Introduction

Natural radioactivity levels and the associated external exposure due to gamma radiation depend mainly on the local geological setting of an area and differ from region to region (United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) 2000; Florou et al. 2007; Jia and Torri 2007). The United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) has studied the exposure due to local geological settings (natural sources) and concluded that it contributes over 98% of the radiation dose to the population (excluding medical exposure). Whereas, the nuclear power production and nuclear weapons testing have only limited contribution. The global average human exposure from natural sources is 3.0 mSv/year (United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) 2008).

Thus, the determination of naturally occurring radionuclides in groundwater is essential in order to assess the environmental and public health impact (Isam et al. 2002). Furthermore, radiological studies of groundwater combined with geological and hydrochemical studies are of high importance, since (i) they provide valuable information leading to an integrated assessment of groundwater quality (United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) 2000; Isam et al. 2002; Jia et al. 2009) and, (ii) they can reveal important geological information, since the presence of natural radionuclides in groundwater is determined by their concentrations in the bedrock (Walencik et al. 2010).

In several cases, hot springs have been identified to be related with high natural radiation areas (Danali et al. 1986; Mishra 1990; Sohrabi 1990). In Greece, occurrence of hot springs is common due to the geological setting of the country. Specifically, presently active volcanic and hydrothermal locations occur mainly along the South Aegean Active Volcanic Arc which is related to the subduction of the African plate beneath the Aegean micro plate. The magmatic and volcanic processes and the tectonic setting which includes active graben and continental rift systems favor the rise of deep groundwater solutions that are discharged at the surface as hot springs. As a result of this geological setting, natural radioactivity in groundwater samples presents elevated concentrations compared with worldwide samples. Concentrations of up to 6289 Bq/L 222Rn, 5 Bq/L 226Rn, and 2.4 Bq/L 228Rn have been reported in the literature (Trampidou 2004; Florou et al. 2007; Athansoulis et al. 2009; Athanasoulis et al. 2016). Inactive volcanism occurs also further north, within the North Euboean Gulf and in a back-arc position with respect to the South Aegean Active Volcanic Arc. Hot spring occurrences in this area are related to the Plio-Pleistocene volcanic center of Lichades (Georgalas 1938; Pe-Piper and Piper 2002).

Here, we report on the radiological and hydrochemical characteristics of thermal and fresh groundwater samples, from the northern part of Euboea island and the neighboring part of the mainland, eastern central Greece. Natural radionuclide concentrations (222Rn, 226Ra, 228Ra, 228Th, and 40K) have been measured and are assessed in combination with temperature, pH, and major and trace ions and elemental concentrations in groundwater. The data are interpreted within the context of local geology in order to provide new information concerning the groundwater circulation and reveal new geological information. Also, a holistic water quality assessment is attempted combining both the radiological and hydrochemical data.

Study area

The study area includes the northern part of Euboea Island and the neighboring part of the mainland, eastern central Greece (Fig. 1). The volcanic center of Lichades islands lies in the center of the investigation area, in North Euboean Gulf. Sampling sites are located on-shore, on both sides of the North Euboean Gulf. Specifically, samples were collected from springs and boreholes at Ilia, Aedipsos, and Agios Georgios located on Euboea Island as well as from Atalanti, Arkitsa, Kamena Vourla, and Thermopylae located on mainland Greece. The selected sampling locations represent all main hot springs in the area, while fresh water samples were obtained from water points (springs and boreholes) that are used mainly for irrigation purposes in the area. Hot springs are scattered on both sides of the North Euboean Gulf (Kanellopoulos et al. 2017a, 2017b), in mainland and northern Euboea Island and they have been exploited for therapeutic spa purposes since the antiquity. Fresh groundwater samples were collected from springs and boreholes in close proximity to hot springs/boreholes.

Fig. 1
figure 1

Simplified geological map (modified after Kanellopoulos 2011)

Full size image

Lichades islands consist of trachyandesite lava flows, dated at 0.5 Ma. Volcanic rocks dating 1.7 Ma also outcrop in the Kamena Vourla area and comprise lavas of trachyandesite composition (Georgalas 1938; Pe-Piper and Piper 2002). The volcanic rocks are related to the neo-tectonic activity, since they are located along one of the main shear zones of the area (Kranis 1999). Innocenti et al. (2010) based on Sr–Nd–Pb isotopic data relate this volcanic center with the large volcanic belt that developed north of the Pelagonian–Attic–Cycladic–Menderes massifs, encompassing a 35-Ma time span. This belt is widespread over a large area from NW Greece–Macedonia to the Aegean–western Anatolia. Karastathis et al. (2011) showed that there is a magma chamber under the North Euboean Gulf area, at 7–8 km depth.

