DETERMINING THE DIRECTIONS AND CHARACTERISTICS OF UNDERGROUND W ATER FLOW IN KARST FOR THE PURPOSE OF TRAFFIC ROUTES CONSTRUCTION: THE CASE OF THE NEW DIV AČA-KOPER RAILW AY LINE (SW SLOVENIA) DOLOČITEV SMERI IN ZNAČILNOSTI TOKA PODZEMNE VODE V KRASU ZA NAMEN IZGRADNJE PROMETNIH POTI: PRIMER NOVE ŽELEZNIŠKE PROGE DIV AČA-KOPER (JZ SLOVENIJA) Metka Petrič 1,2 *, Nataša Ravbar 1,2 , Luca Zini 3 , Chiara Calligaris 3 , Riccardo Corazzi 4 , Zdenka Žitko 5 , Marco Restaino 6 & Martin Knez 1,2 Abstract UDC 556.34(497.47) Metka Petrič, Nataša Ravbar, Luca Zini, Chiara Calligaris, Riccardo Corazzi, Zdenka Žitko, Marco Restaino & Martin Knez: Determining the directions and characteristics of un- derground water flow in karst for the purpose of traffic routes construction: the case of the new Divača-Koper railway line (SW Slovenia) The new railway line between Divača and Koper/Capodistria in south-western Slovenia is being built, a part of which crosses the southern outskirts of the Classical Karst plateaux. It will run through two tunnels, the northern tunnel T1 (6.7 km long) and the southern T2 (6 km long), which partially cross karst aquifer system. A multi-tracer test with injections of fluores- cent dyes uranine and naphthionate, bypassing the karst vadose zone, was carried out to define the directions and dynamics of the underground water flow. The main goals were better un- derstanding of the complex hydrogeological conditions in the area and assessment of possible environmental impacts on the nearby water sources. With tracing of uranine injected into a nearby cave stream, the direction of flow from the northern T1 tunnel mainly towards the Reka-Timavo aquifer system and further towards the Timava/Timavo springs was proved. The peak velocities, as determined from the peaks of the tracer breakthrough curves, range from 29 m/h to 36 m/h. Through Izvleček UDK 556.34(497.47) Metka Petrič, Nataša Ravbar, Luca Zini, Chiara Calligaris, Riccardo Corazzi, Zdenka Žitko, Marco Restaino & Martin Knez: Določitev smeri in značilnosti toka podzemne vode v krasu za namen izgradnje prometnih poti: primer nove želez- niške proge Divača‒Koper (JZ Slovenija) Med Divačo in Koprom v jugozahodni Sloveniji je v izgradnji nova železniška proga, ki deloma seka južno obrobje planote matičnega krasa. Proga ima dva predora, severni T1 (dolg 6,7 km) in južni T2 (dolg 6 km), ki potekata skozi kraški vodo- nosnik. Za določitev smeri in dinamike toka podzemne vode je bil izveden kombinirani sledilni poskus z injiciranjem fluo- rescenčnih sledil uranina in naftionata neposredno v zasičeno cono. Glavna cilja sta bila boljše razumevanje zapletenih hidro- geoloških razmer na tem območju in ocena možnih okoljskih vplivov na bližnje vodne vire. Z injiciranjem uranina v vodni tok v bližnji kraški jami je bila dokazana smer toka podzemne vode z območja severnega predora T1 predvsem proti vodo- nosnemu sistemu Reka-Timava in naprej proti izvirom Timave. Glede na pojav vrha krivulje iztekanja sledila so bile določene navidezne dominantne hitrosti med 29 m/h in 36 m/h. V večjih in dobro povezanih kraških kanalih sistema Reka-Timava lah- ko te hitrosti dosežejo 88 m/h. V Jami 1 v Kanjaducah, ki je del sistema, je bilo zaznanih približno 74 % injiciranega sledila. 1 ZRC SAZU, Karst Research Institute, Titov trg 2, 6230 Postojna, Slovenia; metka.petric@zrc-sazu.si, natasa.ravbar@zrc-sazu.si, martin.knez@zrc-sazu.si 2 UNESCO Chair on Karst Education, University of Nova Gorica, Glavni trg 8, 5271 Vipava, Slovenia 3 Department of Mathematics and Earth Sciences, University of Trieste, Via Weiss 2, 34128 Trieste, Italy; zini@units.it, calligar@ units.it 4 Caving club Commissione Grotte “E. Boegan” - Società Alpina delle Giulie CAI Trieste, via Donota 2, 34121 Trieste, Italy; corazzi@iname.com 5 Speleological Association of Slovenia, Lepi pot 6, 1000 Ljubljana, Slovenia; tajnistvo@jamarska-zveza.si 6 Caving club Società Adriatica di Speleologia, Via Rossetti 59A, 34141 Trieste, Italy; marco.restaino@yahoo.it * Corresponding author Received/Prejeto: 19.04.2020 DOI: 10.3986/ac.v49i1.8582 ACTA CARSOLOGICA 49/1, 63-80, POSTOJNA 2020 COBISS: 1.01 INTRODUCTION Karst aquifers contain large volumes of water, represent- ing water resources of worldwide importance (Stevanović 2019). Due to their specific characteristics they are par- ticularly vulnerable to various human activities on the surface. Traffic is one of the pollution sources that can jeopardize the quality of karst water. Water flowing off of roads and other paved surfaces can quickly percolate into the subsurface and further toward karst water sources (Hötzl 1999; Turpaud et al. 2018). Pollution can be con- stant due to continuous runoff or episodic (catastrophic) due to the spills of pollutants during traffic accidents. The latter represent an immediate hazard as spills can instant- ly release a large volume of pollutants (Kogovšek & Petrič 2011), whereas the former may release lower concentra- tions into the karst system but for a prolonged period of time (Kogovšek 2011). Moreover, when constructing surface or under- ground traffic structures in karst, several significant im- pacts have been encountered so far, such as water inrush in tunnels, or drying of springs and base-flow losses at mountain streams due to drilling, or engineering set- backs due to geomorphological and stability constraints (e.g., Alija et al. 2013; Li & Li 2014; Vincenzi et al. 2014; Zini et al. 2015; Jin et al. 2016; Raithel et al. 2016). As regards water-related issues, a karst-specific wa- ter inrush and runoff management plan should be pre- pared for each section of the traffic route before the con- struction takes place (Zhou & Beck 2005). Furthermore, the monitoring of any potential negative impacts of its construction and operation should be based on a com- prehensive analysis of the characteristics of groundwater flow (Gabrovšek et al. 2015). The standard methods of obtaining this information are numerical groundwater modelling, analytical and empirical approaches. These require diverse geological and hydrogeological input data, whose quality in karst mainly depends on their indirect acquisition and is prone to lower reliability due to high bedrock heterogeneity. Concurrently, these data are often affected by a high level of uncertainty, because they are generally scarce in quantity. Thus, the obtained results could be less reliable for predicting groundwater drainage patterns towards, along and from the proposed route. To most straightforwardly assess this information, tracer tests with artificial tracers are especially useful (Massei et al. 2006; Benischke et al. 2007; Goldscheider et al. 2008; Barberá et al. 2018). This study focuses on the planning and construction of a new railway line between Divača and Koper/Capo- distria (Slovenian/Italian name in the bilingual area) in south-western Slovenia, which will significantly improve the traffic connections between the Slovenian coast and the inland (Knez et al. 2019). Its section in the karst area, 12.7 km of which will run through tunnels, intersects the highly vulnerable trans-boundary karst aquifer of the Classical Karst, which is a strategic water resource for both countries: Italy and Slovenia (Urbanc et al. 2012). As the underground water network is highly branched and the recharge areas of a greater number of springs in- tersect there, several tracer tests have been conducted for this purpose. The article sums up the results of important previous tracer tests and details the results of the latest multi-tracer test, which was conducted in the autumn of 2018 with injection points inside the karst system. As the tracer test was conducted in the cross-border Severni del južnega predora T2 je del širšega bifurkacijskega območja med Osapsko reko in izviri v Boljuncu. V njem so bile z injiciranjem naftionata v eno izmed vrtin v trasi načrtova- ne železniške proge ugotovljene dominantne hitrosti toka med 10 in 13 m/h. Ocenjeno je bilo, da je približno 25 % injicirane- ga naftionata izteklo skozi izvir Osapske reke in približno 5 % skozi izvire v Boljuncu. Z upoštevanjem rezultatov raziskave bo mogoč ustrezen monitoring