An Integrated Approach to Characterising Sulphur Karst Springs: A Case Study of the Žvepovnik Spring in NE Slovenia

Abstract

We present an integrated approach to characterizing the Žvepovnik sulphur spring, comprising detailed basic geological (mapping), geochemical (physico-chemical, elementary), isotopic (δ²H, δ¹⁸O, δ¹³C_DIC, δ³⁴S and ³H), and microbiological analyses. We used a multi-parameter approach to determine the origin of the water (meteoric or deeper infiltration), the origin of the carbon and sulphur, and water retention times. Our special research interest is the origin of the sulphur, as sulphur springs are rare and insufficiently investigated. Our results show that the Žvepovnik spring occurs along the fault near the contact between the dolomite aquifer and overlying shales and volcanoclastic beds. The spring water is the result of the mixing of (1) deeper waters in contact with gypsum and anhydrite and (2) shallow waters originating from precipitation and flowing through the surface carbonate aquifer. The results of δ²H and δ¹⁸O confirm local modern precipitation as the main source of the spring. δ¹³C_DIC originates from the degradation of organic matter and the dissolution of carbonates. We therefore propose four possible sources of sulphur: (1) the most probable is the dissolution of gypsum/anhydrite; (2) barite may be a minor source of sulphur; (3) the microbial dissimilatory sulfate reduction; and (4) the oxidation of pyrite as the least probable option.

1. Introduction

Understanding karst springs in terms of their origin and dynamics poses important research challenges. Characteristics of spring water are the result of the flow of older groundwater and inflows of younger waters from the surface. Therefore, the interaction of groundwater with the aquifer host rocks as well as the characteristics of surface water sources constitute the basis for understanding their origin [1,2]. The karst springs often respond very quickly and intensively to meteorological phenomena; therefore, it is essential also to understand the impact of the composition of precipitation on spring water. Understanding the exchange of water within the geospheres is vital to addressing environmental issues linked to water resource management [3]. Exchanges within the geosphere are also important in karst environments due to the higher degree of vertical connectivity of these systems with the atmosphere and deeper aquifer units through faulting [4,5,6].

To date, only a few natural sulphur springs have been found and characterized in Slovenia, where karst comprises almost half (i.e., 43%) of the country’s territory. The Žvepovnik sulphur spring, locally known as Žviponik, is one of five known natural sulphur springs documented in Slovenia. The name originates from the term “žveplo”, which means sulphur in the Slovenian language. While some sulphur springs such as Žveplenica [7,8] and Žvepleni studenec pri Rihtarju have been known in Slovenia for decades and are protected for their natural value, information on the Žvepovnik sulphur spring was only first published in a tourist guide of the Municipality of Ljubno in 2013. At the initiative of the Municipality of Ljubno, the first expert inspection of the studied spring was carried out in 2016 as part of a research study on sulphur springs in Slovenia. As an exceptional and rare natural phenomenon in the country, the Žvepovnik sulphur spring was also determined to be of protected natural value in 2019. The locations of sulphur springs in the territory of the Republic of Slovenia can be found under the Nature Conservation Atlas (https://www.naravovarstveni-atlas.si: accessed 30 January 2022) Natural Values tab: Žveplenica (ID 686), Žvepleni studenec pri Rihtarju (ID 502), and the Žvepovnik sulphur spring researched herein (ID 80218). In recent years, two more sulphur spring locations are under investigation: submarine and coastal sulphur springs near the city of Izola [9] and Smrdljivec [10].

The integration of different investigative methods is crucial for a comprehensive understanding of spring dynamics. A fundamental geological understanding (lithological and structural characteristics [11,12]) of the wider area provides important information on the backgrounds of the springs for the further planning of water sampling and are crucial for final interpretations. Hydrogeochemistry and isotopic composition provide information about the sources and residence times; water–rock interaction along the flow path helps us understand the sources of the dissolved constituent in water within the aquifer; and the mixing of individual groundwater bodies is also key [13].

Isotopes in the water molecule determined as the stable isotope composition of hydrogen (δ²H) and oxygen (δ¹⁸O) serve as natural fingerprints and allow us to trace the natural origin of the water, which is reflected in its isotope composition as the water circulates through the water cycle [14]. They are commonly used to characterize and trace the source of water, the influence of different processes during infiltration, and the flow of water through the water body or to quantify exchanges of water, solutes, and particulates between hydrological compartments according to different hydrological processes [15,16]. Therefore, δ²H and δ¹⁸O are powerful tools that can be used to track the path of water molecules from precipitation to surface and groundwater, and further to drinking water supply. Many authors have reported on the multi-parameter approach and combine the isotope and other geochemical parameters to address specific questions (e.g., [17] and related papers).

Carbon isotopes are used to assess the origin of dissolved inorganic carbon (DIC), which is the main component in waters draining carbonate watersheds. Concentrations of DIC and its isotopic composition of carbon (δ¹³C_DIC) are governed by processes occurring seasonally in the soil–aquifer system. Changes in DIC concentrations result from the addition or removal of carbon from the DIC pool, whereas changes in δ¹³C_DIC result from the fractionation accompanying the transformation of carbon or from the mixing of carbon from different sources. The major sources of carbon to aquifer DIC loads are the dissolution of carbonate minerals and soil CO₂ derived from root respiration and from the microbial decomposition of organic matter. The major processes removing DIC in aquifer systems is carbonate mineral precipitation [18]. δ¹³C_DIC for deciphering geochemical processes in combination with hydrogeochemcal data and other isotope tracers (e.g., δ¹⁸O, δ²H) have been investigated in many freshwater springs worldwide in different climates [19,20,21,22].

One of the major anions in karst groundwater other than HCO₃⁻ is sulphate (SO₄²⁻). Its origin in groundwater varies, from precipitation, dissolution of evaporites, oxidation of sulphides, and anthropogenic inputs [23]. Different sources of sulphate can be distinguished by the isotope signatures of δ³⁴S_SO4 and δ¹⁸O_SO4 [24,25].

A detailed characterization of the environmental parameters is important to understand the accompanying biota. Underground sulphur habitats can be considered as hotspots of biodiversity [26,27]. The microbiota that colonizes terrestrial springs likely originate from groundwater but may also originate from the surface. The presence and distribution of microorganisms is likely controlled by spring geochemistry and surface and subsurface transport mechanisms [28].

