Introduction

Submarine spring, a term often used to describe point-source conduit flow of submarine groundwater discharge (SGD) in a karst environment (Swarzenski et al. 2001; Fleury et al. 2007; Dimova et al. 2011; Parra et al. 2015; Moosdorf and Oehler 2017; Bakalowicz 2018), is a peculiar coastal feature that provides insight into the functioning of the aquifer system, its relation with seawater, and the geological setup that influences hydraulic flow in the aquifer. Seawater-groundwater mixing, rock-groundwater interaction, and the source of the spring’s water can be investigated by analyzing the water’s geochemical and isotopic data. Knowledge of SGDs is especially important in assessing geochemical matter-related cycles and nutrient input in coastal environments (Slomp and Van Cappellen 2004; Lee et al. 2009; Moore 2010; Pavlidou et al. 2014; Tamborski et al. 2020), SGD discharge and its impact on the global water budget (Taniguchi et al. 2002; Burnett et al. 2006; Kwon et al. 2014), and studying water resources, as SGDs are often useful for public needs (Moosdorf and Oehler 2017).

Karst submarine springs are the most reported form of SGD (Taniguchi et al. 2002) and are particularly common along the Mediterranean coast (Fleury et al. 2007; Bakalowicz 2018). They are usually affected by the intrusion of seawater (Bonacci and Roje-Bonacci 1997; Fleury et al. 2007; Parra et al. 2015; Gemici et al. 2016; De Filippis et al. 2016; Xu et al. 2019; Liso and Parise 2020). As a result, understanding the connection between groundwater and seawater is important for the protection of water resources. The Slovenian coastal aquifer is also influenced by seawater, as Brenčič (2009) reported the mixing of seawater (31%) and groundwater (69%) in the carbonate aquifer under reducing conditions in the Slovenian coastal town of Lucija (Lu-1, Fig. 1b), located some 5 km by air from Izola. Seawater and carbonate groundwater mixing was also reported by Petrini et al. (2013) from nearby thermal waters in Tržič/Monfalcone in Italy (Gulf of Trst/Trieste). Seawater interactions are often a topic in SGD research, where the determination of temperature (T), electrical conductivity (EC), geochemical, and sometimes isotopic composition are common tools used to assess seawater intrusion and to identify main water types (Stüben et al. 1996; Swarzenski et al. 2001; Slomp and Van Cappellen 2004; Charideh and Rahman 2007; Parra et al. 2015; Pétré et al. 2020).

Apart from mixing with seawater, specific geochemical and thermal features are not particularly common in karst submarine springs and are usually associated with volcanic or tectonic activity (Caramanna et al. 2021; Corliss et al. 1979; Michard et al. 1993; Prol-Ledesma et al. 2004; Varnavas and Papavasiliou 2020; etc.). Italy is known for its many thermal sulfuric acid caves formed as a result of the deep circulation of H2S-rich fluids in a karst and evaporite environment (D’Angeli et al. 2019). Based on geochemical and sulfate isotopic data, the Santa Cesarea Therme system in Apulia was found to be recharging from the seawater via localized fault damage zones, circulating at a depth of about 2–3 km at 85 °C, and then interacting with evaporites and limestones, while sulfide is produced by bacterial reduction (Santaloia et al. 2016; D’Angeli et al. 2021). The Crescent Beach springs in Florida are one of the best-known submarine sulfur karst springs. Their recharge is also associated with the fracture zone. Geochemical tracers show that the spring water is the result of a complex mixing of seawater and artesian groundwater (Swarzenski et al. 2001).

Izola, a coastal town in Slovenia (Northern Adriatic Sea), has been known for centuries to have springs of warm, sulfur-rich water (Kramar 2003). Only two decades ago, submarine springs with the same properties were discovered nearby (Žumer 2004, 2008). In addition to the submarine and terrestrial sulfur springs of Izola, only four other sulfur springs are known in Slovenia (all terrestrial): the Žveplenica dolomite spring (Zega et al. 2015; Mulec et al. 2015) with sulfur that may have derived from pyrite (Zega et al. 2015), the Smrdljivec spring with sulfur of possible coal-related origin (Mulec et al. 2021), the Žvepovnik karst spring with sulfur of evaporite or barite origin (Žvab Rožič et al. 2022), and Riharjev studenec (not yet investigated).

The first geochemical research of the submarine sulfur springs of Izola was conducted by Faganeli et al. (2005), who identified the hydrogeochemical characteristics of the spring water. However, said research did not provide sufficient results to fully understand the origin of the water and the hydrogeochemical processes in the Izola karst aquifer.

Isotopic composition can provide further information about the water source and interactions. In addition to physicochemical and geochemical analysis, the isotopic composition of H and O has been used to trace the origin of the water and to compare the isotopic signal with precipitation, as was done in the study of groundwater flow in Zagreb, Croatia (Marković et al. 2013), and Syrian submarine springs (Charideh and Rahman 2007). Stable carbon isotopes are useful indicators of dissolved inorganic carbon (DIC) sources in groundwater systems. DIC is the main species in carbonate environments. Changes in DIC concentrations result from the addition or removal of carbon from the DIC pool, while changes in δ13CDIC values result from fractionation accompanying the transformation of carbon or the mixing of carbon from different sources. Major sources of carbon to aquifer DIC loads consist in the dissolution of carbonate minerals, soil CO2 derived from root respiration, and microbial decomposition of organic matter (Aucour et al. 1999; Li et al. 2005; Kanduč et al. 2007, 2012). Dissolved sulfate in aquatic environments has multiple sources: atmospheric deposition, evaporite dissolution, oxidation of sulfide minerals, soil sulfate, and sewage (Otero and Soler 2002; Rock and Mayer 2009; Gammons et al. 2013). Given that the investigated springs have a strong sulfurous odor and that layers of sulfur-bearing coal are present in the study area (Pleničar et al. 1973; Benedik and Rožič 2002; Brenčič 2009), the sulfur isotope of sulfate (δ34SSO4) was used to identify the source of SO42−, as has been done for some other sulfur springs (Zega et al. 2015; Santaloia et al. 2016; D’Angeli et al. 2021; Žvab Rožič et al. 2022). A combination of δ34SSO4 and δ18OSO4 is widely used and provides more conclusive information for the qualitative tracing of sources of dissolved SO42− in aquatic systems (Zhang et al. 2015, 2021; Cao et al. 2018). The reduction of sulfate to sulfide is usually associated with redox conditions and an acidic environment, which poses a risk to groundwater quality. The origin of sulfur and the contamination of karst groundwater from anthropogenic activities has been successfully determined using hydrogeochemical and isotopic characterization in various studies of karst environments (Dogramaci et al. 2017; Sun et al. 2017, 2021; Tang et al. 2021).