Bedrock geological formations of the Alpine orogenic cycle in the study area belong to the western part of the Geotectonic Units of the internal zones of Greece i.e. Pelagonian and Sub-Pelagonian (Mountrakis 1986; Jolivet et al. 2013). The eastern central part of mainland, were Thermopylae and Kamena Vourla are located, belong to the Sub-Pelagonian Geotectonic Unit. The basement consists mainly of carbonate rocks (limestone and dolomite) of Middle Triassic-Middle Jurassic age. An ophiolitic thrust sheet representing a relic of Thethian oceanic crust is overthrusted onto the carbonates. Drilling and geophysical data in Kamena Vourla indicate that a shale-chert formation rich in serpentinized ophiolite rocks exists at depth of 90–250 m (Orphanos and Sfetsos 1975). Post-Alpine surface deposits include a sequence of fluvio-lacustrine marls, clays, conglomerates, sands, as well as some lignite layers of late Miocene-Pleistocene age (Kranis 2007).

The northern Euboea, consist of both Pelagonian and Sub-Pelagonian Geotectonic Units. In the studied area, a lower series, with a Permian–Triassic volcaniclastic complex overlie a pre-middle to middle Carboniferous metamorphic basement and is overlain by middle Triassic shallow marine clastic and carbonate rocks intercalated with volcanic rocks best developed at the southeast part of Aedipsos (Katsikatsos et al. 1984). On top of this sequence, Jurassic limestone and Late Jurassic–Early Cretaceous ophiolite rocks occur (Katsikatsos et al. 1984; Scherreiks 2000). The Paleozoic and Mesozoic sequences were folded or imbricated as the result of two main tectonic events (Alpine and Eo-Alpine). In the greater area of northern Euboea, lignite layers have been identified inside Neogene-Lower Pleistocene lake sediments (Vakalopoulos et al. 2000).

The whole area is highly faulted due to the prevailing extensional tectonics. The Northern Euboean Gulf occurs at the western extremity of the North Anatolian Fault and is one of the most neo-tectonically active areas in Greece dominated by extensional tectonics as the rest of the Aegean Sea (Pe-Piper and Piper 2002; Jolivet et al. 2013). It is characterized dominantly by ENE-WSW to E-W and to a lesser degree by NNE-SSW to NE-SW trending normal and transtensional faults (Angelopoulos et al. 1991; Vavassis 2001; Kranis 1999, 2007), representing two graben structures, responsible for the formation of northern Euboea coastline (Northern Euboean Gulf Graben–Oreoi Strait) and of the Sperchios basin (Sperchios Graben).

Materials and methods

A total of 12 groundwater samples were collected (Table 1). The thermal groundwater samples were collected from the main hot springs and boreholes providing water for spa therapy and thermal baths in the study areas and the fresh groundwater samples were collected from representative fresh water springs and boreholes providing groundwater, mainly used for irrigation purposes. Fresh water sampling points are in most cases in close proximity to the hot spring/ boreholes.

Table 1 Samples locality, physiochemical parameters measured in situ and hydrochemical type
Full size table

Chemical analysis

The borehole samples were collected after adequate time of pumping in order to flush out the residual groundwater. All samples were collected in a 1-L and 100-mL polyethylene bottles for laboratory analyses. The 100-mL samples intended for the determination of metals were filtered through 0.45-μm membrane filters and acidified to pH < 2, with HNO3 (Suprapur 65%). All samples were stored in a portable cooler containing ice packs, transported to the laboratory and refrigerated at ~4 °C until analysis. The 1-L polyethylene bottles used for major ion determinations were filtered upon arrival at the laboratory, but not acidified.

