Comparative Assessment of Chemical and Isotopic Composition of Geothermal Fluids in the Eastern Part of the Büyük Menderes Graben (Western Türkiye)

Abstract

In this study, we comparatively discuss chemical and isotopic characteristics of thermal waters from several geothermal fields within the eastern part of the Büyük Menderes graben, Western Türkiye. The studied thermal waters with a wide range of temperature (33 to 242 °C) and pH value (6.10 to 9.38) show water types varying from Ca-Mg-HCO₃-SO₄ to Na-HCO₃-SO₄. The chemical composition of waters is controlled by several processes, which include temperature, circulation depth, extent of water–rock interaction, dissolution/precipitation, mixing, cation exchange and microbial activity. All thermal waters are of meteoric origin and generally have deep circulation. δ¹³C data indicate that marine limestone and mantle-derived CO₂ are the major sources of carbon in thermal waters and δ³⁴S values imply that the sulfate is originated from the Neogene gypsums. At discharge temperatures, all thermal waters are saturated with various carbonate, silica and clay minerals, which is supported by the XRD analysis of scaling materials. The REY composition of the scaling samples showed that the limestone is the source rock for the rare earth elements. Thermal waters with a positive ¹⁸O shift of 2.7 to 4.6‰ have reservoir temperatures of 170–245 °C, whilst other waters with a shift of <1‰ have reservoir temperatures in the range of 79 to 166 °C. Regarding the distribution of temperature, electricity production seems to be more suitable in the western part of the region, although relatively low-temperature areas in the east also look promising.

1. Introduction

The investigation of renewable energy sources has become a major research issue due to environmental considerations and the instability of fossil fuel prices. In this context, geothermal resources in regions of intense tectonomagmatic activity have attracted significant attention to producing geothermal energy [1,2]. Geothermal fields occur across the world in a variety of geological settings and each geothermal system may have distinct characteristics which are reflected in the chemistry of thermal fluids and their potential applications [3]. The tectonic setting of Türkiye is represented by protracted seismic activity as manifested by a number of devastating earthquakes [4,5,6,7]. The seismic unrest in Anatolia is closely related to the collision between the Eurasian and the Arabian Plates [8]. This continental deformation has given rise to the formation of several major tectonic structures such as the North and East Anatolian fault zones and the Aegean horst–graben system [9]. As a part of the Alpine–Himalayan orogenic belt, Türkiye has experienced a complex tectonomagmatism since the Neogene, which has caused extensive geothermal activity as manifested by quite a number of geothermal areas across the country [10,11,12]. Mineral Research and Exploration Institute of Turkey (MTA) opened a total of 671 wells with a depth of 472,000 m from which 5000 MWt of heat energy has been produced [13]. The total geothermal electricity installation capacity in Türkiye is around 1691 MWe [14,15].

For the last three decades, Turkey has struggled to increase the rate of renewable energy sources in electricity production [16]. Especially since 2007, private companies have made significant progress with the acceleration of investments in this sector. However, despite Turkey being among the top five countries in the world in electricity production from geothermal energy and direct use, the country has not yet reached the desired level and only 1.54% of total energy consumption (2.75% of renewable energy sources) is met by the geothermal energy [14].

Western Anatolia comprises 78% of Turkey’s geothermal potential [8]. E–W tending graben systems in Western Anatolia are the result of the transitioning of compressional tectonics to the extensional regime during the Late Cenozoic time [7,17,18]. The NW–SE extending Denizli Graben (Denizli Basin) lying in the eastern termination of the Büyük Menderes Graben hosts several thermal waters with temperatures in the range of 29 to 242 °C [11,19]. Chemical and isotope compositions of thermal fluids in the Pamukkale, Karahayıt, Gölemezli, Yenicekent, Tekkehamam and Kızıldere geothermal fields in the Denizli Basin were the subject of intense research over the last decade [20,21,22,23,24,25,26]. In these studies, the evolution and reservoir temperatures of thermal waters were examined using various geochemical models. The Pamukkale area, only 16 km from the center of Denizli city, is home to world-famous travertines and a historical pool, so-called Cleopatra pool, the last Hellenistic queen of ancient Egypt. The increasing demand for hot waters that are utilized for spa centers and greenhouse applications has allowed the discovery of new geothermal fields and the drilling of additional wells.

In the present study, we sampled waters of one thermal spring and nine thermal wells in the Denizli Basin. Among these, waters of two high-temperature wells (97 and 98 °C) in the Tosunlar field, newly drilled wells in Karahayıt (KH-4) and Yenicekent (YK-5), were sampled for the first time. The aim of this study is to examine isotopic and geochemical features, and processes affecting the chemistry and mineral equilibrium of the thermal fluids in the Denizli Basin. The isotopic analyses include oxygen, hydrogen, sulfur and carbon (¹³C/¹²C and ¹⁴C) contents. Mineral compositions and REE concentrations of crust (scaling) samples are evaluated comparatively with water counterparts. Comparison of our chemical and stable isotope data with the Tekkehamam and Kızıldere geothermal fields, located at the western end of the Denizli basin, and with the geochemistry data compiled from previous studies will bring about a better understanding of the water–rock interaction process that controlled the geochemistry of thermal waters in the basin. We believe that this work, which used a multiple fluid–mineral equilibria approach with a variety of geochemical and isotopic data, will be very beneficial for further studies.

2. Geological and Hydrogeological Setting

The compressional regime, which started with the collision of the Eurasian and Arabian plates in the Miocene, gave rise to the westward movement of the Anatolian plate. It was changed to an extensional regime by the Late Miocene, resulting in the formation of horst–graben systems in Western Anatolia [8,9,27,28,29,30,31] (Figure 1a,b). The Denizli Basin with a length of 62 km and width of 7–28 km is located at the intersection of the E–W trending Büyük Menderes graben and the NW–SE trending Gediz graben [32,33,34,35] (Figure 1c).