The aim of the present study was to determine geochemical and biological processes occurring in the Žvepovnik spring using an integrated approach comprising basic geologic, geochemical, isotopic, and microbiological analyses. We combined the use of several isotope analyses to determine water origin (meteoric or deeper infiltration), origin of carbon and sulphur in water, and water retention times. Of special research interest was the origin of sulphur, as sulphur springs are rare, remain insufficiently investigated, and are often interesting to the public owing to their characteristic rotten-egg smell. A holistic approach, including the scientific and public conservation legislative approach, is therefore necessary to characterize the springs properly; here, we present one such example.

2. Materials and Methods

2.1. Location and Basic Characteristics of the Spring

The Žvepovnik sulphur spring is located near the town of Ljubno ob Savinji in the Savinja Valley (NE Slovenia). The spring is located in the forested area on the right bank of the stream of Rajserjev graben (WGS84 coordinates: 46.3500° N, 14.8452° E; local national D96/TM system: e = 488,093 m, n = 134,568 m and Z = 460 m) (Figure 1). The spring water was obtained via a plastic pipe into a wooden drinker from where it flows into the Rajserjev graben stream a few meters on. The smell of the spring cannot be detected from a distance, while the distinctive odor of hydrogen sulfide can be detected in its vicinity (from a distance of a few meters). Directly at the outflow of the spring, white and pale-yellow organic sediments and attached microbial biofilm were observed.

2.2. Geological Characterisation

The geological characterisation of the Žvepovnik spring was performed using several methods. The geological setting was compiled from previous work on regional structure, stratigraphy, and basic geological maps (scale 1:100,000). The surroundings of the spring were further mapped (approximately 0.5 km²) at 1:5000 scale using a combination of boundary tracing and the interpolation of point outcrops. The contact between Anisian dolomite (basement aquifer rock) and Oligocene beds (cover impermeable rock) was studied in detail. The outcrop of approximately 10 m² was analysed, two short laterally positioned sedimentological sections (scale 1:50) were logged, and diverse lithologies were sampled. From rock samples, thin sections were made for further microfacies analysis that was manifested under a polarized Zeiss Axioplan2 (Carl Zeiss AG, Oberkochen, Germany) and a Nikon Eclipse E600 (Nikon Instruments Inc., Melville, NY, USA) petrographic microscope. Thin sections were partly stained with Alizarin Red.

In order to analyse the mineral composition of the rock, samples were powdered in agate mortar and analysed using X-ray diffraction using a Philips PW3710 X-ray diffractometer (PAN-Analytical, Almelo, Nederland) with CuKα₁ radiation (with wavelength 1.54060 Å) and a secondary graphite monochromator. The data was collected at 10 kV with a current of 10 mA at a speed of 3.4° 2θ per minute at a range of 2θ 3.01° to 70.01° with in steps of 2θ 0.02°.

2.3. Field Sampling, Physico-Chemical Measurements, Geochemical and Isotopic Analyses of Groundwater

The first field observations and physico-chemical measurements were carried out in March 2016, while detailed research was performed between November 2017 and July 2018 (Table 1). Geochemical and isotopic analyses were performed up to two times, while the measurements of physico-chemical parameters (temperature (T), pH, specific electroconductivity (EC), and dissolved oxygen in mg/L and % (D.O. and D.O.%) were carried out repeatedly from November 2017 to July 2018 every month with a WTW Multi 3430 Multiparameter probe (WTW GmbH, Weilheim, Germany).

Groundwater sampling for geochemical analyses was performed once (Table S2). Water samples for geochemical analysis were preserved in 50 mL HDPE bottles and kept refrigerated until analysed in the laboratory. For cations, the water was filtered through a 0.45 μm nylon filter and pre-treated with HNO₃⁻ to pH = 2 immediately on site.

Samples for isotopic analyses (δ¹⁸O, δ²H, δ¹³C_DIC) and total alkalinity were collected twice (Table S3). For the analysis of total alkalinity and isotope composition of dissolved inorganic carbon (δ¹³C_DIC), water was filtered through 0.45 µm pore-sized membrane filters and stored in 30 mL high-density polyethylene (HDPE) bottles for total alkalinity and 12 mL Labco glass vials with septum, without headspace for δ¹³C_DIC analyses. Samples for oxygen and hydrogen isotopic composition were collected in 50 mL HDPE bottles that were rinsed with sample water prior to filling. Additional water samples for some ions, hardness, and alkalinity (Table S4) were collected for analyses on a portable field YSI 9300 spectrophotometer (YSI, Yellow Springs, OH, USA).

For tritium (³H) (Table S3), the water was sampled once in 1 L HDPE bottles.

The groundwater sample was taken for analyses of δ³⁴S_SO4 and δ¹⁸O_SO4 (Table S3). The water was filtered through 0.45 µm pore-sized membrane filters, treated with 2N HCl solution to keep water at acidic conditions (pH = 2) at all times, and stored in 50 mL HDPE bottles. The BaSO₄ precipitated with BaCl₂.

Water samples were collected aseptically in sterile bottles for microbiological analysis and transported to a laboratory in a refrigerated box.

Dissolved hydrogen sulphide in the spring water was determined immediately after sampling on site according to the manufacturer’s instructions (YSI).

2.3.1. Hydrogeochemistry

Measurements of major, minor, trace elements, and REE were performed at the ActLabs accredited laboratory (Activation Laboratories Ltd., Ancaster, ON, Canada). Elements were analysed (see Table S2) by Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) (ActLabs Code 6 ICP/MS—Hydrogeochemistry), major anions (SO₄²⁻, NO³⁻, Cl⁻, F⁻) were determined by Ion Chromatography (IC), while some of the major cations (e.g., Ca²⁺, Mg²⁺) were measured using an Overrange Inductively Coupled Plasma-Optical Emission Spectrometry (ICP-OES) owing to their excessively high concentrations for ICP-MS.

Limits of detection for the ICP-MS, ICP-OES, and IC methods were generally in the range of a few μg/L and are reported for each element in detail on the ActLabs homepage (https://actlabs.com/downloads/: accessed 30 November 2021 → Geochemistry Schedule of Services and Fees → Actlabs—Schedule of Services—Euro → Hydrogeochemistry Section). A blank sample was also analysed, and all its measured values were below the detection limit. A duplicate sample was also analysed in the laboratory (not discussed in this study, but was measured in the same batch, yielding good quality measurements).

Additional chemical analyses for hardness, alkalinity, K⁺, NH₄⁺, NO₂⁻, NO₃⁻, Cl⁻, H₂S, SO₄²⁻, PO₄³⁻, Fe(tot), SiO₂, Cu²⁺, Cr(tot), Mn²⁺, Ni²⁺, Zn²⁺ were performed using a YSI 9300 spectrophotometer according to the manufacturer’s instructions (YSI) (see Table S4).

