The present study addresses a fluviokarst region in Vâlcan mountains (South Carpathians range), within the catchment of Sohodol valley (Fig. 1). It is a rugged topography area, with several peaks ranging between 1400 and 1500 m a.s.l., while only one, Şigleul Mare, reaches almost 1700 m a.s.l.). On the other hand, the streambed of the Sohodol main valley descends, at its exit from the considered karst domain, down to 400 m a.s.l..
Most part of the considered catchment is covered by beech forests, grassland usually occupying only some of the highest peaks and ridges. The climate is temperate, with a mean annual temperature that declines from 8–9°C at the mountain front, down to about 2°C on the highest ridges. Average annual rainfall amounts to about 900 mm at the mountain front, increasing to 1200–1400 mm at higher altitudes. Two rainfall peaks occur, totalling 130–150 rainfall days/year: the main peak is recorded during March-May, and the secondary one in November-December. Snow falls occur during the cold season, with the snow layer usually lasting for 100–150 days (all indicated climate data are according to Muică 1995). Sudden temperature increases toward the end of February and in March result (Neamu 1998) in early snowmelt that triggers flash floods.
All rock formations in the Sohodol watershed belong to a single geological unit, the Lainici Nappe (Berza et al. 1994; Liégeois et al. 1996), whose basement includes two polymetamorphic groups (Berza 1978) of Late Precambrian age.
The more recent, metasedimentary Lainici-Păiuş Group includes a lower “Carbonate-Graphitic Formation” underlying an upper “Quartzitic and Biotite Gneiss Formation” (Liégeois et al. 1996). In addition, three types of magmatic intrusions were distinguished within the metamorphic formations: (i) dykes and bodies consisting of Late Precambrian leucogranitoids (Liégeois et al. 1996); (ii) Late Precambrian granites and granodiorites building up large plutons (Savu et al., 1971; Berza, 1978); (iii) Early Paleozoic (Pre-Silurian) dykes and sills (“Motru Dyke Swarm”) which include mainly andesites and basaltic andesites (Féménias et al., 2008). Overall, two thirds of the Lainici-Păiuş Group occurrence area was estimated to be occupied by granitoids (Berza 1978), yet a thorough cartographic representation of this setting was provided only for the large Precambrian plutons (Fig. 1).
The older, metavolcanic Drăgşan Group includes (Berza et al. 1994; Liégeois et al. 1996) an Amphibolite Formation (consisting mainly of banded amphibolites). It is widely developed in the hanging wall of a major reverse fault that brings the amphibolites in contact with the Lainici-Păiuş Group (Stan et al. 1979; Berza et al. 1994). It has been however presumed (Berza, 1978) that Drăgşan metamorphics could be present also underneath the Lainici-Păiuş metasediments.
A Paleozoic cover common to both metamorphic groups includes (Stan et al., 1979) low grade formations consisting of chlorite and sericite schists (belonging to the Ordovician and Devonian time-interval), and of carbonate and graphitic rocks (of Carboniferous age).
The Mesozoic cover of the Lainici Nappe starts (Stan et al. 1979; Berza et al. 1994) with Early Jurassic siliciclastic deposits in Gresten facies (conglomerates, sandstones and shales). They are overlain by a thick carbonate series (various kinds of limestones, occasionally dolomitized – Pop 1973) that was deposited during the Middle Jurassic-Early Cretaceous (Aptian) interval, and which hosts the karst features concerned by the present study. The carbonate series is transgressively covered by Middle-Late Cretaceous clastics, some of which are intruded by ophiolite veins and dykes (Stan et al. 1979). Locally, Middle Jurassic-Early Cretaceous carbonates also occur in a thrust sheet that tectonically overlies Middle-Late Cretaceous clastics.
The watersheds at the headwaters of Sohodol trunk stream and of its main tributaries (Măcriş, Şipot, Gropu Sec, Scărişoara, Gropu cu Apă) are developed, over 72 km2, within terrains that include the Lainici Nappe metamorphic and granitic basement and its Paleozoic low-grade cover formations. The indicated stream courses next enter an extended surface occupied mainly by Mesozoic carbonate rocks, where the subsequent 42 km2 of watersheds occur.
The carbonate rocks domain is yet rendered more complex by of an uplifted, island-shaped granite body, which crops out over about 4 km2 and accordingly imposes restricted pathways to the limestone-hosted underground flows along Sohodol valley (Fig. 1).