The Menderes massif metamorphics are the basement rocks in the study area (Figure 1c). The lower part of the massif consists of alternating bands of Precambrian–Cambrian augen gneiss, quartzite, micaschist and amphibolite. These units are exposed in the western part of the area between Buldan and Kızıldere. They are overlain by the Devonian–Carboniferous schists. The uppermost part of the massif is composed of phyllite, quartzite and marble of the Permian–Carboniferous age. The Menderes massif is covered with a tectonic contact by the Çökelez formation, which is made up of Jurassic–Cretaceous limestone, dolomitic limestone and evaporitic sediments of the Lycian nappes. Neogene units that unconformably overlie the Massive and Çökelez formation consist of four different formations; from bottom to top, (1) conglomerate, sandstone, siltstone and lignite-bearing clayey limestone of the Miocene Kızılburun formation, (2) lacustrine limestone, marl, claystone, siltstone and evaporite sediments (e.g., gypsum) of the Miocene Sazak formation, (3) Miocene–Pliocene Kolankaya formation represented by the alternation of sandstone, siltstone, claystone and marl, and (4) Pliocene Ulubey formation consisting of lacustrine limestones intercalated with claystone and marl [36]. These Neogene units are covered with an angular unconformity by the Quaternary Tosunlar formation that is composed of loosely cemented conglomerate, sandstone, siltstone and mudstone [36]. Travertine and alluvium are the youngest units that set above these formations. Travertines are divided into two groups: fossil and modern travertines. The white travertines precipitated from the Pamukkale thermal spring waters and the yellowish red travertines in the Karahayıt and Yenice areas that precipitated from the thermal well waters represent the modern travertines. The NW-trending Akköy fissure ridge with a length of 1400 m, height of 40 m and width of up to 800 m is the most important fossil travertine [37]. The ages of travertines in the Pamukkale, Karahayıt and Akköy areas are at least 400,000 years [4,38,39] whilst those in Gölemezli attain ages exceeding 600,000 years [40].

Intense seismic activity and rapid extension in the region (30–40 mm/year; [9]) has caused devastating earthquakes around the ancient city of Hierapolis and Denizli [41,42,43,44]. Crustal thinning due to extension has resulted in a high geothermal gradient. Based on the reservoir temperatures calculated by the silica geothermometers, heat flux values of thermal springs in the Denizli region were found 193 mW/m² for Tekkehamam, 170 mW/m² for Kızıldere, 76 mW/m² for Pamukkale and 68 mW/m² for the Yenicekent geothermal fields [45]. In a recent study by Bilim et al. [46], Curie point depth for Denizli was estimated at 10–11 km and the heat flux and the geothermal gradient were reported as 135–160 mW/m² and 55–65 °C/km, respectively. Gradient wells with depths of 80–150 m drilled in the Kızıldere and Tekkehamam geothermal fields yielded gradients from 0.1 to 1.0 °C/m [47,48]. The geothermal gradient in the Karakova uplift 7 km SSW of Pamukkale was measured at 0.14 °C/m [49]. These high gradient values decrease rapidly from a depth of 150 to 200 m (e.g., Avşar and Altuntaş [24]). It can be concluded that the geothermal gradient in the region is at least 2–4 times higher than the world average (~33 °C/km).

Figure 1. (a) Main horst–graben structures in Western Türkiye (modified from Bozkurt [50]), (b) locations of major geothermal fields in the Büyük Menderes Graben (modified from Karakuş and Şimşek [19], Akkuş et al. [51], Şimşek [52]), (c) geological map of the study area (modified and adapted from Alçiçek et al. [23], Koçyiğit [35], Sun [53], Konak and Şenel [54]).

Figure 1. (a) Main horst–graben structures in Western Türkiye (modified from Bozkurt [50]), (b) locations of major geothermal fields in the Büyük Menderes Graben (modified from Karakuş and Şimşek [19], Akkuş et al. [51], Şimşek [52]), (c) geological map of the study area (modified and adapted from Alçiçek et al. [23], Koçyiğit [35], Sun [53], Konak and Şenel [54]).

Geothermal fields in the Denizli basin host more than one reservoir rock. The intensely fractured schist, quartzite and marble of the Menderes massif represent the deeper reservoir. In the Pamukkale area, highly fractured limestone and dolomitic limestones of the Çökelez formation that are naturally prone to dissolution are also considered reservoir rocks. The Miocene Kızılburun formation above the massif acts as cap rock and limestones of the overlying Sazak formation with a limited lateral continuity form a shallow reservoir. The impermeable lithologies of the younger formations are considered the cap rocks of the shallow reservoir (Figure 2).

The conceptual hydrogeological model of the studied areas (except for the Tosunlar area) has been discussed in detail in previous studies [20,21,22,23,24,61,62]. These models briefly include the seeping of rainwater into the ground, being heated by a high geothermal gradient, and ascending to the surface along the fault/fracture systems. Thermal waters emerge as thermal springs at temperatures of 25–57 °C in Yenicekent (especially in the Büyük Menderes River bed), Pamukkale and Gölemezli areas. However, the temperature of springs in Tekkehamam is as high as 98 °C. The geological and tectonic structure of the Tosunlar field is principally similar to other fields and represented by the same hydrogeological model.

In the north of Pamukkale, cold water springs generally discharge from the permeable units of the Ulubey and Kolankaya formations at flow rates ranging from 0.01 to 11 L/s. Wells drilled in the same units produce cold groundwater at a high flow rate. There are also low-discharge springs manifested from the fractured and altered sections of the gneisses around the Buldan area (Figure 1c).

