Research on Dinaric karst aquifers is usually limited to analyses of the geological setting through the study of water occurrences at the surface (springs, ponors, surface streams) and of preferential underground flow paths [1
], but is rarely conducted using boreholes. The research area for the present study is located in the northern part of the Dinaric karst region, which is one of the world’s locus typicus karst landscapes (Figure 1
). The area is characterized by very deep karstification predefined by tectonics, as well as by extremely irregularly distributed dissolution of the carbonate rocks. The most important challenges at defining the direction of water flow in such a large catchment area are related to the delineation of catchments and subcatchments of the Dinaric karst aquifers [3
]. In the water resource management in the Dinaric Karst, according to available quantities, hypsometric position and hydrogeological and morphological characteristics of the terrain, water is often collected in several reservoirs. Then, based on prognostic and runoff models, water is sought to be optimally utilized and, if possible, passed through several hydropower plants (HPP Peruća, HPP Orlovac, HPP Đale, HPP Kraljevac and HPP Zakučac on the Cetina River; Sabljaci and Bukovnik reservoirs for the needs of HPP Gojak and HPP Lešće on the Donja Dobra River). Around half of the Croatian territory is situated in the Dinaric karst belt, with 22 hydroelectric power plants (HPP) in the area and among them 14 large HPP with over 10 MW of available power.
The aim of the research was to increase the share of exploited water to generate more electricity in the power plants of Gojak and Lešće, which are located on the Donja Dobra River. The secondary aim was also to reduce the risk of floods in the area of Ogulin (Figure 1
The Drežničko Polje is a typical karst polje (field), where a natural lake is formed periodically in the ponor (swallow-hole) zone in the southeastern part of the area. In accordance with established underground connections, water from these swallow holes drains to the Ogulinsko Polje through the Velika Kapela Mountain massif. The main spring in the Ogulinsko Polje is that of the Zagorska Mrežnica, which represents one of the focal points of this research. The maximum yield of this spring, together with nearby springs in the narrow zone of discharge, is 45 m3/s. These include three major and several smaller springs along the southern edge of the artificial hydro-technical lake of Sabljaci. Permanent springs include the Zagorska Mrežnica (used for the public water supply of the city of Ogulin, with a flow rate of 200 l/s) and the Bistrac, while the Pećina spring is intermittent.
The distance between the swallow holes in the Drežničko Polje and the Zagorska Mrežnica spring is 10.4 km, and the difference in altitude amounts to 117 m (the swallow holes are at 436.4 m a.s.l. and the spring is at 319.3 m a.s.l.). Immediately downstream of the spring zone the artificial lake of Sabljaci is located, the primary function of which is the production of electric power. The useful volume of the lake is 33 × 106 m3, which is not sufficiently large to receive water from Zagorska Mrežnica spring during the rainy part of the hydrological year. As a result, a considerable portion of water overflows the dam (an annual average of 4.09 m3/s with a maximum of over 80 m3/s, whereas flows of over 40 m3/s are a regular seasonal occurrence) and flows underground into the adjacent catchment of the Tounjčica and Dobra Rivers and does not pass through HPP Gojak and HPP Lešće turbines. An expansion of the volume of the Sabljaci reservoir is not possible due to the topographical features and the urbanization of the surrounding area.
Several times a year the Drežničko Polje turns into a natural karst retention with an average annual total capacity of around 250 × 106 m3 of water, representing around 7.6 times the volume of the Sabljaci reservoir. The maximum level of retention in the Drežničko Polje was 459.5 m a.s.l. (spring, 2013), with an acquired retention volume of 60.9 × 106 m3 and a water column of around 20 m.
In order to increase the amount of exploited water for electricity production, efforts were made to investigate the feasibility of extending the controlled retention of flood water in the natural retention feature of the Drežničko Polje through a variety of hydrotechnical interventions on major swallow holes in the non-vegetation period (from 15 October to 15 April). This would reduce and slow down the underground flow of flood waters from the polje, whilst enhancing the regime of the Zagorska Mrežnica spring. The top of the water wave would be lower and the duration of relatively high levels of water in the Drežničko Polje would be extended. Such levelling out of the flow curve at the springs in the Ogulinsko Polje would enable better energy utilization, while seasonal flooding in the Ogulinsko Polje would be reduced.