Sampling sites selection relying on previous information about karst hydrology
Karst drainage features within the concerned section of Sohodol valley have been discussed in a paper published by Iurkiewicz and Mangin (1994). Based on their investigations, they conjectured that a common groundwater flow system was discharging via two large perennial karst springs, Pătrunsa (PAT) and Picuiel (PIC), both located on the right side of Sohodol valley (Table 1, Fig. 1). It was additionally stipulated that Fuşteica (FUS) impenetrable swallet - where Sohodol stream itself loses in its right-side bank part (and in drought periods, the entirety) of its discharge - contributed to the supply of the same groundwater flow network. Moreover, about 300 m downstream FUS swallet, and also on the right side of the valley, it is situated the Cave Downstream Fuşteica Swallet, whose fossil passages intercept, about 500 m after the entrance, an underground stream (sampling point CFUS - Table 1, Fig. 1), which flows into a sump.
Table 1
The sampling sites within the investigated karst area of Sohodol valley
side of Sohodol | Type | Name | code | Latitude N | Longitude E | elevation (m a.s.l) | estimated flow ratea (L/s) | sampling |
Oct 2023 | May 2023 | Mar 2023 | Nov 2022 | Oct 2022 |
right | impenetrable swallet | Fuşteica | FUS | 45°12'29.29" | 23° 7'54.96" | 516 | 250 | x | x | | | x |
cave intercepting an underground stream | Downstream Fuşteica Swallet | CFUSb | 45°12'21.74" | 23° 7'52.17" | 525 | n.e. | x | | | | x |
major impenetrable spring with assumed allogenic supply | Picuiel | PIC | 45°10'37.97" | 23° 7'59.02" | 425 | 100–200 | x | x | | | x |
Pătrunsa | PAT | 45°10'44.40" | 23° 7'56.43" | 430 | 100–500 | x | | | | x |
small impenetrable spring with no allogenic input | Piva | PIV | 45°11'43.80" | 23° 3'13.61" | 910 | 1–10 | x | x | | | |
left | cave intercepting an underground stream | At the Mouth of Valea Rea | CVARb | 45°10'22.82" | 23° 8'10.81" | 415 | n.e. | x | | | x | |
impenetrable spring with assumed allogenic supply | Valea Rea | VAR | 45°10'23.71" | 23° 8'10.53" | 410 | n.e. | x | | x | x | |
small impenetrable spring with no allogenic input | Travertine at Picuiel | TRAP | 45°10'41.40" | 23° 7'59.90" | 430 | n.e. | x | | x | | |
a - according to Iurkiewicz and Mangin (1994)
b - the indicated coordinates correspond to the cave entrance
n.e. - not estimated
Iurkiewicz and Mangin (1994) have published details of an artificial tracer test which had substantiated the interconnection between PIC and PAT springs, but no specific information was provided about tracer tests having checked if FUS swallet (and/or CFUS cave stream) were linked to PAT and PIC major outflows.
In the framework of the present study, water samples have been collected from all the above-indicated sites located on the right side of Sohodol. There have been sampled in addition two sites situated on the valley left side (Table 1, Fig. 1):
-
Valea Rea (VAR) impenetrable perennial spring, which has a far less abundant flow rate than either PAT, or PIC outflows;
-
the underground stream intercepted in the Cave at the Mouth of Valea Rea, which is positioned quite close to VAR spring and about 5 m above it. A 40 m long streamlet (sampling point CVAR) permanently flows along the cave passage, to eventually discharge by the cave entrance.
Iurkiewicz and Mangin (1994) have not dedicated any discussion to the latter two sites, and no artificial tracer test has addressed possible hydrological links that could involve VAR spring and/or CVAR cave stream.
In an attempt to develop and expand the overall karst drainage model previously formulated by Iurkiewicz and Mangin (1994), the present study aimed to identify hydrochemical evidence that would contribute to elucidate underground hydrological connections which involved the six above-indicated sampling sites.
At the same time, central to the present investigation was the circumstance that the relative contributions of the allogenic and autogenic water inputs differed between one sampling site and another. Specifically, since metamorphic and granitic formations built up most of the watershed supplying FUS swallet (Fig. 1), its sinking flow was considered to have a chemical composition largely representative for allogenic, silicate-derived water. Alternatively, it was reasonable o assume that each of the two springs PAT and PIC, and possibly also CFUS cave stream likely consisted of silicate-carbonate water mixtures. Moreover it was assumed, as a first order approach, that also VAR spring and CVAR cave stream discharged a mixture consisting of various proportions of silicate (allogenic) and carbonate (autogenic) water. Investigating the mixing process between allogenic and autogenic inputs also required some information on the chemistry of the pristine carbonate-derived (i.e. autogenic) water. Identifying sampling sites presumably representative in this respect was however not straightforward. One such site was conjectured to be the small perennial karst spring Piva (PIV), which discharged (Fig. 1) from a limestone body isolated on a mountain top that was devoid of any allogenic water inputs. Since large amounts of travertine were noticed to have precipitated from PIV outflow, analogous deposits associated to another small perennial spring (TRAP), were assumed to indicate that also its water was exclusively carbonate-derived.
