Geochemical patterns and origin of alkaline thermal waters in Central Greece (Platystomo and Smokovo areas)

Abstract

This study investigates the origin and chemical composition of the thermal waters of Platystomo and Smokovo areas in Central Greece as well as any possible relationships of them to the neighboring geothermal fields located in the south-eastern part of Sperchios basin. The correlations between different dissolved salts and the temperature indicate that the chemical composition of thermal waters are controlled by, the mineral dissolution and the temperature, the reactions due to CO₂ that originates possibly by diffusion from the geothermal fields of Sperchios basin and the mixing of thermal waters with fresh groundwater from karst or shallow aquifers. Two major groups of waters are recognized on the basis of their chemistry: thermal waters of Na–HCO₃–Cl type and thermal waters mixed with fresh groundwater of Ca–Mg–Na–HCO₃ type. All thermal waters of the study area are considered as modified by water–rock interaction rainwater, heated in depth and mixed in some cases with fresh groundwater when arriving to the surface. Trace elements present low concentrations. Lithium content suggests discrimination between the above two groups of waters. Boron geochemistry confirms all the above remarks. Boron concentration ranges from 60 μg L⁻¹ to 10 mg L⁻¹, while all samples’ constant isotopic composition (δ¹¹B ≈ 10 ‰) indicates leaching from rocks. The positive correlation between the chemical elements and the temperature clearly indicates that much of the dissolved salts are derived from water–rock interactions. The application of geothermometers suggests that the reservoir temperature is around 100–110 °C. Chalcedony temperatures are similar to the emergent temperatures and this is typical of convective waters in fault systems in normal thermal gradient areas.

Introduction

The study area of Platystomo belongs to the graben structure of Sperchios basin (Central Greece), which owes its formation to the N–S extensive movement of the Aegean plate, due to both the subduction of the African plate under the Eurasian plate and the extrusion of the Anatolian block (Taymaz et al. 1991), while the study area of Smokovo is located at the north margin of the Sperchios Graben. The Basin of Sperchios is an E–W valley ~50 km long by 3–12 km wide, which is drained by the Sperchios river that flows eastward toward to Maliakos gulf (Aegean Sea). The south part of Sperchios basin, between the mid-basin and the Maliakos gulf, hosts several thermal springs, those of Platystomo area and those located at the southern edge such as “Hypati” located 2 km apart from the first relieves of Iti mountain as well as “Kamena Vourla”, “Psoroneria”, and “Thermopylae” thermal springs located in the south–eastern part of the basin near the coast of Maliakos gulf (Fig. 1). Many scientists studied the thermal springs of Sperchios basin, especially those located in the south-eastern edge. However, the geochemical patterns and the origin of the thermal waters of Platystomo and Smokovo areas remain poorly documented.

Fig. 1

A simplified geological map of the study area (Bornovas and Rondogianni-Tsiambaou (1983) modified by Lambrakis; 1 alluvial loose deposits; 2 oligocene molassic formations in the Mesohellenic Trench: clays, conglomerates, sandstones and marls; 3 neogene sediments (undivided): lacustrine deposits: conglomerates, sandstones, marls and glays; 4 flysch sediments of Pindos zone; 5 flysch sediments of Giona–Parnas zone; 6 flysch sediments of Pelagonian zone; 7 limestones of Upper Cretaceous; 8 limestones of Triassic-Lower Jurassic of the Parnassos-Ghiona zone; 9 limestones Upper Jurassic; 10 jurassic shale and cherts formation: radiolarites, sandstones, pelites with white limestone, lenses and enclosed ophiolitic bodies; 11 ophiolitic complex; 12 fault; 13 overthrust; 14: geological boundary; 15 thermal spring

According to Stahl et al. (1974), the high salinity of Sperchios basin thermal water is due to the intrusion of seawater into the karst system fresh water bounding the geothermal field. The geothermal field is located in the hanging wall of the graben and it is associated to tectonic breccias (Gartzos and Stamatis 1996). Dotsika (1991) compared different geothermometers to assess the temperature of reservoirs and she concluded that Sperchios thermal fields are low-enthalpy hydrothermal systems. Psomiadis et al. (2008) showed that most of the thermal waters come from a mixing process between seawater and various fresh waters from different locations and altitudes. According to Duriez et al. (2008), the distribution of the salinity is independent from the distance to the shoreline in the south-eastern part of Sperchios basin.