For the purpose of this project modelling of the runoff from the Drežničko Polje toward the Zagorska Mrežnica spring was carried out using the Hydrologic Engineering Center-Reservoir System Simulation (HEC-ResSim) model. The emergence of extreme flows, the analysis of hydrological trends, definition of catchment area and the impact of an inter-basin water transfer on the river basins of Dinaric karst were the focus of studies by Bonacci and Andrić [7
], Bonacci and Andrić [8
], Pavlić et al. [9
], Pavlić and Jakobović [10
], Pavlić and Parlov [11
], Lončar et al. [12
]. Precisely because of the expected extreme flows, this procedure is justified in order to reduce the flooding of the Ogulinsko Polje. In the Croatian karst, the correlation analysis of time series was performed in the works of Jukić and Denić-Jukić [13
], Žganec [14
], Jukić and Denić-Jukić [15
], in the neighboring karst catchment by Terzić et al. [4
] and in karst of Italy by Fiorillo and Doglioni [16
The paper is based on the research conducted in 2013 and 2017, which are a continuation of research from 2005, 2004 and earlier. The results obtained are not published, but are documentation of Croatian Geological Institute, Faculty of Mining, Geology and Petroleum Engineering at University of Zagreb, Elektroprojekt-Zagreb and others. Due to all of the above, it is difficult to distinguish in the paper what is the result of recent research and what is previously conducted research.
4. Results and Discussion
Underground flow tracings are one of the key sources of data for the definition of catchment borders in karst areas. Tracings in the Jasenačko Polje during medium water levels demonstrated the existence of a groundwater connection between springs of the Zagorska Mrežnica and the spring in the Krakarsko Polje. Repeated tracings in the Jasenačko Polje during low groundwater levels showed a direct connection only with the Zagorska Mrežnica spring. Tracing experiments from other karst poljes showed a direct connection only with the Zagorska Mrežnica spring (Table 2
A new tracing was carried out for the present study from the Bosnića Ponor in the Drežničko Polje (Figure 4
) during the period of high water levels. The tracer was prepared using fire hoses and was directly poured into the ponor, although that particular portion of the polje was flooded. In so doing, the intention was to show how the ponors in this area function under these conditions. As the tracer was not recorded in either of the springs on the opposite part of the Velika Kapela Mountain, and it was greatly deluted in the vicinity of the insertion site (lake above ponor), it was determined that during high groundwater levels the gradients were created from the Velika Kapela Mountain toward the Drežničko Polje while the ponors, then situated at the lake bottom, functioned as springs. In this way, the occurrence of estavelles was demonstrated and proven.
By considering all tracing data, it may be concluded that the apparent groundwater flow rates vary from 0.86 cm/s to 2.9 m/s during medium groundwater levels. During high groundwater levels, groundwater from the catchment flows (drains) only from the mountain massif of the Velika Kapela Mountain in the spring zone (the direct catchment) while the poljes from the other side of the Velika Kapela Mountain, together with the related karst aquifer, are entirely saturated with water. Natural retentions appear in the karst poljes, forming lakes. Due to the lowering of the groundwater levels, these lakes and the surrounding rock mass begin to drain until all the lakes are drained off. Groundwater levels then drop considerably below the poljes and groundwater flows toward springs in the Ogulinsko Polje.
Results of conducted borehole investigations in the Drežničko polje determined the thickness of the Quaternary sediments, the fracturing and karstification of the rock mass beneath them, and the position of major caverns in the carbonate bedrock of the karst polje, as well as the definition of the direction of groundwater flows. Morphological anomalies in the relief in the form of an elongated succession of sinkholes, meanders or abrupt divergences of the bed of the Jaruga stream (Figure 4
) indicate the presence of minor faults in the polje, as well as branches of determined marginal faults and faults with different strikes. From the results of borehole drilling, the dimensions of vertical cross-sections of the conduits clearly vary from 0.2–3.0 m and the majority of these are located in the depth interval between 435 and 399 m a.s.l. (5–41 m below the polje surface).
The water regime of the Drežničko Polje during the hydrological year has two states. In the non-vegetative (colder) period, from 1 October to 31 March, the polje is regularly flooded. In the vegetative (warmer) period, from 1 April to 30 September, the polje is regularly dry, with rare, short-term flooding.
According to the results of the runoff model at the hydrological station in the Drežničko Polje on the Jaruga watercourse, the flows in the Drežničko Polje in the non-vegetative period reach 90 m3/s, with the water level in the field at around 459.5 m a.s.l.
In Figure 5
, cross correlation functions between the flows in the ponor zone of the Drežničko Polje and the Zagorska Mrežnica spring (CCF DM) are shown, as well as between precipitation data and the flow in the ponor zone of the Drežničko Polje, and precipitation data and the flow in the Zagorska Mrežnica spring (CCF RM) for the winter and summer parts of the hydrological year.