It is still important to emphasize that in the case of TRAP spring, a hydrological connection with the other sampled sites is highly unlikely, while in the case of PIV spring, such a link is just impossible: those two small springs can therefore be viewed only as proxies of the carbonate-derived (autogenic) water which was assumed to be actually involved in the supply of PAT, PIC and VAR springs and of the CFUS and CVAR cave streams.
Sampling and analytical procedures
A sampling operation including all the previously discussed sites (Table 1) has been conducted on 14–15 October 2023, during a low water stage. Some of the concerned sites, located on the right side of Sohodol valley, had been sampled also on 6–7 May 2023 (high water conditions) and on 29 Oct. 2022 (low water stage); while other sites, from the valley left side, had previously been sampled on 26 March 2023 (high water conditions) and on 26 November 2022 (low water stage) (Tables 1 and 2). Note that VAR spring could be sampled only during low water periods, since during high water, its outlet was flooded by the Sohodol streamflow.
The water samples to be analysed for major, minor and trace elements were collected in Nalgene HDPE bottles. The analytical determinations on water samples filtered by Thermo Scientific Chromacol Polyether Sulphone Syringe Filters (0.45 µm pore size) were conducted in the Hydrogeochemistry Laboratory of the Emil Racoviţă Institute of Speleology (Bucharest, Romania).
Gran electrometric titration with a HCl solution of 0.05 M concentration was used (Rounds 2006) for determining the total alkalinity. An UV/VIS double beam Lambda 25 spectrophotometer (PerkinElmer, United States) was used for measuring SO4 concentrations according to the ASTM-D516-07 (American Society for Testing and Materials 1995) standard method. The determinations accuracy was measured by means of a SANGAMON-03 (Environment, Canada) certified reference.
Table 2
Chemical characteristics of the sampled waters in Sohodol valley karst area
Side of Sohodol | Sampling site | Sampling date | pH | conductivity | Na | K | Ca | Mg | HCO3 | SO4 | Cl | NO3 | SiO2 | Fe | Sr | Al | Ba | As | Rb |
µS/cm | mg/L | µg/L |
right | FUS | 14-Oct-2023 | 6.31 | 146.1 | 1.757 | 0.806 | 27.0 | 2.17 | 65.6 | 11.20 | 5.21 | 0.898 | 9.39 | 9.0 | 44.2 | bql | 6.40 | 1.349 | 1.014 |
CFUS | 14-Oct-2023 | 5.81 | 152.2 | 1.731 | 0.754 | 27.0 | 2.21 | 80.0 | 11.14 | bql | 1.101 | 9.36 | 36.9 | 43.3 | bql | 6.24 | 1.373 | 0.951 |
PIC | 15-Oct-2023 | 6.12 | 212.0 | 1.446 | 0.475 | 39.4 | 2.64 | 121.9 | 8.98 | bql | 1.488 | 8.18 | 28.0 | 34.7 | 8.46 | 5.73 | 1.332 | 0.618 |
PAT | 15-Oct-2023 | 5.78 | 210.2 | 1.417 | 0.545 | 43.5 | 2.55 | 122.2 | 8.24 | bql | 1.425 | 7.90 | 23.7 | 33.6 | 9.04 | 5.52 | 1.295 | 0.584 |
PIV | 14-Oct-2023 | 6.78 | 334.5 | 0.279 | 0.120 | 59.1 | 8.26 | 223.2 | 4.82 | bql | 1.051 | 2.43 | 58.7 | 24.2 | 13.95 | 4.74 | 1.142 | bql |
FUS | 7-May-2023 | 5.75 | 115.6 | 1.807 | 0.804 | 22.6 | 1.92 | 61.9 | 11.89 | bql | 1.189 | 9.44 | 34.9 | 36.7 | 20.77 | 4.77 | 0.547 | 1.021 |