The thermal springs of Platystomo and Smokovo areas are known for their curative properties especially for stomach and liver diseases. For some of the thermal springs, chemical analyses comprising major elements, gases and isotopes, together with field studies undertaken by the Greek Institute of Geological and Mineral Exploration (Papastamataki and Leonis 1982; Barnes et al. 1986; Kakavas 1986; Papadeas 1992) gave interesting indications on water properties and origin.

The main objective of this study is the investigation of the origin and chemical composition of the alkaline springs of Platystomo and Smokovo areas. The study is mainly based on the evaluation of major trace elements and environmental isotopes. The chemical characteristics of “Hypati” thermal spring located in the south edge of Sperchios basin are also examined for comparison with the alkaline springs of the study areas. “Hypati” spring exhibits both thermal (32.6 °C) and mineral (TDS > 1,000 mg L⁻¹) character.

Geological and hydrogeological setting

The study areas belong to the north surrounding mountainous region of the Sperchios Graben and are built up by rocks of (a) Subpelagonian zone, (b) Parnassos–Ghiona zone, and (c) Pindos zone (Fig. 1) (Bornovas and Rondogianni-Tsiambaou 1983). The Subpelagonian zone lying in the eastern part of the area consists of Cretaceous age flysh, Mesozoic limestone of Upper Cretaceous and Upper Jurassic age, shale, and cherts. Ophiolitic thrust sheets cover a significant part of the zone. They consist of basic and ultrabasic igneous rocks, such as diabases, peridotites, pyroxenites, gabbros, serpentinites, laves, and tuffs. The peridotites are usually serpentinized. The Parnassos–Ghiona zone in the southernmost part consists of Paleocene flysch as well as limestone and dolomite of Triassic to Lower Jurassic. The Pindos zone covers the western part of the study area. This zone consists mainly of Paleocene age pelitic and psammitic flysch, which is characterized by the intense presence of organic material, as well as limestones and radiolarites of Jurassic to Cretaceous age. The lower part of the Sperchios Graben is covered by unconsolidated sediments of Pleistocene–Holocene age (Fig. 1).

Tectonically the Sperchios Graben is related with the Anatolia fault that passes through the north Turkey enters the Aegean Sea and ends to the Sperchios Graben. The micro-plate of the eastern Mediterranean region proposed by McKenzie (1972) is northwards delimitated by an important extension zone. This zone coincides with the Anatolian fault. Hence, the Platystomo area is characterized by intense neo-tectonic activity with large faults and rift structures. Generally, the deep extensional faults enhance the formation of geothermal activity.

The dominant rift structure of the study area is the E–W large normal faults. Another major group of faults are those of NW–SE direction that represent the tectonic contacts of the different stratigraphic sequenced geological formations (Fig. 1). Thermal spring emergencies are facilitated by faults. Electrical resistivity prospecting suggests that the Thermi spring (PL-T-5) rises through a fault with an altitude of at least 1,000–1,100 m (Kakavas 1986).

The thermal springs of Paleovracha (PA-10), Platystomo (PL-T-5); (PL-K-6); (PL-M-7), Smokovo (SM-2), Soulada (SO-1) issue from recent alluvial deposits that consist of sands and clays with pebbles and gravels. This formation, presenting limited extension and thickness, overlay the eastern Pindos flysch, Danian to Eocene age, with a total thickness of 2,000 m (Kallergis et al. 1970). Pindos flysch consists of fractured thick-bedded sandstones and greywacke with conglomerate, clayey and silty intercalations. Shallow fresh water aquifers have been developed in the conglomerate intercalations. Ekkara (EK-4) and Archani (AR-9) spring issue from peridotites, which belong to the ophiolitic complex and host less productive aquifer. Kaitsa spring (KA-3) issues from limestones, which host a local aquifer.