Parameters such as pH, electrical conductivity (E.C.), total dissolved solids (T.D.S.), and temperature were measured in situ using portable multiparameter instruments of the Department of Geology and Geoenvironment, National and Kapodistrian University of Athens (N.K.U.A.). Analyses for total concentrations of four major ions (HCO3−, NO3−, Cl−, SO42−) and some major and trace elements (Cd, Co, Cr, Fe, K, Mg, Mn, Na, Ni, Pb, Zn) were carried out at the Department of Geology and Geoenvironment, N.K.U.A. according to the Standard Methods for the Examination of groundwater and wastewater (Clesceri et al. 1989). Alkalinity (as HCO3−) was measured using the titration method [2320-Β]. Nitrate ions were measured by the cadmium reduction method [4500-NO3 E], SO4 2− by the turbidimetric method [4500-SO4 E] and Cl− by the mercuric nitrate method [4500-Cl C], using a spectrophotometer (Hack DR/2000 or DR/4000). Multi-element standard solutions prepared by serial dilution of single certified standards were used for calibration of analytical instruments. The concentrations of K+ and Na+ were determined by flame emission photometry (Jenway PFP 7, see 3500-K B and 3500-Na B, Clesceri et al. 1989). Calcium ions and Mg2+ were determined by flame atomic absorption spectroscopy (see 3500-Ca and 3500-Mg, Clesceri et al. 1989) using a Perkin Elmer 603 instrument and trace metal elements by graphite furnace atomic absorption spectroscopy (Perkin Elmer 1100B). Elements Cd, Co, Cr, Mn, Pb, Ni, Fe, and Zn were determined by Atomic Absorption Spectrometry (AAS, Perkin Elmer 1100B), with Graphite Furnace (HGA-400). When concentrations were high (> 0.1 mg/L), Flame Atomic Absorption Spectrometry (F-AAS) was used instead for these elements (Table 2). Multi-element standard solutions prepared by serial dilution of single certified standards were used for calibration of analytical instruments.

Table 2 Geochemical analysis of groundwater samples
Full size table

Additionally, all samples were analyzed by inductively coupled plasma–atomic emission spectroscopy (ICP-AES) and by inductively coupled plasma-mass spectrometry (ICP-MS) at the accredited ACME Analytical Laboratories Ltd., Canada, for a series of major and trace elements.

Radiological analysis

Groundwater samples were collected in 1-L Marinelli beakers for the determination of 222Rn. The pH was adjusted to 1 by adding nitric acid and the beakers were sealed air tightly in order to avoid the diffusion of radon. About 4.5 L of groundwater samples were collected in plastic bottles for the determination of 226Ra, 228Ra, 40K, and 228Th and the pH was adjusted to 1 by adding nitric acid.

The radiological analyses were carried out at the National Centre for Scientific Research “Demokritos” (NCSRD)—Institute of Nuclear Technology—Radiation Protection. In the laboratory, Gamma-spectroscopy was applied, in order to determine the composition of these samples in 226Ra, 228Ra, 228Th, 222Rn, and 40K radionuclides (see Florou et al. 2007). A high-resolution gamma-spectrometry system was used to measure gamma-ray spectra in an energy range between 50 and 2000 keV. The system comprises a high-purity germanium (HPGe) detector with an efficiency of 20% relative to a 3" x 3" NaI (Tl) scintillator. The calibration of the HPGe detector was made by use of simulated multi gamma sources (in the same geometry). The multi-channel analyzer emulation software Maestro2 (ORTEC) allowed data acquisition, display, storage, and analysis of the acquired spectra. The samples were measured for 70,000 s, whereas the relative statistical error (1σ) was kept always to within 10%. So, the radionuclides activity concentrations were determined from the analysis of gamma lines of their decay products (Florou 1992; Trampidou 2004; Florou et al. 2007). This required a parent daughter secular equilibrium within the samples that was achieved through their sealing in containers and a time delay prior to gamma-spectroscopy. The quality assurance of the laboratory is ensured per year as Environmental Radioactivity Laboratory (ERL) participates to the Analytical Laboratories for the Measurement of Environmental Radioactivity of International Atomic Energy Agency (ALMERA-IAEA) proficiency tests.

Estimation of groundwater circulation depth and residence time

The geothermal reservoirs in most of the studied areas have not been verified by drilling data yet, so any information concerning the underground circulation of the geothermal fluids such as the depth of the groundwater circulation and residence time is of high interest.

The depth of groundwater circulation (Zcirc) has been estimated by the following equation (Wang et al. 2015):

$$ {Z}_{cicr}=-\frac{T_r-{T}_o}{GradT}+{Z}_o $$

where T r  = reservoir temperature (in oC); T o  = temperature of initial fresh groundwater (in oC); Zo = depth of the constant-temperature zone (in m), and GradT = temperature gradient (in m/oC).