The aim of this study was to expand on our knowledge of the sulfur springs of Izola, which is important in better understanding the conditions of the coastal environment, geochemical cycles, and for planning the use and management of water resources. The main objectives were to determine a geological and hydrogeochemical characterization of the submarine and terrestrial karst sulfur springs of Izola, to identify the potential sources of H, O, C, and S, to determine the origin of the spring water, and to create a conceptual model of water flow in the area. The findings in this study emphasize the importance of using diverse methods in assessments of submarine sulfur springs in order to obtain more conclusive results when mixing with seawater.

Research area

Previous research on the thermal and sulfidic groundwater of Izola

The first mentions of thermal and sulfidic waters in Izola date back to the seventeenth century by Bishop Tommasini. The first analysis of terrestrial spring water was carried out in the nineteenth century by Eduard Papp, who noted the presence of ammonium, sulfur, magnesium, nitrate, copper, and chlorate, as well as a distinctive smell of sulfur. Soon after, thermal baths were built, but they were not in operation for long (Kramar 2003). In 2002, a 501 m deep borehole LIV-1 (Fig. 1c) was drilled in Izola for geological and hydrological analyses of the aquifer and the thermal water. The report noted the presence of coal and organic-smelling water, with a temperature of 23.2 °C at a depth of 500 m (Benedik and Rožič 2002).

Submarine springs with similar physical and chemical properties were discovered in 2002 by Žumer (2004, 2008). He reported that the submarine springs have a temperature of 29.9 °C and a strong sulfurous odor. Water analysis was performed by Faganeli et al. (2005), who reported higher Na, K, Mg, and Sr concentrations than those of groundwater and 10-times lower than seawater. The spring water was enriched in DIC, which, together with lower δ13CDIC values, suggests oxidation of organic compounds in the spring. The presence of sulfide points to the reducing conditions in the submarine spring water (Faganeli et al. 2005).

Location and shape of the submarine sulfur springs of Izola

Submarine sulfur springs of Izola are located in the SE part of the Gulf of Trst/Trieste, about 400–1000 m from the coast of the town of Izola (SW Slovenia) and are clustered into three groups: Izola (springs notations M01, M02, M03), Bele skale (M04, M05), and Ronek (M06, M07, M08, M09, M10, M11, M12) (Fig. 1c, Table 1); the smell of sulfur is also present in the terrestrial sulfur spring (K04).

Table 1 Submarine sulfur spring locations, depths, and sizes; data from Slavec (2012)
Full size table

Submarine sulfur springs flow out of Quaternary sediments in funnel-shaped depressions in the seafloor at depths of 19–32.2 m below sea level, with the deepest depression (M03; 11.1 m) having a diameter of 41 m at the top (Slavec 2012; Table 1) and about 2 m at the bottom (Žumer 2004). The mixing of spring water, seawater, and sediment occurs in the funnel.

Coordinates of K04, VK01, SW01, SW02 and SW03 are available in the Pangaea data repository (Žvab Rožič et al. 2023).

Geological setting

SW Slovenia and Istria are characterized by two major rock associations (Fig. 1b). The basal unit consists of Cretaceous to Lower Eocene limestones of the Dinaric/Adriatic carbonate platform with intercalations of coal layers in the Liburnia Formation of Late Cretaceous–Early Paleogene (Pleničar et al. 1969, 1973; Jurkovšek et al. 2013). The limestones gradually pass through transitional marls into Eocene Flysch, which consist of alternating marl and turbiditic sandstone beds. Sporadic calciturbidites also occur within the flysch, mainly as thin- to medium-bedded, graded calcarenites, but several calciturbiditic/calcidebritic megabeds are also intercalated (Pavšič and Peckmann 1996; Placer et al. 2004; Vrabec and Rožič 2014) and can be used for detailed geological mapping (e.g. Rožič and Žvab Rožič 2023).

Fig. 1
figure 1

a Map of the location of springs on the N Adriatic Sea. b General geological map and cross-section of investigated area (simplified after Pleničar et al. 1973 and Placer et al. 2010). Note that the lithology colors also mark their hydrogeological characteristics, with limestone (green) representing a karst aquifer and flysch (ocher) generally acting as a groundwater barrier. c Detailed geological map of the Izola area with the schematic stratigraphic column: sulfur springs are located at the stratigraphic contact between limestone and flysch on both limbs of the Izola limestone core in areas with thin Quaternary sediment cover (adopted from Rožič and Žvab Rožič 2023)

Full size image

Structurally, the investigated area belongs to the External Dinaric Imbricated Belt, which is characterized by structural shortening in the SW direction (Fig. 1b). The most prominent structural element is a Palmanova/Črni kal thrust fault that runs along the SW slopes of the Kras (Karst) Plateau, which is otherwise dominated by limestone formations. Rather a large number of thrust-sheets are associated with the main thrust fault and, in this zone, limestone and flysch successions are quickly alternating. Further to the SW, the area of the Slovenian coast and Istria shows only minor internal tectonic deformation and, with the exception of the Izola anticline limestone core, consists entirely of Eocene flysch. This area is known as the Istria-Friuli Underthrust Zone. The recognized structures are, from the NE towards the SW, the Buzet thrust fault, the already mentioned Izola anticline, and the Sv. Križ thrust fault (Fig. 1b). This structural unit is confined further to the SW by the Buje fault in Croatia, while the transition from flysch back to limestones occurs slightly earlier in the Sečovlje area. This contact is partly stratigraphic and partly in the form of a vertical fault (Pleničar et al. 1969, 1973; Placer 2005, 2007; Placer et al. 2010).