3. Site Description

3.1. General

The studied geothermal fields are located in the western part of the Denizli Basin, at 38° latitude and 29° longitude (Figure 1b,c). The lithology logs of the geothermal wells opened in the Karahayıt, Gölemezli, Yenicekent, Tosunlar, Tekkehamam and Kızıldere fields are shown in Figure 2. The change in the thickness of units and reservoir depths in the Karahayıt and Gölemezli wells is probably due to faulting. In the study area, the high elevations are generally comprised of the horsts where the Menderes massif metamorphics are exposed, which attain 1400 m around Buldan. Sampled thermal waters and well elevations are 364–369 m in Pamukkale, 330 m in Karahayıt, 160–200 m in Gölemezli, 165 m in Yenicekent and 145 m in Tosunlar. Büyük Menderes River and its tributary Çürüksu Stream have an average flow rate of 44.3 and 9.36 m³/s, respectively [63]. According to meteorological data between 1991 and 2020, the average annual air temperature in the region is 16.9 °C. The hottest month is July (28.3 °C) and the coldest month is January (6.2 °C). Total annual precipitation is 573.8 mm, the rainiest month is January with 85.7 mm, and the driest month is August with 12.1 mm. For this region, potential evapotranspiration is calculated as 930 mm and actual evapotranspiration is 399 mm.

3.2. Geothermal Fields

In the Pamukkale geothermal field (PGF), there are 13 thermal water ponors and four main thermal springs; PAM (also known as Ancient Pool and Özel İdare), Jandarma, Beltes and İnciraltı (Figure 3 and Figure 4a). Pamukkale thermal waters are manifested along the surface ruptures of the Pamukkale Fault (NW–SE trending). Tracer and pumping tests carried out to determine the hydraulic relationship, flow direction and interactions between the main outlet points of thermal waters indicated the presence of two main flow systems in the area. There was a water loss of approximately 40 L/s along the waterway under the surface from the outlet to the natural channels carrying water to the travertine area [64]. Although the total flow rates of Pamukkale thermal springs fluctuated between approximately 150 and 400 L/s in the last 30 years, they tended to decrease especially in the last 10 years. Results of previous studies have revealed that the chemical compositions of Pamukkale thermal springs do not change significantly over time [23,25,62,65]. Temperatures of Pamukkale thermal waters and Çukurbağ spring (elevation 285 m) are about 35 and 56 °C (Figure 3 and Figure 4(a7)). Pamukkale thermal spring waters are used for bathing, travertine precipitation and agricultural irrigation after flow on the travertines. In this study, we sampled the spring PAM, which is discharging from the ancient pool.

The Karahayıt geothermal field (KGF) is a thermal tourism center located 21 km N of Denizli city. The first thermal well in Karahayıt was drilled in 1981, and the number of thermal drillings increased rapidly in the following years due to increasing tourism activity. As of 2006, there were more than 200 hot water wells in an area of about 1 km². The excessive thermal water exploitation from these wells caused the drying of all thermal springs (including the “Red water” spring, famous for the red travertines it precipitated) and some shallow wells. In order to make sustainable use of the Karahayıt geothermal fluid, thermal waters with temperatures ranging from 58.0 to 61.5 °C and flow rates between 12 and 58 L/s were produced from four geothermal wells (KH-1, KH-2, KH-3 and KH-4) opened by the Denizli Governorship in 2007 (Table 1, Figure 4b).

The Gölemezli geothermal field (GGF) is located 25 km NW of the city center of Denizli. In 1993, there were an old spa facility and four thermal springs with temperatures varying from 34 to 58.3 °C [65]. In the area, seven geothermal wells with temperatures up to 88 °C (DG-coded drillings in Table 1 and Figure 4c) were drilled by government agencies. Today, in addition to the official wells, this area hosts a thermal spring (Figure 4(c3)), and ten private thermal boreholes, only one of which is active. For this study, the waters of Göl-1 and Göl-2 wells were sampled.

The Yenicekent geothermal field (YGF) is located 32 km NW of the Denizli city. This field is home to Çizmeli (42 °C) and Kamara (55 °C) springs [66]. The Kamara thermal spring that was manifested through a fissure ridge was dried up in 1998 because of a thermal well (Kamara well) drilled near the spring which produced water with a temperature of 55.1 °C (Figure 4d). In the following years, another five wells (YK-1 to YK-5) opened in the YGF by governmental organizations yielded promising results (Table 1, Figure 4d). For the present study, we sampled water from the well YK-5.

The Tosunlar geothermal field (TGF) is located 25 km NW of the city center of Denizli (Figure 1c). Two deep wells (T-3 and T-4) drilled by private companies in 2007 and 2008 are currently used as reinjection wells. Another two wells (T-1 and T-2) opened in 2013 produced thermal water with a temperature of about 100 °C. The high temperature and CO₂ gas content of the thermal fluids enabled integrated use (Table 1). In this study, water samples were collected from T-1, T-2 and T-3 wells.

The Tekkehamam geothermal field (TeGF) is located 29 km west of Denizli city, at the eastern end of the Büyük Menderes graben (Figure 1c). TH-1 (615 m, 116 °C) in 1968 and TH-2 (2001 m, 168 °C) in 1997 are the first wells opened in this field, which is home to several thermal springs with temperatures varying between 29 and 98 °C [51]. In the following years, several other wells drilled by the public and private companies produced fluids with temperatures up to 170 °C.

Table 1. Information on some thermal wells in the study area and types of their use.

