatively short Rjecina River watercourse (18.5 km). Despite that fact, it is not possible to make detailed and reliable hy-drological analyses, due to the gaps in measurements and end of operation of two stations. The river hydrological regime downstream of the Valid Dam (river km 11.9) ACTA CARSOLOGICA 47/2-3 - 2018 125 OGNJEN BONACCI, MAJA OSTRIC & TANJA ROJE-BONACCI t (year) Fig. 6: Mean annual discharges measured at the hydrological station Rjecina Spring divided into two time series for periods: (1) 1948-1979 (missing 1960-1965) and (2) 1980-2016 (missing 2001). Solid lines indicate trend: blue line for the first subperiod, red line for the second. Fig. 7: Linear regressions between mean annual discharges, Q (m3s-1), measured at Rjecina Spring and annual precipitation, P (mm), measured at the Marcelji gauging station during two subperiods: (1) 1966-1979 and (2) 1980-2016 (missing 2001). n fp.ipr.iwû RPRiNr 3)-1 3 /MAPr.Pi .in 1966-1979 • fi = 0.765 \ • • p AQ = 1500 » 1.96 mm \ « • • • • * ■ • * ^—" • « } ii 1 i \ if 10 MB n na S"1 1969-2016 1 i |t 2OOÛJOOI; 200?; M12-2Û14 f «a vi 1 1 I- Ï 194 7-1 Ï6 8 < \l 4 1 i a »1 = 9.1 2 m S" \l t J - i i i i 1 i rf- i - Ï • 7 r 4 t- I 1 * 4 >< r > « * 1 CiBOMfrOOJOM^OfflOM^PiOSOhiPffiOJOMS^OlOMSOOOhl^ai t (year) Fig. 10: Mean annual discharges measured at the Grohovo hydro-logic station divided into two time series for periods (1) 1947-1968 and (2) 1968-2016 (missing 1976-1979, 1995-1997, 2000-2001, 2007, 20122014). ACTA CARSOLOGICA 47/2-3 - 2018 125 OGNJEN BONACCI, MAJA OSTRIC & TANJA ROJE-BONACCI Tab. 1: Characteristics ofhydrological gauging stations along the Rjecina River in four different subperiods. Station name Rjecina Spring Kukuljani Zoretici Martinovo selo Drastin Grohovo Susak Elevation (m a.s.l.) 325.214 288.720 284.240 273.046 234.761 194.315 2.069 Start of operation 1.1.1948 17.1.1946 1.6.1987 24.9.1964 1.6.1986 18.10.1946 1.6.1948 End of operation - 31.12. 1975 - - - - 4.12.1966 Missing year 1960-1965, 2001 - - 2000, 20072014 2010, 2012, 2014 1976-1979, 1995-1997, 20002001,2007, 20122014. - Q (m3 s-1) 19491958 7.41 7.88 - - - 8.36 11.9 19661968 7.9 7.9 - 8.43 - 9.12 13.7 19691973 7.63 8.32 - 7.64 - 2.33 - 19901999, 20022005, 2015-2016 5.79 - 6.40 6.62 7.97 *1.42 - *missing 1995-1996 GROUNDWATER IN THE CATCHMENT AREA Piezometers represent an exceptionally important source of wide-range information, necessary for all types of investigations related to the regime of water circulation in karst. In the studied area eleven deep piezometer boreholes were drilled, in the period of app. 30 years, in order to provide better understanding of groundwater behaviour and the hydrogeology of the area. Fig. 1 shows locations of 9 analysed (P1, P2, P2B, P3, P4, P5; B1, B2, B3) piezometers. Groundwater monitoring was performed in three periods: (1) 1974-1986, (2) 1994-1996 and (3) 2012-2013. We analysed only the last two periods, because only these data (Biondic et al. 1997, Kuhta et al. 2014) were available to the authors. In the first period (1994-1996), portable meters were used for measurements of water levels every three days, and for measurements of temperature and conductivity once in a month. In the second period, water level data loggers (HOBO Water Level Logger, Onset computers) were installed in the boreholes and continuous water level and temperature measuring was carried out in two hour intervals. Tab. 2 provides characteristics of seven piezometers measured during the second period (1994-1996) (Biondic et al. 1997). Tab. 3 is a matrix of the coefficient of linear correlation between groundwater levels measured at seven piezometers during the same period (1994-1996). From the data given in both tables, it is possible to