Cretaceous to Paleogene rock formations are overlain by Pliocene to Quaternary deposits. Below the sea, these deposits are divided into two units. The lower unit consists of the Pliocene–Pleistocene fluvial deposits (mainly sand, gravel, and silty clay) and is occasionally interrupted by marine and brackish sediments. The upper unit is represented by Holocene marine fine-grained clastic sediments (sandy silt, clayey silt, silt, and silty sand) with frequent foraminifera, bivalve, and gastropod shells (Ogorelec et al. 1981, 1987, 1991, 1997; Covelli et al. 2006). The upper unit of the Holocene marine deposits varies in thickness, ranging from 0 m in coastal areas to several tens of meters in the central part of the gulf (Romeo 2009; Slavec 2012; Vrabec et al. 2014; Trobec et al. 2018). At the same time, on the mainland, the Holocene is characterized by fluvial deposits covering valley floors with brackish or (anthropogenic) saline environments at river mouths (Pleničar et al. 1973; Ogorelec et al. 1981; Kovač et al. 2018).

Recent detailed geological mapping of the Izola area has revealed that the Izola submarine sulfur springs occur on both limbs of the Izola anticline limestone core with an axial orientation of WNW-ESE (Fig. 1c). All of them occur at the stratigraphic boundary between limestone and flysch. The springs are found in those areas where the Quaternary deposits are thin enough to be washed out, resulting in a funnel-shaped depression morphology of the springs. The Izola group is situated on the NE limb, while the Bele Skale and Ronek groups are located along the SW limb.

Hydrogeological setting

The research area is positioned at the coastal karst aquifer. Outcropping limestones of the Izola anticline are covered by the town’s infrastructure, therefore the recharge by precipitation is negligible here. The surrounding flysch covers the underlying limestones and represents a barrier with very low permeability. The groundwater in Izola is most likely connected to and recharged from a karst aquifer of Classical Karst in the NE area in Fig. 1b (Petrič et al. 2002). The flysch layers covering the limestones are several hundred meters thick, therefore deep groundwater of the Izola area presents deep regional flow with artesian pressure. There is very little hydrological data on groundwater in the Izola area, and no water table map or groundwater flow pattern are yet available. The piezometric level in the Izola area rises towards the sea and is found at a depth of 36 m in the Izola town center (Benedik and Rožič 2002).

Sampling and analyses

Field sampling and measurements of physicochemical parameters

Field samples of sulfur springs were collected from three submarine sulfur springs (one from each group; M03, M05, M11) and one terrestrial sulfur spring (K04; Fig. 1c). Groundwater samples from a well (VK01) and seawater (SW01, SW02, SW03) were also collected for comparison and interpretation of the source of the submarine sulfur springs. Physicochemical measurements were accompanied by sampling and were conducted in three sampling campaigns: in June and July 2020 on water samples from M05, M11, K04, and SW01, in October 2020 on samples from M03, M05, K04, VK01, and SW02, and in April 2021 on samples from M03, M05, M11, K04, VK01, and SW03 (Žvab Rožič et al. 2023).

Submarine spring water samples were collected by professional divers who sampled the water with 100-ml syringes inserted into the spring opening in the seabed sediment. The seabed sediment covering the limestones from which the springs emerge is silty and loose. The seawater and the spring water mix in the loose sediment. The divers could not reach the spring opening at the limestone because working deeper in the loose sediment is dangerous for divers.

Measurements of the physicochemical parameters of the water samples, such as temperature (T), pH, electroconductivity (EC), and dissolved oxygen (DO), were performed in the field as soon as the divers brought the samples on board using a calibrated Multi 3430 (WTW GmbH, Weilheim, Germany).

Samples for oxygen and hydrogen isotope analysis (δ18O and δ2H) were collected in high-density polyethylene (HDPE) bottles. Samples for hydrogeochemical (cations and anions), dissolved inorganic carbon isotope (δ13CDIC), total alkalinity (TA) and sulfur isotope (δ34SSO4, δ18OSO4) analysis were filtered through 0.45 µm pore-sized membrane filters and stored in 30-mL HDPE bottles and 12-mL Labco glass vials with septum, without headspace for the δ13CDIC analyses. Samples for hydrogeochemical analysis were acidified with HNO3. Samples for sulfur isotope analysis were acidified on site with a few drops of 2N HCl and then treated with BaCl2 in the laboratory to collect filtered BaSO4. All samples were refrigerated until further analysis.

Analysis of major, minor and trace elements

Hydrogeochemical analysis was performed at the Actlabs laboratory (ActLabs 2021) using inductively coupled plasma mass spectroscopy (ICP-MS) for major, minor, and trace elements (ActLabs Code 6-HR-ICPMS), and ion chromatography for anions (ActLabs Code 6B-Ion Chromatography). Reference materials, independent quality control, and detection limits for each element or compound for the methods ICP-MS and IC are described and reported on the ActLabs website (ActLabs 2021).