2.3.2. Geochemical Modelling

The PHREEQC for Windows program [30] was used to calculate saturation indexes (SI) for minerals and partial pressure (P_CO2). In this program, the LLNL database was used; compared to other databases, it contains the highest number of species considered.

2.3.3. Determination of Isotope Composition of Hydrogen and Oxygen

The isotope composition of hydrogen (δ²H) and oxygen (δ¹⁸O) was determined using the H₂–H₂O [31] and CO₂–H₂O [32,33] equilibration technique. Measurements were performed on a dual inlet isotope ratio mass spectrometer (DI IRMS, Finnigan MAT DELTA plus, Finnigan MAT GmbH, Bremen, Germany) with an automated CO₂–H₂O and H₂–H₂O HDOeq 48 Equilibration Unit (custom built by M. Jaklitsch). The temperature of the water bath was 18 °C. The water vapour trap was cooled to −55 °C. The CO₂ (Messer 4.5, Messer, Ruše, Slovenia) and H₂ (IAEA) gases were used as working standards for water equilibration. Samples of water were measured as independent duplicates. All measurements were performed together with laboratory reference materials (LRM) calibrated periodically against primary IAEA calibration standards VSMOW2 and SLAP2 to VSMOW-SLAP scale [34]. The results are expressed in the standard δ notation (in per mil, ‰) as deviation of the sample (sp) from the standard (st) as:

δ^yX (‰) = (R_sp/R_st −1) × 1000

where δ^yX is hydrogen (δ²H_VSMOW-SLAP, abbreviated as δ²H) or oxygen (δ¹⁸O_VSMOW-SLAP, abbreviated as δ¹⁸O) and R the ²H/¹H or ¹⁸O/¹⁶O, respectively. Results were normalized to the VSMOW-SLAP scale using the LIMS programme. In order to normalize results, we used LRMs, namely W-3869 with defined isotope values and estimated measurement uncertainty δ²H = +2.5 ± 0.7‰ and δ¹⁸O = +0.36 ± 0.04‰ and W-3871 with values of δ²H = −148.1 ± 0.7‰ and δ¹⁸O = −19.73 ± 0.02 ‰. For independent quality control, we used LRM W-45 with defined isotope values and an estimated measurement uncertainty of δ²H = −60.6 ± 0.7‰ and δ¹⁸O = −9.12 ± 0.04‰, and commercial reference materials USGS 45 (δ²H = −10.3 ± 0.2‰, δ¹⁸O = −2.238 ± 0.006‰) and USGS 46 (δ²H = −235.8 ± 0.4‰, δ¹⁸O = −29.80 ± 0.02‰). The average sample repeatability for δ²H and δ¹⁸O was 0.2 and 0.02‰, respectively.

2.3.4. Determination of Total Alkalinity after Gran

In order to measure alkalinity, the water sample was passed through a 0.45 m nylon filter into an HDPE bottle and kept refrigerated until analysed. First, the pH was measured in the lab with a pH meter (Mettler Toledo AG 8603, Schwerzenbach, Switzerland). The total alkalinity was then determined within 24 h after sample collection using the Gran titration method [35] with a precision of ±1%. Approximately 8 g of the water sample was weighed into a plastic container and placed on a magnetic stirrer. A calibrated pH electrode (7.00 and 4.00 ± 0.02) was placed in the sample and the initial pH was recorded. Reagencon HCl 0.05 N (0.05 M) was used for titration. The titration performed using a CAT titrator (Ingenierbüro CAT, M. Zipperer GmbH Ballrechten-Dottingen, Germany). The method has been described in detail by Zuliani et al. [36].

2.3.5. Determination of Isotopic Composition of Dissolved Inorganic Carbon

Samples were stored in glass serum bottles filled with no headspace and sealed with septa caps immediately after collection on the field. The δ¹³C_DIC values were then determined using continuous flow IRMS (Europa Scientific 20–20) with an ANCA-TG preparation module (Sercon Limited, Crewe, UK) for the sample taken in July 2018, while with IsoPrime 100 coupled with the Multiflow preparation module (Elementar, Manchester, UK) for the sample that was taken in November 2017. Phosphoric acid (100%) was added (100–200 µL) to a septum-sealed vial, which was then purged with pure He. The sample (6 mL for ANCA-TG 20–20 and 1 mL for MultiflowBio preparation modules) was then injected, and the headspace CO₂ measured (modified after [37,38]). In order to determine the optimal extraction procedure for bottled water samples, a standard solution of Na₂CO₃ (Carlo Erba reagents, Val de Reuil, France) and a Scientific Fisher sample with a known δ¹³C_DIC value of −10.8 ± 0.2‰ and −4.8 ± 0.1‰ were used. Messer reference gas for the calibration of measurements with known δ¹³C_CO2 −35.4 ± 0.2‰ was also used. The reference material (Carlo Erba solution) was used to convert the analytical results to the Vienna Pee Dee Belemnite (VPDB) scale. The average sample repeatability was 0.2‰.

2.3.6. Determination of Tritium

The tritium (³H) was measured with a liquid scintillation counting (LSC) TriCarb 3170 TR/SL (PerkinElmer, Waltham, MA, USA). Before analyses, the sample was treated and prepared using electrolytic enrichment, which consists of primary distillation, electrolyte enrichment, and secondary distillation. The error of the analysis was ±0.5 TU, with a detection limit of 3 TU.

2.3.7. Determination of Isotopic Composition of Sulphur and Oxygen in Sulphate

The isotopic composition of sulphur and oxygen were measured in precipitate that was obtained from water samples according to standard procedures. Measurements of δ³⁴S_SO4 in δ¹⁸O_SO4 were performed at the University of Arizona in the USA (https://www.geo.arizona.edu/node/153: accessed 30 November 2021). For analysis of isotopes, sulphides and sulphates were converted to SO₂, and the CO₂ and H₂O were removed. The δ³⁴S_SO4 was analysed with a continuous-flow isotope gas-ratio mass spectrometer (CFIRMS, ThermoQuest Finnigan Delta PlusXL, Thermo Fisher Scientific Inc., Waltham, MA, USA). Samples were combusted at 1030 °C with O₂ and V₂O₅ using an elemental analyzer (Costech) coupled to the mass spectrometer. Standardization of the method was based on several other sulfide and sulfate materials from different laboratories and international standards OGS-1 and NBS123, which were used to convert the analytical results to the Vienna-Canyon Diablo Troilite (VCDT) for sulphur in sulphate and Vienna Standard Mean Ocean Water (VSMOW) for oxygen in sulphate. The precision was ± 0.15‰ (1σ) based on repeated internal standards. δ¹⁸O_SO4 was measured as CO with CFIRMS (Thermo Electron Delta V, Thermo Fisher Scientific Inc., Waltham, MA, USA). Samples were combusted with excess C at 1350 °C using a thermal combustion elemental analyzer (ThermoQuest Finnigan, Thermo Fisher Scientific Inc., Waltham, MA, USA) coupled to the mass spectrometer. Precision was ± 0.4‰ (1σ), based on repeated internal standards.