Analytical procedures and methods

Sampling was undertaken on September 2005. Samples were collected at the outlets of spring, except for PL-T-5 samples that were collected some meters away from the spring using a pump. They were kept in two polyethylene bottles. The first bottle of 0.5 L volume was filtered on-site through 0.45 μm pore size Millipore filters. It was then acidified to pH about 2 with 65 % ultra pure HNO₃ and used for major cation analysis, trace element determination, and boron isotopes. The second non-acidified aliquot (1 L) was retained to determine major anions and cations. Temperature (T), electrical conductivity (E.C.), Redox potential (Eh), pH, and sulfites hydrogen (H₂S) were measured from untreated samples in situ. Since dissolved oxygen could not be measured, Eh was used as the main tool to identify aerobic and anaerobic conditions. The major cations Ca²⁺, Mg²⁺, Na⁺, and K⁺ were determined by atomic absorption spectroscopy (GBC Avanta). Trace element determinations and boron isotope analysis were conducted using ICP-MS Perkin Elmer, ELAN 6100. For major cation and trace element determinations, accuracy was controlled using appropriate laboratory standards. For boron isotopes, a standard deviation of less than 1.5 % was determined by repeat analysis of NIST SRM-951 standard (¹¹B/¹⁰B = 4.043 ± 0.005). Isotope ratios are reported as δ¹¹Β = [(¹¹Β/¹⁰Β)_sample/(¹¹B/¹⁰B)_{SRM 951r} − 1] × 1,000.

Alkalinity was measured by volumetric titration using bromocresol green-methyl red indicator by Hach^® titration kits. Similarly, chlorides (Cl⁻) were measured using the AgNO₃ method by Hach^® titration kits while sulfates (SO₄ ²⁻), nitrates (NO₃ ⁻), and ammonium (NH₄ ⁺) were determined using a spectrophotometer (Hach^®, DR/4000). Temperature (T), pH, electrical conductivity (E.C.), and redox potential (E _h) were determined using an ion/E.C. meter (Consort^® C533) with combined electrodes. All analyses were conducted at the Laboratory of Hydrogeology, University of Patras immediately after collection. The charge balance error for all chemical analyses is within a highly acceptance range (±5 %). All chemical calculations, including those for saturation indices and geothermometers, were undertaken using the PHREEQC software package (Parkhurst and Appelo 1999). A sample from Hypati thermal spring (YP-11), which issues from the Pleistocene–Holocene unconsolidated sediments of the south part of Sperchios Graben, was taken for comparison purposes. Additionally, a fresh water sample from flysch conglomerates (PL-X-8) was also taken for comparison purposes.

Groundwater chemistry

Classification of thermal waters

Chemical data for all samples are provided in Tables 1 and 2. As it is shown in Table 1, all water samples of Platystomo and Smokovo areas are alkaline with pH ranging between 7.4 and 11.2. They are characterized by negative values of redox or strong reducing conditions. Some of the samples present high amounts of ammonia (apart from PL-X-8, PL-K-6, EK-4 and KA-3), low concentrations of NO₃, high concentration of H₂S (apart from PL-X-8, EK-4 and AR-9), and low salt concentrations. Hypati spring shows different chemical characteristics with pH value 6.41, low concentration of ammonia, and high salt concentrations similar to that of all thermal springs in the eastern edge of Sperchios valley (Stahl et al. 1974; Psomiades et al. 2008).

Table 1 Physicochemical parameters and major constituent analyses of Platystomo and Smokovo thermal waters

Table 2 Trace element analyses of Platystomo and Smokovo thermal waters

Gas analyses from Platystomo thermal waters (Papastamataki and Leonis 1982; Minissale et al. 1989) show that the N₂/O₂ ratio values range between 4.3 and 13.1. These values are higher than those of the atmosphere (3.71) indicating the presence of non-atmospheric nitrogen in the thermal water circuits. Radon (²²²Rn) and Thorium (Th) content is low and varies from 0.5 to 6 Bq kg⁻¹ and from 0.1 to 0.8 Bq kg⁻¹, respectively. Kamena Vourla thermal spring, which is situated at the eastern part of Sperchios Basin, exhibits much higher Radon content (up to 100 Bq kg⁻¹) than Platystomo thermal springs. Hence, a relationship based on Radon and Thorium content of waters of the thermal springs of Platystomo area and Kamena Vourla seems not to exist.

Classification of thermal water was made using Piper diagram. As shown in Fig. 2, thermal springs can be mainly classified in three groups (two major groups and a third one). The first one includes all spring waters hosted in flysch formation, collected from Thermi (PL-T-5), Morfoneri (PL-M-7), Palaiovracha (PA-10), Soulada (SO-1), and Smokovo (SM-2) springs presenting Na–HCO₃–Cl water type. This group is characterized by low concentrations in calcium and magnesium about 2.5 and 0.3 mg L⁻¹, respectively. The alignment of the samples on a line (Fig. 2) between (PT-T-5) and (SM-2) depends on their Cl⁻ content. Concerning the water samples, PA-10, PL-T-5, and PL-M-7 the increased chloride concentration (app. 120 mg L⁻¹) in relation to the other samples (app. 20 mg L⁻¹) could possibly be attributed to leachate from local saline deposits in the sediments (see Sect. 5).