conclude that groundwater in different piezometers reacts individually, yet homogeneously in the whole analysed system. For example, in P1 the groundwater level ranges only 27.55 m, while in piezometer P2 it ranges 122.35 m, which is app. 4.5 times higher. At the same time, the coefficient of linear correlation between groundwater levels of these two piezometers is statistically significant, r = 0.762. Extremely different groundwater reaction in piezometers, drilled within a small distance, less than 10 m away from each other, is common in karst environment (Drogue 1980, 1985; Bonacci & Roje-Bonacci 2012). Piezometers connected to active karst conduits react more rapidly than piezometers drilled in a karst matrix. Tab. 4 provides characteristics of five piezometers measured during the third period (2012-2013) (Kuhta et al. 2014). Fig. 11 shows time series of the groundwater level measured at those five piezometers and discharges of the Rjecina Spring (modified from Kuhta et al. 2014). It should be mentioned, that piezometer P2B was drilled in 2005, just 10 m from the existing piezometer P2 that was measured in the second period. New piezometer P2B was drilled due to the intensive dynamics of groundwa-ter level measured in P2 in the second period. Most of the existing piezometers were not fully penetrable as they were in the first and second period, so water level data loggers could not be installed at a proper depth. This is the reason why the lowest groundwater levels could not be measured (B1, B3 in Fig. 10, Tab. 4). 124 ACTA CARSOLOGICA 47/2-3 - 2018 WATER RESOURCES ANALYSIS OF THE RJECINA KARST SPRING AND RIVER (DINARIC KARST) Tab. 2: Characteristics of piezometers measured during the second period (1994-1995). P1 P2 P3 P4 P5 B1 B2 Elevation of the borehole mouth (m a.s.l.) 580 524 518 550 505 332 295 Start of operation 15.08.1994 18.12.1994 11.03.1995 01.05.1995 21.07.1995 15.08.1994 15.8.1994 End of operation 30.12.1996 30.12.1996 30.12.1996 30.12.1996 29.12.1996 02.12.1996 30.11.1996 H (m a.s.l.) 469.20 352.75 358.44 404.90 410.44 241.01 232.83 H (m a.s.l.) 496.75 475.10 499.90 503.20 442.80 323.00 295.00 H -H . (m) 27.55 122.35 141.46 98.3 32.36 81.99 62.17 r P1 P2 P3 P4 P5 B1 B2 P1 1 0.762 0.584 0.618 0.811 0.450 0.660 P2 1 0.862 0.685 0.536 0.629 0.711 P3 1 0.825 0.584 0.613 0.615 P4 1 0.715 0.434 0.392 P5 1 0.699 0.600 B1 1 0.797 B2 1 Tab. 3: Matrix of coefficient of linear correlation between groundwater levels measured during the second period (1994-1996). The groundwater in the piezometer P1 above Rjecina Spring (325.24 m a.s.l.) is active in all hydro-logical conditions, even during the summer period when Rjecina Spring dried out (middle of July - middle of September 2013). This indicates that P1 does not measure or represents the actual groundwater level in the aquifer, but local changes in groundwater due to the local influence. Groundwater levels in piezometer P2B act similar as in piezometer P1. It is visible that the level of P2B started a gradual decrease from June 2013 with a sudden drop of app. 40 m in August followed with the sudden increase after an insignificant rainfall in the end of August. This sudden increase was not followed by the increase of Rjecina spring discharge. This leads to a conclusion that P2B is drilled in rocks of lower permeability. Comparing the data from Tab. 2 and 4, it is obvious that groundwater level oscillations are different for P2B and P2, although they are located within a 10 m distance. All other piezometers show similar range of groundwater level measured in those two different periods. Tab. 5 is a matrix of the coefficient of linear correlation between groundwater levels measured at five piezometers during the third peri- Tab. 4: Characteristics of piezometers measured during the third period (2012-2013). P1 P2B** B1 B2 B3 Elevation of the borehole mouth (m a.s.l.) 580 523 325 295 320 Start of operation 02.10.2012 02.10.2012 01.10.2012 01.10.2012 01.10.2012 End of operation 15.10.2013 15.10.2013 17.4.2013 15.10.2013 15.10.2013 Water temperature (°C) 9.18 9.28-9.77 8.38-8.78 7.98.