2.4. Microbiological Analyses

Ready-to-use microbiological media Compact Dry plates (Nissui Pharmaceutical, Tokyo, Japan) were used to determine the concentration of heterotrophic aerobic bacteria (Compact Dry TC), coliforms and Escherichia coli (Compact Dry EC), and enterococci (Compact Dry ETC) directly from 1 mL of the water sample. EC, ETC, and TC plates were incubated at 37 °C for 48 h; in addition, TC plates were incubated at 20 °C for 7 days. Concentrations of bacteria were expressed as Colony-Forming Units (CFU) per millilitre.

AquaSnap Total (Hygiena, Santa Ana, CA, USA) was used to measure total ATP content in a water sample to determine both microbial ATP (living cells and particulate matter) and free ATP (non-microbial or dead cells). Results were expressed in Relative Light Units (RLU), where 1 RLU equals 1 fmol ATP per millilitre.

Table 1. Summary list of monitoring parameters, sampling periods, and methodologies.

Parameter	Sampling Period	Method, Instrument Used
Physico-chemical parameters	Monthly, November 2017–July 2018	WTW Multi 3430 Multiparameter probe (WTW GmbH, Weilheim, Germany)
Geochemical analyses (anions, cations)	21 November 2017, 23 July 2018	Inductively Coupled Plasma-Mass Spectrometry (ICP-MS), Ion Chromatography (IC), Overrange Inductively Coupled Plasma-Optical Emission Spectrometry (ICP-OES) (for details see https://actlabs.com/: accessed 30 November 2021) and YSI 9300 spectrophotometer (YSI, Yellow Springs, OH, USA)
Dissolved hydrogen sulphide	21 November 2017 and 23 July 2018	YSI 9300 spectrophotometer (YSI, Yellow Springs, OH, USA)
Isotopic composition of hydrogen and oxygen (δ18O, δ2H)	21 November 2017, 23 July 2018	H2–H2O [31] and CO2–H2O [32,33] equilibration technique; dual inlet isotope ratio mass spectrometer (DI IRMS, Finnigan MAT DELTA plus, Finnigan MAT GmbH, Bremen, Germany) with an automated CO2–H2O and H2-H2O HDOeq 48 Equilibration Unit (custom built by M. Jaklitsch)
Isotopic composition of dissolved inorganic carbon (δ13CDIC)	21 November 2017, 23 July 2018	Continuous flow IRMS (Europa Scientific 20–20) with an ANCA-TG preparation module (Sercon Limited, Crewe, UK)
		IsoPrime 100 coupled with the Multiflow preparation module (Elementar, Manchester, UK)
Total alkalinity after Gran	21 November 2017	CAT titrator (Ingenierbüro CAT, M. Zipperer GmbH Ballrechten-Dottingen, Germany), pH meter
		(Mettler Toledo AG 8603, Schwerzenbach, Switzerland)
Tritium (3H)	21 November 2017	Liquid scintillation counting (LSC) TriCarb 3170 TR/SL (PerkinElmer, Waltham, MA, USA)
δ34SSO4 and δ18OSO4 in sulphate	21 November 2017	Continuous-flow isotope gas-ratio mass spectrometer (ThermoQuest Finnigan Delta PlusXL and Thermo Electron Delta V; Thermo Fisher Scientific Inc., Waltham, MA, USA)
Microbiological analysis	23 July 2018	Biomass (total ATP content), cultivation of microbial indicators (heterotrophic aerobic bacteria, E. coli/coliforms, enterococci)

Table 1. Summary list of monitoring parameters, sampling periods, and methodologies.

Parameter	Sampling Period	Method, Instrument Used
Physico-chemical parameters	Monthly, November 2017–July 2018	WTW Multi 3430 Multiparameter probe (WTW GmbH, Weilheim, Germany)
Geochemical analyses (anions, cations)	21 November 2017, 23 July 2018	Inductively Coupled Plasma-Mass Spectrometry (ICP-MS), Ion Chromatography (IC), Overrange Inductively Coupled Plasma-Optical Emission Spectrometry (ICP-OES) (for details see https://actlabs.com/: accessed 30 November 2021) and YSI 9300 spectrophotometer (YSI, Yellow Springs, OH, USA)
Dissolved hydrogen sulphide	21 November 2017 and 23 July 2018	YSI 9300 spectrophotometer (YSI, Yellow Springs, OH, USA)
Isotopic composition of hydrogen and oxygen (δ18O, δ2H)	21 November 2017, 23 July 2018	H2–H2O [31] and CO2–H2O [32,33] equilibration technique; dual inlet isotope ratio mass spectrometer (DI IRMS, Finnigan MAT DELTA plus, Finnigan MAT GmbH, Bremen, Germany) with an automated CO2–H2O and H2-H2O HDOeq 48 Equilibration Unit (custom built by M. Jaklitsch)
Isotopic composition of dissolved inorganic carbon (δ13CDIC)	21 November 2017, 23 July 2018	Continuous flow IRMS (Europa Scientific 20–20) with an ANCA-TG preparation module (Sercon Limited, Crewe, UK)
		IsoPrime 100 coupled with the Multiflow preparation module (Elementar, Manchester, UK)
Total alkalinity after Gran	21 November 2017	CAT titrator (Ingenierbüro CAT, M. Zipperer GmbH Ballrechten-Dottingen, Germany), pH meter
		(Mettler Toledo AG 8603, Schwerzenbach, Switzerland)
Tritium (3H)	21 November 2017	Liquid scintillation counting (LSC) TriCarb 3170 TR/SL (PerkinElmer, Waltham, MA, USA)
δ34SSO4 and δ18OSO4 in sulphate	21 November 2017	Continuous-flow isotope gas-ratio mass spectrometer (ThermoQuest Finnigan Delta PlusXL and Thermo Electron Delta V; Thermo Fisher Scientific Inc., Waltham, MA, USA)
Microbiological analysis	23 July 2018	Biomass (total ATP content), cultivation of microbial indicators (heterotrophic aerobic bacteria, E. coli/coliforms, enterococci)

3. Results and Discussion

3.1. Geological Characterisation of the Žvepovnik Spring

The spring is situated in the Upper Savinja Valley at the eastern outskirts of the Kamnik–Savinja Alps. Structurally, the spring area belongs to the easternmost segment of the Southern Alps formed by Miocene south-directed thrusting and post-thrusting, eastward-oriented strike-slip movements [39,40]. Today, this segment of the Southern Alps is bounded by the Periadriatic fault zone to the north and the Sava fault to the south. The area between these W–E to WNW–ESE major strike-slip fault zones is characterized by NE–SW-oriented connecting faults that show a relative lowering of the eastern tectonic blocks [41]. A cross-cutting of these tectonic directions is evident also in the wider area of the Žvepovnik spring (Figure 2).