Geothermal Field	Well No	Date	Depth (m)	T (°C)	Q (L/s)	Utilization
Karahayıt	KH-1 a	2007	468	61.0	15	Thermal bath and spa, hotel and pension heating.
	KH-2 a	2007	452	61.0	40
	KH-3 a	2007	570	61.0	80
	KHR-1 b	2011	900	54.0	27
	KH-4 c	2019	650	54.5	30
Gölemezli	DG-1 d	2001	1500	88.0	15	Greenhouse heating (192,000 m2) and bathing.
	DG-2 d	2002	597	73.0	140
	DG-3 e	2002	549	66.0	110
	DG-4 e	2003	750	70.0	45
	DG-5 e	2003	750	62.0	30
	Göl-1 e	2008	605	65.0	120
	Göl-2 e	2011	137	68.0	136
Yenicekent	Kamara f	1998	146	55.1	-	Kamara well water is used for thermal bath.Greenhouse heating (55,000 m2).
	YK-1 g	2002	54	57.0	20
	YK-2 g	2002	238	67.0	140
	YK-3 g	2002	250	36.0	4
	YK-4 g	2013	300	53.0	34
	YK-5 h	2015	245	61.0	35
Tosunlar	T-1 i	2013	2463	132.0	144	Electricity production (3.8 MWe), CO2 production andgreenhouse heating (197,000 m2).
	T-2 i	2013	2653	148.8	39
	T-3 i	2007	1265	106.6	91
	T-4 i	2008	918	85.4	73
Tekkehamam	TH-2 d	1997	2001	168.0	12	Electricity production (106 MWe), thermal bath and spa, greenhouse heating.
Kızıldere	R-1 d	1998	2261	240.6	81	Electricity production (260 MWe), CO2 production, thermal bath, district heating, greenhouse heating.

Note: The temperature and flow rates in the Karahayıt, Gölemezli and Yenicekent fields are obtained from well tests that were carried out just after the drilling. The wells, which were artesian type at the time of their drilling, are currently operated with pumping. The temperatures of the Tosunlar, Tekkehamam and Kızıldere fields are bottom-hole temperatures. ^a Gökgöz et al. [55], ^b Alçiçek et al. [22], ^c Kocairi [56], ^d Akkuş et al. [51], ^e Gökgöz [67], ^f Bülbül [66], ^g Akdemir et al. [68], ^h Çelik [59], ⁱ Subay [60].

Table 1. Information on some thermal wells in the study area and types of their use.

Geothermal Field	Well No	Date	Depth (m)	T (°C)	Q (L/s)	Utilization
Karahayıt	KH-1 a	2007	468	61.0	15	Thermal bath and spa, hotel and pension heating.
	KH-2 a	2007	452	61.0	40
	KH-3 a	2007	570	61.0	80
	KHR-1 b	2011	900	54.0	27
	KH-4 c	2019	650	54.5	30
Gölemezli	DG-1 d	2001	1500	88.0	15	Greenhouse heating (192,000 m2) and bathing.
	DG-2 d	2002	597	73.0	140
	DG-3 e	2002	549	66.0	110
	DG-4 e	2003	750	70.0	45
	DG-5 e	2003	750	62.0	30
	Göl-1 e	2008	605	65.0	120
	Göl-2 e	2011	137	68.0	136
Yenicekent	Kamara f	1998	146	55.1	-	Kamara well water is used for thermal bath.Greenhouse heating (55,000 m2).
	YK-1 g	2002	54	57.0	20
	YK-2 g	2002	238	67.0	140
	YK-3 g	2002	250	36.0	4
	YK-4 g	2013	300	53.0	34
	YK-5 h	2015	245	61.0	35
Tosunlar	T-1 i	2013	2463	132.0	144	Electricity production (3.8 MWe), CO2 production andgreenhouse heating (197,000 m2).
	T-2 i	2013	2653	148.8	39
	T-3 i	2007	1265	106.6	91
	T-4 i	2008	918	85.4	73
Tekkehamam	TH-2 d	1997	2001	168.0	12	Electricity production (106 MWe), thermal bath and spa, greenhouse heating.
Kızıldere	R-1 d	1998	2261	240.6	81	Electricity production (260 MWe), CO2 production, thermal bath, district heating, greenhouse heating.

Note: The temperature and flow rates in the Karahayıt, Gölemezli and Yenicekent fields are obtained from well tests that were carried out just after the drilling. The wells, which were artesian type at the time of their drilling, are currently operated with pumping. The temperatures of the Tosunlar, Tekkehamam and Kızıldere fields are bottom-hole temperatures. ^a Gökgöz et al. [55], ^b Alçiçek et al. [22], ^c Kocairi [56], ^d Akkuş et al. [51], ^e Gökgöz [67], ^f Bülbül [66], ^g Akdemir et al. [68], ^h Çelik [59], ⁱ Subay [60].

The Kızıldere geothermal field (KıGF), the first discovered high enthalpy field in Türkiye, is located at the eastern end of the Büyük Menderes graben, 30 km northwest of Denizli city (Figure 1c). Geothermal studies in Kızıldere started with the collaboration between the MTA and the United Nations Development Program (UNDP) and the first thermal well was opened in 1968. The deep geothermal wells drilled between 1968 and 1986 produced geothermal fluid with temperatures of 198 to 212 °C [69]. R-1 reinjection well which was drilled in 1998 to a depth of 2261 m produced fluid of 242 °C, and since then it has been used as a production well [70]. Many production and reinjection wells have been opened in this field since 2008. The installed power of the plants increased from 17.4 MW in 1984 to 260 MW [26].

4. Materials and Methods

Physiochemical parameters of waters (e.g., pH, temperature, and electrical conductivity-EC) were measured at sampling points with a Hach Lange HQ40D multi-parameter meter (Hach Company, Loveland, CO, USA). Before the measurements, the device was calibrated using standard solutions. The batch for the anion analysis was filtered and collected into 250-milliliter high-density polyethylene (HDPE) containers. Another batch for cation and trace element analyses was sampled into 100-milliliter HDPE bottles and suprapure HNO₃ was added to prevent precipitation. The sample taken for SiO₂ analysis was diluted with pure water.