Based on the 226Ra and 222Rn concentrations, the residence time (Tres) of the geothermal fluids has been estimated (Cherdyntsev 1971; Zhou et al. 2008; Liu et al. 2015), by applying the following equation:

$$ {T}_{res}=-\frac{1}{\lambda}\ln \left(1-\frac{N_{Ra}}{N_{Rn}}\right) $$

where λ = decay constant of 226Ra, i.e., 0.00043; NRa and NRn = 226Ra and 222 Rn contents (in Bq/L) respectively.

Radiological dose estimation

In order to estimate the equivalent dose to humans from ingestion of radionuclides by water drinking, the following parameters are required: (i) the concentration of the radionuclides in water (measured in Bq/L), (ii) the daily consumption rate of groundwater (L/day), and (iii) the dose conversion factor for each radionuclide.

According to WHO, adults consume an average of 2 L of water per day. So, the annual dose of radionuclide consumed in groundwater for an adult is calculated using the following equation:

Annual dose (mSv/year) = dose per unit intake (mSv/Bq) × annual groundwater consumption (L/y) × radionuclide concentration (Bq/L).

Results

Hydrochemical characteristics of groundwater samples

The physiochemical parameters measured in situ at each location and the hydrochemical types of each sample are presented in Table 1, while concentrations of major and trace ions and elements are presented in Table 2.

The chemical analyses data were plotted on a Piper diagram (Fig. 3), in order to evaluate the hydrochemistry of the studied groundwater samples. All thermal groundwater samples are plotted in the same area and belong to the Na-Cl hydrochemical type (Table 1). In contrast, the fresh groundwater samples are plotted separately and are scattered, indicating a variety of hydrochemical types.

Temperatures of the thermal groundwater samples range from 33 up to 80 °C; the maximum values were observed for the Euboea samples (Table 1, Fig. 2a). The fresh groundwater samples show distinctly lower values, varying from 12 up to 20 °C.

Fig. 2
figure 2

Chemical composition of groundwater samples plotted in Piper diagram. (With turquoise color are symbolized the fresh groundwater samples from eastern central Greece, with blue color the fresh groundwater samples from Euboea, with pink color are symbolized the hot groundwater samples from eastern central Greece and with red fresh the fresh groundwater samples from Euboea)

Full size image

Total dissolved solids (T.D.S.) and electrical conductivity (E.C.) in thermal groundwater samples vary from 7.5 to 54.5 g/L and from 11.3 to 27.1 mS/cm respectively. Aedipsos samples present the maximum values (Table 1). The fresh groundwater samples present noticeable lower values to T.D.S. and E.C. (0.3–1.2 g/L, 0.3–2.4 mS/cm, respectively).

Similarly, the physicochemical and chemical parameters show distinctive differentiation between the thermal and the fresh groundwater samples (Tables 1 and 2, Fig. 3a–c). All thermal groundwater samples show pH values below 7 (5.9–6.4) and all fresh groundwater samples approximate to and mostly above 7 (6.9–7.4, Table 1). All thermal groundwater samples show high Cl concentrations from 4400 up to 16,100 mg/L and the fresh groundwater samples from 24 up to 478 mg/L (Table 1). All thermal groundwater samples show high Ca concentrations from 470 up to 1140 mg/L and the fresh groundwater samples from 62 up to 120 mg/L (Table 1).

Fig. 3
figure 3

Diagrams presenting the a temperature vs pH values, b temperature values vs Cl concentrations, c temperature values vs Ca concentrations, d temperature values vs U concentrations. (With turquoise color are symbolized the fresh groundwater samples from eastern central Greece, with blue color the fresh groundwater samples from Euboea, with pink color are symbolized the hot groundwater samples from eastern central Greece and with red color the fresh groundwater samples from Euboea)

Full size image

All samples from Euboea show low concentrations in Ni (most of them even below detection limit), in contrast with the samples from eastern central Greece (up to 35 μg/L, Fig. 3d). Similarly, even though all samples, present low concentration of U, the samples from Kamena Vourla, both fresh and thermal ones, present distinctive higher uranium concentration (Table 2). These results indicate that for the specific elements of Ni and U, the lithological characteristics of the aquifer play the major role in determining water sample composition irrespectively their high or low temperature.

Radiological characteristics of groundwater samples

The radiological analysis results and selected ratios between natural radionuclides are presented in Table 3. The concentrations of naturally occurring radionuclides in the studied groundwater samples vary widely, with higher concentrations recorded in the thermal groundwater samples.