The Kamnik–Savinja Alps are dominated by a thick carbonate succession that originated from the Upper Permian to the end-Triassic. The carbonates built up a near-continuous succession with just a few formations in the Lower and Middle Triassic that additionally contain interlayers of fine-grained clastics [42,43,44]. The local occurrence of Middle Triassic volcanic rocks has also been reported [45,46]. With the prominent disconformity, Mesozoic carbonates are overlain by Oligocene beds. They begin with a locally present Okonina conglomerate that passes upwards into the widespread, deep-marine shales and volcanic rocks of the Smrekovec Volcanic Complex [47,48].

Due to its tectonic structure, Mesozoic succession dominates in the central part of the Kamnik–Savinja Alps, whereas, towards the east dolomite, outcrops in the fault-bounded carbonate massifs are surrounded by Oligocene clastic and volcanic rocks [41,49]. Such a situation is also observed in the Žvepovnik spring area that is dominated by Oligocene fine clastics, though close to the spring Anisian Dolomite outcrops in a small erosional window. Further to the NE, carbonates form the extensive Golte massif [47].

In the near vicinity of the Žvepovnik spring, a sulphidic ore deposit of proposed hydrothermal origin is present (Figure 2). The Lepa njiva antimony deposit, where stibnite is associated with barite and chert, is well studied [50,51]. In older publications, the ore is considered to occur in Upper Permian limestone, but Mlakar [52] pointed out that it also spreads to the Triassic beds close to the Oligocene volcanic and clastic rocks. Associated barite enrichment is documented also in the closely situated Anisian carbonates (Figure 2) [47].

Our detailed mapping of the surroundings of the Žvepovnik spring revealed that Anisian dolomite is in erosional contact with the overlying Oligocene beds that are dominated by fine-grained clastic rocks in this area. The boundary is stratigraphic in the north and the west, whereas, in the south and the east, there are fault contacts. Tectonically disturbed Oligocene beds at the microlocation of the Žvepovnik spring indicate another W–E fault situated there. It runs parallel to the faults that mark the southern margin of the Anisian Dolomite. No cinematic indices were recognized in this fault; however, considering the kinematic behaviour of the parallel faults, a relative downward displacement of the southern block is most probable. A normal fault seems more probable, but strike-slip movements (with beds dipping with strikes different from the strike of a fault plane) cannot be excluded.

Sedimentological study of the contact between Anisian dolomite and Oligocene clastics yields important information on the potential source of the sulphur. The study revealed that the uppermost part of the dolomite is a clast-supported dolomite breccia composed solely of dolomite clasts equal to the underlying massive dolomite. In some clasts, a primary limestone structure is still recognizable. They belong to peloidal/bioclastic wackestone and calcimicrobic boundstone with fenestrae, both microfacies typical for Anisian carbonates [53]. Several foraminifera were recognized, i.e., Earlandia sp., Aulotortus sinuosus, Reophax sp., and Citaella dinarica, the latter indicating the Anisian age of the dolomite (dolomite breccia). Oncoids were also detected. Matrix in breccia is microsparitic dolomite, locally sparitic dolomite, and calcite. Breccia contains veins filled with saddle dolomite and calcite with pronounced twin lamellae. Anisian dolomite originated in shallow seas and was later dolomitized. The uppermost part of the succession was brecciated, presumably due to long exposure prior to sedimentation of overlying beds. The transport of dolomite clasts was negligible. Dolomite and calcite veins probably mineralized from the hydrothermal waters.

Above the disconformity, an Oligocene fine clastic succession material was deposited. These beds consist largely of marl/shale that contain fine quartz grains, muscovite, and plankton foraminifera. Silt laminae and thin beds also occur. A thin, matrix-supported breccia lies in the laterally discontinuous, i.e., channelized, beds (up to 20 cm) that are intercalated within fine clastic rocks. The matrix in breccia is shale/marl with quartz and muscovite. Clasts are polimict and belong to diverse carbonates such as bioclastic packstone, dolostone, and ooidal grainstone, the latter of which typical for the Lower Triassic Werfen formation [53]. Additionally, plasticlasts occur and indicate the erosion of fine-clastic sediment. These clasts show partial dolomitization. The matrix and clasts are replaced by pyrite (proved by XRD analysis) that comprises up to 15% of the rock volume. It occurs in microcrystals that are dispersed inside the matrix, concentrated in clusters, or that replace (partially or completely) particular clasts in breccia. We note that pyrite impregnation is observed solely in basal meters of the Oligocene beds. Calcite and dolomite veins similar to those of the underlying dolomite also occur. The age of the Oligocene beds was assessed in previous studies [48]. This formation originated in a deep marine environment, fine clastic by pelagic sedimentation, and probably by diluted turbidites, or breccia by debris flows. Clasts in breccia indicated that older formations (than Anisian dolomite) were also exposed at the time of the deposition laterally from the investigated outcrop. Carbonate veins indicate hydrothermal waters that run through the cracks in these beds. Hydrothermal origin is also implied for the intense pyrite impregnation of the basal Oligocene beds.

Figure 2. Geological characteristics of the Žvepovnik sulphur spring: (a) Geological map of the wider area with the location of the spring and Sb mineralisation (simplified from [47]), (b) detailed geological map and geological A–B cross-section of the Žvepovnik spring of the nearby surroundings (author: T.P.), (c) photograph and sketch of the sedimentologically analysed contact between Anisian dolomite/dolomite breccia (violet) and Oligocene clastics (light brown—shale/marl, blue—breccia) with positions of investigated samples (red stars); outcrop is located approximately 100 m upwards in the same valley as the Žvepovnik sulphur spring.