morebitnih negativnih vpli- vov nove železniške proge na podzemne vode. Študija prinaša izboljšan pristop k načrtovanju nadzora nad nevarnostmi pro- metnih poti na območjih zelo zakraselih kamnin s prepletan- jem podzemnih vodnih tokov. Ključne besede: kraška voda, železniška proga, sledilni poskus, zasičena cona, aktivni vodni tokovi, matični kras. the wider and well-connected conduits of the Reka-Timavo system, the peak velocities can reach up to 88 m/h. The recov- ery of uranine in an intermediate cave, i.e., Jama 1 v Kanjad- ucah, amounted to approximately 74 %. The northern section of the southern T2 tunnel is a part of a wider bifurcation zone between the Osapska Reka and the Boljunec/Bagnoli springs, where peak flow velocities between 10 and 13 m/h have been determined by tracing of naphthionate injected into a borehole located in the line of the planned tunnel. It has been estimated that about 25 % of the injected naphthionate flew out through the Osapska Reka spring and about 5 % through the Boljunec/ Bagnoli springs. Based on this research, proper monitoring of any potential negative impacts of the new railway line will be made possible. The study presents an approach to better plan- ning of hazard control of traffic routes in complex and highly karstified rock settings. Key words: karst water, railway route, tracer test, phreatic zone, active cave streams, Classical Karst. METKA PETRIČ, NATAŠA RAVBAR, LUCA ZINI, CHIARA CALLIGARIS, RICCARDO CORAZZI, ZDENKA ŽITKO, MARCO RESTAINO & MARTIN KNEZ ACTA CARSOLOGICA 49/1 – 2020 64 Slovenian-Italian region and its implementation was very demanding due to the great number of sampling points and a few hard-to-reach injection and sampling points, several organizations were involved in its planning and implementation: Karst Research Institute ZRC SAZU, Speleological Association of Slovenia, Department of Mathematics and Earth Sciences, University of Trieste, and the caving clubs Commissione Grotte “E. Boegan” and Società Adriatica di Speleologia from Trieste. MATERIALS AND METHODS STUDY AREA The karst area, through which the planned new Divača – Koper/Capodistria railway line will run, is characterized by an imbricate thrust structure, where bands of Eocene flysch (medium to low permeable layers on Fig. 1A) are interspersed among carbonate rocks – predominantly limestone of Upper Cretaceous, Palaeocene and Eocene age (well permeable layers on Fig. 1A). The surface karst features and the number of registered karst caves indi- cate well-karstified carbonate rocks and typical charac- teristics of karst aquifer systems. Flysch consists of very poorly permeable marls, silicate sandstones and breccias (Placer et al. 2010; Gabrovšek et al. 2015; Jurkovšek et al. 2016). The flysch layers between karst aquifers present lo- cal hydrogeological barriers, along which water may ac- cumulate and flow. The flysch rocks on the surface enable surface flows that sink into the underground at the con- tact with the karst. On the other hand, flysch also acts as a hydrological barrier which causes underground karst waters to discharge onto the surface in karst springs. The route of the new railway runs from Divača about 3 km across the karst surface and then enters the northern tunnel (T1), 6.7 km long, which ends in the up- per part of the valley of the Glinščica/Rosandra River. After a 250 m long bridge the railway enters the southern tunnel (T2), 6 km long. The last part of the second tunnel passes from the karst area to the viaduct and then runs almost entirely through tunnels across the flysch area to the Port of Koper/Capodistria. The route in karst mostly runs at a depth of about 200 m below the surface, in some sections it is more than 300 m deep. In case of very high water, the water table may rise above the planned railway line in some sections (Gabrovšek et al. 2015). The northern part of the route intersects the area of the Classical Karst with the underground flow of the Reka River, which sinks from the surface into the Škocjanske jame caves approximately 3.5 km east of the route and then outflows in the spring belt between Nabrežina/ Aurisina and Timava/Timavo spring (the Reka-Timavo conduit system). In the 1989-2018 period, the hydro- logical station Cerkvenikov mlin (roughly 7 km ahead of the ponor) measured the lowest flow rate of the Reka River at 0.25 m 3 /s, while the mean flow rate was 7.75 m 3 /s (ARSO 2020). During very high water levels, its flow rate can increase to over 300 m 3 /s. The Reka-Timavo conduit system can be reached in several deep caves on the Slove- nian and Italian side, including the caves Jama 1 v Kanja- ducah and Labodnica/Abisso di Trebiciano. The main springs of the area are the Timava/Timavo and Sardoč/Sardos springs. The first has a minimum dis- charge of 7.4 m 3 /s, a maximum of 158 m 3 /s and a mean of 29.3 m 3 /s (Gemiti 1995) and the latter a minimum dis- charge of 0.6 m 3 /s, a maximum of 5.6 m 3 /s and a mean of 1.9 m 3 /s (Gemiti 2004). At the Sardoč/Sardos spring the first water supply system in Italy was constructed in 1922 and at present it supplies about 250,000 inhab- itants with drinking water. The third spring in order of discharge magnitude is the Nabežina/Aurisina spring with a mean value of 0.3 m 3 /s. Between Timava/Timavo and Nabrežina/Aurisina springs, several small coastal springs arise, their average estimated discharge is 0.5-1 m 3 /s (Gemiti 1984a, 1995; Zini et al. 2014; Cucchi et al. 2015). The springs west of the Sardoč/Sardos spring are not influenced by the waters of the Reka-Timavo aquifer system (Calligaris et al. 2019). The southern part of the railway route in the karst area is located in the catchment area of the Rižana spring, which was already in use in the early 19 th cen- tury, while in 1935 a regional water supply system was constructed to supply Slovenia’s coastal municipalities with drinking water. Today a majority of the inhabitants of this area (86,000 permanent inhabitants, increasing to 120,000 during the tourist season) are connected to this water supply network (Fig. 2). Based on the con- ducted hydrogeological research and numerous tracer tests (Krivic et al. 1987, 1989) the catchment area of the spring was estimated at 247 km 2 . Most of it is karstic, however, the spring is also recharged by the sinking rivers at the southern edge of the flysch Brkini Hills. The location of the spring is linked to the contact of the carbonate aquifer with the very poorly permeable DETERMINING THE DIRECTIONS AND CHARACTERISTICS OF UNDERGROUND W ATER FLOW IN KARST FOR THE PURPOSE OF TRAFFIC ROUTES CONSTRUCTION: THE CASE OF THE NEW DIV AČA-KOPER RAILW AY LINE (SW SLOVENIA) ACTA CARSOLOGICA 49/1 – 2020 65 METKA PETRIČ, NATAŠA RAVBAR, LUCA ZINI, CHIARA CALLIGARIS, RICCARDO CORAZZI, ZDENKA ŽITKO, MARCO RESTAINO & MARTIN KNEZ Fig. 1: A) Hydrogeological map of the broader area of the planned route with the results of previous tracer tests; B) map of the recent tracer test results and C) location of sampling points in the Glinščica/Rosandra valley (for names of the springs see Tab. 3). ACTA CARSOLOGICA 49/1 – 2020 66 flysch rocks, across which the Rižana River flows into the Adriatic Sea. The flow rates of the Rižana River are between 70 L/s and 63.2 m 3 /s, while the mean flow rate is 3.5 m 3 /s (ARSO 2020). To the north of Rižana, several smaller springs are located at the contact between flysch and limestone. The intermittent spring of the Osapska Reka can reach the flow rate of several m 3 /s after intense precipitation. It is the overflow of the waters from the catchment area of the Rižana spring in times of high water conditions (Krivic et al. 1989). On the Italian side of the border, in the Glinščica/Rosandra valley at an altitude of around 60 m a.s.l., in the village of Boljunec/Bagnoli della Rosan- dra there are several springs known under the common name of Boljunec/Bagnoli (Fig. 1). The springs Na placu/ Abbeveratoio and Pri pralnici/Lavatoio (water supply of a fish farm) are permanent; occasionally, water also flows out of a karst cave Jama/Antro di Bagnoli as an overflow of the Pri pralnici/Lavatoio spring in times of high water conditions (Fig. 3). More