Concentrations of HCO3− were not determined directly by laboratory analysis but were obtained from total alkalinity values. It should be noted that while the two parameters are related, total alkalinity is not strictly equivalent to HCO3− concentration but is a good estimate, and the difference between the two values in the measured pH range in the field is less than 1%.

Analysis of isotopic composition of hydrogen and oxygen

The isotopic composition of oxygen and hydrogen was determined at the Jožef Stefan Institute using H2–H2O (Coplen et al. 1991) and CO2–H2O (Epstein and Mayeda 1953; Avak and Brand 1995) equilibration on a Finnigan MAT DELTA plus dual inlet isotope ratio mass spectrometer with an HDOeq 48 automatic CO2–H2O and H2–H2O equilibrator. Cu was added to the sulfur springs samples prior to measurements. The CO2 and H2 gasses were used as working standards for water equilibration. Results are expressed as δ-values in per mil (‰). Two laboratory reference materials (LRM) calibrated to the international scale VSMOW-SLAP were used to normalize the data. Reference materials LRM W-45, W-53, USGS 45, USGS 46 and USGS 47 were used for independent quality control of the measurements (Kanduč et al. 2020, 2021; Žvab Rožič et al. 2021). The average repeatability of the samples was 0.02‰ for δ18O and 0.3‰ for δ2H.

The isotopic composition of monthly composite precipitation samples is regularly monitored at the Portorož station (Fig. 1b) by the Jožef Stefan Institute as part of the Slovenian Network of Isotopes in Precipitation (SLONIP) since 2000 (Vreča et al. 2022); in the past, it was also monitored at Kozina. Both stations are located in the recharge area (Fig. 1b) and represent the source of investigated groundwater, therefore we used the available data for the interpretation of our results.

Analysis of total alkalinity

The water sample was passed through a 0.45 µm nylon filter into an HDPE bottle and kept refrigerated until analysis. First, pH was measured in the laboratory using a pH meter (Mettler Toledo AG 8603, Schwerzenbach, Switzerland). Within 24 h of sample collection, the Gran titration method (Gieskes 1974) was used to determine TA with an accuracy of ±1%. Approximately 8–10 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 the titration. Titration was performed using a CAT titrator (Ingenierbüro CAT, M. Zipperer GmbH Ballrechten-Dottingen, Germany). The method is described in detail by Zuliani et al. (2020). Two replicates were measured.

Analysis of the isotopic composition of dissolved inorganic carbon

The isotopic composition of carbon from dissolved inorganic carbon was determined using the Spötl procedure (Spötl 2005; Kanduč 2006). Ampoules of saturated phosphoric acid (100–20 μL) were flushed with helium, 5 mL of the water sample was added according to the alkalinity, and then CO2 was generated. δ13CDIC was measured using the Europa Scientific 20–20 isotope mass spectrometer (Sercon Limited, Crewe, UK) with the ANCA-TG preparation module. Standard solution Na2CO3 (Carlo Erba CE, Val de Ruil, France aqua preparation (8 mg Na2CO3/12 mL distilled water) with a known δ13CDIC value of −10.8‰ ± 0.2‰ and tap water (Reactor Center Podgorica) with a value of −13.0‰ ± 0.1‰ were used as control materials. The CE standard was used for normalization of the results. Two replicates of each sample were measured.

Analysis of isotopic composition of carbonates (δ 13CCaCO3)

The isotopic composition of carbon from carbonate (δ13CCaCO3) was determined using an Europa Scientific 20–20 isotope-ratio mass spectrometer (Sercon Limited, Crewe, UK) with ANCA-TG preparation module. Samples were first ground and homogenized. 10 mg of limestone, 20–30 mg of marl and 50 mg of sandstone were weighed into 12 mL glass ampoules and flushed with helium. Then, 0.5 ml of H3PO4 was added (McCrea 1950). Reference materials: NBS 18 (δ13C = −5.014‰ ± 0.035, δ18O = −23.2‰) and IAEA CO-1 (δ13C =  +2.492 ± 0.030, δ18O = −2.4‰ ± 0.1) were used for the normalization of δ13C measurements. Five replicates for each sample were performed.

Analysis of isotopic composition of sulfur and oxygen in sulfate

The water was filtered through a 0.45-µm filter membrane to remove any dirt in the water samples and the clean, clear water was then transferred to a beaker. We added a few drops of 2N HCl solution to the water to prevent the precipitation of BaCO3. While heating, we added a few drops of a 0.5 M BaCl2 solution to precipitate BaSO4. The BaSO4 powder was then carefully collected from a dry filter for δ34SSO4 and δ18OSO4 measurements.

The sulfur isotope composition of sulfate (δ34SSO4) was analyzed at The University of Arizona with SO2 gas using a ThermoQuest Finnigan Delta PlusXL continuous-flow gas-ratio mass spectrometer with coupled elemental analyzer (Costech). Samples were combusted at 1030 °C with O2 and V2O5. Standardization is based on international standards OGS-1 and NBS123, and several other sulfide and sulfate materials that have been compared between laboratories. Calibration is linear in the range −10 to +30‰. Precision is estimated to be ±0.15‰ or better based on repeated internal standards (The University of Arizona 2021). The oxygen isotope composition of sulfate (δ18OSO4) was measured on CO gas using a continuous-flow gas-ratio mass spectrometer (Thermo Electron Delta V) with coupled elemental analyzer (ThermoQuest Finnigan). Samples were combusted with excess C at 1350 °C. Standardization is based on international standard OGS-1. Precision is estimated to be ±0.3‰ or better based on repeated internal standards (The University of Arizona 2021).

Geochemical modelling

Saturation indices of calcite (SIcal), dolomite (SIdol), anhydrite (SIanh), gypsum (SIgyp) and CO2 (SICO2) were calculated using major cations and anions determined by ActLabs and the PHREEQC program (WATEQ4F database). Other elements and phases were not included in the calculations, since they were not detected in all samples. Saturation indices were calculated for each sample and then averaged for each group (submarine, terrestrial sulfur springs, groundwater, seawater).