Figure 2. Geological characteristics of the Žvepovnik sulphur spring: (a) Geological map of the wider area with the location of the spring and Sb mineralisation (simplified from [47]), (b) detailed geological map and geological A–B cross-section of the Žvepovnik spring of the nearby surroundings (author: T.P.), (c) photograph and sketch of the sedimentologically analysed contact between Anisian dolomite/dolomite breccia (violet) and Oligocene clastics (light brown—shale/marl, blue—breccia) with positions of investigated samples (red stars); outcrop is located approximately 100 m upwards in the same valley as the Žvepovnik sulphur spring.

3.2. Physico-Chemical Parameters of the Žvepovnik Spring

The measurements of physicochemical parameters are presented in Table S1 and Figure 3. The error of the analysis is very low, at 0.26%. The groundwater temperature at the spring was very constant during the complete measuring period, ranging from 10.4 °C to 10.8 °C (on average 10.6 °C), and therefore showed no significant seasonal fluctuations. pH varied from 7.14 to 7.44, with an average value of 7.30. The specific electroconductivity was rather constant, ranging from 635 μs/cm to 723 μs/cm. The only significantly lower value (380 μs/cm) was measured on 12 December 2017—one day after heavy rain with a spring discharge three times higher than usual (Table S1), indicating mixing with precipitation water. Values for dissolved oxygen were relatively low, mostly around 50% (again, the exception is the day with the highest discharge, when saturation was close to 100%; Table S1). However, the values were much higher than in the Žveplenica spring, where very low oxygen saturation was observed (0.93 mg/L or 10% [8]).

Anoxic conditions, i.e., D.O. < 0.5 mg/L, were not observed in the water of the Žvepovnik spring during our investigation (Table S1); on the contrary, the water was oxygenated to some extent, but still contained a detectable amount of hydrogen sulphide (Table S4).

3.3. Hydrogeochemistry of the Žvepovnik Spring

The major geochemical results for the Žvepovnik groundwater are presented in Tables S2 and S3 (supplementary material). Spring groundwater is dominated by HCO₃⁻, SO₄²⁻, Ca²⁺, and Mg²⁺ ions. The Ca²⁺/Mg²⁺ molar ratio (1.4) indicates that dolomite (the major dissolving rock in the catchment area) weathering can be a source of major solutes within the aquifer; however, this mineral is not the only one dissolving, as the ratio greater than 1 indicates higher calcite content, corresponding also to limestone and/or gypsum/anhydrite dissolution. The SO₄²⁻ ion concentration is relatively high (109–116 mg/L). Most trace elements were found to be below the detection limit (Table S2).

In the Piper diagram (Figure 4), the geochemical composition of the Žvepovnik spring is placed very close to the 201 m-deep borehole TO-1 (located approximately 3 km southeast of the spring [54]) and slightly farther than the Žveplenica spring [8], indicating a much higher source of sulphate ion and slightly higher Ca²⁺ content than in the Žveplenica spring. The water in the Žvepovnik spring therefore comes not only from the dissolution of dolomite, as was expected from the geological mapping. We supposed four sources of sulphur, which are discussed further at the end of this chapter, along with the isotopic data results: (1) oxidation of pyrite (enrichment detected at the disconformity between Triassic and Oligocene beds), (2) microbial-driven dissimilatory sulphate reduction, (3) dissolution of gypsum/anhydrite (occurring along with the Upper Permian carbonates and in Lower Triassic rocks) in our case study, and (4) dissolution of barite. The latter is possible locally due to occurring mineral deposits of barite (explained in the geological setting). In Figure 5, the concentrations of Ba²⁺ and SO₄²⁻ ions fall directly among the published water samples [55], making the origin of sulphur from this mineral feasible.

There are some interesting differences between the chemical analyses we conducted in November 2017 and July 2018. In the July 2018 sampling, ammonium was present in the spring, chloride levels were higher, and chlorine was also detected (Table S4). The results of the chemical analyses indicate that the spring may be impacted to some degree by anthropogenic factors or agricultural activities. There is a pasture with grazing livestock in the vicinity.

3.4. Geochemical Modelling of the Žvepovnik Spring

The calculated saturation index (SI) for the most abundant minerals shows a slight undersaturation of calcite and oversaturation of dolomite (

Table 2. Six PHREEQC scenarios (S1 to S6) and spring data with calculated saturation indices for calcite, dolomite, anhydrite and gypsum, and partial pressure of CO₂ gas, given as 10^SI bar. Saturation indices of these minerals and CO₂ (g) are also presented for the speciation of Žvepovnik groundwater. Data for the Žveplenica spring are taken from Zega et al. [8].

Scenario	Minerals	SICAL	SIDOL	SIANH	SIGYP	SIBAR	SICO2(g)
S1	Dol	−0.46	0	−1.94	−1.62	0.54	−1.67
S2	Cal + Dol	0	0	−1.73	−1.4	0.54	−1.8
S3	Dol + Gyp	0.04	0	−0.32	0	1.29	−1.63
S4	Cal + Dol + Gyp	0	0	−0.32	0	1.3	−1.62
S5	Dol + Bar	−0.46	0	−1.94	−1.62	0	−1.67
S6	Cal + Dol + Bar	0	0	−1.73	−1.4	0	−1.8
/	Rajserjev graben	0.18	1.31	−1.91	−1.58	0.51	−2.03
/	Žveplenica	−0.17	0.71	−2.87	−2.55	−0.53	−2.1

). The partial pressure of CO₂ in the water was 10^−1.95 bar (equal to 11,220 ppm), which is 27 times the saturation of atmospheric CO₂ (410 ppm = 10^−3.39 bar). Relatively large undersaturation values were recorded for gypsum and anhydrite, therefore the Žvepovnik groundwater is far from an equilibrium state for these two minerals and the origin of the sulphur cannot be determined from these two minerals alone (Table 2).