PIC | 7-May-2023 | 5.93 | 194.9 | 1.329 | 0.482 | 40.0 | 2.52 | 118.8 | 7.32 | bql | 2.082 | 7.24 | 63.2 | 26.1 | 27.75 | 5.18 | 0.238 | 0.630 |
PIV | 6-May-2023 | 6.45 | 306.5 | 0.307 | 0.358 | 58.2 | 7.51 | 198.8 | 3.27 | bql | 1.153 | 1.71 | 71.5 | 19.2 | 7.43 | 6.37 | bql | 0.300 |
FUS | 29-Oct-2022 | 6.08 | 144.2 | 1.880 | 0.918 | 26.6 | 2.21 | 71.9 | 9.95 | 1.19 | bql | 9.58 | 40.0 | 47.6 | 6.49 | bql | 0.503 | 0.955 |
CFUS | 29-Oct-2022 | 6.09 | 159.5 | 1.845 | 0.914 | 29.0 | 2.32 | 100.8 | 9.57 | bql | 1.261 | 9.15 | 44.3 | 46.4 | bql | bql | 0.484 | 0.911 |
PIC | 29-Oct-2022 | 6.20 | 209.9 | 1.516 | 0.644 | 40.2 | 2.66 | 120.1 | 6.56 | bql | 1.737 | 8.13 | 77.7 | 35.6 | 7.46 | bql | 0.326 | 0.657 |
PAT | 29-Oct-2022 | 5.99 | 209.5 | 1.504 | 0.630 | 41.6 | 2.63 | 123.6 | 5.06 | bql | 1.709 | 8.01 | 77.3 | 35.9 | 10.13 | bql | 0.345 | 0.626 |
left | CVAR | 15-Oct-2023 | 6.20 | 268.6 | 0.830 | 0.260 | 58.6 | 1.91 | 167.9 | 9.58 | bql | 1.595 | 5.67 | 42.0 | 39.8 | 6.87 | 14.78 | 0.942 | bql |
VAR | 15-Oct-2023 | 5.89 | 353.7 | 0.677 | 0.288 | 78.9 | 2.03 | 220.3 | 8.64 | bql | 2.030 | 5.07 | 60.9 | 46.0 | bql | 11.40 | 1.138 | bql |
TRAP | 15-Oct-2023 | 6.34 | 286.7 | 0.295 | 0.146 | 75.7 | 1.16 | 212.6 | 9.24 | bql | 1.086 | 3.73 | 60.6 | 26.8 | bql | 6.05 | 1.256 | bql |
CVAR | 26-Mar-2023 | 6.24 | 261.0 | 0.815 | 0.333 | 53.9 | 1.76 | 153.1 | 9.72 | 0.58 | 1.679 | 4.69 | 52.2 | 31.4 | bql | 10.42 | bdl | bql |
TRAP | 26-Mar-2023 | 6.39 | 359.1 | 0.309 | 0.328 | 77.4 | 1.23 | 223.1 | 8.77 | bql | 1.690 | 3.10 | 84.4 | 24.6 | bql | 4.05 | bdl | 0.233 |
CVAR | 26-Nov-2022 | 6.29 | 255.4 | 0.936 | 0.470 | 58.3 | 1.87 | 166.9 | 7.06 | bql | 2.028 | 6.07 | 121.3 | 40.3 | 13.54 | 9.68 | 0.116 | bql |
VAR | 26-Nov-2022 | 6.17 | 324.7 | 0.755 | 0.455 | 74.4 | 1.98 | 216.8 | 7.97 | bql | 2.626 | 5.44 | 156.3 | 46.5 | bql | 7.76 | 0.131 | bql |
Relative analytical uncertainty (%) | 5.0 | 5.0 | 7.6 | 9.3 | 6.5 | 5.0 | 10.0 | 10.0 | 5.8 | 10.0 | 9.3 | 6.7 | 3.4 | 11.6 | 9.8 | 10.3 | 13.6 |
bql – below quantification limit
bdl – below detection limit
For measuring the concentrations of all other chemical constituents there was used a NexION 300S (PerkinElemer, Shelton, CT, USA) ICP-MS system, provided with an S10 Autosampler. The determinations have followed the U.S. EPA (2014) 6020B standards. Determinations of Ca were conducted in dynamic reaction cell (DRC) mode, by using NH3 as a reactive gas. The concentrations of Na, K, Mg and Sr were measured in kinetic energy discrimination (KED) mode by using He as inert gas. The chloride concentration determinations followed the standard mode (U.S. EPA 2014). Reference materials purchased from High-Purity Standards™ (Charleston, SC, USA) were used for calibrations. The contents of Na, K, Mg, Ca and Sr for the laboratory control samples were measured by using NIST Standard Reference Material®1640a, and the Cl contents for analogous samples were measured by using Simulated Seawater Standard (HighPS). The instructions of the ISO 11352:2012 standard were followed for estimating the analytical measurements uncertainty.