-8.48 7.28-7.78 Diver depth (m a.s.l.) 474.12 327.70 250.24 219.16 265.73 H (m a.s.l.) 469.31 371.04 * 232.75 * Hmax (m a.s.l.) 502.34 432.98 316.27 294.85 305.10 H - H . (m) 33.03 61.94 - 62.10 - * groundwater level below the measuring instrument; **piezometer drilled in 2005 Tab. 5: Matrix of coefficient of linear correlation between groundwater levels measured during the third period (2012-2013). r P1 P2B B1 B2 B3 P1 1 0.539 0.611 0.682 0.574 P2B 1 0.484 0.548 0.462 B1 1 0.949 0.982 B2 1 0.858 B3 1 ACTA CARSOLOGICA 47/2-3 - 2018 125 OGNJEN BONACCI, MAJA OSTRIC & TANJA ROJE-BONACCI Fig. 11: Groundwater level measured at five deep piezometers located in the Rjecina River catchment in the third period (2012-2013) and discharges of the Rjecina Spring (modified from Kuhta et al. 2014). Fig. 12: Graphical presentation of the groundwater temperature measured monthly in two piezometers (P2 and B1 in Fig. 1) (modified from Biondic et al. 1997). od (2012-2014). The conclusion is practically the same as in the previous case for measurements during the second period (1994-1996). Because of steep terrain configuration, groundwater has intensive flowing gradient. Due to highly developed surface and underground karst forms the catchment area recharges and stores deep circulating water with high storage capacity. Fig. 12 represents the groundwater temperature in piezometers P2 and B1 (Biondic et al. 1997) measured monthly in the period 1994-1996. The temperature in piezometer P2, located at a higher altitude than Rjecina Spring, has a constant decrease downwards approaching to the spring water temperature. The piezometer B1 located in the Grobnik Polje at a lower altitude than Rječina Spring showed different behaviour. The groundwater temperature constantly increases. Using thermal signatures, Doucette & Peterson (2014) delineate flow components in the karst aquifer of the springs Copperhead and Langle (Northwest Arkansas, USA). They revealed three distinct reservoirs (epikarst, shallow groundwater and deep groundwa-ter), which have different thermal relationships with the air temperature. During a more humid period, the epikarst water temperatures follow the air temperature trend more closely. During drier conditions, the shallow groundwater temperatures are more similar to air temperature. Deep groundwater temperatures show no relationship to variations in surface air temperature. In our case (see Tab. 4) the temperatures of groundwater 124 ACTA CARSOLOGICA 47/2-3 - 2018 WATER RESOURCES ANALYSIS OF THE RJECINA KARST SPRING AND RIVER (DINARIC KARST) measured on the same depth (installed loggers for measuring water levels) showed range within 0.5°C during period of measurement (October 2012- October 2013) with the exception of P1 where constant temperature was measured during the period. Data in Tab. 4 indicate deep groundwater temperatures with no variations with surface air temperature. Values of electrical conductivity measured in piezometer P2 change in narrow range between 270 and 287 ^S cm-1, and have decreasing trend from the surface to the deep underground (Biondic et al. 1997). Hundred meters below the surface, electrical con- ductivity is around 17 ^S cm-1 (or around 6%) lower than near the surface. Decrease of electrical conductivity with depth indicates lower salt content, that is faster flow. The higher the water