COBISS: 1.01 MONITORING THE fLOOD PULSES IN THE EPIPHREATIC ZONE Of KARST AQUIfERS: THE CASE Of REKA RIVER SySTEM, KARST PLATEAU, SW SLOVENIA SPREMLJANJE POPLAVNIH VALOV V EPIfREATIČNI CONI KRAŠKEGA VODONOSNIKA: PRIMER REKE REKE, KRAS, JZ SLOVENIJA franci GABROVŠEK1 & Borut PERIC2 Abstract UDC 556.3 (497.4-14) Franci Gabrovšek & Borut Peric: Monitoring the food pulses in the epiphreatic zone ofkarst aquifers: Te case ofReka river sy stem, Karst plateau, SW Slovenia Te Reka river sinking into Škocjan caves (Škocjanske jame) is the main allogenic input into the aquifer of Classical Karst. So far the subsurface fow of the Reka river between Škocjan caves and the spring of Timava in Italy has been reached in fve caves. Two were recently discovered in Slovenia. Continuous logging of water levels and temperatures in four of these caves was established in spring 2005. Te paper presents and briefy discusses the frst results obtained in three of them. Te results are indicating a fast passage of a food wave along a well devel-oped conduit system. key words: karst hydrology, aquifer, food pulse, Reka, Kras, Slovenia. Izvleček UDK 556.3 (497.4-14) Franci Gabrovšek & Borut Peric: Spremljanje poplavnih valov v epifreatični coni kraškega vodonosnika: Primer reke Reke, Kras, JZ slovenija Reka Reka, ki ponika v Škocjanske jame, je najpomembnejši alogeni vir, iz katerega se napaja kraški vodonosnik. Doslej so med Škocjanskimi jamami in izviri Timave v Italiji našli pet jam, kjer je mogoče priti do podzemnega toka Reke. Dve so pred kratkim odkrili v Sloveniji. Pomladi leta 2005 se je začelo stalno spremljanje nivoja in temperature vode v štirih od teh jam. Članek predstavlja in na kratko obravnava prve rezultate, pridobljene na treh merilnih mestih. Rezultati kažejo na hitro potovanje poplavnega vala po dobro razvitem sistemu kraških kanalov. ključne besede: kraška hidrologija, vodonosnik, poplavni val, Reka, Kras, Slovenija. INTRODUCTION Te Kras (Classical Karst) plateau has been attracting re-searchers for more than a century. Its aquifer is as com-plex as a karst aquifer can get. A more than 300 m deep vadose zone, huge underground cavities, all possible fow regimes, complex recharge and discharge conditions and complex evolution, enough to believe that the system is far from being resolved. Tis paper presents the frst results of an ongoing efort to put a new stone into the puz-zle of the aquifer of Kras. fig. 1 presents a generalized map and a cross-sec-tion of the Kras plateau and its surroundings. It shows the main geological formations, caves with the active underground fow and the measurement points presented in this paper. Kras belongs to the Adriatic-Dinaric tectonic plate in the region of the Outer Dinarids (Kranjc, 1997). Te folds sink towards NW under Soča alluvium. Te same direction is also followed by the main draining conduits from SE of the plateau. Te area is mostly composed of Cretaceous and Tertiary carbonate sediments. Te depth of the unsaturated zone reaches more than 300 m. Many caves, remains of an old drainage net-work, can be found along its complete vertical profle. 1Karst Research Institute, ZRC SAZU, Titov trg 2, SI-6230 Postojna, Slovenia 2Park Škocjanske jame Public Agency , Škocjan 2, SI-6215 Divača, Slovenia Received / Prejeto: 20.06.2006 ACTA CARSOLOGICA 35/1, 35–45, LJUBLJANA 2006 fRANCI GABROVŠEK & BORUT PERIC Underground Rekacoursa Caveswlth underground Rekaflovv Flysch Statlons / / / Fig. 1: Simplifed map and cross-section of the Kras plateau with main geological formations, caves and measurement points presented in this study. Below the piezometric surface, the structure of the aquifer is largely unknown. An indicator of a well devel-oped conduit system was a sudden collapse in the Reka stream near Gornje Vreme in 1980, at the fysch-lime-stone contact, where around 1 m3/s still disappears underground (Brilly et al., 2002). Te focus of our study is epiphreatic zone, charac-terized by a high fow variability of the Reka river which