Table 3 Radiological analysis and radionuclide ratios
Full size table

In hot springs, the concentrations of 222Rn, 226Ra, and 228Ra range from < 0.5 up to 179 Bq/L for 222Rn, from <0.025 up to 5.2 Bq/L for 226Ra and from < 0.025 up to 2.91 Bq/L for 228Ra. In all samples, 222Rn concentrations are higher than the parental 226Ra, with only exception the T-AED-2 sample. In Kamena Vourla area, that phenomenon is particularly intense. The 222Rn/226Ra ratio varies from 1 to 917.

Circulation depth and groundwater residence time

In order to estimate the thermal groundwater circulation depth (Zcirc), for reservoir temperatures, representative estimated temperatures of geochemical and isotopic geothermometers were applied (Dotsika 2015; Kanellopoulos et al. 2016). Based on the study of Kanellopoulos and Mitropoulos (2013) and using representative fresh groundwater samples from the area near Aedipsos and Ilia, the average temperature of the local fresh groundwater is considered to be 14.5 °C. For Thermopylae and Kamena Vourla, based on Kanellopoulos (2011), the average temperature of the local fresh groundwater is assumed to be 11 and 19 °C, respectively.

As far as the geothermal gradients are concerned, data from the most recent and representative geothermal drills conducted by the Greek Institute of Geology and Mineral Exploration (I.G.M.E.) were calculated and applied. So, the average geothermal gradient is 18.7 °C/100 m for Aedipsos (based on AD-L-4 borehole, Chatzis et al. 2008), 10.4 °C/100 m for Ilia (based on G-18 borehole, Gkioni-Stavropoulou 1998), 20.08 °C/100 m for Thermopylae (based on SP-D-2 borehole, Metaxas et al. 2008), and 38.37 °C/100 m for Kamena Vourla (based on G11/72 borehole, Orfanos and Sfetsos 1975).

The estimated circulation depths (Table 4) are (i) ~1240 m for Ilia, (ii) ~800 m for Aedipsos, (iii) ~250 m for Kamena Vourla, and (iv) ~480 m for Thermopylae area.

Table 4 Groundwater circulation (Zcirc) time of thermal groundwater samples
Full size table

Based on the 226Ra and 222Rn concentrations, the estimated residence time (Tres) for the geothermal fluids (Table 5) is ~555 years for Ilia, ~27 years for Kamena Vourla, and ~ 403 years for Thermopylae. Because of the higher concentration of 226Ra to 222Rn, the residence time for the Aedipsos thermal water could not be estimated.

Table 5 Thermal groundwater residence time (Tres)
Full size table

Radionuclide dosimetry

Since groundwater contains more than one radionuclide, the doses arising from each radionuclide must be summed to give the total dose.

Different dose per unit intake (dose factors) were utilized for the ingestion of each radionuclide, i.e., for 226Ra is 2.8 × 10−4 mSv/Bq, for 228Ra is 6.9 × 10−4 mSv/Bq, for 228Th is 7.2 × 10−5 mSv/Bq, and for 40K is 6.2 × 10−6 mSv/Bq (International Atomic Energy Agency (IAEA) 1996; World Health Organization (WHO) 2002). For 222Ra dose per unit intake, there are several models estimating radon doses due to ingestion (Lopez et al., 2004). The main problem in order to estimate the committed effective dose and the dose equivalent in stomach is the fact that there are no internationally accepted dose factors. For adults, a committed effective dose per unit intake of groundwater dissolved radon has been estimated at 10−8 Sv/Bq (United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) 1993) and slightly higher for children. The estimation has been calculated by the application of a modified ICRP model for the ingestion of radon in groundwater (Kendall et al. 1988). The National Research Council of USA (Wallström 2001) proposed a conversion factor of 0.38 × 10−8 Sv/Bq for the committed effective dose. The estimates for the dose equivalent to stomach per unit activity ranged between 5 × 10−8 and 2 × 10−7 Sv/Bq and an average value of 10−7 Sv/Bq has been adopted (Oliveira et al. 2001; Lopez et al. 2004).

This estimation is believed to be an appropriate average for Greece, giving an annual consumption of 730 L for each Greek adult, assuming a 2 L oral consumption of water per day. Based on the abovementioned parameters, the dose estimations from ingestion were calculated and presented in Table 6.

Table 6 Dose estimations for ingestion
Full size table

Discussion

Geological origin of radionuclides

Based on the combined radiological and hydrochemical data, the thermal groundwater samples usually show small variation in their values and compared with the fresh groundwater samples show distinctive higher concentrations in natural radionuclides and major and trace ions and elements, indicative of an immiscible regime between fresh and thermal aquifers.