hardness, the greater its electrical conductivity is. Bakalowicz & Mangin (1980) measured the electrical conductivity of water of several karst springs recharged from the deep karst aquifer. The values ranged between 220 and 470 ^S cm-1. Electrical conductivity of water emerging from Jadro Spring (Croatia) ranged between 250 and 350 ^S cm-1, and of water in the Blue Lake (Croatia) between 301 and 438 ^S cm-1 (Bonacci et al. 2013). CONCLUSIONS The analysis of climatic data indicates an increasing trend of air temperature and a decreasing trend of rainfall in the analysed region. The reduction of rainfall, the temperature and consequently evapotranspiration increase and the future anthropogenic interventions will strongly and dangerously affect Rječina Spring and the Rječina River water availability. The need for a better understanding of the impact of the development of the Rijeka HEPP and new pressures is of crucial importance. Despite great scientific efforts and fast technological development, scientist and engineers still have not been able to reliably foresee mostly dangerous consequences of very different human activities in karst terrains. Some of them occurred instantaneously but some emerge after long time (years or decades). Vulnerable karst ecosystems and environment are more prone to those dangerous and difficultly predictable consequences. Analysing hydrological changes within the karst aquifer (Lurbach system, Austria) Mayaud et al. (2016) concluded that the observed changes are caused by changes within the karst system due to the modification of hydraulic conductivity and storage within the conduit network. They concluded that very probably the main reason is the plugging of the drainage conduits with sediments rather than by varying hydro-meteorological conditions. Process of plugging and clogging of karst conduits is a highly dynamic natural process, which cannot be controlled. Due to this fact scientist and engineers rarely consider it. Anthropogenic interventions can substantially influence the above mentioned natural process. Very probably these processes exist in the Rječina Spring and River karst aquifer, but problem is that our knowledge about them is insufficient. Groundwater extractions through three controlled locations (Rječina Spring, Zvir Spring and Zvir 2 Intake gallery) of 23.8 x 106 m3 per year (average in the 19972015 period, that is app. 755 l/s average daily for the same period) from the karst aquifer can cause the groundwater level to fall. Because of an insufficient number of deep piezometers and short monitoring period it is not possible to precisely define the value and range of the groundwa-ter decrease. In order to ensure sustainable development of a broader basin area of the Rjecina River, its surface and the groundwater quantities and levels should be continuously monitored, in more river profiles and with many more deep piezometers than today. This would be of special importance due to the plans to construct new Ku-kuljani HEPP. Moreover, it is realistic to expect that the use of the groundwater will increase in the near future, especially during the tourist season. Assessment of the groundwater recharge potential zone and definition of the Rjecina Spring and River catchment areas are extremely important for the effective management of groundwater systems, sustainable development of environment and society. Because of the fact that the Rjecina River headwater is extremely important for the water supply and ecology of a larger densely populated region, this area of a relatively small extension should be more intensely studied and managed. Naturally, the variability of climate in mountain areas is high. It should be taken into