Mixing with seawater was estimated using a simple mixing model based on Cl− concentration and EC values to estimate the seawater content in the springs (Eq. 1):

$$w \left( \% \right) = \frac{{x_{{({\text{sample}})}} }}{{x_{{({\text{seawater}})}} }} \times 100\% ,$$
(1)

where w percent of seawater in the sample, x Cl− or EC.

The conservative mixing equation assumes that the parameters (i.e., Cl−) on which the calculation is based are conserved during the mixing processes, therefore, the chloride ion was used in the calculation of the effect of mixing with seawater. Seawater sample S03 was used for this calculation, as its location is independent and does not coincide with the location of the submarine sulfur spring sampling (as is the case for SW01 and SW02, Fig. 1c).

Results and discussion

All results of the physicochemical parameters, geochemical and isotopic analyses are freely available in the Pangaea data repository (Žvab Rožič et al. 2023). The results presented in this paper focus on submarine sulfur springs (M03, M05, M11) and a terrestrial sulfur spring (K04) compared to non-sulfurous groundwater (VK01) and seawater samples (SW01, SW0, SW03).

Physicochemical parameters of submarine and terrestrial sulfur springs, groundwater and seawater

The temperature of the submarine sulfur springs was not constant throughout the seasons, with mean values as follows: 22.2 °C, 20.5 °C and 14.2 °C in the summer and autumn of 2020 and the spring of 2021, respectively. The temperature of the terrestrial sulfur spring was more constant, with a mean value of 18.5 °C for all three sampling events. The mean pH value of 7.5 for the submarine and terrestrial sulfur springs is lower than the mean pH of seawater (8.2) and higher than the groundwater from well VK01 (7.0). The EC values of the submarine sulfur springs are high and vary significantly, from 16,900 to 58,000 μS/cm. The EC values of the terrestrial sulfur spring are less variable and lower, ranging from 3370 to 5410 μS/cm. The mean DO content of 7.46 mg/L of all sulfur springs was lower than that of seawater (9.69 mg/L) and higher than that of groundwater (3.88 mg/L) (Fig. 2, Table 2).

Fig. 2
figure 2

Mean values of physicochemical characteristics of investigated submarine sulfur springs (M03, M05, M11), terrestrial sulfur spring (K04), groundwater (VK01) and seawater (SW01, SW02, SW03)

Full size image
Table 2 Temperature, pH, electrical conductivity, and dissolved oxygen values for submarine sulfur springs (M03, M05, M11), terrestrial sulfur spring (K04), groundwater (VK01) and seawater (SW01, SW02, SW03)
Full size table

Although we observed large temperature fluctuations in submarine sulfur springs throughout the year and temperatures in submarine sulfur springs were not as high as previously reported (29.6 °C; Žumer 2004, 2008), we measured temperatures of up to 4.2 °C higher for submarine sulfur springs in April 2021 than seawater and we also see elevated mean temperatures of the spring water as well. Assuming that submarine sulfur springs and a terrestrial sulfur spring represent the outflow of the same aquifer, the higher temperatures of the terrestrial sulfur spring also indicate the thermal nature of the springs. High EC values indicate mixing with seawater, which is more present in the submarine sulfur springs than in the terrestrial sulfur spring.

Hydrogeochemical characteristics of submarine and terrestrial sulfur springs, groundwater and seawater

Submarine sulfur spring water is dominated by Na+ (3.77–11.10 mg/L) and Cl− (7.14–23.20 mg/L) ions. High concentrations of Mg2+ (300–1.05 mg/L) and SO42− (1.10–3.00 mg/L) are also present in the submarine sulfur springs. These results point to mixing with seawater, since seawater contains the highest concentrations of Na+, Mg2+ and Cl− ions. The Ca/Mg ratio is below 1 for all submarine sulfur springs, while the terrestrial sulfur spring has a Ca/Mg ratio of 0.98–1.74. In addition to the predominant Na+ and Cl− ions, the terrestrial sulfur spring also shows high levels of Ca2+, Mg2+, HCO3− and SO42−. In contrast to the submarine sulfur springs, which are characterized by the Na–Cl water type, the terrestrial sulfur spring has a higher proportion of the Ca–HCO3 type, characteristic of groundwater (VK01) that is dominated by Ca2+ and HCO3− ions and which is typical of limestone dissolution.

High Na+ and Cl− concentrations in both the submarine and terrestrial sulfur springs indicate mixing with seawater. The results of the mixing model based on the Cl− concentration and the EC values show that the lowest calculated seawater content was observed in samples M03 from October (44% based on EC values, pure seawater based on Cl−), M03 from April 2021 (48% seawater based on Cl− and 71% of seawater based on EC values), M05 from July 2020 with 31–33% seawater, M05 from October 2020 with 40–47% seawater, and M11 from April 2021 with 29–35% seawater (Table 3). Since other samples (M05 from April 2021, M11 from July 2020) show seawater contents of greater than 80% we suspect that the syringe sampling is not optimal, and we sampled spring water mixed with seawater in the sediment covering the spring opening. In the future, it may be worth developing a more sophisticated sampling method in order to generate more accurate and conclusive results.

Table 3 Calculated seawater content in the samples based on Cl− concentrations and EC values
Full size table

The results of other ions are freely available in the Pangaea data repository (Žvab Rožič et al. 2023) and largely reflect seawater mixing or are below the detection limit. Higher levels of Fe and Mn ions were detected in the sulfur springs and groundwater, with the submarine sulfur springs containing up to 34.4 µg/L of Mn, the terrestrial sulfur spring with Mn up to 361 µg/L, and groundwater containing the highest Mn concentration of 540 µg/L. 50 µg/L of Fe ions were detected in the terrestrial sulfur spring and up to 990 µg/L of Fe in the groundwater sample. Fe in the submarine sulfur springs was below the detection limit, which was raised following sample dilution due to high TDS. Elevated Fe and Mn concentrations may point to redox conditions in the aquifer, comparable to that of submarine sulfur springs in Florida, since the reduced Fe and Mn are more soluble (Swarzenski et al. 2001).