Fig. 2

Piper’s diagram for the classification of waters

The second group includes the samples Klouvio (PL-K-6), Xenia (PL-X-8), and Kaitsa (KA-3) of Ca–Mg–Na–HCO₃ water type. This water type could be considered as thermal and fresh water mixture, which is also indicated by the presence of nitrate ions. The PL-K-6 and PL-X-8 samples of the second group exhibit higher calcium and magnesium content related to the PL-T-5 and PL-M-7 samples of the first group, although all these samples originate from Platystomo area. The low Ca and Mg concentration of the samples of the first group could be probably explained by carbonate precipitation (see Sect. 4.3).

The third group includes the most alkaline samples of Archani (AR-9) and Ekkara (EK-4) springs that originated from the ophiolitic complex and presents no dominant cation.

Aqueous geochemistry

The solid line, in Fig. 3a–f, reflects the concentration–dilution characteristic line for seawater, while the dotted line is the linear regression line of the samples. These lines do not converge indicating that for all water samples of the study areas, the ratios of Cl to the elements Na, SO₄, B, Ca, and Mg present different values from those of a diluted seawater. All thermal waters are enriched in Na in relation to seawater. This enrichment could be attributed to the weathering of silicate minerals according to the reactions showing in section 4.3. Exception is the YP-11 sample (Hypati spring), which lies on the concentration–dilution characteristic line and presents high salt content. As is shown in Fig. 3, the chemical species for this sample, sea and rain have a common origin indicating that the chemical composition of Hypati spring has influenced from seawater.

Fig. 3

Logarithmic plots of relationships between water chemical species: Cl versus Na (a), SO₄ (b), B (c), Ca (d), Mg (e), and Arithmetic plot B/Cl versus δ ¹¹B (f)

The thermal waters are even more enriched in sulfate (Fig. 3b) in relation to seawater. Sulfate could be probably produced by anoxic oxidation of S²⁻ by oxidized forms of metallic cations as Fe³⁺, thus producing high Fe contents (Andrews et al. 1994). This could actually happen, since samples with higher Fe concentrations correspond to higher sulfate concentrations.

Additionally, sulfate concentrations may be resulted from the oxidation of H₂S at, or close to, the thermal water output according to the reaction: H₂S + 2O₂ → SO₄ ²⁻ + 2H⁺ (Valentino and Stanzione 2003). This reaction, which produces two moles of H⁺ for the oxidation of each mole of H₂S, contributes to the lowering of the pH of solution. However, as the pH increases, the above-mentioned process becomes strongly limited.

Sulfate could also be probably produced from the dissolution of the rocks of molassic formations (conglomerates, sandstones). As is shown in Table 1, SO ²⁻₄ concentration and H₂S content (1 and 0.1 mg L⁻¹ respectively) in PL-X-8 sample are low compared with the thermal springs PL-T-5 and PL-M-7, which exhibit higher sulfate concentration.

The B/Cl ratio of the samples is enriched compared with that of seawater (Fig. 3c). All samples are aligned between the two end members PA-10 and PL-K-6. The next Fig. 3d, e confirms that Ca/Cl and Mg/Cl ratios, which are completely different from those of seawater, range between the same end members, which belong to the first and second group of the classification of thermal waters. Hence, the chemical composition is a mixture between these end members so that some samples of the alkaline thermal waters of the study area include a fresh water component that masks their initial pure thermal character. This could also mean that some of the thermal springs are in the first stages of interaction with rocks.

The linear regression line of Fig. 3f shows that most of the samples have a constant B/Cl ratio, which corresponds to a value of δ ¹¹B equal to about 10 ‰ and is perpendicular to the line connecting samples with a marine composition. This perfect divergence along with the precedents suggests that thermal water composition is only related with water–rock interaction.

The analyses of oxygen and hydrogen isotopes (Papastamataki and Leonis 1982), shown in Table 3, indicate that the thermal water of Platystomo is closely related to the rainfall in the region.