consideration that in mountain areas the impact of climate change on water resources is very uncertain and variable (Buytaert & De Bievre 2013). Calo & Parise (2006) stress that "Determining the karst disturbance can be very difficult because of the inherent complexity of karst systems and subjective because it requires interpretation of the karst environment by the experts, depending upon their dif- ACTA CARSOLOGICA 47/2-3 - 2018 125 OGNJEN BONACCI, MAJA OSTRIC & TANJA ROJE-BONACCI ferent background." In complex karst surface and underground environment settings, the combined use of geophysical imaging, surface waters hydrological analysis, groundwater measurements (water level, water temperature, electrical conductivity etc.), geochemical measurements, and tracing techniques can provide insights into the local karst hydrology and groundwater processes. As the evidence suggests, it is realistic to expect that the occurrence of water shortage in the nearby future will be much more frequent and severe. The conflicts between water supplies and environment will be stronger. Better understanding of complex interrelations between surface water and groundwater, as well as the influence of designed anthropogenic actions, can help in negative consequences mitigation (Dragoni et al. 2015). REFERENCES Bakalowicz, M. & A. Mangin, 1980: L'aquifère karstique. Sa définition, ses caractéristiques et son identification.- Mémoires Hydrogéologique Série Société Géologique de France, 11, 71-79. Ban, D., 2013: Brana Valici za vrijeme velikog vodostaja: [Online] Available from: https://www.flickr.com/ photos/150439487@N03/33698666113/in/photo-stream/ [Accessed 3rd March 2018]. Barberá, J.A. & B. Andreo, 2015: Hydrogeological processes in a fluviokarstic area inferred from the analysis of natural hydrogeochemical tracers: The case study of eastern Serranía de Ronda (S Spain).- Journal of Hydrology, 523, 500-514. DOI: 10.1016/j.jhy-drol.2015.01.080 Benac, C., Rubinic J. & N. Ozanic, 2003: The origin and evolution of coastal and submarine springs in Bakar Bay.- Acta Carsologica, 32/1, 157-171. Benac, C., Dugonjic, S., Vivoda. M., Ostric, M. & 2. Arbanas, 2011: A complex landslide in the Rjecina Valley: results of monitoring 1998-2010.- Geologia Croatica, 64, 3, 239-249. DOI: 10.4154/GC.2011.20 Biondic, B., Dukaric, F., Kuhta, M. & R. Biondic, 1997: Hydrogeological exploitation of the Rjecina river spring in the Dinaric karst.- Geologia Croatica, 50, 2, 279-288. Bonacci, O., 2001: Analysis of the maximum discharge of karst springs.- Hydrogeology Journal, 9, 4, 328-338. DOI:10.1007/s100400100142 Bonacci, O., & T. Roje-Bonacci, 2012: Impact of grout curtains on karst groundwater behaviour: an example from the Dinaric karst.- Hydrological Processes, 26, 18, 2765-2772. DOI:10.1002/hyp.8359 Bonacci, O., Andric, I. & Y. Yamashiki, 2013: Hydrology of Blue Lake in the Dinaric karst.- Hydrological Processes, 28, 4, 1890-1898. DOI:10.1002/hyp.9736 Bonacci, O., Ostric, M. & T. Roje-Bonacci, 2017: Prilog hidrologiji krskog izvora Rjecine.- Hrvatske Vode, 25, 100, 99-108. Brkic, 2., Kuhta, M. & T. Hunjak, 2018: Groundwater flow mechanism in the well-developed karst aquifer system in the western Croatia: Insights from spring discharge and water isotopes.- Catena 161, 14-26. Buytaert, W. & B. De Bièvre, 2013: Assessing climate change impacts on water resources in remote mountain regions.- Geophysical Research Abstracts, 15, EGU2013-10412. Calo, F. & M. Parise, 2006: Evaluating the human disturbance to karst environments in southern Italy.- Acta Carsologica, 35/2, 47-56. Doucette, R. & E.W. Peterson, 2014: Identifying water sources in a karst aquifer using thermal signatures.