Geochemical modelling of submarine and terrestrial sulfur springs

Calcite saturation indices of the submarine sulfur springs and the terrestrial sulfur spring are close to the equilibrium state, while SIdol values show a slight oversaturation of dolomite in submarine sulfur springs; dolomite oversaturation in the seawater is even higher. Such oversaturation of dolomite in karst groundwaters is typical due to the problematic precipitation of dolomite (Verbovšek and Kanduč 2016). All samples are undersaturated with gypsum and anhydrite (Table 4). The partial pressure of CO2 (pCO2) in the submarine sulfur springs (10–2.24 atm) is almost 20 times the normal atmospheric value (10–3.39 atm; Žvab Rožič et al. 2022), while the pCO2 of the terrestrial sulfur spring is 33-times the atmospheric pCO2. Other phases and minerals were not included in the geochemical modelling, as their element concentration were low and the presence of these minerals not expected.

Table 4 Mean values of saturation indices SIcal, SIdol, SIanh, SIgyp of investigated waters in the Izola area
Full size table

Isotopic characteristics of submarine and terrestrial sulfur springs, groundwater and seawater

The isotopic value of the submarine sulfur springs had a range of 44.6‰ for δ2H, with the lowest negative value of −38.8‰ to the highest positive value of +5.8‰. The δ18O of submarine waters showed a range of 6.60‰, from −5.88 to +0.72‰ (Fig. 3). The hydrogen and oxygen isotopic composition of the terrestrial sulfur spring, groundwater, and seawater is less variable. The mean δ2H and δ18O values of the terrestrial sulfur spring K04 were −41.5‰ and −6.59‰, respectively (Fig. 3). Local groundwater samples from the well VK01 are also presented for comparison, with mean δ2H and δ18O values of −41.0‰ and −6.12‰, respectively, and seawater samples with mean δ2H and δ18O values of +7.5‰ and +0.94‰, respectively (Fig. 3). The isotopic composition of precipitation at Portorož with weighted mean δ2H and δ18O values of −40.5‰ and −6.38‰ for the period 2011–2020 is characteristic for the coastal areas of the northeastern Adriatic and similar to the mean values observed for the 2001–2003 period at Portorož (Vreča and Malenšek 2016). All δ2H and δ18O data related to the investigated samples are plotted near the global meteoric water line (GMWL; δ2H = δ18O × 8 + 10; Craig 1961) and the reduced major axis precipitation amount weighted local meteoric water line (LMWL Portorož, 2011–2020; δ2H = 7.87 × δ18O + 9.77; Šušmelj et al. 2022). Therefore, the main source of water in all samples is local precipitation. However, the submarine sulfur springs and groundwater data plot slightly below the GMWL and LMWL along the mixing line (ML) between the seawater and the weighted mean δ2H and δ18O values for Kozina (Vreča and Malenšek 2016) (Fig. 3). Values for the terrestrial sulfur spring are positioned between the precipitation values characteristic for Portorož and Kozina, reflecting the infiltration of modern precipitation in the recharge area of the springs. The submarine sulfur spring samples M03 from October 2020, M05 from July 2020, M05 from October 2020, and M11 from April 2021 plot closer to the terrestrial waters and LMWL, indicating less mixing with seawater and a higher proportion of groundwater originating from precipitation. The isotopic composition of the groundwater is also very similar to the isotopic composition of the Smrdljivec sulfur karst spring, which is located in the Classical Karst (Mulec et al. 2021), which is in one of the presumed recharge areas of the Izola sulfur springs.

Fig. 3
figure 3

(modified after Karolytė et al. 2017)

a δ2H versus δ18O, values of submarine sulfur springs of Izola, terrestrial sulfur spring, groundwater, seawater (this study in squares) and other studies: Smrdljivec sulfur spring (Mulec et al. 2021), Žvepovnik sulfur spring (Žvab Rožič et al. 2022), Žveplenica sulfur spring (Zega et al. 2015) and groundwater from central Slovenia (Verbovšek and Kanduč 2016). b δ2H and δ18O fractionation processes: 1. low-T mineral reactions, 2. hydration of silicates, 3. H2S exchange, 4. evaporation from surface, 5. high-T exchange with minerals

Full size image

All investigated water samples are enriched with heavier isotopes (i.e. 2H and 18O) due to the proximity to the Mediterranean and the Adriatic Sea. In comparison, other karst and fractured aquifers in central Slovenia (Verbovšek and Kanduč 2016; Žvab Rožič et al. 2022) (Fig. 3) are affected by the continental isotopic effect, as an increasing depletion of 2H and 18O is observed in the precipitation in Slovenia with distance from the sea (Vreča et al. 2006; Kern et al. 2020).

The mean δ13CDIC values of submarine sulfur springs (M03, M05, M11), terrestrial sulfur spring (K04), and groundwater (VK01) were −1.4‰, −12.2‰, and −12.3‰, respectively. The seawater in Izola (SW01, SW02, SW03) had an average δ13CDIC value of +0.7‰. The measured δ13CDIC of the submarine sulfur spring M03 from the study of Faganeli et al. (2005) was −3.5‰ at a salinity of 5‰ (lower salinity than this research, approximately 14% of seawater in the sample), while δ13CDIC of the seawater was −0.9‰ and salinity 35‰.

The results of δ13C from different rock types from Izola Bay (sandstone in the flysch formation, marl in flysch formation and nummulite limestone) vary from −0.6‰ (carbonate from the flysch formation) to +2.2‰ (limestone) with an average value of +0.4‰.