to the north is the Glinščica/Rosandra River watershed. The river originates at the confluence of the Krvavi potok/Rio del Sangue and Grižnik/Grisa torrents and is recharged by several small springs with a discharge of few L/s. The Zroček 1 and Zroček 2 springs flow into the Krvavi potok/Rio del Sangue, while Klinšca/Oppia 1 and Klinšca/Oppia 2 feed the downstream Glinščica/ Rosandra River (Zini et al. 2011). In the area of Dolina/ San Dorligo village there are three smaller karst springs which used to be or still are used as local water sources (Na Kaluži, Moganjevac, Zgurenc). DETERMINING THE DIRECTIONS AND CHARACTERISTICS OF UNDERGROUND W ATER FLOW IN KARST FOR THE PURPOSE OF TRAFFIC ROUTES CONSTRUCTION: THE CASE OF THE NEW DIV AČA-KOPER RAILW AY LINE (SW SLOVENIA) Fig. 2: The Rižana spring is an extremely important source for drinking water supply; A) the overflow of the spring at the reser- voir; B) the spring in the reservoir (Photos: J. Kogovšek, M. Petrič). Fig. 3: In the village of Boljunec/Bagnoli della Rosandra, the springs Na placu/Abbeveratoio (A) and Pri pralnici/Lavatoio (B) are permanent; there is also the intermittent spring of Jama/Antro di Bagnoli (C) (Photos: M. Petrič). ACTA CARSOLOGICA 49/1 – 2020 67 PREVIOUS TRACER TESTS AND REASONS FOR NEW RESEARCH The first tracings were conducted in the Classical Karst already at the beginning of the twentieth century and were followed by many others, which showed the con- nections of the Reka River from the Škocjanske jame caves to the Timava/Timavo springs and some other smaller springs in the Gulf of Trieste (Timeus 1928; Mo- setti 1965; Gemiti 1984b, 1998; Galli 2012; Peric 2012). The tracers were also recovered in various water caves inbetween (e.g., Jama 1 v Kanjaducah, Labodnica/Abis- so di Trebiciano; Fig. 1). The determined flow velocities (with regard to duration of the transfer of the tracer and the linear distance between the injection point and the point in which the tracer appeared), which depend on the hydrological conditions, amount from 47 to 204 m/h (Turpaud et al. 2018). Between 1985 and 1987 (Krivic et al. 1987, 1989) numerous tracer tests were conducted in the sinking streams at the foothills of the Brkini Hills with the in- tention of determining the recharge area of the Rižana spring; certain connections with the Osapska Reka spring were also established. Some indications of groundwater connections between the most western ponors and the springs of Timava/Timavo and Boljunec/Bagnoli were assessed as uncertain. The tracer tests described above were not enough to determine more precisely the directions and characteris- tics of the flow of water away from the area of the planned railway route. Therefore, three additional tracer tests were conducted in 2001, 2009 and 2010 with injections into a sinking stream to the Beka-Ocizla Cave system, into a fissure on the surface and into two boreholes (Kogovšek & Petrič 2004; Petrič & Kogovšek 2011; Gabrovšek et al. 2015). The results are summed up in Tab. 1. Summing up the findings of the tracings described in this chapter, it can be said with certainty that the sink- ing river Reka flows towards the springs in the Gulf of Trieste, especially towards the Timava/Timavo springs. The tracing conducted in the T1-8 borehole with 0.3 kg of amidorhodamine G, proved a secondary connection with the springs in Boljunec/Bagnoli, with a highly prob- able main direction towards the springs in the Gulf of Trieste. Confirming this direction would require the use of a much greater quantity of tracer for the tracing in the T1-8 borehole due to the great distance and higher flow rates of the Timava/Timavo spring; however, that was impossible under the given conditions because of the proximity of the Boljunec/Bagnoli spring and the danger of colouring it with too high concentrations of the tracer. Several tracings were conducted in the area of the T2 tunnel; in the northern part, the main direction of the underground flow towards the Boljunec/Bagnoli springs has been confirmed, and in the southern part the direc- tion towards the Rižana and Osapska Reka springs has been confirmed. The watershed between them could not be delineated more precisely without conducting an ad- ditional tracing in the intermediate area. That is why in the final phase of the research, before beginning the construction of the new railway line, it was decided that another multi-tracer test should be conduct- ed to further explain the directions and characteristics of the underground flow in both above-mentioned sections of the line. Because a great part of the railway route is running through the tunnels, injection points bypassing vadose zone have been selected in order to assess simi- lar flow conditions to possible contamination from con- struction or operation of tunnels. In the area of the T1 tunnel the karst cave Davorjevo brezno was chosen (Fig. 1). The location of the cave is roughly 3 km east of the tunnel, between the Škocjanske jame caves and the T1-8 borehole, which were already used as injection points in previous tracings. In order to avoid the need to use a greater quantity of the tracer to confirm the connection to the Timava/Timavo springs, the Jama 1 v Kaniaducah METKA PETRIČ, NATAŠA RAVBAR, LUCA ZINI, CHIARA CALLIGARIS, RICCARDO CORAZZI, ZDENKA ŽITKO, MARCO RESTAINO & MARTIN KNEZ Tab. 1: Results of tracer tests in 2001, 2009 and 2010. For location of injection points see Fig. 1B. Injection point Date of injection Tracer Sampling point Proved connection Peak velocity [m/h] (Recovery) Beka-Ocizla Cave system 29/3/2001 uranine Pri pralnici/Lavatoio Jama/Antro di B. Rižana main main secondary 33 29 (91%*) 6 (2%) Črnotiče (fissure) 1/12/2009 uranine Rižana Osapska Reka Pri pralnici/Lavatoio main main secondary 22 (87%) 33 (11%) 10 (1%) Borehole T2-12 18/11/2010 uranine Rižana Osapska Reka Pri pralnici/Lavatoio main main secondary 62 (41%) 23 (33%) 48 (2%) Borehole T1-8 18/11/2010 amidorhodamine G Pri pralnici/Lavatoio secondary 61 (7%) * Sum of recovered tracer through the Pri pralnici/Lavatoio and Jama/Antro di Bagnoli springs. ACTA CARSOLOGICA 49/1 – 2020 68 and Labodnica/Abisso di Trebiciano caves in the Reka- Timavo karst aquifer were chosen as sampling points. The T2-18 borehole was chosen as the injection point in the area of the T2 tunnel; the borehole was drilled in a zone with the least reliable prediction of the underground flow direction based on the existing results, i.e., the possible direction towards all three main springs in the area (Rižana, Osapska Reka, Boljunec/Bagnoli). METHODS Daily precipitation data (ARSO 2019) and flow rates of the sinking river Reka at the hydrological station Cerkvenikov mlin and of the Rižana River at the hy- drological station Kubed in 30-minute intervals (ARSO 2020) were obtained from ARSO websites. At the Boljunec/Bagnoli Spring water levels have been moni- tored since 2010 in 30-minute intervals with the Onset HOBO (U20) programmable autonomous instruments; occasionally, the total flow rate of the joint stream was also measured with the current meter Ott HydroMet C20 or SonTek FlowTracker 2. Additionally, the flow rate of the Na placu/Abbeveratoio spring was measured sepa- rately, as according to previous research (Gabrovšek et al. 2015), it has a different recharge area than the springs Pri pralnici/Lavatoio and Jama/Antro di Bagnoli. Based on these measurements, the total flow rate of the springs Pri pralnici/Lavatoio and Jama/Antro di Bagnoli was deter- mined as the difference between the measured values. By comparing the measured flow rates to the simultaneously measured water levels in Jama/Antro di Bagnoli, the rat- ing curve of dependence between the water levels and flow rates was obtained, and based on the equations ob- tained the flow rates were calculated for all the measured levels. Unpublished data from periodic measurements of flow rates in Jama 1 v Kanjaducah using the salt dilution method were provided by colleagues from the Karst Re- search Institute ZRC SAZU. A comparison with the flow rates data for the Reka River showed very similar values, with a delay of approximately 12 hours. Davorjevo brezno