-Environmental Earth Sciences, 72, 12, 5171-5182. D0I:10.1007/s12665-014-3387-2 Dragoni, W., Giontella, C., Melillo, M., Cambi, C., Di Matteo, L. & D. Valigi, 2015: Possible response of two water systems in Central Italy to climatic changes.- In: Moramarco, T., Barbetta, S. & L. Brocca, (eds) Advances in watershed hydrology.-Water Resources Publications, pp. 397-424, Highlands Ranch, USA Drogue, C., 1980: Essai d'identification d'un type de structure de magasins carbonates fissures.- Mémoires Hydrogéologique Série Société Géologique de France, 11, 101-108. Drogue, C., 1985: Geothermal gradients and ground water circulation in fissured and karstic rocks: the role played by the structure of the permeable network.-Journal of Geodynamics, 4, 219-231. Fiorillo, F. & M. Pagnozzi, 2014: Recharge processes of karst massifs: examples from Southern Italy.- In: Andreo, B., Carrasco, F., Duran, J.J., Jiménez, P. & J.W. LaMoreaux, (eds) Hydrogeological and environmental investigations in karst systems. Springer, pp. 103-111, Berlin, Heidelberg. Ford, D.C. & P.W. Williams, 2007: Karst hydrogeology and geomorphology.- John Wiley & Sons, Chiches-ter, 562 pp. 124 ACTA CARSOLOGICA 47/2-3 - 2018 WATER RESOURCES ANALYSIS OF THE RJECINA KARST SPRING AND RIVER (DINARIC KARST) Gajic-Čapka, M., Ožanic, N. & N. Krvavica, 2014: Estimation of maximum short-term precipitation over the Rijeka region.- E-journal Civil Engineering Faculty, Osijek University, 5, 9, 49-59. DOI:10.13167/2014.9.6 Garbrecht, J. & G.P. Fernandez, 1994: Visualization of trends and fluctuations in climatic records.- Water Resources Bulletin, 30, 2, 297-306. Goldscheider, N., 2005: Karst groundwater vulnerability mapping: application of a new method in the Swa-bian Alb, Germany.- Hydrogeology Journal, 13, 4, 555-565. Kneževic, R., 1999: Osnovne značajke režima porječja Rječine.- Acta Geographica Croatica ,34, 73-88. Kogovšek, J., 2011: Threats to the karst water sources from traffic in normal conditions.- In: Knez, M. et al. (eds), Karstology And Development Challenges On Karst I - Water. Karst Research Institute ZRC SAZU, Postojna, pp. 47-64. Kuhta, M., Brkic, Ž. & L. Kuhta, 2014: Izvor Rječine i sjeverozapadni rub Grobničkog polja: hidrogeološki radovi-hidrokemijska i izotopna istraživanja. Croatian Geological Survey, Unpublished study, Zagreb. Report number:010/14. Mance, D., Hunjak, T., Lenac, D., Rubinic, J. & Z. Roller-Lutz, 2014: Stable isotope analysis of the karst hydrological systems in the Bay of Kvarner (Croatia).- Applied Radiation and Isotopes, 90, 23-34. DOI:10.1016/j.apradiso.2014.03.001 Mayaud, C., Wagner, T., Benischke, R. & S. Birk, 2016: Understanding changes in the hydrological behaviour within a karst aquifer (Lurbach system, Austria).- Carbonates and Evaporites, 31, 4, 357-365. DOI:10.1007/s13146-013-0172-3 Munda, B., Trutin, M. & D. Pribeg, 2009: Novelacija granica zastitnih zona vodocrpilista rijeckog vodoop-skrbnog sustava.- GeoAqua, Unpublished study, Zagreb. Parise, M. & J. Gunn (eds.), 2007: Natural and Anthropogenic Hazards in Karst Areas: Recognition, Analysis and Mitigation. Special Publications Vol. 279. Geological Society, pp. 202, London. DOI: 10.1144/ SP279. Petric, M., Ravbar, N. & J. Kogovsek, 2011: Characteristics of karst aquifers, their vulnerability and endan-germent.- In: Knez, M. et al. (eds), Karstology And Development Challenges On Karst I - Water. Karst Research Institute ZRC SAZU, Postojna, pp. 7-19. Velic. I. & I. Vlahovic, 2009: Geological map of the Republic of Croatia 1:300.000. Croatian Geological Survey, Zagreb. Zupan Hajna, N., Mihevc, A. & M. Prelovsek, 2010: Land use.- In: Zupan Hajna, N. et al. (eds), Introduction To The Dinaric Karst. Karst Research Institute at ZRC SAZU, Postojna. ACTA CARSOLOGICA 47/2-3 - 2018 125