Basic geochemical modelling in PHREEQC software was performed for six equilibrium scenarios: S1: water equilibrium with dolomite only, S2: calcite + dolomite, S3: dolomite + gypsum, S4: calcite + dolomite + gypsum, S5: dolomite + barite, and S6: calcite + dolomite + barite (Table 2). Other minerals were not considered due to their relatively low abundance and low solubility, and thus cannot contribute greatly to the major dissolved ions in groundwater. Saturation indices from the Žveplenica spring [8] were also included in this table for comparison. From the results, it is possible to observe that none of the scenarios modelled can definitively explain the minerals in equilibrium, but some models are more realistic than others. For instance, dolomite in the spring water is oversaturated with no calcite or dolomite precipitating. There are many reasons for the non-precipitation of dolomite [56], with the most important ones related to kinetics and crystal structure. The problematic precipitation rates of dolomite compared to calcite are referred to as “the dolomite problem”, with most being described in sedimentological references (a good review was given by [57]). In comparison to the vast amounts of dolomite beds in former geological periods, recent dolomites are forming in very restricted areas. The reasons for this are many, but most authors point to the slow reaction kinetics [58,59]. Consequently, dolomite can be far more oversaturated than the calcite before precipitating. Therefore, the most probable of the four scenarios (Table 2) seems to be the dissolution of dolomite and/or calcite (scenarios S1 and S2). This corresponds to the geological setting and molar ratio of Ca to Mg, which is approximately 1.4, indicating more calcitic dolomites than pure dolomites. The SIs of gypsum and anhydrite in scenario S1 are most similar to the actual SIs recorded in the Žvepovnik spring, with all being quite undersaturated.

However, the dissolution of gypsum and anhydrite cannot be excluded, despite their undersaturation, as discussed further on. The undersaturation of both minerals can be explained by the mixing of deeper-sourced waters, which are in contact with the sulphate-bearing minerals (Upper Permian and/or Lower Triassic), where the water reaches equilibrium with both minerals, and of shallow waters, which are infiltrated as meteoric waters and flow through the Middle Triassic carbonate, predominantly dolomite aquifer. The latter are undersaturated with gypsum and/or anhydrite, though are in equilibrium with calcite/dolomite. As a result, deeper sulphate-dissolving waters are mixed with shallow carbonate-equilibrated waters.

Differences in the partial pressure of CO₂ (Table 2) could be related to biological activity and not only to the dissolution or precipitation of carbonates. A comparison with the Žveplenica sulphur spring shows lower undersaturation with gypsum/anhydrite in the Žvepovnik spring and higher sulphate concentrations. Žveplenica water is also related to greater dolomite dissolution.

3.5. Isotopic Characteristics of the Žvepovnik Spring

The isotopic composition of hydrogen and oxygen of the Žvepovnik spring was determined during two different seasons, namely in autumn 2017 and summer 2018 (Table 1 and Table S3). No significant difference was observed between the seasons (Figure 6), with mean δ²H and δ¹⁸O values of −60.3‰ and −8.85‰, respectively, which fit along the global meteoric water line (GMWL; δ²H = 8 × δ¹⁸O + 10; [60]). No monitoring of isotopes in precipitation is performed regularly in this area, but some data do exist for the nearby Velenje basin for the period 2012–2015 [61]. The isotopic values of the Žvepovnik spring fall into the range of Velenje precipitation and were very similar to the average weighted values of Velenje precipitation (−64.9‰ and −9.14‰ for δ²H and δ¹⁸O, respectively). Results indicate slightly modified local modern precipitation as a source of the Žvepovnik without modifications due to water evaporation or other geochemical processes that would change the hydrogen and oxygen isotopic signal. We also compared our results with other available data for similar sulphur springs [7,8,10] and groundwater from karst-fissured aquifers from central Slovenia [62] and Pliocene and Triassic boreholes from the Velenje area [63]. Isotopic data from the Žvepovnik spring were similar to the mean data values from karst-fissured aquifers in central Slovenia [62] and the Sovra artesian borehole [8] and are typical for groundwater in shallow Slovenian aquifers where precipitation represents the primary source of recharge [64].

Boreholes made in Pliocene sediments from the Velenje area [63] typically have more negative values that plot slightly higher left above the GMWL and may be due to recharge under different (cooler) climatic conditions (e.g., paleorecharge), recharge of predominantly winter precipitation, or from higher elevations [61]. δ²H and δ¹⁸O groundwater values from Triassic boreholes from Velenje area were slightly more negative than our data indicated [63], which likely reflects the mixing of precipitation and shallow groundwater. In contrast, average δ²H and δ¹⁸O values of the Žveplenica sulphur karst spring [8], the Studenec karst spring [7], the Izola submarine and terrestrial sulphur springs [9], and the Smrdljivec sulphur spring [10] (located in the western part of Slovenia, Figure 1) are typically more positive and also plot all along the GMWL, indicating the predominant influence of local precipitation that changes across Slovenia in space and time [65,66].

δ¹³C_DIC for groundwater from the Žvepovnik spring was measured in two seasons: November 2017 (autumn) and July 2018 (summer). Alkalinity (measured only in November 2017) was measured as 5.4 mM and δ¹³C_DIC value was −12.5‰. Alkalinity and δ¹³C_DIC values (Figure 7) indicate that the Žvepovnik spring has a δ¹³C_DIC value characteristic of karstic and fractured aquifers in Central Slovenia [63]. The Žveplenica sulphur karst spring from the Trebuša valley had similar δ¹³C_DIC values as the Žvepovnik spring, which was enriched with ¹²C isotope by 0.6‰ and had a higher alkalinity (about 1 mM) than the Žveplenica karstic spring [8]. The Izola sulphur submarine and terrestrial springs sampled in 2020–2021 in spring and autumn yielded different δ¹³C_DIC and alkalinity values (Figure 7); the submarine springs originated in carbonate rocks, while the terrestrial springs originated in flysch rocks [9]. Biogeochemical processes were calculated as follows: line 1 (with a value of 1.2‰): dissolution of carbonates according to the average δ¹³C_CaCO3 (2.2‰) value—predicted value [19] resulting in 1‰ ± 0.2 enrichment in ¹²C in DIC [67], line 2 (with a value of −12.5‰) non-equilibrium carbonate dissolution by carbonic acid produced from soil zone CO₂ [19], and line 3 (with a value of −18.2‰) open-system equilibration of DIC with soil CO₂ originating from the degradation of organic matter with δ¹³C_soil = −27.2‰ [19]. The Žvepovnik groundwater falls within line 2 (Figure 7) due to the non-equilibrium carbonate dissolution caused by the carbonic acid produced from soil zone CO₂.

The tritium (³H) concentration was 3.85 TU, which is close to the detection limit of the method. The value indicates the mixing of recent and older water, which means a deep source of water mixed with precipitation. The recorded ³H concentrations were comparable with those measured at Žveplenica [8].