is the main allogenic input to the aquifer. Te river fows about 50 kilometers on impermeable fysch rocks, con-tinues for another 7 kilometers as a surface fow on a limestone terrain, and starts its underground course at Škocjan caves. It emerges at the Timavo springs in Tri-este Bay Te air distance between Škocjan caves and springs of Timavo is around 33 km. Based on the data of the Environmental Agency of the Republic of Slovenia for the period 1961-1990, the average discharge of the Reka River is 8.26 m3/s. Te ratio between low and high waters reaches 1 to 1700 with the maximum measured discharge 305 m3/s, and minimum 0.18 m3/s. Te springs of Timavo have an average discharge of 30.2 m3/s. Beside main spring, the aquifer discharges through the many 36 ACTA CARSOLOGICA 35/1 - 2006 other smaller springs in the vicinity, many of which are bellow the sea surface. Eforts to reach Reka between Škocjan caves and the springs of Timavo have long history (Kranjc, 1997). At the moment we know fve caves leading to the active sub-surface fow: Kačna cave and Labodnica/Grotta di Trebi-ciano are well known and have already been thoroughly investigated. Recently, three additional caves were pushed down to the depths of active Reka fow: Lazzaro Jerko in Italy; Jama 1 v Kanjaducah and Brezno v Stršinkni dolini in Slovenia. Te river has also been reached through Brezno 3G, which turned out to be another entrance of Kačna cave. In Škocjan caves and Kačna cave it is possible to fol-low several kilometers of the underground river, while only small fragments are accessible in other caves as the confning siphons are not far apart, therefore further ex-ploration is lef to cave divers. for more information on geology, speleology, hy-drogeology and history of exploration and research of Classical Karst and its aquifer refer to (Cucchi et al., 2000; Galli, 1999; Kranjc, 1997; Mihevc, 2001). MONITORING THE fLOOD PULSES IN THE EPIPHREATIC ZONE Of KARST AQUIfERS Cuchi and Zinni (2002) reported on the continu-ous monitoring of physical parameters of the subsurface Reka fow. Tey have measured level, temperature and specifc electric conductivity in Škocjan caves, Grotta di Trebiciano/Labodnica, Lazzaro Jerko and Timavo springs. Based on more than 100 food events, a distinc-tion was made between three principal types of food waves, characterized by the presence or absence of in-fow from the sources that feed the Timavo river system: namely the Reka River, Brestovica basin and Soča-Vipava rivers basin. Te temperature and conductivity changes between Škocjan caves, Labodnica/Trebiciano and Laz-zaro Jerko cave indicate that the underground fow is fast (even more than 800 m/h) and continuous. Teir results indicate a “direct drainage” along Škocjan-Labodnica/ Trebiciano-Lazzaro Jerko. Tey proposed that the aqui-fer of Classical Karst is at the 3rd state of the ford-Ewers’s speleogenetic model (ford & Ewers, 1978). OBSERVATION POINTS Except in Škocjan caves, the underground fow of Reka is not easy accessible. To reach it one must descend between 250 and 330 m down vertical shafs and steep galleries. New entrances are rarely found. During the food events when the water rises and squeezes the air from the voids and fractures, an intense airfow can be detected at some spots at the surface. Tese are the spots where cav-ers start following narrow leads through the vadose zone and hope to enter an easy passage to the river. Normally it takes years of digging and climbing to succeed. Except in Škocjan caves and Kačna cave, all the discoveries have been done this way Between february and October 2005, data loggers were placed in Škocjan caves, Kačna cave, Jama 1 v Kan-jaducah and Brezno v Stršinkni dolini. Data from the frst three caves have been retrieved so far and are presented Škocjanske jame - Martelova dvorana Škocjan caves - Martel's hali Kačna jama ¦ Škocjanski kanal Kačna cave ¦ Škocjan channel in this paper. Te distance between Škocjan caves and Kačna cave is about two kilometres. Te