is a 303 m deep and just over 5 km long karst cave with an entrance at the altitude of 510 m a.s.l. (Mesarec 2011); the injection point is located 279 m deeper in the terminal siphon at the altitude of 231 m a.s.l. (Tab. 2). The flow rate of the stream upstream of the siphon was measured first, using the salt dilution meth- od. At a flow rate of 53 L/s, a solution of 3 kg of uranine was poured into it on 27 November 2018 at 12:05 (Fig. 4). The T2-18 borehole in the area of the T2 tunnel was chosen as the second injection point (Fig. 1; Tab. 2). While drilling, zones of greater fissuration and good water permeability were detected in the borehole in the limestone of Eocene age, which guarantees a relatively fast underground outflow of the injected tracer. In the borehole with a well top at the height of 408 m, levels of groundwater were measured at altitudes between 249 DETERMINING THE DIRECTIONS AND CHARACTERISTICS OF UNDERGROUND W ATER FLOW IN KARST FOR THE PURPOSE OF TRAFFIC ROUTES CONSTRUCTION: THE CASE OF THE NEW DIV AČA-KOPER RAILW AY LINE (SW SLOVENIA) Fig. 4: Injecting 3 kg of uranine on 27 November 2018 at 12:05 into the stream (A) upstream of the terminal siphon (B) in Davorjevo brezno (Photos: M. Blatnik). Tab. 2: Coordinates of injection points (in Davorjevo brezno Z is the altitude of the terminal siphon, while in the borehole it is the highest measured water level) and the amounts of the tracers used. Injection point GKX GKY Z (m a.s.l.) Tracer Amount of tracer Davorjevo brezno 419990 55609 231 uranine 3 kg Borehole T2-18 413818 50096 254 naphthionate 3 kg ACTA CARSOLOGICA 49/1 – 2020 69 and 254 m a.s.l. A solution of 3 kg of naphthionate was poured into the borehole on 27 November 2018 at 12:20 (Fig. 5) and washed away with 5.5 m 3 of water from a tank. Based on the known hydrogeological characteris- tics, a total of 19 points were chosen for the sampling, namely 13 springs, 2 cave streams and 4 surface streams. Several blank samples were collected beforehand at all the points (Tab. 3) in October and November 2018. As the connection between Jama 1 v Kanjaducah, Labodnica/Abisso di Trebiciano and Timava/Timavo springs is well known (Boegan 1938; Cucchi et al. 2015; Calligaris et al. 2018; Calligaris et al. 2019), it was de- cided not to sample the Timava/Timavo springs, but only the waters flowing in the two caves. In this way it was possible to use a smaller amount of tracer. In the two caves, the samples were collected manually. Between 30 November and 20 December 2018, 11 samples were tak- en in the cave Jama 1 v Kanjaducah; with initially a daily frequency, and later on every two days. Between 27 No- vember and 21 December 2018, 18 samples were taken in the Labodnica/Abisso di Trebiciano cave. Until 9 Decem- ber 2018, the sampling was conducted once a day, and after that every two days. In this cave, for uranine detec- tion also a field fluorimeter (GGUN-FL24 fluorimeter), having a sampling rate of 15 minutes, was installed. Due METKA PETRIČ, NATAŠA RAVBAR, LUCA ZINI, CHIARA CALLIGARIS, RICCARDO CORAZZI, ZDENKA ŽITKO, MARCO RESTAINO & MARTIN KNEZ Tab. 3: Coordinates of sampling points (CS-cave stream, SP-spring, SS-surface stream) and linear distances between the two injection points (L1 for Davorjevo brezno and L2 for the T2-18 borehole) and the sampling points. No. Sampling points Type GKY GKX Z (m) L 1 (km) L 2 (km) 1 Jama 1 v Kanjaducah CS 412875 62010 15 9.9 12.0 2 Labodnica/Abisso di Trebiciano CS 409140 60708 12 12.3 11.6 3 Rižana SP 413334 43215 69 13.9 6.9 4 Osapska Reka SP 411330 48050 105 11.5 3.1 5 Boljunec/Bagnoli (joint) SP 411392 52671 60 9.2 3.5 6 Pri pralnici/Lavatoio SP 411407 52668 61 9.2 3.5 7 Jama/Antro di Bagnoli SP 411442 52657 62 9.2 3.5 8 Na placu/Abbeveratoio SP 411404 52689 62 9.2 3.5 9 Glinščica/Rosandra SS 412094 53470 89 8.3 3.8 10 Klinšca/Oppia 1 SP 412271 53455 92 8.0 3.7 11 Klinšca/Oppia 2 SP 412287 53458 93 8.0 3.7 12 Glinščica/Rosandra (above Oppia) SS 412306 53454 94 8.0 3.7 13 Krvavi potok/Rio del Sangue SS 413192 52903 178 7.8 2.9 14 Grižnik/Grisa SS 413191 52883 178 7.8 2.9 15 Zroček 1 SP 413633 52883 250 7.2 2.3 16 Zroček 2 SP 413706 52854 241 7.4 2.4 17 Moganjavec SP 411470 51179 142 9.7 2.6 18 Zgurenc SP 411409 51171 145 9.7 2.6 19 Na Kaluži SP 411345 51391 103 9.7 2.8 Fig. 5: A solution of 3 kg of naphthionate was injected into the T2-18 borehole on 27 November 2018 at 12:15 (Photos: M. Ćuk). ACTA CARSOLOGICA 49/1 – 2020 70 to some instrumental problems, it has only been possible to recover the data from 13 December to 21 December 2018. The ISCO 6712 automatic samplers were installed on 26 November 2018 at the Rižana (Fig. 6) and Osap- ska Reka springs, and at the joint stream of the Boljunec/ Bagnoli springs. The sampling frequency ranged from every 6 hours at the beginning of the test to twice a day when the sampling was completed on 13 February 2019. For the detection of uranine, continuous fluorescence measurements were carried out in parallel at the Rižana spring in 30-minute intervals from 27 November 2018 to 28 January 2019 using the Götschy Optotechnik LLF-M field fluorometer. Between 26 November 2018 and 20 December 2018, periodic manual sampling was conducted (once a day or every two days, depending on the precipitation conditions) at seven springs (Na placu/Abbeveratoio, Pri pralnici/Lavatoio, Jama/Antro di Bagnoli, Klinšca/Op- pia 1 and 2, Zroček 1 and 2) and at five sections of the surface streams in the Glinščica/Rosandra valley and of three springs in the area of Dolina/San Dorligo village (Na Kaluži, Moganjevac, Zgurenc). All collected samples were properly stored at ~4°C in dark vials. In order to establish the presence of trac- ers, they were analysed in the laboratory of the Karst Research Institute ZRC SAZU with the PERKIN EL- MER LS 45 luminescence spectrometer; the uranine at E ex =491nm, E em =512nm and the naphthionate at E ex =320nm, E em =430nm (Käss 1998, Meus et al. 2006). Analyses were conducted soon after the sampling. After decantation, the entire batch of samples was analysed again as the sampling was completed at individual sam- pling points. This was necessary because high turbidity levels were detected in most of the samples taken from the springs, especially after intense precipitation, which interfered with the fluorescence measurements. Due to the poorer stability of the naphthionate (Goldscheider et al. 2001, Benischke et al. 2007) the samples were analysed for the presence of this tracer within 14 days. According to the literature (Benischke et al. 2007) the instrumental detection limit for uranine is 0.001 µg/L. The naphthi- onate has less favourable properties, especially in waters containing organic contamination; thus, the literature mentions diverse detection limits, ranging from 0.1 to 1 µg/L; the appearance of higher values is not excluded (Meus et al. 2014). With luminescence spectrometer LS45 measured concentrations of uranine in 5 samples taken manually in Labodnica/Abisso di Trebiciano from 13 December to 21 December 2018, were used for calibration of field fluorimeter GGUN-FL24. For the calibration of LS45 and LLF-M, samples of different concentrations were pre- pared with the used uranine and water from the Rižana spring before the injection. DETERMINING THE DIRECTIONS AND CHARACTERISTICS OF UNDERGROUND W ATER FLOW IN KARST FOR THE PURPOSE OF TRAFFIC ROUTES CONSTRUCTION: THE CASE OF THE NEW DIV AČA-KOPER RAILW AY LINE (SW SLOVENIA) Fig. 6: The sampling point at the Rižana spring; A) The ISCO 6712 automatic water sampler and the LLF-M field fluorometer; B) the setting up of probes and pumping pipes (Photos: M. Petrič). ACTA CARSOLOGICA 49/1 – 2020 71 RESULTS AND DISCUSSION PRECIPITATION AND HYDROLOGICAL CONDITIONS To illustrate the precipitation and hydrological condi- tions in the period prior to and during the testing, the results are summarized in Fig. 7, where the data for the period between 20 October 2018 and 14 February 2019 are presented. After abundant rainfall in late October 2018, the flow rates of the Reka and Rižana rivers were high (more than 229 m 3 /s for Reka and 35 m 3 /s for Rižana); in early Novem- ber the water level slowly