The autocorrelation functions of the Zagorska Mrežnica spring (ACF M) show that the signal is autocorrelated with itself for the winter and summer periods. The winter ACF M shows no memory loss, whereas the summer ACF M shows memory loss after a lag time of 26 days, where the value of the ACF falls below 0.2, according to Mangin [38
]. This result is consistent with the fact that the karst aquifer is constantly saturated with groundwater during the winter part of the year, and that it is not uniformly saturated with water during the summer part of the year. A lag time of 26 days, after which the memory effect is lost, indicates the existence of underground caverns that have a retention effect on the flow of water during the summer part of the year. During low groundwater levels, the main water table (the phreatic zone) significantly lowers, amounting to a decrease of tens of meters in the Velika Kapela Mountain massif (Figure 2
). In the vadose zone, this probably results in many “hanging” aquifers (limited aquifers above the water table, or caverns filled with water that drains slowly through fractures within the rock mass) that slowly infiltrate towards the phreatic zone.
Cross correlation functions between flows in the ponor zone of the Drežničko Polje and the Zagorska Mrežnica spring discharge for the summer part of the year have a steeper decrease with lag time, as the flows through the ponors are abrupt. A maximum CCF value of around rxy(k) = 0.94 is generally obtained for a lag time of one day. For the winter part of the year, CCF do not show as steep a decrease with lag time, because flows through the ponor zone are more permanent. A maximum CCF value of around rxy(k) = 0.85 is generally observed for a lag time of two days. The CCF DM reveals one common property in that the CCF never reaches a zero value during the winter and summer parts of the year, which is explainable by noting the fact that the Zagorska Mrežnica spring never dried up. This observation, in conjunction with the results of dye tracer experiments during high groundwater levels (which demonstrated that these ponors are estavelles and act as sub-lake springs in that period), suggests that this is mainly the result of the limited conductivity of the karst conduits near the main springs, which causes a significant groundwater level rise in the entire Velika Kapela Mountain massif.
Cross correlation functions between the rainfall data and the Zagorska Mrežnica spring (CCF RM) show a maximum correlation coefficient for lag times of around two days for the winter part of the year, and from 2–4 days for the summer part of the year.
In accordance with the great number of mapped ponors, suffosion depressions and sink holes in the Drežničko Polje, as well as with results of exploratory drilling, the rock substrate of the polje may be qualified as strongly karstified. In addition, from the dynamics of flooding and dry periods, the karstification may be estimated to decrease considerably with depth. An abrupt increase of the gradient is related to the polje margins with the main swallow holes.
The maximum capacity of the main swallow hole zone, measured during the surface flow in the Jaruga stream with free sinking into the swallow holes, amounts to around 22 m3
/s. The total maximum recharge of water into the Drežničko polje during the rainy period and/or in the snow melting season, is estimated to be 90 m3
/s according to the model. In the period of high inflows (i.e., the increasing portion of the discharge curve) the spring capacity exceeds the capacities of ponors. The karstified underground rock mass is filled first, and following this the polje is flooded and turns into a lake which can contain up to 60.9 × 106
of water. According to the hydrological analysis, the biggest floods in the field occur from the middle of October to the end of February, and minor floods occur during the spring months of April and May. Floods usually last from a few days to around twenty days, with an increase of water in the field to a maximum of 20 m (459.5 m a.s.l., April and May of 2013). During the greater part of the year when the field is not flooded, the carbonate underground rock mass is more or less totally saturated by groundwater. From the observations of the groundwater levels in the 26 piezometers (Figure 4
), and from the water level measurements of the Jaruga stream, the flooding of the polje begins after the filling of the “underground karst retention”. When the water level of the Jaruga stream rises from 436.4 m a.s.l. (the bed of the stream) towards 440 m a.s.l. (as the SE lower part of the polje begins to flood), the karst underground rock mass is abruptly saturated, which is why the surface and groundwater levels align. The entire field is flooded at the level of 445 m a.s.l. The measured water levels indicate that surface flow occurs relatively quickly in the swallow holes of the Potočak and Kolovoz. In accordance with the analysis of the corresponding levels of the flood waters and their descending trend, as well as groundwater levels in the piezometers, it was determined that after the discharge of the surface part of the retention, the discharge of the underground retention begins first in the area of the main ponor zone, and then gradually occurs from the southern part towards the northern part of the polje. The natural retention of the Drežničko Polje functions as a single unit in its underground and surface portions, in a hydraulic sense. This is shown by the overlapping of the water levels of the Jaruga stream above the elevation of 440 m a.s.l. and the levels in all piezometers, regardless of either ascending or descending trends of the flow curve. The greatest difference between the maximum and minimum groundwater levels is found in the zone ahead of the main ponors on the SE margin of the Drežničko Polje. This amounts to 37.72 m and the lowest difference is found along the eastern edge of the middle area of the polje, and amounts to 9.2 m. Regardless of the difference in the water levels, the level curves of all piezometers show very similar frequencies and trends, meaning that the rock mass forms a single unit in a hydraulic sense, while the changes of groundwater levels have a broader regional character (Figure 6
Conduits in the karst underground rock mass are interconnected and their orientations and distribution are in line to a great extent with registered fault zones. The greater portion of the caverns and fractured zones in the rock mass are located in the depth interval between 435 m a.s.l. and 399 m a.s.l. Based on the analysis of the water levels in piezometers (Z-4, Z-5, Z-6 in the field and Z-8, Z-9 on the eastern carbonate boundary), as well as on structural-tectonic investigations, the main zone of discharge from the Drežničko Polje was located in the SE margin where the main zone of ponors were formed. Based on the low water levels in piezometers (i.e., the levels in summer 2003 and summer 2004) the lowest position may be concluded to have decreased to an altitude of 399 m a.s.l. at those times.