direct distance from Kačna cave to Jama 1 v Kanjaducah is about seven kilometers (see fig. 1). Te instruments were fxed to the underground river banks. In Škocjan caves and Kačna cave the micro location was chosen so that the instruments cannot be damaged by larger pieces of fotsam. fig. 2 shows simplifed sketches of the caves with the positions of measurement points. In Škocjan caves it was fxed at Martel’s lake (P1) at the end of 2.2 mil-lions m3 large Martel’s chamber, in Kačna cave at rapids in Škocjanski kanal passage (P2), in Kanjaduce (P3) and Brezno v Stršinkni dolini it is located at terminal sumps at the end of the caves. Jama 1 v Kanjaducah Fig. 2: Simplifed sketches of the caves with positions of the measuring points. Flood levels based on fotsam occurrence (mihevc, 2001) are marked in Kačna cave. maps of Škocjan caves and Kačna cave were obtained from the Karst Research Institute archive and Slovenian cave cadastre. map of Jama 1 v Kanjaducah was obtained from the web site of Sežana caving club (http://www.brlog.net/jds/kanjaduce.htm). ACTA CARSOLOGICA 35/1 - 2006 37 fRANCI GABROVŠEK & BORUT PERIC INSTRUMENTS To log water level and temperature we use TD Diver pro-duced by Van Essen, a Schlumberger company, Holland (fig. 3). Instruments measure and record pressure and temperature. Recently we have introduced instruments (CTD Diver) which additionally log specifc electric con- range of 100 m. Precision of the level measurements is 0.1% of the full range, i.e. 10 cm in our case. Precision of temperature sensor is 0.1°C. Data from the instrument can be retrieved to computer via optical bridge as shown in fig. 3. Fig. 3: Datalogger in the ofce, connected to the computer via optical reader (lef) and fxed to the wall of the terminate lake of Jama 1 v Kanjeducah (P3) (right). ductivity. Te instruments are autonomous, the autono-my being guaranteed by the life span of internal batteries which is 8-10 y (depending on measurement frequency) and internal memory which can hold up to 24000 read-ings (TD). Tey are easy to program by the enclosed sofware. Te sampling interval is between 5 seconds and 99 hours, sampling can be linear, logarithmic or event based. We used the instruments with the measurement Te pressure sensor is a ceramic transducer there-fore the measured level value is the sum of water and air pressure converted to a water column. for small level fuctuation (e.g. decimeter scale), the levels should be compensated with the barometric measurements of the surrounding atmosphere. Since we are interested in the large scale oscillations, we have not done that. RESULTS AND DISCUSSION Te main input to our system is the sinking stream of Reka therefore the fow hydrograph at the station Cerkvenikov mlin provided by Slovene Environmental Agency is tak-en as an input data into the system. As mentioned, there is a considerable leakage from Reka into the karst aquifer before to the arrival at Škocjan caves which is neglected in the course of discussion. One should also consider the dispersed and concentrated input to the conduits from the karst surface along the entire pathway Te latter was reported by divers who conducted research in the terminal siphons of Škocjan caves. Results for the entire period are shown in fig. 4. During the spring, four food events with the level rise of several meters occurred. A dry period followed with some small scale events in August. fig. 5 presents larger events in a weekly time win-dow. Upper graphs show levels, whilst lower graphs their time derivatives, i.e. rising and dropping rates in m/h. Top axes give dates, bottom axes give time in hours elapsed since the recording started in Škocjan caves (feb-ruary 18th, 2005). Note that the fow at Cerkvenikov mlin is in units of 10 m3/s. 38 ACTA CARSOLOGICA 35/1 – 2006 MONITORING THE fLOOD PULSES IN THE EPIPHREATIC ZONE Of KARST AQUIfERS 18C 5" 16-"fc 14: o 12: c 10: O 8: i- x 2: 22° 20 18 16 14 U12 s-10 H 8 6 4 2 Mar Apr May Jun Jul Aug Sep Date [morrth/day/year] Fig. 4: Upper fgure presents fow rates at Cerkvenikov mlin (C) and levels at points P1-P3 during the whole period. lower fgure shows temperatures at measurement points. Colour code is valid for both graphs. ¦ I I =~ - Cerkvenikov mlin (C) -Skocjan caves (P1) - Kacna cave (P2) : - Jama 1 v Kanjaducah (P3) : u I 11 ^.i-------f^i.L^-jjj 5^^*^ ~ \/ i/f" First event (Fig. 5a), starting on March 27th, is compiled of two food pulses with time lag of a day with peak fow rates at 15 and 25 m3/s respectively. Te level response to the frst pulse was of similar magnitude in all three caves. Second pulse with 25 m3/s did not make a considerable diference in Škocjan caves (P1) and Kan-jeduce (P3), yet the level in Kačna cave (P2) rose for al-most 7 meters. Slow increase of the input fow resulted in a slow increase of the levels. In Kačna cave (P2) the rate of level increase reached 0.9 m/h. Second event (Fig. 5b) on April 9th , started with a 60 m3 pulse which dropped to 20 m3/s in 3 days. Long re-cession of input fow resulted in long recessions of levels in P1 and P2. Tis is the only event when the levels at P3 are above those at P2. One could attribute this to the un-known recharge along the pathway between Kačna cave and Jama 1 v Kanjaducah. 3C&2005 Date [mtrVdcl/yyyy] 3/28/2005 3/30/2005 4/1/2005 B80:00 900:00 920:00 940:00 960:00 980:00 1000:00 1020:00 Elapsed tirne [hh:mm] 1200:00 1220:00 1240:00 1260:00 1280:00 1300:00 1320:00 1340:00 1360:00 Elapsed time [hh:mm] 4/gj/2005 Date [mm/dd/yyyy] 4/27/2005 4/29/2005 5/1/2005 Cerkvenikov mlin (C) (Pl) Kacnacave(P2) Jama 1 v Kanjaducah (P3) 1600:00 1620:M 1640:00 1660:00 1680:00 1700:00 172a00 1740:00 Elapsed time [hh:mm] 5/^8/2005 Date [mm/dd/yyyy] 5/20/2005 5/22/2005 5/24/2005 2140:00 2160:00 2180:00 2200:00 2220:00 2240:00 2260:» 2280:00 2300:00 Elapsed time [hh:mm] Fig. 5: Te evolution of levels and level changes during for major food events. Colour codes are valid for upper and lower graphs. ACTA CARSOLOGICA 35/1 – 2006 39 fRANCI GABROVŠEK & BORUT PERIC Tird event (Fig. 5c) on May 25th was the largest. It comprises a single pulse with a maximum fow of 120 m3. Te responds at P1 and P2 is vigorous. At P2 (Kačna cave) the rate of level rise reached 9 m/h. Maximum level at P1 (Škocjan caves) is 4 m, at P2 18 m and at P3 14 m. How undisturbed the food wave passed the way to P2 can be seen from the kink in the rising limb of hydro-graph which is nicely preserved in the level hydrograph at P2. Fourth event (Fig. 5d) is similar but smaller com-pared to the third event and needs no further discussion at this point. Tere were several small food events following the dry period in August. One of them is shown in the fig. 15. All leve ls show a sharp rise suggesting that the pas-sage of the pulse through the system is little afected by the restrictions. Te next step we take is to plot our results versus input. Terefore, fig. 6 presents the levels at all points in dependence on the fow rates at Cerkvenikov mlin. We shifed the levels back in time with respect to fow to consider the travel time between Cerkvenikov mlin and 40 60 80 Q[mVs] 100 120 18 16 14 12 g 10 = 8 6 4 2 0 rt P- rfirhdš tL • Jama 1 v Kanjaducah 20 40 60 80 100 120 Q[m3/s] Fig. 6: levels in Škocjan caves and Kačna cave (a) and Jama 1 v Kanjaducah (b) in dependence on the fow rates at Cerkvenikov mlin. 40 ACTA CARSOLOGICA 35/1 – 2006 particular measuring point. Terefore, the fow rates at time t are plotted with the levels at time t+ t, where t is the average value given in the Tab. 1. Te choice is rather intuitive and although dubious for several reasons, the results are satisfactory for the frst estimation. Levels in Škocjan caves and Kačna cave show similar behavior below 10 m3/s. Te level rises as the recharge increases according to the relations valid for the open channel fow (Dingman, 2002). When the fow exceeds 15 m3/s, the curve in Kačna cave deviates. We attribute this to the constrictions downstream from the P2 in Kačna cave, which becomes fully fooded when