began to drop. Then on 25 and 26 November 2018 a new rainfall event of 34 mm slightly increased the flow rates again. The tracers were injected on 27 November 2018. This was followed by a few minor pre- cipitation events in early December and by heavier rainfall on 8 December 2018 with 42 mm of precipitation. Even after the first tracer breakthrough curve, the sampling was continued until February 2019, when very high flow rates were reached after several days of precipitation (more than 300 m 3 /s for Reka and 68 m 3 /s for Rižana). TRACING WITH URANINE (injection into Davorjevo brezno) Uranine was detected for the first time in a sample taken in the cave Jama 1 v Kanjaducah on 8 December 2018 at 20:00 (Fig. 8). Since the concentration was very low (0.012 µg/L), it can be deduced that this sample shows the beginning of tracer appearance. The peak of the trac- er breakthrough curve can be determined less reliably, because in that period the sampling was conducted every 2 days. The highest concentration of 0.56 µg/L was meas- ured in a sample taken on 11 December 2018 at 14:40. As expected, the tracer also appeared, with a shorter delay, in the Labodnica/Abisso di Trebiciano cave; the first ap- pearance (0.022 µg/L) was detected on 9 December 2018 at 12:40, while the highest concentration of 0.58 µg/L was measured in the sample taken on 13 December 2018 at 12:00. The decrease of tracer concentrations was meas- ured also by the field fluorimeter (Fig. 8). Taking into account the linear distance between the injection point and Jama 1 v Kanjaducah (Tab. 4) and the time interval from injection to the first tracer appearance, the maximum flow velocity of 37 m/h was calculated; by taking into account the time interval from injection to the appearance of the highest tracer concentration, the peak velocity of 29 m/h was calculated (Tab. 3). A temporally displaced flow rates curve for the Reka River was used to assess the flow rates in the cave Jama 1 v Kanjaducah and, based on this, the proportion of re- covered tracer was calculated (Fig. 8). Due to such ap- proximate method of determining the flow rate and the small sampling frequency it is only a rough estimate, but it nevertheless enabled an evaluation of the tracing re- sults. In the first flood pulse after injection, about three- METKA PETRIČ, NATAŠA RAVBAR, LUCA ZINI, CHIARA CALLIGARIS, RICCARDO CORAZZI, ZDENKA ŽITKO, MARCO RESTAINO & MARTIN KNEZ 0 10 20 30 40 50 60 70 80 0 50 100 150 200 250 300 350 20/10/18 8/11/18 27/11/18 16/12/18 4/1/19 23/1/19 11/2/19 Q_Rižana (m 3 /s) Q_Reka (m 3 /s) Time of injection Reka Rižana B 0 10 20 30 40 50 60 70 P (mm) A Fig. 7: A) Daily pre- cipitation (P) at the ARSO Škocjan sta- tion; B) the flow rates (Q) of the Rižana River at the ARSO Kubed station and of the Reka River at the ARSO Cerkvenikov mlin station, meas- ured at 30-minute in- tervals for the period prior to injection and throughout the sam- pling. ACTA CARSOLOGICA 49/1 – 2020 72 DETERMINING THE DIRECTIONS AND CHARACTERISTICS OF UNDERGROUND W ATER FLOW IN KARST FOR THE PURPOSE OF TRAFFIC ROUTES CONSTRUCTION: THE CASE OF THE NEW DIV AČA-KOPER RAILW AY LINE (SW SLOVENIA) Fig. 8: A) Precipitation (P) at the ARSO Škocjan sta- tion; B) the measured flow rates (Q) of the Reka River at the ARSO Cerkvenikov mlin station and the es- timated flow rates (Q) of the Reka River in Jama 1 v Kanjaducah; C) uranine concentrations in samples from Jama 1 v Kanjaducah and Labodnica/Abisso di Trebiciano (LAB/TRE_ GGUN: measured with field fluorimeter; LAB/ TRE_LS45: measured in samples in laboratory); D) the estimated share of re- covered tracer in Jama 1 v Kanjaducah.  0 0.1 0.2 0.3 0.4 0.5 0.6 Uranine conc. (µg/l) Jama 1 v Kanjaducah LAB/TRE_GGUN LAB/TRE_LS45 Time of injection C 0 50 Q (m 3 /s) B 0 20 40 P (mm) 0 20 40 60 80 100 26/11/18 1/12/18 6/12/18 11/12/18 16/12/18 21/12/18 26/12/18 Tracer recovery (%) Jama 1 v Kanjaducah D A 0 0.1 0.2 0.3 0.4 0.5 0.6 26/11/18 6/12/18 16/12/18 26/12/18 5/1/19 15/1/19 25/1/19 4/2/19 14/2/19 Uranine conc. (µg/l) Ri žana Osapska reka Boljunec/Bagnoli Time of injection C 0 20 40 60 80 Q (m 3 /s) B 0 20 40 60 80 P (mm) A Fig. 9: A) Precipitation (P) at the ARSO Škocjan sta- tion; B) the measured flow rates (Q) of the Rižana River at the ARSO Kubed station; C) the measured uranine concentrations in Rižana, Osapska Reka and Boljunec/Bagnoli. ACTA CARSOLOGICA 49/1 – 2020 73 METKA PETRIČ, NATAŠA RAVBAR, LUCA ZINI, CHIARA CALLIGARIS, RICCARDO CORAZZI, ZDENKA ŽITKO, MARCO RESTAINO & MARTIN KNEZ quarters of the injected uranine passed through Jama 1 v Kanjaducah. It can be deduced that this is the main flow direction from Davorjevo brezno along a system of inter- connected and highly permeable karst fissures and con- duits. Owing to the high proportion of recovered tracer and the demanding fieldwork in the deep karst caves, the sampling was terminated after the first flood pulse had passed through. Based on experience from previous trac- ings, it can be assumed that the tracer would continue to appear in both caves in smaller concentrations in the subsequent flood pulses, thus further increasing the pro- portion of the recovered tracer. The linear distance between the injection point in the Davorjevo brezno cave and the Labodnica/Abisso di Trebiciano cave is 12.3 km (Tab. 3); taking this distance into account, the maximum and peak flow velocities of 43 m/h and 32 m/h, respectively, were calculated (Tab. 4). However, judging by the characteristics of the Reka- Timavo aquifer system proved by other tracer tests, the tracer previously detected in Jama 1 v Kanjaducah also appeared in Labodnica/Abisso di Trebiciano. It there- fore travelled from Davorjevo brezno to the conduits of the Reka-Timavo system somewhere between the Škocjanske jame caves and the Jama 1 v Kanjaducah cave, and then through the latter cave towards the Labodnica/ Abisso di Trebiciano cave and finally towards the springs in the Gulf of Trieste. Taking this path into account, the maximum and peak velocities are 48 m/h and 36 m/h, re- spectively; while the peak velocity between Jama 1 v Kan- jaducah and Labodnica/Abisso di Trebiciano is estimated at 88 m/h. However, the flow velocities are strongly de- pendent on the rainfall regime which, in turn, influence the hydraulic gradient. Its increase consequently results in an increase of flow velocities. The velocities by tracer tests represent the conditions at the time of the test and higher velocities can be expected. In addition, changes in flow directions due to changes in hydrological conditions are possible. All this should be taken into account when assessing the vulnerability of the springs used for water supply. In the initial period of the tracer test, before the first tracer appearance, the samples were analysed as soon as they were delivered to the laboratory; the first increase in the measured fluorescence was detected in the joint stream of the Boljunec/Bagnoli springs on 7 December, and the highest value of 0.36 µg/L on 9 December 2018. Soon afterwards, the increases in fluorescence were also detected in the samples from Jama 1 v Kanjaducah and Labodnica/Abisso di Trebiciano. The analyses conducted after decantation confirmed the uranine concentrations that had originally been measured in both caves; the signal measured later on in the samples from Boljunec/ Bagnoli was much smaller and the measured concentra- tions were up to 0.046 µg/L (Fig. 9). On the basis of these results the flow of water from Davorjevo brezno towards the springs in Boljunec/Bagnoli cannot be reliably con- firmed. V ery similar concentrations were measured in the Pri pralnici/Lavatoio and Jama/Antro di Bagnoli springs, and even lower in the Na placu/Abbeveratoio. However, based on these results no definite conclusions on differ- ent recharge areas of individual springs in Boljunec/Bag- noli can be made. Until 14 February 2019, the Rižana and Osapska Reka springs were also sampled (Fig. 9). It has been de- duced from the measured fluorescence values that the groundwater from Davorjevo brezno do not flow to- wards these two springs. In the samples of the surface stream of the Glinščica/ Rosandra River and