Table 3 Isotopic data

Additionally, available Tritium data with low to null Tritium content suggest that thermal water of some springs (PL-T-5, PL-M-7) were isolated from atmosphere for many years, while thermal water of other springs (PL-K-6, AR-9) for fewer. This could confirm that the springs PL-T-5 and PL-M-7 belong to the same group as it is already mentioned, while thermal waters are modified by rock–water interaction rainwater, heated in depth, and mixed with fresh water when arriving to the surface. The mixed waters are characterized by higher calcium and magnesium concentrations in relation to the non-mixed waters.

Carbonate system chemistry of the Platystomo-Smokovo areas

The low Ca and Mg content of thermal water imply that any Ca or Mg supply from plagioclase or calcite dissolution has to be removed from the solution.

Two mechanisms make this possible: cation exchange and removal of calcium due to calcite precipitation, and CO₂-induced silicate hydrolysis.

Cation exchange processes are described by the equations below (Andrews et al. 1994):

$$ \begin{aligned} 2{\rm N}a{\rm X} \, + {\text{ Ca}}^{2 + } &\to {\text{ CaX}}_{2} + \, 2{\text{Na}}^{ + } \hfill \\ {\text{CaCO}}_{3} + {\text{ H}}_{2} {\text{O}} &\to {\text{ Ca}}^{2 + } + {\text{ HCO}}_{3}^{ - } + {\text{ OH}}^{ - } \hfill \\ \end{aligned} $$

where Χ− denotes the ion exchanger that may be clay minerals.

The above equations result in the continuous removal of calcium that leads to solutions under calcite saturation. Table 4 compares the saturation index values for the various spring waters. Some of the samples are undersaturated in respect to calcite, aragonite, and dolomite.

Table 4 Saturation indices of aqueous solutions with respect to relevant minerals

Carbon dioxide diffusion into the aquifer of thermal water is probably attributed to the groundwater content in CO₂ (approximately 1.000 mg L⁻¹) of Hypati spring (YP-11), which is located in the vicinity of the area. The presence of recently active and normal EW faults would assist this diffusion. Gartzos and Stamatis (1996), based on carbon stable isotope data, suggested that CO₂ in the wider area is liberated from metamorphic reactions (decarbonisation) of carbonate at depth, as it is the case for many areas worldwide (Bailey 1989; Ceron and Pulido-Bosch 1999). Silicate hydrolysis of plagioclases (albite, anorthite and anaclite) claims consumption of CO₂, according to the reactions below (Andrews et al. 1994):

$$ \begin{gathered} {\text{NaAl}}_{2} {\text{Si}}_{3} {\text{O}}_{8} + \, 5\raise.5ex\hbox{$\scriptstyle 1$}\kern-.1em/ \kern-.15em\lower.25ex\hbox{$\scriptstyle 2$} {\text{ H}}_{2} {\text{O }} + {\text{ CO}}_{2} \to {\text{Na}}^{ + } + {\text{HCO}}_{3}^{ - } + \, \raise.5ex\hbox{$\scriptstyle 1$}\kern-.1em/ \kern-.15em\lower.25ex\hbox{$\scriptstyle 2$} {\text{ Al}}_{2} \left( {{\text{Si}}_{2} {\text{O}}_{5} } \right)\left( {\text{OH}} \right)_{4} + \, 2{\text{H}}_{4} {\text{SiO}}_{4} \hfill \\ {\text{CaAl}}_{2} {\text{Si}}_{2} {\text{O}}_{8} + \, 3{\text{H}}_{2} {\text{O }} + \, 2{\text{CO}}_{2} \to {\text{Ca}}^{2 + } + \, 2{\text{HCO}}_{3}^{ - } + {\text{ Al}}_{2} \left( {{\text{Si}}_{2} {\text{O}}_{5} } \right)\left( {\text{OH}} \right)_{4} \hfill \\ {\text{NaAlSi}}_{2} {\text{O}}_{6} + \, 3\raise.5ex\hbox{$\scriptstyle 1$}\kern-.1em/ \kern-.15em\lower.25ex\hbox{$\scriptstyle 2$} {\text{ H}}_{2} {\text{O }} + {\text{ CO}}_{2} \to {\text{Na}}^{ + } + {\text{HCO}}_{3}^{ - } + \, \raise.5ex\hbox{$\scriptstyle 1$}\kern-.1em/ \kern-.15em\lower.25ex\hbox{$\scriptstyle 2$} {\text{ Al}}_{2} \left( {{\text{Si}}_{2} {\text{O}}_{5} } \right)\left( {\text{OH}} \right)_{4} + {\text{ H}}_{4} {\text{SiO}}_{4} \hfill \\ \end{gathered} $$