This clear differentiation, between the fresh and thermal samples and at the same time the close spatial relation between the thermal samples, could be explained by accepting the hypothesis that the hot springs in eastern central Greece (Kamena Vourla and Thermopylae) and northern Euboea (Aedipsos and Ilia) belong to a single geothermal system, as suggested by previous studies such as Kanellopoulos (2011) and Kanellopoulos et al. (2017a, 2017b). All hot springs from both Euboea island and eastern central Greece areas are the surface manifestation of a medium-, possible high- enthalpy geothermal field (Chatzis et al. 2008; Kanellopoulos et al. 2016, 2017b). Their occurrence is the result of the combination of the active tectonic regime and the Plio-Pleistocene magmatic activity in the area. The thermal groundwater is circulating through the major fault systems of the northern Euboea graben. All thermal groundwater samples show a Na-Cl hydrochemical type, irrelevant of their distance from the sea, indicating the high participation of seawater in that geothermal system. This is in agreement with previous isotopic studies (e.g., Dotsika 2015). In contrast, the fresh groundwater samples show several different hydrochemical types reflecting the lithological variety of the related aquifers.

Based on the geochemical analyses, elements like Ni and U present similar concentrations, at both fresh and thermal groundwater samples from some areas, which differs from the rest of the studied samples. In these cases, the local geology controls the presence and concentration of these trace elements in both fresh and thermal groundwater aquifers.

All samples from Euboea show low concentrations of Ni, in contrast with samples from eastern central Greece (i.e., Kamena Vourla, Thermopylae and Atalanti), which show concentrations up to 35 μg/L, as a result of the large occurrences of ultramafic rocks of the ophiolitic sequence in the greater area (Fig. 1) and their impact on the hydrochemical composition of the thermal and fresh groundwaters. Kanellopoulos et al. (2014, 2015) have identified similar geochemical anomalies in the groundwater-soil-plant system in the neighboring area of Atalanti and have attributed them to the weathering of ophiolite rock occurrences.

Combining the radiological and geochemical results (Tables 2 and 3), it is revealed that uranium and a number of natural radionuclides, show maximum concentrations in both, fresh and thermal groundwater samples from Kamena Vourla. The measured concentrations of natural radionuclides from Kamena Vourla hot springs are among the highest concentrations recorded in Greece (Trampidou 2004; Florou et al. 2007; Athansoulis et al. 2009; Athanasoulis et al. 2016). In previous studies, the highest concentrations of natural radionuclides have been recorded in hot springs located mainly in the Aegean Islands, where the hot springs are closely related to recent volcanic activity and plutonic rock emplacement, such as in Santorini and Ikaria Islands.

Uranium in the fresh and thermal groundwater samples of Kamena Vourla could be attributed to a plutonic intrusion rich in uranium minerals, unknown until now. This hypothesis is supported even farther by the radiological data, especially for the samples collected in Kamena Vourla, as 222Rn concentrations are higher than the parental 226Ra. This high imbalance between 226Ra and 222Rn could be explained by the high diffusion factor of 222Rn, especially when it is compared with the low solubility of 226Ra in groundwater. Consequently, the existence of 222Rn in groundwater indicates the surrounding rocks that are enriched in uranium minerals as its sources. Also, a small percentage of 222Rn probably emanates from the decay of the parental 226Ra.

Hot groundwater boreholes and springs in Kamena Vourla presenting the highest values of radioactivity are in close proximity to boreholes and springs with very low values (Orfanos and Sfetsos 1975). This observation is suggesting a significant influence of radiation with small spatial impact radius. Small occurrences (few m3) of traciandesite lava in the seaside of Kamena Vourla area, related to the volcanic center of Lichades provide further evidence of the presence of a plutonic intrusion at depth.

The low concentrations of 226Ra, especially in the thermal groundwater samples, could be associated with the presence of high concentrations of SO4−2 and Ca+ in these samples (Tables 2 and 3, Fig. 4), as suggested by Lopez et al. 2004. The presence of these ions in concentrations close to the CaSO4 solubility degree causes precipitation of the CaSO4 and consequently parallel co-precipitation of the RaSO4. It is noted that the vast thermogenic travertine deposits at Aedipsos and Thermopylae, characterized as the biggest active thermogenic travertine systems in Greece, also contain gypsum (Kanellopoulos et al. 2017a).