fow exceeds 15 m3/s. fig. 7 presents an extended elevation of the section of the Kačna cave which includes P2. Grey area gives the passage height which is 4-7 m. Te passage is about 15 m wide. Vertical scale above P2 shows levels, each bar rep-resenting 2 m. Dotted line gives the level of P2. To see what happens when part of the channel becomes completely fooded, we have employed a simple numerical model of sloping channel system with restrictions as shown on fig. 8. A system of four rectangular channels, each 400 m long, 5 m wide is subjected to the water input from the lef. Channel 1 and 3 are more than 50 m high, while channel 2 and 4 representing restrictions are 3 and 1.5 m high. We used Storm Water Management Model (V.5) obtained from the US Environmental Protection Agency (http://www.epa.gov/ednnrmrl/models/swmm/index. htm). Te model allows calculations of fow and transport through the system of opened and closed channels as a response to a direct input or an input from the prescribed catchments area. One can apply static, kinematic or dy-namic routing method and thus simulate many scenarios which are relevant for karst, when matrix fow could be neglected. Te model has good potential for further in-depth exploration of food wave passage through a well developed karst system (Campbell & Sullivan, 2002). To a system presented on fig 8. we introduced linear increasing fow rates. Te Q-H graphs at points p and p 1 2 are presented on fig. 9. Initially an open channel fow exists along the whole domain (fig. 8a). Te relation between fow and level for a uniform fow in an open channel can be obtained from the Chezy equation (Dingman, 2002). flow and level are related by the power law ? h ? Q , where n = 6/10 for a uniform rectangular channel (see fig. 9). As the channel 4 becomes fully fooded (fig. 8b) the level at p2 deviates. Now the mass balance at point p2 in channel 3 requires dH A\H)-----= Qin (t) — Qout, (1) dt where h is the level, A(h) is the area of water surface in the channel 3 and Q (t) is the infow into the channel MONITORING THE fLOOD PULSES IN THE EPIPHREATIC ZONE Of KARST AQUIfERS Penasti rov Brzice Škocjanski kanal lOm P2(z=174m) Fig. 7: Extended elevation of the section of Kačna cave that includes our measurement point. grey area presents the channel height, dotted line the level of P2. Remapped from the original survey tables. (Source: Slovenian cave cadastre) 200 400 600 800 1.000 1.200 1.400 1 Distance (m) 0 200 400 600 800 1.000 1.200 1.400 1.E Distance (m) 0 200 400 600 800 1.000 1.200 1.400 1.6 Distance (m) Fig. 8: System of sloping channels with restrictions. Restriction between points 4 and 5 is smaller than one between 2 and 3. A) Open channel fow is active along the whole domain. B) Restriction between points 4 and 5 is fooded. C) Both restrictions are fooded. Fig. 9: levels at points p1 (red line) and p2 (black line) in dependence on the input fow rates Q. letters denote the situations a, b and c presented on Fig. 8. 3. Qout is the outfow to the channel 4, given by Darcy-Weisbach (Beek et al, 1999) equation ^^ — ~ Č2J *^out ~ K*Lout (2) f is the friction factor, L the length of the conduit, S its cross-section, d its hydraulic diameter and g gravitational acceleration. Ah is the diference between entrance and exit levels of channel 4 and Az the elevation diference between the two sides. Combining Eqs. 1 and 2 we get dH(t) J / .. ¦^ T ' Tt S \r) + AZ — L/,„ v J (3) Solution of Eq.3 gives the time dependence of the level, assuming that Qin(t), k and A are known. for arbi-trary input we can fnd numerical solutions. We present model results when a food wave recorded at Cerkve-nikov mlin on May 25th (see fig. 5) passes the system on the fig. 8. fig. 10a presents the level hydrograph at p2 and the input fow hydrograph Qin , whilst fig. 10b the fow-level curve. Te dashed blue line on both fgures denote the rising limb of the hydrograph. Te fow-level curve (fig. 10b) exhibits a hysteresis which can be also observed on the fow-level