its tributaries, increased fluores- cence values were measured after the precipitation on 8 December 2018. Considering that the water was very turbid at the time and the increase was detected in one sample only, it was concluded that the increased fluores- cence was not the result of uranine coming from Davor- jevo brezno, but was induced by increased turbidity. TRACING WITH NAPHTHIONATE (Injection into the T2-18 Borehole) For the area of the T2-18 borehole the possibility of an underground flow towards the springs in Boljunec/Bag- noli and towards the Rižana and Osapska Reka springs was predicted. Confirming this connection with the Rižana spring would require the use of a greater trac- er quantity due to the higher flow rates and greater distance; however, that would mean that the nearby springs at Boljunec/Bagnoli, which have a much lower Tab. 4: Uranine appearance; L – linear distance between injection point and sampling point; t 1 – time interval from injection to the first tracer detection; v max – maximum flow velocity; c max – highest measured tracer concentration; t p – time interval from injection to recording the highest tracer concentration; v p – peak velocity; R – estimated proportion of recovered tracer. Uranine L (km) t 1 (h) v max (m/h) c max (mg/m 3 ) t p (h) v p (m/h) R (%) Jama 1 v Kanjaducah (KAN) 9.9 272 37 0.56 338 29 74 Labodnica/Abisso di Trebiciano (LAB/TRE) 12.3 288 43 0.58 384 32 LAB/TRE (through KAN) 13.9 288 48 0.58 384 36 Between KAN and LAB/TRE 4.0 45 88 ACTA CARSOLOGICA 49/1 – 2020 74 flow rate, might be dyed visible. For this reason, the fluorescent tracer naphthionate was chosen, which is invisible to the naked eye even in the concentration of 1 g/L (Benischke et al. 2007). However, this tracer has a few unfavourable properties that hinder the interpreta- tion of the results of analyses. The natural background values may be increased by the appearance of organic contamination alone, as has been demonstrated by the analysis of blank samples taken during high water con- ditions. Regardless, when conducting the sampling after injection, the concentrations measured in the springs of Boljunec/Bagnoli (up to 6.9 µg/L) and Osapska Reka (up to 3.5 µg/L) at the time of a smaller flood pulse in early December 2018 deviate (Fig. 10) and indicate an underground water connection between the T2-18 borehole and both springs Due to the problem with the background values, only a rough assessment of the trac- er recovery was possible. It was estimated that approxi- DETERMINING THE DIRECTIONS AND CHARACTERISTICS OF UNDERGROUND W ATER FLOW IN KARST FOR THE PURPOSE OF TRAFFIC ROUTES CONSTRUCTION: THE CASE OF THE NEW DIV AČA-KOPER RAILW AY LINE (SW SLOVENIA) 0 20 40 60 80 P (mm) 0 1 2 3 4 5 6 7 26/11/18 6/12/18 16/12/18 26/12/18 5/1/19 15/1/19 25/1/19 4/2/19 14/2/19 Naphtionate conc. (µg/l) Boljunec/Bagnoli Rižana Osapska reka Time of injection C 0 20 40 60 80 Q_Rižana (m 3 /s) Rižana Boljunec/Bagnoli B A 0 20 40 60 80 P (mm) A 0 1 2 3 4 5 6 7 26/11/18 1/12/18 6/12/18 11/12/18 16/12/18 21/12/18 26/12/18 Naphthionate conc. (µg/l) Boljunec/Bagnoli Pri pralnici/Lavatoio Jama/Antro di Bagnoli Na placu/Abbeveratoio Time of injection B Fig. 10: A) Precipitation (P) at the ARSO Škocjan station; B) the measured flow rates (Q) of the Rižana River at the ARSO Kubed station and the flow rates (Q) of the Boljunec/Bagnoli spring; C) the measured naphthionate concentra- tions in Rižana, Osapska Reka and Boljunec/Bag- noli. Fig. 11: A) Precipitation (P) at the ARSO Škocjan station; B) the measured naphthionate concentra- tions of samples from the individual springs and joint stream in the Boljunec/Bagnoli village. ACTA CARSOLOGICA 49/1 – 2020 75 mately 5% of the tracer passed through the Boljunec/ Bagnoli spring, and approximately 25% through the Osapska Reka spring. The smaller proportions of recov- ered tracer were anticipated on account of injection into the borehole and the poorer properties of naphthionate compared to uranine. The comparison of naphthionate concentrations in individual springs of Boljunec/Bagnoli indicates differ- ent characteristics (Fig. 11). Despite the deficiencies of the tracer used, it is possible to conclude that naphthion- ate was not detected in the Na placu/Abbeveratoio spring and the underground water connection between the T2- 18 borehole and this spring doesn’t exist. This finding is in agreement with the results of previous tracer tests (Gabrovšek et al. 2015) and supports the idea that the Na placu/Abbeveratoio spring is recharged from the area of Glinščica/Rosandra valley. On the other hand, the appearance of naphthion- ate in the Moganjavec spring in the village of Dolina/San Dorligo on the Italian side of the border is convincing. The highest concentration 19.3 µg/L was measured in a sample taken on 7 December 2018, i.e., before a distinct increase in discharge (Fig. 12). The concentrations grad- ually decreased until 11 December 2018; the values were very low before and after this flood pulse. In the mea- surements for uranine, the fluorescence values were con- stantly below the detection limit. As the sampling point represents only one part of the Moganjavec spring, it is not possible to assess the total flow rate and the propor- tion of recovered tracer in this spring. The maximum and peak velocity of the under- ground flow was calculated from the T2-18 borehole to the Boljunec/Bagnoli spring at 13 m/h, to the Osapska Reka spring at 11 and 10 m/h, and to the Moganjavec spring at 12 and 11 m/h (Tab. 5). In all the other monitored points, the naphthionate concentrations were not high enough to confirm a con- nection to the T2-18 borehole. METKA PETRIČ, NATAŠA RAVBAR, LUCA ZINI, CHIARA CALLIGARIS, RICCARDO CORAZZI, ZDENKA ŽITKO, MARCO RESTAINO & MARTIN KNEZ Tab. 5: Naphthionate appearance; L – linear distance between injection point and sampling point; t 1 – time interval from injection to the first tracer detection; v max – maximum flow velocity; c max – highest measured tracer concentration; t p – time interval from injection to re- cording the highest tracer concentration; v p – peak velocity; R – estimated proportion of recovered tracer. Naphthionate L (km) t 1 (h) v max (m/h) c max (mg/m 3 ) t p (h) v p (m/h) R (%) Boljunec/Bagnoli 3.5 270 13 6.9 277 13 ∼5 Osapska Reka 3.1 297 11 3.5 316 10 ∼25 Moganjavec 2.6 214 12 19.3 240 11 ∼0.1 Fig. 12: A) Precipitation (P) at the ARSO Škocjan station; B) the measured naphthionate concentra- tions in the Moganjavec spring in the village of Dolina/San Dorligo near Trieste. 0 20 40 P (mm) A 0 5 10 15 20 26/11/18 1/12/18 6/12/18 11/12/18 16/12/18 21/12/18 26/12/18 Naphtionate conc. (µg/l) Moganjavec Time of injection B ACTA CARSOLOGICA 49/1 – 2020 76 CONCLUSIONS In light of the vulnerability of karst water sources, a com- prehensive analysis of the characteristics and dynamics of groundwater flow in their recharge areas should be performed prior to starting construction of traffic routes. Water in karst systems may flow from a single point to- wards multiple springs in different directions. The under - ground pathways and the location of the watersheds be- tween the catchment areas of individual springs may also change depending on the hydrological conditions. This should be taken into consideration while planning regu- lar monitoring and implementing measures in the event of a possible accident. For the assessment of groundwater drainage patterns along the proposed routes, tracer tests with artificial tracers are especially useful. Several tracer tests have been conducted in the area in question, but especially the older ones were not tied to the planned construction of a new railway line. Regardless, they provide important information for as- sessing the potential impact of its construction and op- eration. Based on the collected older data and the find- ings of the new tracer test, the main characteristics of the flow of water in the karst area of the railway route between Divača and Koper/Capodistria can be summed up. From the area of the northern T1 tunnel the under- ground flow is directed through the aquifer of the Clas- sical Karst mainly towards the Timava/Timavo springs