Fig. 4
figure 4

Concentrations of 226Ra (in Bq/L; with blue color), Ca (in mg/L; with red color) and SO42− (in mg/L; with green color)

Full size image

In the studied samples, the 222Rn/226Ra ratio varies from 0.3 to 917, which is among the highest values recorded in Greek hot springs (Trampidou 2004; Florou et al., 2007). Ikaria’s hot springs which are among the highest radioactive in Greece and worldwide, show 222Rn/226Ra ratios up to 967 (Trampidou 2004; Florou et al. 2007). In other Greek hot springs such as Loutraki, the values are as low as 105–160 (Trampidou 2004; Florou et al. 2007). Among the studied samples, the highest 222Rn/226Ra ratios were recorded in the fresh groundwater samples, compared to thermal groundwater sample ratios of 0.5 to 86.9. The high imbalance of 222Rn/226Ra indicates that only a small percentage of the dissolved 222Rn emanates from groundwater 226Ra. The 222Rn is inactive and easily dissoluble gas. So, most of the 222Rn originates from the surrounding rocks. It is released in the groundwater from the surrounding rocks with rate that depends on their porosity and other characteristics of the rocks (Bettencourt et al. 1988).

The 226Ra/228Ra ratios also present large variation, from 0.6 to 343 (Table 3). But in this case, the highest ratios, i.e., from 0.6 to 343, are observed for the thermal groundwater samples. In Ikaria’s hot springs, the 226Ra/228Ra ratio ranges from 1 to 7 and in Loutraki’s hot springs from 0.8 to 16 (Trampidou 2004; Florou et al. 2007). That large variation of ratios could be attributed to the geochemistry of parent isotopes. According to Cherdyntsev theory, 226Ra and 228Ra exist in groundwater, specifically in old-fossil groundwater, in ratios equal to the median concentration of 226Ra and 228Ra in bedrock. In regions where uranium enrichment exists, the 226Ra prevails. When the uranium concentration is low, 228Ra prevails.

Consequently, occurrences of uranium, which are not linked with important concentrations of thorium, give a high ratio of 226Ra/228Ra (Asikainen 1981). In the studied samples, most of the 226Ra/228Ra ratios are higher than unity, so it can be assumed that they are related with bedrock lithologies, that contain more uranium than thorium minerals. The highest 226Ra/228Ra ratio is 343 and was found in a thermal groundwater sample (T-KMV-2) from Kamena Vourla. Also, a high imbalance of 222Rn/226Ra ratio was recorded in the same sample. However, high concentrations of 222Rn and 226Ra or high ratios of 226Ra/228Ra (Table 3, Fig. 5) are not related essentially to high concentrations of uranium minerals in the aquifer’s surrounding area, but could also originate from intense migration of Ra or from the parent 238U in the greater region.

Fig. 5
figure 5

Histogram presenting activities (in Bq/L) of the analyzed radionuclides

Full size image

The geothermal reservoirs in the studied areas have not been validated by drilling data yet. So, any estimations concerning the depth of the geothermal fluid circulation and residence time, is valuable. With respect to the estimated circulation depths, it seems that the geothermal fluids in northern Euboea are circulating in greater depth (up to ~1240 m) compared to eastern central Greece’s geothermal fluids and consequently they present longer residence times (up to ~550 years). However, the circulation depth and the residence time at Thermopylae are comparable (~480 m and ~400 years). Exception to these estimations is the area of Kamena Vourla, where the estimated residence time is ~30 years and the circulation depth is considerable lower, i.e., 255 m. It is interesting that stable isotopic studies (Dotsika 2015) have suggested that the heating of the geothermal systems at Sperchios (e.g., Thermopylae) could be due to strong and deep circulation, probably with low velocity to allow the water time for the chemical change in the major fracture zone of the area. This hypothesis is in accordance with the Thermopylae circulation depths and residence time, but it contrasts with the Kamena Vourla estimations. The geothermal fluids of Kamena Vourla probably circulate using mainly the costal E-W fault system, and especially its intersections with some local fault systems. It could be assumed that the even though Kamena Vourla belong to the greater geothermal system of Euboea-Sperchios, it is presenting some distinctive characteristics. The empirical method of radioactive isotope tritium could be also applied in the future, in order to estimate the residence time from the recharge area to the discharge point.