curves of our real recorded hydrographs (fig. 6), particularly for point P3, but also at P2. One reason is diferent location of fow and level hydrograph. Going downstream, the error we make by applying a constant time lag between the fow at Cerkvenikov mlin and the level at the station increases. We suspect that this is the main reason for the large areas of the hysteresis curves for Jama 1 v Kanja-ducah. Another reason for the hysteretic behavior is food-ing caused by restrictions. for an in-depth study of this behavior we would have to analyse the solutions of Eq. 3, what we are not about to do. To demonstrate this we employ the even simpler model shown on fig. 11. It com-prise of a large 50 m high and 20 m wide channel ending with a 2 m high restriction of the same width. Input is at ACTA CARSOLOGICA 35/1 - 2006 41 fRANCI GABROVŠEK & BORUT PERIC 140 120 100 Sj" 80 51 60 ¦f I 40 20 0 r\\ a) / •~ , > -~ * • i i \ < ^ ~-~ , , Te results are given in figs. 12 and 13. fig. 12a, presents the level response to a linear increase and drop of the fow rates (dashed line) for L = 1 km and diferent 15:00 20:00 25:00 30:00 Elapsed tirne [hh:mm] 35:00 140 120 100 To" "E 80 % 60 E, x 40 20 a) , i ' i i , * , > , a=0.7 rti , v , \ / a=0.9m v\ ' \\ ' // a~1 * \ ' \ x ' —¦? =1.b m^ v \ '^y a 00:00 05:00 10:00 15:00 20:00 25:00 30:00 35:00 Elapsed tirne [hh:mm] . a=0.7 m Fig. 10: model results of the through the system on Fig. 8. a) Dashed line presents the input fow rates, full lines give the level hydrogram h2 b) Flow-level curve. 20 40 60 80 100 120 140 160 Fig. 11: Simple model of a large channel ending with restriction. Te length of the entrance channel is l, a is the height of the restriction, A the surface area of the water prior to it and Q (t) the input fow hydrograph. the lef-hand side and increases linearly from 0 to 150 m3/s between 0 - 24 hours and decreases linearly with the same slope during the second 24 hours. qn[m3/s] Fig. 12: a) Flow and level hydrographs for linearly increasing/ decreasing food wave through the model given in Fig. 11. l=1 km. Diferent lines present results for various apertures of restriction as denoted on the fgure. b) Q-h plots for diferent restriction apertures. heights of the restriction as given on the graph. for a = 2 m we see that the shape of level curve resembles that of the fow. As the height of the restriction decreases the fow through it is more and more inhibited and the level curves become distorted. fig. 12b shows Q-H plots for these cases. Te areas of the curves increase as the height of the restriction decreases. fig. 13 presents case where the restriction height is constant, but the length of the input channel changes from 1 km to 20 km. Te geometry of restriction is con-stant, with the height of 1 m. Te level curves now de- 42 ACTA CARSOLOGICA 35/1 – 2006 MONITORING THE fLOOD PULSES IN THE EPIPHREATIC ZONE Of KARST AQUIfERS 140 120 100 S? "E 80 L 60 E, 1 40 20 a ) 1 X ---- q / 1 X -L= 1km , \ -L=2km , X -L-5km ' \ L-9nkm , , v \ , \ , , x , x ' 1 00:00 05:00 10:00 15:00 20:00 25:00 30:00 35:00 Elapsed tirne [hh:mm] 20 40 60 80 100 120 140 CUm3/s] Fig. 13: a) Flow and level hydrographs for linearly increasing/ decreasing food wave through the model given in Fig. 11. a=1 m. lines present results for diferent lengths of the input channel as denoted on the fgure. b) Q-h plots for diferent lengths of the input channel. pend on the fow-restriction relation given by Eq. 3, and also on the distortion of the input food wave due to its propagation in the open channel. Even though the models give at least qualitative ex-planation of what might happen in the real system, the latter is (un)fortunately not as simple as that. Explorers have been following fow directions and food remains in Kačna cave since the discovery of Reka in the cave back in 1972. Tere are many bypass and overfow passages, two of them just a few meters above and few meters be-low the P2, both joining together and leading to a system of galleries