Waters for oxygen–hydrogen isotope and tritium analyses were sampled into 50-milliliter and 500-milliliter HDPE bottles. For carbon isotope analysis, waters were collected by filtering into brown glass containers and transported within a cold chain. Waters taken for δ³⁴S and δ¹⁸O (sulfate) analyses were filtered and acidified. Before the analyses, barium sulfate was precipitated by adding barium chloride dihydrate (BaCl₂.2H₂O).

Anion (Cl⁻, SO₄²⁻, HCO₃⁻ and F⁻), cation (Na⁺, K⁺, Ca²⁺, Mg²⁺ and Li⁺) and tritium analyses of samples were carried out at the Water Chemistry Laboratory of the Hacettepe University. Cation and anion concentrations were measured using a Thermo Scientific Dionex Ion Chromatography device (Waltham, MA, USA). An ICS 1000 and LC 25 were used for analyzing anions and cations, respectively. HCO₃⁻ was determined by titration method using methyl orange as indicator solution and 0.01 N H₂SO₄ as titrant solution. Charge-balance error of waters is in the range of −3.2 to 4.6%. Tritium was measured with the liquid scintillation (Perkin Elmer Quantulus, Waltham, MA, USA) method. Al, B, Ba and Sr elements were analyzed by ICP-MS technique at Bureau Veritas Mineral Laboratories (Canada) with the laboratory’s method detection limits ranging from <0.01 to <5 ppb. SiO₂ was analyzed with the spectrophotometric method (HACH DR/4000, Ames, IA, USA) at the Geochemistry Laboratory of the Pamukkale University.

δ¹⁸O and δD analyses were conducted at the SIRFER Laboratory of the University of Utah using a Picarro cavity ring-down spectrometer (Picarro, Santa Clara, CA, USA) with an analytical precision of 0.1‰ Vienna Standard Mean Ocean Water (VSMOW) for δ¹⁸O and 0.2‰ VSMOW for δD. The sulfur (δ³⁴S) (VCDT) and oxygen (δ¹⁸O) isotope analyses of sulfate and carbon (δ¹³C) isotope analysis of dissolved inorganic carbon (DIC) were determined at the Environmental Isotope Laboratory of the University of Waterloo, Canada. For δ³⁴S analysis, sample material was thermally converted to gas through an elemental analyzer coupled with a continuous flow isotope ratio mass spectrometer (CFIRMS). For δ¹⁸O measurement, barium sulfate was converted to CO gas through high temperature (1450 °C) pyrolysis combustion using an elemental analyzer coupled to an Isoprime (GV Instruments, Manchester, UK) CFIRMS. The analytical precision for both analyses is 0.3‰ (2σ).

For δ¹³C and ¹⁴C (DIC) measurements, the inorganic carbonate batch is converted to carbon dioxide which is captured and reduced to graphite in Pyrex glass tubes. Graphitized samples are analyzed at the National Electrostatics Corporation model 1.5SDH-1 Pelletron Accelerator (DirectAMS, Bothell, WA, USA). The δ¹³C-DIC result obtained is within the specification of ±0.2‰. Measurements of the radiocarbon concentration have a precision and accuracy of 0.3% pmC.

Rare earth element analyses of scale samples were conducted at the Bureau Veritas Mineral Laboratories (Canada) by ICP-MS technique. XRD analysis of scale samples was conducted with Shimadzu (Kyoto, Japan) XRD-6000 model X-ray diffractometer (Ni filtered, CuKα radiation) at the Technology Application Research Center of the Afyon Kocatepe University.

5. Results and Discussion

The physicochemical and isotope compositions of thermal water samples collected from PGF, KGF, GGF, YGF and TGF are given in

Table 2. Physicochemical compositions of thermal and cold water samples from geothermal fields in the east of the Büyük Menderes graben.