Health implications of groundwater chemistry

In order to assess the possible suitability of the studied groundwater samples, for human consumption, the analytical values were compared with the parametric levels, imposed by the current relative legislation (Directive 98/83/EU). It is noted that the National Greek law is in agreement with European Union (EU) Directive 98/83/EU. All the thermal groundwater samples are exceeding the As and B parametric values and two of them are exciting also the Ni parametric value (Table 2). So, all of them are not suitable for drinking. Concerning the fresh groundwater samples, only the F-GLT-1, from Gialtra area, exceeds the NO3 parametric value.

Health and state agencies such as World Health Organization (WHO), United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) in an attempt to assess the impact of radiation in human health, have recommended specific guideline limits for isotopic concentration in drinking groundwater. These limits were based on the possible impact on human health and they must be considered as trigger values for further investigations (World Health Organization (WHO) 2011). For that reason, these values vary, for example, in relevance to 222Rn concentrations in groundwater, the United States Environmental Protection Agency (US-EPA) has defined a value of 11 Bq/l (Environmental Protection Agency (EPA) 2000), the UNSCEAR, has defined a value of 40 Bq/l (United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) 1993) and the WHO and EU have defined a value of 100 Bq/l (World Health Organization (WHO) 2011; European Union (EU) 2001; see Table 3). The studied samples are below these limits, with only exceptions being one fresh water sample (F-KMV-1) and one thermal water sample (T-KMV-2), both from Kamena Vourla area.

In order to achieve a more realistic and integrated approach about the radiological impact to human health, since we are recipients of a combination of radionuclides, the estimated annual dose has been calculated, revealing high estimated doses, in some cases. The global average annual dose per person was estimated by the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) (2008) to be about 3.0 mSv/year, with 19.6% derived from medical diagnosis and 0.4% from other anthropogenic sources. So, European Union (EU) (2013) has suggested to the member states to set the limit on the effective dose for public exposure at 1 mSv/year. United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) (1988) has estimated the total effective dose equivalent, from the natural origin radiological impact through all the pathways of internal irradiation, in areas of normal background for 222Rn to be 0.85 mSv/year. Although, water supplies with high 222Rn concentration could be used after special treatment, since 222Rn gas is characterized by high volatility and quickly dissipates when exposed to the atmospheric conditions. The calculated mean dose of 222Rn for the fresh and thermal groundwater samples is below the suggested value. Exception to that are both fresh and thermal groundwater samples from Kamena Vourla, which show values of 2.98 and 13.07 mSv/year 222Rn respectively.

The elevated values in natural radionuclides are closely related to the geological background of the studied areas, further studies ought to be conducted focusing especially on Kamena Vourla groundwaters, in order to further understand the geochemical behavior of radioisotopes. Also, further research is needed in order to better assess the levels of internal irradiation taking into account food consumption in this area, e.g., consumption of fish. However, no such data are available at the moment to allow a comparison between different regions in the world and Greece.

Conclusions

Combining the results from the radiological and hydrochemical analysis and the geology of the studied areas, the following conclusions arise:

  • All the studied thermal groundwater samples, from all regions, share similar hydrochemical characteristics, to such degree that could be categorized as one group, clearly distinguished from the fresh groundwater samples. The thermal and fresh groundwater samples show no mixing.

  • The hydrochemical characteristics of all groundwater samples were associated with the local rock formations.

  • High concentrations of 222Rn are mainly due to the enrichment of groundwater from the surrounding rocks.

  • A U-rich plutonic body at depth probably controls the natural radiation of groundwaters in the greater area of Kamena Vourla. Radionuclide ratios indicate that it has high-influence factor with small spatial radius impact.

  • The geothermal fluid in Euboea makes a deeper journey (up to ~1240 m) during its circulation, compared to eastern central Greece. The shallowest depth circulation was estimated for Kamena Vourla area (255 m).

  • The estimated residence time of the geothermal fluid is quite long (up to ~550 years), with only exception the Kamena Vourla areas (~30 years). These estimations are considered as indicative, until drilling data reveal the details of the geothermal fluid circulation.

  • Based on the Directive 98/83/EU and the estimated effective radiation doses for ingestion by adults, most of the fresh groundwater samples are suitable for drinking, with only few exceptions.

  • Based on the Directive 98/83/EU, all thermal groundwater samples are exceeding the parametric values of As and B and some of them exceed the Ni parametric value. Additionally, estimated effective radiation doses for ingestion of several thermal groundwater samples, by adults are in general above 1 mSv/year, making them not suitable for drinking.