in Italy. Through the most permeable conduits of the Reka-Timavo aquifers system the velocity of the water flow (calculated based on linear distance) ranges from 47 to 204 m/h during different hydrological conditions. The new tracer test, in which uranine was injected into the stream in the Davorjevo brezno cave east of the T1 tunnel in late November 2018, has confirmed the main flow direction from the area of the T1 tunnel towards the Jama 1 v Kanjaducah and Labodnica/Abisso di Trebiciano caves and onwards towards the Timava/Ti- mavo springs. During medium-high water conditions the flow velocities along this system of interconnected and highly permeable karst fissures and conduits ranged from 30 to 50 m/h. Previous tracing in the T1-8 bore- hole has shown that a very small portion of groundwa- ter also flows towards the springs in Boljunec/Bagnoli, but only from the southern part of the T1 tunnel and under favourable hydrological conditions. In order to obtain additional information on the characteristics of the underground flow in the area of the southern T2 tunnel a different tracer was used for the tracing conducted in November 2018. The used tracer naphthionate proved to be less suitable for re- searching large karst systems. Furthermore, the results indicate that the amount of tracer injected was too low under the given conditions. Nevertheless, new informa- tion was obtained regarding the position and character- istics of the wide bifurcation zone between the springs of Rižana, Osapska Reka and Boljunec/Bagnoli. It has been confirmed that velocity and the dominant flow di- rection depend on the position within the tunnel. In the northern part, the groundwater flows mainly towards the springs Pri pralnici/Lavatoio and Jama/Antro di Bagnoli in Boljunec/Bagnoli, whereas no tracer was detected in the Na placu/Abbeveratoio spring. T owards the south the proportion of the flow towards the Rižana and Osapska Reka springs begins to increase. The flow velocities de- termined through various tracer tests range from 10 to 60 m/h. As the tracer test was conducted in the cross-border Slovenian-Italian region and because the work was very demanding due to the hard-to-reach streams in deep karst caves and the large number of sampling points, the successful implementation of the test depended on the cooperation between the researchers and speleologists on both sides of the border. Due to the frequently long duration and high costs of tracer tests it was important to prepare in advance a detailed plan of the test and to ad- just this plan to the real conditions during its implemen- tation. The conclusions are beneficial to the monitoring of any potential negative impacts of the construction and operation of the new Divača – Koper/Capodistria rail- way line. For individual sections of the planned railway line, the most probable directions of the groundwater flow were determined and the times of possible occur- rence of contamination in the springs were evaluated ac- cording to worst-case scenarios. As a basis for an early warning system, a monitoring scheme was established in which the dynamics of sampling are adapted to precipi- tation and hydrological conditions. The study example could also be useful for any further research on planning the hazard control of traffic routes in karst. DETERMINING THE DIRECTIONS AND CHARACTERISTICS OF UNDERGROUND W ATER FLOW IN KARST FOR THE PURPOSE OF TRAFFIC ROUTES CONSTRUCTION: THE CASE OF THE NEW DIV AČA-KOPER RAILW AY LINE (SW SLOVENIA) ACTA CARSOLOGICA 49/1 – 2020 77 ACKNOWLEDGEMENTS This study was supported by the Slovenian Infra- structure Agency/Railways Division of the Ministry of Infrastructure. The authors acknowledge that the pro- gramme Karst Research, No. P6-0119, and the project Environmental effects and karst water sources: impacts, vulnerability and adaptation of land use, No. J6-8266, were financially supported by the Slovenian Research Agency, and the project Accordo attuativo di collabora- zione, Individuazione e delimitazione degli acquiferi carsici del Friuli Venezia Giulia by the Geological Sur- vey of the FVG Region. Our thanks go out to Matej Blat- nik, Leon Drame, Franjo Drole, Franci Gabrovšek, Blaž Kogovšek, Cyril Mayaud, Mitja Prelovšek, Mateja Za- del (ZRC SAZU, Karst Research Institute), Sara Biolchi, Chiara Boccali, Christian Leone, Luca Terribili (Depart- ment of Mathematics and Earth Sciences, University of Trieste), Paolo de Curtis, Louis Torelli, Spartaco Savio, Marco Armocida (Commissione Grotte E. Boegan), Gre- ga Maffi, Tina Bizjak, Bogomir Remškar, Mitja Mršek, Jerica Koren, Miha Staut, Ines Klinkon, Matej Zalokar, Boštjan Vrviščar (Spelelogical Association of Slovenia), Edgardo Mauri, Davio Fabris, Paolo Camerino, Paolo Cossi, Francesco Degrassi, Enrico Iuorio, Piero Slama, Giuseppe Masarin, Alessandra Mautone, Mauro Catta- rini, Dennis Dugulin, Claudio Bratos, Giulio Dagostini, Bruno Vojtissek, Manuel Laspada (Società Adriatica di Speleologia), who were involved in the injection, sam- pling and analyses alongside the authors of the article; to Darija Bak, Edi Zobec and Rižanski vodovod Koper for their assistance in organizing sampling on their proper- ties; to Joerg Prestor for his participation in the discus- sion regarding which injection points to choose; and to Mitja Prelovšek for providing the data on the flow rates in Jama 1 v Kanjaducah. REFERENCES Alija, S., Torrijo, F.J. & M. Quinta-Ferreira, 2013: Geo- logical engineering problems associated with tun- nel construction in karst rock masses: the case of Gavarres tunnel (Spain).- Engineering geol- ogy, 157, 103-111. https://doi.org/10.1016/j. enggeo.2013.02.010 ARSO, 2019: Slovenian Environment Agency, Interactive weather.- [Online] Available from: https://meteo. arso.gov.si/met/en/app/webmet/ [Accessed 16th February 2019]. ARSO, 2020: Slovenian Environment Agency, Archive hydrological data.- [Online] Available from: http:// vode.arso.gov.si/hidarhiv/pov_arhiv_tab.php [Ac- cessed 6th January 2020]. Atlas okolja, 2020: Environmental Atlas, Tracer tests data.- [Online] Available from: http://gis.arso.gov. si/atlasokolja/profile.aspx?id=Atlas_Okolja_AXL@ Arso [Accessed 20th January 2020]. Benischke, R., Goldscheider, N. & C. Smart, 2007: Tracer techniques.- In: Goldscheider, N. & D. Drew (eds) Methods in karst hydrogeology. Taylor&Francis, pp. 161-184, London. Boegan, E., 1938: Il Timavo. Studio sull’idrologia carsica subaerea e sotterranea.- Mem. Ist. It. Speleol., pp. 251, Trieste. Barberá, J.A., Mudarra, M., Andreo, B. & B. De la Torre, 2018: Regional-scale analysis of karst underground flow deduced from tracing experiments: examples from carbonate aquifers in Malaga province, south- ern Spain.- Hydrogeology Journal, 26, 23–40. htt- ps://doi.org/10.1007/s10040-017-1638-5 Calligaris, C., Mezga K., Slejko F.F., Urbanc J. & L. Zini, 2018: Groundwater characterization by means of conservative (d 18 O and d 2 H) and non-conservative ( 87 Sr/ 86 Sr) isotopic values: the Classical Karst Re- gion aquifer case (Italy–Slovenia).- Geosciences, 8, 9, 321. https://doi.org/10.3390/geosciences8090321 Calligaris, C., Galli M., Gemiti F., Piselli S., Tentor M., Zini L. & F. Cucchi, 2019: Electrical Conductiv- ity as a tool to evaluate the various recharges of a Karst aquifer.- Rendiconti Online della Società Ge- ologica Italiana, 47, 13-17. https://doi.org/10.3301/ ROL.2019.03 Cucchi, F., Zini, L. & C. Calligaris (eds), 2015: Le acque del Carso Classico / Vodonosnik klasičnega Krasa. Progetto/Projekt HYDROKARST.- Edizioni Univer- sità di Trieste, pp. 181, Trieste. Gabrovšek, F ., Knez, M., Kogovšek, J., Mihevc, A., Mulec, J., Otoničar, B., Perne, M., Petrič, M., Prelovšek, M., Slabe, T ., Šebela, S., T urk, J. & N. Zupan Hajna, 2015: The Beka-Ocizla Cave System (Cave and Karst Sys- tems of the World).- Springer, pp. 102, Cham [etc.]. https://doi.org/ 10.1007/978-3-319-04456-9_2 Galli, M., 2012: I traccianti nelle ricerche sul Timavo.- Edizioni Università di Trieste, pp. 253, Trieste. Gemiti, F., 1984a: La portata del Timavo alle risorgive di METKA PETRIČ, NATAŠA RAVBAR, LUCA ZINI, CHIARA CALLIGARIS, RICCARDO CORAZZI, ZDENKA ŽITKO, MARCO RESTAINO & MARTIN KNEZ ACTA CARSOLOGICA 49/1 – 2020 78 S. Giovanni di Duino.