Geochemical processes were calculated as follows: Line 1 (with a value of 1.2‰) dissolution of limestone according to the average δ13CCaCO3 (2.2‰) value—predicted value (Kanduč et al. 2012) causing 1‰ enrichment in 12C in DIC (Romanek et al. 1992), line 2 with a value of −12.5‰ non-equilibrium carbonate dissolution by carbonic acid produced from soil zone CO2 (Kanduč et al. 2012), and line 3 (with a value of −18.2‰) open system equilibration of DIC with soil CO2 originating from degradation of organic matter with δ13Csoil = −27.2‰ (Kanduč et al. 2012) (Fig. 4).

Fig. 4
figure 4

δ13CDIC versus TA, values of submarine sulfur springs, terrestrial sulfur spring, groundwater, seawater, and other studies: submarine sulfur springs of Izola (Faganeli et al. 2005; *TA values as HCO3− concentration), Žvepovnik sulfur spring (Žvab Rožič et al. 2022), Žveplenica sulfur spring (Zega et al. 2015), springs from North Slovenia (Kanduč et al. 2012; **data from all seasons) and groundwater from central Slovenia (Verbovšek and Kanduč 2016). Please refer to the text for an explanation of dotted lines 1, 2, and 3, indicating biogeochemical processes

Full size image

Seawater falls around equilibration line 1, while samples from the submarine sulfur springs (from limestone) fall between equilibration lines 1 and 2. Terrestrial sulfur spring K04 (from nummulite limestone and flysch rocks in the hinterland (Fig. 1a)) fall close to equilibration line 2 and show carbonate and soil CO2 contribution. Groundwater in Izola (VK01) is also near equilibration line 2 but has higher total alkalinity (longer retention time) in comparison to the terrestrial sulfur spring (K04), but a similar δ13CDIC value (Fig. 4). Groundwater (VK01) and the terrestrial sulfur spring (K04) fall along the same fractionation line 2 as the previously studied Žvepovnik (Žvab Rožič et al. 2022) and Žveplenica sulfur karst spring (Zega et al. 2015) (Fig. 4). These two previously studied sulfur springs, VK01 and K04, have similar δ13CDIC to karst groundwater from SW Slovenia (Kukar 1998) and most karst and fractured aquifers in central Slovenia (Verbovšek and Kanduč 2016) (Fig. 4). Springs from North Slovenia have a more variable δ13CDIC due to their different geological backgrounds (Kanduč et al. 2012).

Alkalinity and δ13CDIC values (Fig. 4) indicate that the seawater near Izola has the highest δ13CDIC, while the submarine sulfur springs near Izola have total alkalinity and δ13CDIC values between those of seawater and the terrestrial sulfur spring and the groundwater in Izola (Fig. 4). The higher δ13CDIC value of the submarine sulfur spring is also due to the mixing with seawater with δ13CDIC of +0.7‰ ± 0.3‰, in addition to the dissolution of carbonates (fractionation line 1). It is impossible to say how much of a contribution comes from the dissolution of carbonates (with value of δ13CDIC of +1.2‰) and how much from the mixing with seawater (with an average value of +0.7‰), which in our case is considerable.

Sulfate isotope analysis was performed once, in the spring of 2021. The δ18OSO4 and δ34SSO4 values of the submarine sulfur spring sample M03 are +9.1‰ and +22.2‰, respectively. Sample M11 shows slightly higher values, δ18OSO4 =  +11.2‰ and δ34SSO4 =  +23.6‰. The δ18OSO4 and δ34SSO4 values of the terrestrial sulfur spring K04 are +11.0‰ and +20.3‰, respectively. Sulfate concentrations in the submarine sulfur springs ranged from 1100 to 2770 mg/L, reaching concentrations similar to those in seawater (2640–3000 mg/L), whereas they were much lower in the terrestrial sulfur spring, ranging from 182 to 283 mg/L.

Although the δ34SSO4 values for the submarine sulfur springs of Izola are typical of seawater and evaporites (Krouse and Mayer 2000), the negative SIanh and SIgyp and the local geologic setting do not indicate an evaporative origin. Since high proportions of seawater were detected in the springs, we can conclude that the marine sulfate in the sulfur springs water samples reflects the mixing of seawater with spring water in the samples. The origin of the sulfate from which the sulfide is formed remains undefined. According to the geological data, a highly probable source of sulfur is the coal of the Liburnian Formation, which underlies the alveolinic-nummulitic limestones in Izola. We did not detect sulfate from FeS2 oxidation as found in the karst groundwater in China, with a negative δ34SSO4 value (Tang et al. 2021), therefore any particular potential risk to the marine environment is excluded.

Characterization and origin of the groundwater in the sulfur springs of Izola

The submarine and terrestrial sulfur springs of Izola exhibit elevated mean temperatures and a distinctive smell of sulfur, which is consistent with the reports on the use of Izola thermal waters for spa purposes in the nineteenth century (Kramar 2003) and the first reports on submarine sulfur springs by Žumer (2004, 2008). Elevated temperatures indicate deep water circulation, presumably more than 100 m deep, in accordance with the expected temperature at a depth of 100 m of around 17 °C in this area (Rajver 2016). The spring water has a higher mean temperature than the groundwater (VK01) at a depth of 180 m and thus rises from a greater depth. The spring water flows deep beneath the flysch that surrounds and covers the limestones of Izola, separating them from the limestones of Karst and the Croatian border area (Fig. 5). At depths of several hundred meters, the water warms up and then rises and springs at the stratigraphic boundary of the limestones of the Izola anticline and the surrounding flysch (Rožič and Žvab Rožič 2023). Heat, lower salinity, and artesian pressure may also play a role in water rise and spring formation.