Site	Sample ID	T °C	EC	pH	Ca2+	Mg2+	Na+	K+	Cl−	SO42−	HCO3−	F−	Li+	Al	B	Ba	Sr	SiO2
This study
PGF	PAM	33.9	2390	6.19	381	93	40	5.1	9.5	653	932	0.9	0.13	0.004	0.96	0.019	6.32	46.6
KGF	KH-2	56.0	3000	6.35	512	119	138	24	25	1089	1208	1.2	0.27	0.010	2.13	0.062	10.68	40.6
	KH-3	60.7	3060	7.01	367	122	131	23	19	948	908	1.6	0.33	0.011	1.97	0.053	9.43	52.7
	KH-4	53.4	3150	7.28	484	92	153	33	29	630	1414	1.9	0.31	0.026	2.44	0.084	5.98	30
GGF	Göl-1	63.3	3780	7.32	318	93	335	56	36	470	1537	1.5	0.32	0.024	3.59	0.079	7.12	71.7
	Göl-2	69.0	4470	6.85	432	124	496	56	58	1812	1165	3.3	1.52	0.012	6.48	0.037	11.30	136
YGF	YK-5	61.5	3140	6.84	254	35	442	70	32	497	1476	2.5	0.66	0.017	3.42	0.056	5.11	57.9
TGF	T-1	97.0	3850	7.48	232	32	752	123	45	638	2063	2.4	0.59	0.020	4.33	0.104	7.50	138
	T-2	98.0	3770	7.77	215	40	688	132	44	651	2063	2.2	0.61	0.012	4.46	0.122	8.09	162
	T-3	84.1	4820	7.27	243	25	940	139	58	971	2332	3.5	1.54	0.030	7.28	0.069	7.16	113
Previous studies
PGF	PAM a,b	35.0	2300	6.37	455	87	40	5.9	12	688	1124	2.1	0.10	0.043	1.00	0.028	6.09	25
	Beltes b	34.7	2340	6.10	446	86	39	5.2	9.6	660	945	1.5	0.14	0.046	1.04	0.024	6.37	47.8
	Çukurbağ b	56.1	3010	6.90	527	114	128	26	27	980	1235	2.8	0.32	0.084	2.72	0.061	7.62	44
KGF	KH-2 b	59.5	2710	7.70	537	117	115	24	22	1078	1085	1.5	0.32	0.073	1.95	0.060	10.70	50
	KH-2 c	47.0	2500	6.25	486	153	145	29	33	1062	1225	na	na	na	na	na	na	51
	KH-3 d	58.2	3050	6.52	512	119	121	25	22	946	1155	na	na	na	na	na	na	na
GGF	DG-1 a	62.0	4030	6.60	390	121	505	55	73	1504	1369	4.9	1.50	na	7.90	na	na	94.1
	DG-3 e	53.0	2930	6.70	393	55	224	56	29	432	1470	na	na	na	na	na	na	47
	Göl-1 e	58.5	3760	6.69	591	124	346	76	42	552	2564	na	na	na	na	na	na	56
	Göl-2 e	65.0	4730	6.34	559	166	602	71	79	1936	1548	na	na	na	na	na	na	132
YGF	YK-1 f	51.5	3080	6.65	266	46	524	94	57	639	1780	na	na	na	na	na	na	46
TekGF	İnaltı a	96.0	4170	7.71	54	7.4	876	84	91	1560	676	17.7	1.30	na	22.0	na	na	164
	W-3 g	99.1	4080	9.38	66	7.9	867	107	76	1077	732	11.1	3.89	1.019	14.2	0.677	na	188
KıGF	R-1 h,i,*	242.0	5570	8.40	4.7	0.2	1380	162	122	717	2388	32	4.40	na	28.0	na	na	470
Cold waters	CS-1 c	11.5	390	7.62	62	4	6	1.2	4	12	203	na	na	na	na	na	na	17
	CS-2 c	13.4	380	7.68	76	8	3	0.5	1	12	264	na	na	na	na	na	na	24
	CS-3 c	8.7	460	8.03	101	5	35	3.4	9	15	316	na	na	na	na	na	na	19

Note: ^a Yıldırım and Güner [71], ^b Kele et al. [25], ^c Alçiçek et al. [22], ^d Özkul et al. [72], ^e Alçiçek et al. [21], ^f Alçiçek et al. [20], ^g Avşar and Altuntaş [24], ^h Tut Haklıdır et al. [26], ⁱ Şimşek [52]. * Bottom-hole temperature; Chemical analysis results of R-1 reflect reservoir water chemistry because of correction by its steam fraction [26]. EC electrical conductivity (µS/cm), concentrations are in mg/L, na: not analyzed.

and Table 3, respectively. These tables also show the analysis results of some thermal and cold waters in the Tekkehamam and Kızıldere areas that were reported in previous studies.

5.1. Physicochemical Characteristics

The studied thermal waters can be divided into six types according to their physicochemical properties (Figure 5 and Figure 6):