- Annali del Gruppo Grotte dell’ Associazione XXX Ottobre, 7, 23-41. Gemiti, F., 1984b: Nuova e originale prova di marcatura delle acque del Timavo.- Annali del Gruppo Grotte dell’ Associazione XXX Ottobre, 7, 43-62. Gemiti, F., 1995: Portata liquida e portata solida del Timavo alle risorgive di S. Giovanni di Duino.- Hy- drores, 13, 75-88. Gemiti, F., 1998: Marcatura delle acque del Timavo a seguito di un versamento di idrocarburi nella valle della Recca e interpretazione dell’evento mediante l’utilizzo di dati meteorologici, idrologici, idrochim- ici.- Annali del Gruppo Grotte dell’Associazione XXX Ottobre, 10, 93–104. Gemiti, F ., 2004: Le sorgenti Sardos e l’ approvvigionamento idrico della Provincia di Trieste.- Atti e Memorie della Commissione Grotte “E. Boegan” , 39, 67-80. Goldscheider, N., Hötzl, H. & K. Kottke, 2001: Microbio- logical decay of naphthionate in water samples as a source of misinterpretation of tracer tests.- In: Seil- er, K.P . & S. W öhnlich (eds.) Characterizing Ground- water Flow: Proceedings of the XXXI IAH Congress, 10 th -14 th September 2001, Munich. CRC Press, 77- 81, Munich. Goldscheider, N., Meiman, J., Pronk, M. & C. Smart, 2008: Tracer tests in karst hydrogeology and spele- ology.- International Journal of Speleology, 37, 27- 40. https://doi.org/10.5038/1827-806X.37.1.3 Hötzl, H., 1999: Industrial and urban produced impacts.- In: Drew, D. & H. Hötzl (eds.) Karst hydrogeology and human activities, Impacts, consequences and Im- plications. International contributions to hydrogeol- ogy, A.A. Balkema, pp. 81-185, Rotterdam. Jin, X., Li, Y., Luo, Y. & H. Lui, 2016: Prediction of city tunnel water inflow and its influence on overlain lakes in karst valley.- Environmental Earth Scienc- es, 75, 1162. https://doi.org/10.1007/s12665-016- 5949-y Jurkovšek, B., Biolchi, S., Furlani, S., Kolar-Jurkovšek, T., Zini, L., Jež, J., Tunis, G., Bavec, M. & F. Cucchi, 2016: Geology of the Classical Karst Region (SW Slovenia – NE Italy).- Journal of maps, 12, S1, 352- 362. http://dx.doi.org/10.1080/17445647.2016.1215 941 Käss, W., 1998: Tracing technique in geohydrology.- A.A, Balkema, pp. 581, Brookfield, Rotterdam. Knez, M., Umek, U., Gabrovšek, F., Mihevc, A., Mulec, J., Otoničar, B., Petrič, M., Pipan, T., Prelovšek, M., Ravbar, N., Slabe, T., Šebela, S., Zupan Hajna, N., Blatnik, M., Kogovšek, B., Kukuljan, L., Mayaud, C., Aljančič, M., Drame, L., Drole, F. & M. Zadel, 2019: Krasoslovna interpretacija raziskav : zbirno poročilo : izvedba dopolnilnih strukturno geoloških, hidrogeoloških, krasoslovnih in geotehniških raziskav za PZI drugega tira železniške proge med Divačo in Koprom : drugi tir železniške proge Divača-Koper.- Gradbeni inštitut ZRMK, Ljubljana. 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, Water. ZRC Publishing, pp. 47-64, Ljubljana. Kogovšek, J. & M. Petrič, 2004: Advantages of longer- term tracing —three case studies from Slovenia.- Environmental Geology, 47, 76-83. https://doi. org/10.1007/s00254-004-1135-8 Kogovšek, J. & M. Petrič, 2011: Spillages of hazardous substances endanger karst waters.- In: Knez, M. et al. (eds.) Karstology and development challenges on karst, Water. ZRC Publishing, pp. 76-83, Ljubljana. Krivic, P., Bricelj, M., Trišič, N. & M. Zupan, 1987: Sledenje podzemnih vod v zaledju izvira Rižane.- Acta carsologica, 16, 83-104. Krivic, P., Bricelj, M. & M. Zupan, 1989: Podzemne vodne zveze na področju Čičarije in osrednjega dela Istre.- Acta carsologica, 18, 265-295. Li, X. & Y. Li, 2014: Research on risk assessment system for water inrush in the karst tunnel construction based on GIS: case study on the diversion tunnel groups of the Jinping II Hydropower Station.- Tun- nelling and Underground Space Technology, 40, 182-191. https://doi.org/10.1016/j.tust.2013.10.005 Mesarec, D., 2011: Davorjevo brezno.- Jamar, 7, 20-26. Massei, N., W ang, H.Q., Field, M.S., Dupont, J.P ., Bakalo- wicz, M. & J. Rodet, 2006: Interpreting tracer break- through tailing in a conduit-dominated karstic aq- uifer.- Hydrogeology Journal, 14, 6, 849-858. https:// doi.org/10.1007/s10040-005-0010-3 Meus, P ., Käss, W . & P .A. Schnegg, 2006: Background and detection of fluorescent tracers in karst groundwa- ter.- Hidrogeologia y Aguas Subterraneas, 18, 65-75. Meus, P., Schnegg, P.A., Frippiat, C. & J. Monfort, 2014: Promises and limitations in the use of sulfonates colourless tracers in hydrogeology.- Geologica Bel- gica, 17, 1, 90-95. Mosetti, F., 1965: Nuova interpretazione di un esperi- mento di marcatura radioattiva del Timavo.- Bolle- tino di Geofisica teorica et applicata, 7, 27, 218-243. Peric, B., 2012: Karst water course tracing between ponor and springs: the Reka river example, Kras/Carso, SW Slovenia-NE Italy.- In: Šebela S., Peric B., Fab- bricatore A. & D. Cergna (eds.) International Con- gress on “Scientific Research in Show Caves”, 13th to 15th September 2012 Postojna, pp. 32-33. Petrič, M. & J. Kogovšek, 2011: Assessment of the possi- ble impact of the construction of the Divača-Koper rail-way line on the quality of karst waters.- In: DETERMINING THE DIRECTIONS AND CHARACTERISTICS OF UNDERGROUND W ATER FLOW IN KARST FOR THE PURPOSE OF TRAFFIC ROUTES CONSTRUCTION: THE CASE OF THE NEW DIV AČA-KOPER RAILW AY LINE (SW SLOVENIA) ACTA CARSOLOGICA 49/1 – 2020 79 Prelovšek, M. et al. (eds.) Pressures and protection of the underground karst: cases from Slovenia and Croatia. Karst Research Institute ZRC SAZU, pp. 138-146, Postojna. Placer, L., Vrabec, M. & B. Celarc, 2010: The bases for understanding of the NW Dinarides and Istria Pen- insula tectonics.- Geologija, 53, 1, 55-86. https://doi. org/10.5474/geologija.2010.005 Raithel, M., Baumbusch, J. & S. Kielbassa, 2016: Con- struction of the New High-Speed Railway Line Ulm–Wendlingen in Karstifed Rock.- Procedia en- gineering, 143, 1144-1151. https://doi.org/10.1016/j. proeng.2016.06.139 Stevanović, Z., 2019: Karst waters in potable water sup- ply: a global scale overview.- Environmental Earth Sciences, 78, 662. https://doi.org/10.1007/s12665- 019-8670-9 Timeus, G., 1928: Nei misteri del mondo sotterraneo.- Alpi Giulie, 29, 1-38. Turpaud, P., Zini, L., Ravbar, N., Cucchi, F., Petrič, M. & J. Urbanc, 2018: Development of a Protocol for the Karst Water Source Protection Zoning: Appli- cation to the Classical Karst Region (NE Italy and SW Slovenia).- Water Resources Management, 32, 1953-1968. https://doi.org/10.1007/s11269-017- 1882-4 Urbanc, J., Mezga, K. & L. Zini, 2012: An assessment of capacity of Brestovica – Klariči Karst Water Supply (Slovenia).- Acta carsologica, 41, 1, 89-100. https:// doi.org/10.3986/ac.v41i1.50 Vincenzi, V ., Gargini, A., Goldscheider, N. & L. Piccinini, 2014: Differential Hydrogeological Effects of Drain- ing Tunnels Through the Northern Apennines, Ita- ly.- Rock Mechanics and Rock Engineering, 47, 947- 965. https://doi.org/10.1007/s00603-013-0378-7 Zhou, W. & B.F. Beck, 2005: Roadway construction in karst areas: management of stormwater runoff and sinkhole risk assessment.- Environmental Geology, 47, 1138-1149. https://doi.org/10.1007/s00254-005- 1252-z Zini L., Visintin L., Cucchi F . & W . Boschin, 2011: Poten- tial impact of a proposed railway tunnel on the karst environment: The example of Rosandra valley, Clas- sical Karst Region, Italy-Slovenia.- Acta carsologica, 40, 1, 207-218. https://doi.org/10.3986/ac.v40i1.38 Zini L., Calligaris C. & E. Zavagno, 2014: Classical Karst hydrodynamics: a shared aquifer within Italy and Slovenia.- In: Castellarin, A. et al. (eds.) Evolving Water Resources Systems: Understanding, Predicting and Managing Water‐Society Interactions. Interna- tional Association of Hydrological Sciences Publi- cation, 364, pp. 499‐504. https://doi.org/10.5194/ piahs-364-499-2014 Zini, L., Calligaris, C. & F. Cucchi, 2015: The challenge of tunnelling through Mediterranean karst aq- uifers: the case study of Trieste (Italy).- Environ- mental Earth Sciences, 74, 281-295. https://doi. org/10.1007/s12665-015-4165-5 METKA PETRIČ, NATAŠA RAVBAR, LUCA ZINI, CHIARA CALLIGARIS, RICCARDO CORAZZI, ZDENKA ŽITKO, MARCO RESTAINO & MARTIN KNEZ ACTA CARSOLOGICA 49/1 – 2020 80