In saturation conditions, the release of calcium from the hydrolysis of anorthite can lead to its removal from the solution in the form of calcite according to the follow equation:

$$ {\text{Ca}}^{{ 2 { + }}} {\text{ + CO}}_{ 3}^{{ 2 { - }}} \to {\text{CaCO}}_{ 3} $$

Both mechanisms described above lead to waters poor in calcium and magnesium and rich in sodium, as is the case for the most springs of the study area. The most probable mechanism seems to be the last, although due to the presence of large amounts of calcite in the flysch sediments, the first mechanism of calcium removal could not be excluded.

Trace elements

Trace elements were determined to investigate the quality and origin of thermal waters (Table 2). All trace elements, except boron and strontium, present low concentrations, below the maximum admissible limits of potable water standards (E.U Council 1998, European Council Directive 98/83/EC on the quality of water intended for human consumption). Several redox sensitive trace elements including Mn, U, Cr, As, Se, and Pb exhibit null concentrations. However, in pure reducing conditions, as for example in PL-T-5 and PL-M-7 samples, Mn, U, Cr, As, Se, and Pb concentrations show higher values than those of PL-K-6 and PL-X-8, which comprise mixed waters. Iron concentration varies between 10 and 145.5 μg L⁻¹ for thermal waters, while for fresh waters (PL-X-8) it is 28 μg L⁻¹ approximately. In reduced environments, iron is found in the form of soluble Fe²⁺ and in oxidized environments the dominant form is the increased valence of insoluble form, Fe³⁺, as also for the previously mentioned metals (Hem 1988). Iron shows increased concentrations in water samples with low Eh.

Strontium and Barium elements are related with calcareous minerals. They are also found in K-feldspars and some aluminosilicate minerals, from where they pass to the waters after incongruent dissolution (Edmunds et al. 2002). These elements show increased concentrations in fresh water PL-X-8 and the mixed water PL-K-6. Rubidium and Lithium are also found in feldspars and aluminosilicate minerals and their dissolution in water is facilitated by temperature. In the Platystomo and Smokovo thermal waters, Lithium has higher concentrations in hot than in cold waters. Edmunds and Smedley (2000) used this element as a reliable indicator of groundwater residence time. It has been also used by Carrillo-Rivera et al. (1996) to discriminate between thermal and shallow groundwaters. According to this consideration, thermal waters of (PL-K-6), (PL-X-8), and (EK-4) samples with low Li concentration should follow a short circuit before emerging to the land surface. This is consistent for the first two samples that comprise fresh water and consequently may be valid for the third (EK-4). Copper, Lead, and Zinc concentrations do not show systematic trends and it appears that they are neither related to residence time nor geothermal influence.

The geochemistry of boron element

Chemical and isotopic studies suggested that most of the dissolved boron in natural waters derived from rocks dissolution depends on the nature of host rocks (Palmer 1991). Boron is one of the most mobile cations during hot fluid–rock interactions (Ellis and Mahon 1967). It exhibits high concentrations in seawaters, thermal waters and wastewaters due to the detergents and consequently all above may contribute to the groundwater composition of an aquatic system. In silicicate material groundwater, unaffected by seawater intrusion and wastewater, the boron content varies as happens in Platystomo and Smokovo areas. Pennisi et al. (2000) reported that the alkaline (pH 7.2–8) and cold (T 12–18.2 °C) groundwater from flysch-filled basement depressions of the western and south western flank of Mt. Etna have a boron concentration of about 0.5–1.45 mg L⁻¹. Andrews et al. (1994) reported much higher concentration of boron that varies between 1.4 and 6.2 mg L⁻¹ in the alkaline (pH 7.25–9.25) and thermal (T 27–39 °C) groundwater of the continental intercalary aquifer of the Irhzer Plain at Niger. In Platystomo and Smokovo areas, boron content of thermal waters varies between 60 μg L⁻¹ and 10 mg L⁻¹. High concentrations (2–10 mg L⁻¹) present the samples of the first group (in flysh hosted thermal springs), while lower concentrations (60–1,000 μg L⁻¹) exhibit the mixed flysh waters and the spring waters issue from ophiolitic complex. Figure 4a shows that the boron content of the thermal water of Platystomo and Smokovo areas is increases with the increase in temperature.