• Ca-Mg-HCO₃-SO₄ type thermal waters: This type of thermal spring water is unique to PGF. These waters, which precipitate the white travertine, are distinguished from thermal waters in other areas by their low temperature and low electrical conductivity and slightly acidic nature. Measurements made between 2002 and 2018 showed that the temperatures of these springs and other thermal springs in the area (Figure 4a) varied between 31.5 and 35.1 °C [23]. The thermal springs and private well waters in the plain have slightly lower temperatures falling in the range of 20 to 32 °C. The EC and pH values of PGF thermal waters vary within a narrow range between 2300 and 2390 µS/cm, and 6.10 and 6.37, respectively (Table 2). The dominant cation is Ca²⁺ (381–455 mg/L, 19.1–22.8 meq/L) and the dominant anion is HCO₃⁻ (932–1124 mg/L, 15.1–18.4 meq/L) (Figure 5 and Figure 6). This type of water has the lowest Na⁺, K⁺, Cl⁻, F⁻, Li⁺, B and SiO₂ contents among other types of thermal waters (Table 2).
• Ca-Mg-SO₄-HCO₃ type thermal waters: KGF thermal well waters characterize this type. The temperatures of this group of waters are between 47.0 and 60.7 °C, EC values are from 2500 to 3080 µS/cm and pH varies from 6.25 to 7.70. The dominant cation and dominant anion are Ca²⁺ and SO₄²⁻ although SO₄²⁻ concentrations are generally slightly higher than HCO₃⁻ (in meq/L). The EC is similar to PGF thermal waters; however, Mg²⁺, Na⁺, K⁺, Cl⁻, Li+, B and SiO₂ values are higher (Table 2, Figure 6). KH-4, one of the thermal wells in the KGF field, is of Ca-SO₄-HCO₃ type and distinguished by higher Na⁺, HCO₃⁻ and less Mg²⁺ and SO₄²⁻ contents (Table 2).
• Ca-Na-HCO₃-SO₄ and Ca-Na-SO₄-HCO₃ type thermal waters: This type of thermal spring water is unique to GGF. Most of the thermal springs in GGF have dried up after the intense drilling, and the temperature of one active spring (Eski Hamam) was measured as 48 °C (Figure 4(c3)). Göl-1 and DG-3 well waters are of Ca-Na-HCO₃-SO₄ type. The temperatures of these thermal well waters are between 53.3 and 63.0 °C, EC is from 2930 to 3780 μS/cm and pH is in the range from 6.69 to 7.32. Ca-Na-SO₄-HCO₃ type Göl-2 and DG-1 thermal well waters have higher temperatures (62.0–69.0 °C; 88 °C at bottom-hole in DG-1 [51]) and EC (4030–4730 μS/cm). Na⁺, Cl⁻ and SO₄²⁻, F⁻, Li⁺, B and SiO₂ values are higher and HCO₃⁻ contents are lower than Ca-Na-SO₄-HCO₃ type waters (Table 2, Figure 5 and Figure 6).
• Na-Ca-HCO₃-SO₄ type thermal waters: YGF and TGF thermal well waters are included in this group. The bottom-hole temperatures of the TGF waters, sampled for the first time in this study, vary between 88 and 149 °C [73] and the discharge temperatures are from 84 to 98 °C. YGF waters (YK-1 and YK-5) have lower temperature (~50–60 °C) and EC than TGF (T-1, T-2 and T-3) waters and lower B and SiO₂ values except Ca²⁺ and Mg²⁺ (Table 2).
• Na-SO₄-HCO₃ type thermal waters: Thermal springs and well waters in TeGF fall into this group. The temperatures of these thermal waters are reported between 23.4 and 168 °C [24,71,74]. Deep wells opened by private companies produced thermal waters with temperatures of about 170 °C. In these waters (in meq/L), 85% of the cations consist of Na⁺ (up to 876 mg/L), and 60–70% of the anions are comprised of SO₄²⁻ (up to 1560 mg/L). These thermal waters have the lowest HCO₃⁻ contents among other group waters. Except for the KıGF samples, Cl⁻, F⁻, Li⁺, B and SiO₂ values are significantly higher than other group waters (Table 2).
• Na-HCO₃-SO₄ type thermal waters: The KıGF thermal waters constitute this group, which are characterized by high temperature (137–245 °C) and EC values (3010–5520 μS/cm) and generally alkaline (pH: 6.3–8.7) thermal fluids. In this group, waters with temperature above 200 °C (24 thermal wells) have the lowest Ca²⁺ (<10.3 mg/L) and Mg²⁺ (<1.79 mg/L), and the highest Na⁺ (1035–1538 mg/L), K⁺ (83–347 mg/L), alkalinity (1806–2990 mg/L), Cl⁻ (93.²⁻254 mg/L), F⁻ (8–33 mg/L), Li⁺ (2.6–5.7 mg/L), B (11.1–31.0 mg/L) and SiO₂ (239–697 mg/L) contents [26]. The physicochemical properties of R-1 thermal well water (Table 2) best represent the KıGF waters.

Figure 5. Piper diagram of the waters.

Figure 5. Piper diagram of the waters.

Figure 6. Semi-logarithmic Schoeller diagram of the thermal waters.

Figure 6. Semi-logarithmic Schoeller diagram of the thermal waters.

Cold spring waters are generally of Ca-HCO₃ type (Figure 5). They are slightly alkaline (pH:7.62–8.03) and have low EC values (380–460 μS/cm) (Table 2).

5.1.1. Factors Affecting the Chemical Composition of Waters

If the Na⁺/Cl^- ratio in water is around one, the source of sodium in the water is halite dissolution; if it is greater than one, sodium is typically derived from silicate weathering [75]. In the Na–Cl graph (Figure 7a), there is a positive relationship between sodium and chloride contents of the studied waters (R² = 0.86). The Na⁺/Cl^- ratio increases with the dissolution of sodium-containing silicate minerals, especially through weathering and cation exchange [76]. The Na/Cl ratios of thermal and cold waters vary between 8.2 and 17.6. In addition, the base exchange index (BEI = [Cl⁻ − (Na⁺ + K⁺)]/Cl⁻, in meq/L, [77]) values are from −1.6 to −27.6. BEI values are generally negative in waters interacted with metamorphic rocks since the extent of contribution of alkali ions is much more than chloride that is released by the weathering of silicates [78]. The deep (Paleozoic and Mesozoic) reservoir rocks of the geothermal fields in the study area are composed of marble, schist, dolomitic limestone and limestones, which is supported by the negative BEI values of the waters.

Limestone, dolomitic limestone and evaporites of the Çökelez formation, limestones and gypsums of the Neogene units, and marble and schists of the Menderes massive are the major sources of Ca²⁺, Mg²⁺ and SO₄²⁻ ions in the thermal waters. The Neogene units roughly comprise 11 to 97% calcite, 3 to 37% dolomite and 0.5 to 39% gypsum [36]. In addition, marble and schists of the Menderes massive host some amount of dolomite [74]. Cretaceous ophiolitic mélange, consisting of serpentinized harzburgite, is another source particularly for Mg²⁺. Ca²⁺, Mg²⁺ and SO₄²⁻ ions, originating from the following dissolution reactions for limestone/marble (1), dolomite (2), gypsum (3) and serpentinite (4):

CaCO₃ + H₂O + CO₂ → Ca²⁺ + 2HCO₃

(1)

CaMg(CO₃)₂ + 2H₂O + 2CO₂ → Ca²⁺ + Mg²⁺ + 4HCO₃

(2)

CaSO₄.2H₂O + H₂O → Ca²⁺ + SO₄²⁻ + 3H₂O

(3)

Mg₃Si₂O₅(OH)₄ + 6CO₂ + 5H₂O → 3Mg²⁺ + 6HCO₃⁻ + 2H₄SiO₄

(4)

In the Ca²⁺ versus Mg²⁺ diagram (Figure 7b), there is a strong positive correlation between Ca²⁺ and Mg²⁺ contents of waters (R² = 0.84) and all samples lie to the right of the 1:1 line. The Ca²⁺/Mg²⁺ ratio of waters is greater than one indicating that calcite rather than dolomite is the primary carbonate mineral that dissolves [79]. For the studied thermal waters, this ratio varies between 1.8 (KH-3 well) and 13.8 (R-1 well), showing a calcite-rich carbonate source rock (e.g., marble and limestones). The dissolution of plagioclase in the metamorphic rocks of the Menderes Massif and gypsum in the Çökelez and Sazak formations also significantly contributes to the Ca²⁺ concentration in waters. The dissolution of dolomitic limestones and serpentinites is likely to be the main source of Mg²⁺.