The combined data from geological mapping, exploratory drilling and hydrogeological monitoring presented in this study have significantly improved the hydrogeological map of the Drežničko Polje, especially the ponor zone. Tracing experiments in the ponor zone during very high groundwater levels have demonstrated that in such hydrological conditions the ponors act as springs at the bottom of the lake, and are therefore estavelles. Further tracings should therefore be performed during lower groundwater levels, when surface water inflows into the ponors (i.e., medium groundwater levels in the middle of the recession period), and during low groundwater levels using artificial washing of dye into the ponor. The connection of other karst poljes from the SW side of the Velika Kapela Mountain with the Drežničko Polje or directly with springs on the other side of the mountain must be investigated in much more detail and in different hydrological conditions, using different tracers simultaneously.
Hydrological analyses of stage-discharge curves led to the division of the year into two main parts, corresponding approximately to two halves of the year. Auto-correlations gave significantly different results for these two periods, which was explained by the very large groundwater level changes in the karst massif of the Velika Kapela Mountain and the fact that in heterogeneous karst aquifers such as this one, many small limited aquifers and pockets of water remain in the vadose zone when the water table drops rapidly. A similar result was obtained with cross-correlations, which showed that the two periods were also clearly separated and results pointed to limitations of the karst conduits, which cannot conduct such a large quantity of water during high groundwater levels, both on the Zagorska Mrežnica spring side and near the ponors (estavelles) in the Drežničko Polje. The rock mass conductivity is naturally much lower and contributes less in a karstified environment such as this one than karst voids, caverns, ponors, dissolution conduits, or, commonly, preferential flow paths. Similar rain events during high groundwater levels would have a completely different influence than during low groundwater levels. Due to this heterogeneity, standard hydro(geo)logical modelling cannot provide plausible results and numerous different methods should be applied to minimize ambiguity.
From groundwater level monitoring in and near the Drežničko Polje, it was concluded that natural accumulations, such as a lake in the polje, and the groundwater beneath and around it, represent a completely interdependent and connected system. The intensity of the groundwater discharge from the Drežničko Polje at the beginning of water recession could be significantly reduced by placing an injection (i.e., a grout curtain) in front of the main zone of the swallow holes, to the west of piezometers Z-4, Z-5 and Z-6, to a depth of 390 m a.s.l., ending at the southern and eastern margins of the field. The results of this research program, which included all the former findings on the area, have illuminated the very complex karst system of the Velika Kapela Mountain and its surroundings. The main purpose of this work, which focused on the possibility and feasibility of a grout curtain that would partially separate the ponor zone from the rest of the Drežničko karst Polje underground, was thus achieved. Designing such an injection curtain with mobile closers above the three largest ponors in the Main ponor zone and the surface sanation of suffosions of the terrain between the grout curtain and the Main ponor zone in the field would: (1) prolong a relatively high groundwater wave which would increase electrical power production in the Gojak and Lešće HPP; (2) enhance the discharge curve shape on the Zagorska Mrežnica water supply spring (which is comparatively less important as only a small proportion of the discharge is used); and (3) cause a slight decrease in the highest part of the discharge curve, which could have an effect in decreasing the maximal floods in Ogulin and its surroundings.