Fig. 4

Arithmetic plots of relationships between the thermal water temperatures and boron concentrations (a); δ¹¹Β values (b); H₂BO₃ ⁻/H₃BO₃ ratio (c), pH versus H₂BO₃ ⁻/H₃BO₃ ratios (d), δ¹¹Β values, (e) and δ¹¹Β values versus H₂BO₃ ⁻/H₃BO₃ ratios (f)

In most groundwaters, boron occurs in the forms of Β(ΟΗ)₃ and Β(ΟΗ) ⁻₄ . In groundwater with pH values less than 8.8 (25 °C), the undisassociated boric acid Β(ΟΗ)₃ dominates and boron behaves as an inert element. Boron isotope ratios vary by more than 80 ‰ in geological materials (Palmer and Swihart 1996). From sediments hosted basins δ¹¹B varies between−10.1 and 23.3 ‰, (Palmer 1991). Siliciclastic sediments typically have low δ¹¹Β values (–15 to +5 ‰), (Palmer and Swihart 1996). Pennisi et al. (2000) reported δ¹¹Β values that vary between 4.3 and 9 ‰, corresponding to low salt content (Cl 95–140  mg L⁻¹) alkaline cold groundwater from flysch-filled basement depressions of the western and south western flank of Mt. Etna. In similar materials at the Tecopa Hot Springs, waters contain 8–10 mg L⁻¹ boron and have δ¹¹Β values that vary between−4.5 and−5 ‰ (Larsen et al. 2001). The seawater δ¹¹Β values approach 40 ‰ (Spivack and Edmond 1986), while the sea salt aerosols also, near ocean have a similar value. Thermal fluid of continental geothermal areas has low δ¹¹B values (Yellowstone National Park, δ¹¹B varies between−9.3 and 4.4 ‰, Palmer and Sturchio 1990) according to the isotopic composition of the source rocks. Marine carbonate is characterized by the dominance of the heavier isotope in relation to most igneous and sedimentary rocks, because boron is incorporated into marine water during carbonate precipitation. Between boron isotopes and under the condition that boron is in the form of Β(ΟΗ)₃ and Β(ΟΗ) ⁻₄ , the latter is enriched in the lighter isotope. Kakihara et al. (1977) computed that at 25 °C the enrichment factor is 1.0194. Then during sorption reactions or precipitation that removes Β(ΟΗ) ⁻₄ from the water, B(OH)₃ is enriched in the heavier isotope.

For the Platystomo and Smokovo areas, increase of temperature seems to contribute to the fractionation of boron isotopes resulting in the enrichment of the heavier isotope (Fig. 4b). Data processing showed that a similar process takes place, considering the evolution of the temperature versus the H₂BO₃/H₃BO₃ ratio. A general negative trend suggests clearly that H₃BO₃ specie enhanced relative to the H₂BO₃ ⁻ along with the temperature increases. The sample YP-11 is differentiated from this general trend. That could be attributed to the completely different nature of the rocks hosting the thermal waters and the fact that probably, YP-11 is influenced by volcanic phenomena as many authors propose (Marinos and Frangopoulos 1973; Garagunis 1978; Gartzos and Stamatis 1996; Fyticas et al. 1984). The pH strongly affects the H₂BO₃−/H₃BO₃ ratio of thermal waters. In Fig. 4d all samples, included rain, sea and YP-11, lie on the regression line that exhibits a high correlation coefficient, which is up to 98 %. The evolution of this ratio is independent from geological conditions, unlike boron isotope ratio, which is independent from thermal water pH, and H₂BO₃−/H₃BO₃ ratios (Fig. 4e, f).