Sulphate concentrations in the Žvepovnik spring ranged from 109 to 116 mg/L. Sulphate isotopic data were measured once. δ³⁴S_SO4 and δ¹⁸O_SO4 values were 17.7‰ and 9.4‰, respectively. We compared our results with the available data for similar sulphur springs in Slovenia [8,10,68] and data from the nearby Velenje basin [61]. The results from previous studies, such as the Žveplenica spring, indicated lower sulphate concentrations and negative δ³⁴S_SO4 values compared to Žvepovnik and revealed that the sulphur in the sulphate is generated via the oxidation of igneous sulphate [8]; the results of δ³⁴S_SO4 and δ¹⁸O_SO4 from the springs at Izola generally yielded values characteristic of seawater, and some can be associated with anaerobic bacterial activity [10,68]. δ³⁴S_SO4 values of Velenje basin groundwater returned a wide range of isotopic concentrations, indicating different sources of sulphate [61]. The δ³⁴S_SO4 and δ¹⁸O_SO4 values of the Žvepovnik spring, however, fell into the range of groundwater values from the Velenje basin, and these values offer several possible explanations for the sulphur and the mixing of waters (deep and shallow) in the spring. The deep water at the Žvepovnik spring gives us information about the origin of lithospheric sulphur. The isotopic values of sulphate fell within the range characteristic for the dissolution of gypsum/anhydrite (evaporites) [69] occurring along with the Upper Permian carbonates and in Lower Triassic rocks in our case study (Figure 8).

3.6. Microbiological Characteristics of the Žvepovnik Spring

ATP as a criterion for microbial biomass indicated concentrations of microorganisms in the range of 10⁵ CFU/mL (24 RLU) based on Hygiena Biomass Estimation (

Table 3. Results of microbiological analyses (date of sampling 23 July 2018).

Parameter	Unit
Adenosine triphosphate-total	RLU	24
Bacteria (37 °C)	CFU/ml	33
Bacteria (20 °C)	CFU/ml	142
E. coli	CFU/ml	0
Coliforms	CFU/ml	7
Enterococci	CFU/ml	3

). Isolates indicative of enterococci were detected in the spring water. Their presence in water usually indicates poor sanitary conditions, but they can also be found on plants and are capable of reproducing in extra-enteric environments [70]. Future studies should provide a more extensive insight into the microbiology of the spring, focusing on faecal indicators and possible measures to protect the spring recharge area.

3.7. The Origin of Sulphur in the Žvepovnik Spring

Complementing the regional geological, geochemical, and isotopic data, the results show the main source of the Žvepovnik spring to be local precipitation, which undergoes different infiltration. Therefore, the spring water represents a mixing of (1) deeper waters, which are in contact with the gypsum and/or anhydrite (proven by the results of δ³⁴S_SO4 and δ¹⁸O_SO4), as both minerals appear regionally in both Upper Permian carbonate rocks and Lower Triassic carbonate–clastic successions [42,71], and (2) shallow waters, which flow through the Middle Triassic carbonate aquifer, which outcrops on the surface in the immediate proximity of the Žvepovnik spring. Both types of water are likely in geochemical equilibrium with gypsum and/or anhydrite in the deeper aquifer and also with dolomite (and calcite) in the shallow carbonates and become undersaturated with gypsum and/or anhydrite when mixed near the surface, as the shallow carbonates are more permeable [72,73,74]. Carbon in the water comes from the non-equilibrium carbonate dissolution by carbonic acid produced from soil zone CO₂, as confirmed by δ¹³C_DIC values. The dissolution of barite is also a possible source of sulphur, as the concentrations of both Ba and SO₄ plot directly among the samples of water-dissolving Ba as a function of dissolved sulphate [55]. Barite was found in the Lepa njiva ore deposit, which is expected to occur on a larger scale in carbonate rocks just under oligocene vulcanoclastes, including areas closer to the Žvepovnik spring [50,51,52] (Figure 9). Some hydrogen sulphide in the spring can also be linked to microbial dissimilatory sulfate reduction. One direct proof of the sulfate reduction process is the formation of reduced inorganic sulfur compounds (e.g., as sulfide minerals FeS₂) which were found in the geological profile (Figure 1). The fourth possible origin of the sulphur is the oxidation of pyrite into sulphate. We assumed that the last two might contribute to the total amount of sulphur compounds in smaller quantities, despite the large amounts of pyrite found in the outcrops and detected in thin sections.

4. Conclusions

Based on the holistic approach used, including the geological, geochemical, isotopic, and microbiological data, the characteristics of the Žvepovnik spring were determined. The results represent an important upgrade in our understanding of the complexity of groundwater flow and mixing within carbonate environments. Namely, the Žvepovnik spring indicates a permanent, deep sulphur-rich source that is diluted by precipitation-responsive, near-surface dolomitic groundwater.

The Žvepovnik spring occurs near the contact between the Triassic dolomite (carbonate aquifer) and the overlying Oligocene shales and volcanoclastic beds (barrier). The microlocation of the spring is bound to what we assume to be a normal fault of unknown displacement, but one with a hydrogeological role (either as a barrier or a zone of increased permeability).

The physico-chemical parameters of the Žvepovnik spring are generally constant. The mixing of sulphide-containing water with the near-surface groundwater is indicated by lower specific electroconductivity and higher oxygen saturation after a rain event.

Geochemically, the Žvepovnik spring indicates that dolomite (and/or calcite) weathering could be the main source of the major solutes within the aquifer. Based on the geochemical modelling, we can conclude that the spring groundwater is a mixture of sulphate-dissolving waters (the water is in equilibrium with gypsum and/or anhydrite) with carbonate-equilibrated waters (undersaturated waters with both minerals, far from equilibrium).

δ²H and δ¹⁸O values indicate local precipitation as the main source of water for the Žvepovnik spring. δ¹³C_DIC originates from the degradation of organic matter and the dissolution of carbonates. ³H concentrations indicate the mixing of recent and older water. The isotopic values of sulphate (δ³⁴S_SO4 and δ¹⁸O_SO4) indicate the dissolution of evaporites.

We therefore propose four possible sources of the sulphur compounds in the Žvepovnik spring. The first and most likely is the dissolution of gypsum/anhydrite, which occurs in the Upper Permian and Lower Triassic succession. As a result of dissolution from barite mineral deposits (known in the wider area), the barite may represent a second source of sulphur—possible, though contributing only small amounts. Although significant quantities of pyrite were detected in the outcrops around the spring there, are two potential sources for the sulphur compounds—microbial driven dissimilatory sulphate reduction and oxidation of pyrite into sulphate. Considering both the geological and isotopic compositions, all four alternatives are possible.

The coordinated and integrated geochemical investigation of the Žvepovnik sulphur spring provided crucial information about the spring’s properties that will serve science and the local community. It is important to conduct such investigations as soon as such natural features are discovered—especially in vulnerable karst areas—in order to adopt appropriate management plans.