at a higher elevation. Based on the fotsam occurrence. Mihevc (2001) found that extreme foods in Kačna cave reach 130 meters above P2. As the monitoring of flood waves continues it will be very interesting to observe the dynamics of foods that may be of a larger scale than those in 2005. In 2002 an instrument was put into Škocjan caves that measures the water level. It recorded floods reaching 40 meters, much larger compared to a 5-meter flood in the present monitoring. The fastest rise recorded in Škocjan caves was 9 m/h. Unfortunately no such measurements were con-ducted in Kačna cave before 2005, where much steeper rises may be expected. Environment Agency data show that the Reka river discharge varied largely at individual foods. We analyzed seven flood pulses to estimate the ve-locity of their propagation through the observed system by taking the peak to peak distance of the level deriva-tive, i.e. the points of the highest rising rate. Not all foods could be easily analysed this way, as the input hydrograph was rather complex. Results are presented in Tab.1. Date Q [m3/s] Time C-P1 [h] Time C-P2 [h] Time C- P3 [h] 27-Mar-05 15 2.5 3 6.3 25-Apr-05 120 2 3.5 10 1-Jul-05 19 2 2.5 5 5-Jul-05 10.1 2.5 3.5 8 7-Aug-05 10 3.5 4.5 9.5 11-Aug-05 26 2.5 4 7.5 29-Aug-05 19 2.5 3 5 Average travel time: 2.5 3.4 7.3 Tab. 1: travel times of selected food pulses from the hydrograph at Cerkvenikov mlin (C) to martelova dvorana in Škocjan caves (P1), brzice (rapids) in Kačna cave (P2) and terminate lake in Jama 1 v Kanjeducah (P3). As can be seen from the table, about 2.5 hours is needed for the food pulse to reach P1 in Škocjan caves, less than an hour more for P2 in Kačna cave and ad-ditional 4 hours for P3 in Jama 1 v Kanjaducah. An interesting point is that there are no big variations in the speed along the way Kačna cave, which is approximately on the half way between Cerkvenikov mlin (C) and Jama 1 v Kanjaducah (P3) is also approximately at the half time of food pulse travel between C and P3. ACTA CARSOLOGICA 35/1 – 2006 43 fRANCI GABROVŠEK & BORUT PERIC TEMPERATURES Temperature is a parameter which carries much infor-mation on hydraulic and thermal conditions in the karst interior (Genthon et al., 2005; Liedl et al., 1998). Water exchanges its temperature with surrounding rocks on its underground course. The heat flux is proportional to the temperature gradient normal to the water-rock bound-ary. Assuming a good mixing of water, the water-rock temperature difference should decrease exponentially with the length of its underground fiow. The exponential factor is a function of fiow rate, geometry of the channel and the normal temperature gradient. fig. 14 presents the temperature evolution during recession after a large fiood event (see fig. 5b). Daily tem- rates drop and the peak to peak distance increases. Oscil-lations at P3 are hardly observed and vanish when the fow is low enough (e.g. smaller than 0.5 m3/s). further data and analysis are needed to understand the temperature dynamics upon arrival of the food puls-es. Nevertheless, we see from fig. 15 that the levels and temperatures respond simultaneously to a small event on August 11th. Along completely fooded parts the level signal is pressure transferred, and therefore faster than the temperature signal. Simultaneous arrivals of both signals show the absence of such segments, leading to a conclu-sion that an open surface fow of Reka along most of the | 1.4 I 12 — 10 i 0.8 ^0.6 I. 0.4 0.2 S Cerkvenikov mlin (C) -Skoqan caves (P1) -Kacnacave(P2) - Jama 1 v Kanjaducah (P3) &9/2005 5/6/2005 5/13/2005 Date [month/day/year] 8?11/2005 8/12/2005 8/13/2005 8/14/2005 Date [month/day/year] 8/15/2005 Fig 14: Flow, level and temperature following a food event. Fig. 15: level and temperature responds to a small food event Numbers at the temperature curves indicate peak to peak following a period of dry conditions. diference in hours between the temperatures at P1 and P2. way could be expected at least for event of