Geothermometry

The geothermometers are important tools in estimating the temperature of the thermal water reservoir. They are based on the assumption that chemical equilibrium between minerals in the reservoir rocks and groundwater has been attained. After a good methodology followed by Minissale et al. (1997), the temperature is first calculated using geothermometers for known temperature reservoir. The coincidence of measured and calculated temperature reflects attainment of full equilibrium in the reservoir. In this case, the geothermometer could be used for estimation of the temperature of similar geothermal areas. The above-described methodology cannot be used for the estimation of the reservoir temperature of Platystomo and Smokovo areas and therefore, the state of the saturation of thermal water to different minerals was computed. Relationships between the measured temperature in the outlet of the springs and the ion composition of the waters have been also taken into consideration. Data processing (Table 4) showed that except calcite, aragonite, dolomite, chalcedony, and quartz with which thermal waters are almost in chemical equilibrium, for the rest of the minerals the saturation indices show undersaturated waters.

Figure 5 shows the dependence between temperature and the chemical species SiO₂, and Na. These positive correlations clearly indicate that the dissolved salts of the non-marine thermal waters are derived from water–rock interactions (Vengosh et al. 2002) confirming the proposed suggestions of Sect 4.2 regarding the origin of thermal waters. Then, the reservoir temperature was estimated using the selected quartz, chalcedony, and Na, K, Ca geothermometers. For the sample EK-4 with negative saturation indices in Chalcedony and Quartz the corresponding geothermometers could not be applied.

Fig. 5

Arithmetic plots of relationships between water temperature and chemical species: a T versus SiO₂, b T versus Li, c T versus Sr, d T versus Ca, e T versus Na, f T versus Cl

Table 5 shows that using the 4, 5 and 6 geothermometers, the flysch hosted thermal waters exhibited similar calculated temperatures of about 100 °C. For the sample PL-K-6 of mixed water, the geothermometers 4 and 5 gave a shifted temperature, about 30–40 °C higher than the precedents. This different behavior can be explained by the poor sensitivity of the Na/K ratio to mixing process. A same shift is shown for samples AR-9, EK-4, and KA-3 originated from the ophiolitic complex. The geothermometer 6 showed more homogenized temperatures for all samples ranging from 100 to 110 °C. The Chalcedony and Quartz geothermometer showed the lowest temperatures of the thermal water reservoir ranging from 70 to 100 °C. However, the Quartz geothermometer cannot be considered as a reliable estimation because all samples are saturated to Quartz, while the chalcedony temperatures are similar to the emergent temperatures. This is typical of convective waters in fault systems in normal thermal gradient areas.

Table 5 Temperature evaluated through chemical geothermometers for aqueous solutions (^oC)

Conclusions

The alkaline thermal water of the Platystomo–Smokovo area are characterized by strong reducing conditions that are related with high amounts of ammonia, low concentrations of NO_3, and high concentration of H₂S. They can form two major groups: the first with thermal waters of a Na–Cl–HCO₃ type from flysch formations correspond to the most pure ones. The second with samples of Ca–Mg–Na–HCO₃ water type comprise mixed water from flysch as well as the samples from the ophiolitic complex. Scatter plot diagrams show that the thermal water composition is mainly developed by rock–water interactions. Rock–water interactions are enhanced under the presence of CO₂ that is probably originates from metamorphic reactions (decarbonisation) of carbonate at depth, or through the recently active east–west fault systems from the Hypati (YP-11) spring located in the geothermal field of the south part of Sperchios Basin, where the groundwater content in CO₂ is about 1.000 mg L⁻¹. It is further suggested that the thermal waters of Platystomo and Smokovo areas are modified by rock–water interaction rainwater, heated in depth, and mixed with fresh water when arriving to the surface. A main characteristic of these waters is the significant low amounts in earth alkaline that may be explained by silicate hydrolysis of plagioclases. Although the concentrations of trace elements are strongly low until null, for the redox sensitive trace elements, this is related to the Eh conditions. Lithium content could be differentiated in thermal waters of a short circuit. Boron behavior depends on the temperature and the pH of the water. Although is not very clear, boron isotopes show enrichment in the lighter member when temperature increases. The decrease of the H₂BO₃ ⁻/H₃BO₃ ratio relative to the temperature suggests enrichment of H₃BO₃ specie with the temperature increases. Divergences of the YP-11 sample from the above-described general trend confirm that this sample origins from different thermal water reservoirs. The temperature of the thermal water reservoir was computed with the use of selected geothermometers. The selection was based on the study of the element concentration relative to the temperature evolution as well as to the study of the saturation indices of minerals. The selected Chalcedony, Quartz, and Na, K and Ca geothermometers showed that the reservoir temperature ranges from 100 to 110 °C.