Fig. 5
figure 5

a General geological/hydrogeological cross-section between the Istra Peninsula and the Kras (Karst) Plateau with possible regional groundwater flow pathways (modified from Placer et al. 2010; major modification is in Izola area with the representation of the outcropping Izola anticline limestone core and in Sečovlje area, where the data from the coal mine (Turk 1955) does not show major fault displacement); b hydrogeological model of the submarine sulfur springs near Izola with the indicated possible groundwater flow paths; the springs occur near the boundary between the karst aquifer (limestone core of the Izola anticline) and the hydrogeological barrier (flysch limbs of the Izola anticline). The micro-locations of the springs could be governed by fracture zones and the thickness of the overlying Quaternary deposits; c detailed hydrogeological cross-section of a spring: groundwater springs from a limestone, penetrates a thin cover of loose sediments, and mixes rapidly with seawater within the lower part of the funnel-shaped depression.

Full size image

δ2H and δ18O show that the source of the freshwater component in the sulfur springs is local precipitation, with a limited influence of evaporation and/or geochemical processes that could modify the stable isotopic composition of the water, since the terrestrial sulfur spring and local groundwater plot on or close to the LMWL (Fig. 3). Limestones in Izola are largely covered by the urban area, where water drains into sewers, therefore greater infiltration is not possible. The water from the Izola sulfur springs could originate from the same recharge area as that of the isotopically similar Smrdljivec sulfur spring (Mulec et al. 2021), in the Classical Karst region, about 25 km E to 30 km NE from Izola. Outcrops of alveolinic-nummulitic limestones at the Croatian border, near the Dragonja river and the Sečovlje coal mine, represent other possible recharge areas. The isotopic composition from this area, recorded at station Portorož and Kozina (Fig. 1b), confirms these assumptions.

The sulfur spring water is highly mineralized as it mixes with seawater in the sediment at the spring depression. Groundwater–seawater mixing within the aquifer, as present in Lucija (Lu-1), 5 km from Izola (Fig. 1b; Brenčič 2009), is unlikely, since groundwater in Izola (VK01) with a Ca–HCO3 facies shows no signs of mixing with seawater at a depth of 180 m.

δ13CDIC and alkalinity confirm that the spring water is coming from a limestone aquifer, yet the terrestrial sulfur spring water is also in contact with organic matter, as some carbonate dissolution occurs due to carbonic acid. The presence of coal in the layers below the alveolinic-nummulitic limestones of Izola (Pleničar et al. 1973; Benedik and Rožič 2002; Brenčič 2009) is a possible source of the organic matter in contact. High Fe and Mn concentrations also indicate mild redox conditions, which are likely the cause of the hydrogen sulfide production by bacterial reduction of evaporite or marine sulfate, as is the case in the thermal springs of the Santa Cesarea Therme system in Italy (Santaloia et al. 2016; D’Angeli et al. 2021), or by the reduction of organic (coal-derived) sulfates. δ18OSO4 and δ34SSO4 suggest that the sulfate is of marine origin; however, said results are affected by the mixing with seawater (Fig. 5).

Conclusions

Physical, geochemical, and isotopic data were used to characterize submarine and terrestrial sulfur springs on the Slovenian coast. The results provide us with a better understanding of the source areas of water and interactions within the aquifer, all of which play an important role in water management (use of water on the coast or management of discharges into the sea). Namely, the presence of seawater, sulfide, and other special properties of the investigated springs may represent a harmful or beneficial contribution to the environment and community.

The applied methodologies provide us with valuable results from which we can draw the following conclusions:

  1. 1.

    Seawater mixing with submarine sulfur springs in the sediments was confirmed by high EC values, high concentrations of Cl−, Na−, and other elements characteristic of seawater, such as enrichment with 13CDIC. Not all springs mixed with seawater equally, as samples of submarine sulfur springs consisted of 29–98% of seawater. The terrestrial sulfur spring also mixed with seawater and consisted of 4–9% seawater. Although natural mixing with seawater is to some degree present in the sulfur springs, the very high proportions of seawater in some submarine spring samples point to problematic sampling.

  2. 2.

    δ2H and δ18O have shown that the origin of the water in the terrestrial sulfur spring can be traced to the infiltration of local modern meteoric water. The source of the freshwater component in the submarine sulfur springs is the same as that in the terrestrial sulfur spring, as they plot on the mixing line between the terrestrial sulfur spring and the seawater.

  3. 3.

    δ13CDIC and alkalinity in the terrestrial and submarine sulfur springs of Izola show that most of the dissolved inorganic carbon in the groundwater and the terrestrial sulfur springs with δ13CDIC around −12‰ originates from the dissolution of nummulite limestone and the flysch formation with soil CO2 contribution, while in the submarine sulfur springs the dissolution of nummulite limestone prevails with seawater mixing, which is not negligible. Even though mixing with seawater was present in the submarine sulfur springs, we can conclude from this and previous investigations that carbonate dissolution constitutes the main process at work in the recharge area of submarine sulfur springs.

  4. 4.

    δ18\({\text{O}}_{{{\text{SO}}_{{4}} }}\) and δ34\({\text{S}}_{{{\text{SO}}_{{4}} }}\) indicate that the sulfate is of marine origin, which reflects spring water mixing with seawater, thus the exact origin of the sulfur is still unclear. Another possible source of sulfur is the coal layers of the Liburnian formation underneath the alveolinic-nummulitic limestones, which are rich in organic matter and sulfur. The higher concentrations of soluble Fe and Mn in the sulfur springs are consistent with the reducing conditions and the bacterial reduction of marine or organic sulfates, which could explain the sulfurous odor. However, further analyses is required to better understand the sulfur cycle in the spring’s aquifer. Moreover, new sampling methods have to be developed to sample submarine karst springs under the unconsolidated silty sediment.