comparable perature oscillation at Kačna cave follows that of Škocjan scales. cave, but its amplitude progressively decreases as the fow CONCLUSION AND fURTHER PERSPECTIVES Te intention of this paper was to present the frst results; therefore conclusion will be rather short. Tere is a fast passage of a food wave through the presented part of the system. To give the relation between the travel time of the water parcel and that of the food pulse further data and analysis are required. Nevertheless some data indicate that the these times are similar at least for small food events (see fig. 15). During low fow, travel times become order of magnitude larger (fig. 14). from the passage of food waves through the system we anticipate a (continu- 44 ACTA CARSOLOGICA 35/1 - 2006 ous) system of large conduits also in the parts which are inaccessible at the moment. Many assumptions are still to be proved. Data are being recorded at all measuring points. Valuable sets of data is expected from Brezno v Stršinkni dolini and caves on the Italian side of the plateau, which are located in the area where the gradient becomes practically f at. further actions include in-depth time series analy-sis, integration of precipitation data, dye tracing of main Reka fow with feld fuorimeters positioned in caves and MONITORING THE fLOOD PULSES IN THE EPIPHREATIC ZONE Of KARST AQUIfERS further numerical modelling of events passing through the conduit-restriction systems with open channel and pressure fow. flooding could be an important factor for the gen-esis of large voids in the studied caves (Mihevc, 2001). Large oscillations of water levels could be important if not crucial factor for the genesis of large voids like sub/ vertical galleries in Jama 1 v Kanjaducah and Brezno v Stršinkni dolini and Lindner’s hall in Grotta di Trebi- ACKNOWLEDGMNETS Tis research would not be possible without an invalu-able work of generations of cavers in all investigated cave. Tanks to cavers from Divača and Sežana for the help with our work. Te study has been supported by the In- terreg project “Monitoring of the underground fow of reka Reka” (Vzpostavitev monitoringa podzemnega toka Reke). REfERENCES Beek, W. J., Muttzall, K. M. K., &Van Heuven, J. W., 1999: transport phenomena. John Wiley & Sons, 329 pp, Chichester, New york. Brilly, M., Mikoš, M., Petkovšek, G., Šraj, M., Kogovšek, J., Drobne, D., & Štravs, L., 2002: Te experimental monitoring of water regime in the Reka river.- Acta carsologica, 31, 65-74, Ljubljana. Campbell, C. W.Sullivan, S. M., 2002: Simulating time-varying cave fow and water levels using the Storm Water Management Model- Engineering Geology 65, 133-139. Cucchi, f.Zinni, L., 2002: Underground Timavo river monitoring.- Acta carsologica, 31, 75-84, Ljubljana. Cucchi, f., forti, P., Marinetti, E., &Zinni, L., 2000: Re-cent developments in knowledge of the hydrogeol-ogy of the classical karst- Acta carsologica, 29, 55-78, Ljubljana. Dingman, S. L., 2002: Physical hydrology. Prentice Hall, 646 pp, Upper Saddle River, N.J. ford, D.C.Ewers, R., 1978: Te development of limestone caves in the dimensions of length and depth.- Cana-dian Journal of Earth Sciences, 15, 1783-1798. Galli, M., 1999: Il timavo, esplorazioni e studi. Vol. 23, Atti e mem. Comm. grotte “E. Boegan”, Societa Al-pina delle Giulie, 195 pp, Trieste. Genthon, P, Bataille, A., fromant, A., D’Hulst, D., & Bourges, f., 2005: Temperature as a marker for karstic waters hydrodynamics. Inferences from 1 year recording at La Peyrere cave (Ariege, france).-Journal of Hydrology, 311, 157-171. Kranjc, A., Ed., 1997: Slovene Classical Karst - “Kras”. ZRC Publishing, 321 pp. Liedl, R., Renner, S., & Sauter, M., 1998: Obtaining infor-mation about fracture geometry from heat fow data in karst systems. In: Jeannin, P.-y.& Sauter, M.(Ed.): modelling in karst systems, Université de Neuchâtel, 143-157. Mihevc, A., 2001: Speleogeneza Divaškega krasa. Vol. 27, ZRC, ZRC Publishing, 180 pp, Ljubljana. ACTA CARSOLOGICA 35/1 – 2006 45