In the Ca²⁺ versus HCO₃⁻ diagram (Figure 7c), thermal waters can be divided into two groups. The first is comprised of PGF, KGF and some GGF thermal waters (Göl-1 and Göl-2) with Ca²⁺/HCO₃⁻ ratio higher than one plotted in the grey area below the 1:1 line. The second group includes KıGF, TGF, YGF and some GGF thermal waters with Ca²⁺/HCO₃⁻ ratio lower than one plotted on and around the regression line. The Ca²⁺ and HCO₃⁻ of the first group of waters indicate derivation from carbonate rock dissolution. Gypsum dissolution is another mechanism that might increase the Ca²⁺ content of these thermal waters. In the second group of waters, there is a negative correlation between Ca²⁺ and HCO₃⁻ contents. In the thermal well waters of KıGF (R-1 well), TeGF, TGF, YGF and GGF, Ca²⁺ constitutes 0.35%, ~7%, ~21%, 34% and 40% of the total cations, respectively, while calcium content attains up to 60% in Pamukkale and Karahayıt waters. With increasing temperatures from PGF to KıGF, concentrations of Na⁺ (R² = 0.75), Cl⁻ (R² = 0.67), F⁻ (R² = 0.80), B (R² = 0.73), Li⁺ (R² = 0.66) and SiO₂ (R² = 0.93) increase whereas a decrease is recorded in Ca²⁺ (R² = 0.48) content. The high Na⁺ and low Ca²⁺ concentrations especially in KıGF and TGF thermal waters might be attributed to (a) a decrease in CaCO₃ solubility with increasing temperature, (b) calcite and aragonite precipitation as a result of boiling (CO₂ escape) during the rise of thermal waters to the surface, (c) increasing feldspar (NaAlSi₃O₈) dissolution with increasing temperature and (d) cation exchange between Na⁺ and Ca²⁺ (Figure 7d). Therefore, thermal waters in these fields have evolved into the Na–HCO₃ type (Figure 7e).

GGF (DG–1 and Göl–2) and TeGF thermal waters have high SO₄²⁻ and relatively low Ca²⁺ concentrations (especially in Tekkehamam thermal waters) (Table 2, Figure 5, Figure 6 and Figure 7f). According to Alçiçek [36], vein-type gypsums were crystallized within dissolution pores of the limestones of the Sazak formation that interact with sulfur-rich hydrothermal solutions. The quite high sulfate concentrations in Gölemezli and Tekkehamam thermal waters can be attributed to one of the reasons listed below:

• Gypsum dissolution within Neogene formations (particularly Sazak formation). In the Gölemezli and Tekkehamam fields, the Neogene cover above the Menderes massif is quite thick. For example, in the MDO-1 well drilled at Tekkehamam to a depth of 2401 m, the cover thickness is 1950 m (Sazak formation progresses between 814 and 1143 m) [74].
• Dissolution of sulfide minerals. Geothermal wells drilled in the region indicated that Menderes massif (especially schist) contains pyrite [51,73]. However, the amount of sulfate originating from sulfite oxidation in the thermal waters is considered insignificant.
• Oxidation of H₂S (or SO₂). The gas component of thermal fluids in the Kızıldere field is composed of 0.008 to 0.052% H₂S [26]. H₂S possibly occurs in the Tekkehamam and Gölemezli fields. Oxidation of H₂S results in the addition of large amounts of sulfate to the water in these fields. The elevated sulfate concentrations in the Tekkehamam springs could also have resulted from further oxidation of H₂S or SO₂ by the oxygen-rich groundwaters which might precipitate native sulfur [19].

Tekkehamam and Gölemezli thermal waters have an intense odor of H₂S, which is less likely to originate from the mantle. More likely, this gas may be formed by bacterial reduction of sulfate at shallow depths. During the drilling of the KB-5 well in the Tekkehamam field, a significant amount of petroleum-like material (PLM) along with 120 °C thermal water gushed to the surface from the clay and marl units of the Kolankaya formation at depths of 120–132 m [80]. Results of analyses showed that PLM from this well is indeed petroleum, and that the source rock is composed of Tertiary carbonate, clay and bacterial organic matter accumulated under terrestrial + anoxic conditions. A similar case is also recognized in the wells of the Gölemezli field (in the DG-1 and Göl-2 localities). When evaporitic gypsum comes into contact with organic matter-rich sediments or hydrocarbon-rich fluids, it allows the development of microbial communities dominated by sulfate-reducing bacteria, which dissolve gypsum at temperatures below 80 °C and promote carbonate and sulfur precipitation [81]. We believe that this mechanism plays a major role in the formation of native sulfur deposits in the Tekkehamam area. Native sulfur is observed at many points that precipitate from H₂S that rises through fractures to higher elevations than the thermal spring manifestation site.

Finally, the Ca²⁺ + Mg²⁺ versus HCO₃⁻ + SO₄²⁻ diagram shows that chemical compositions of PGF, KGF and GGF thermal waters might be best explained by the dissolution of carbonate, silicate and evaporite, while silicate weathering is the dominant factor defining the chemistry of waters in other areas (Figure 7g).

A linear relationship between δ¹⁸O and chloride content may be good evidence for a possible mixing between the thermal waters and warmer or colder groundwater [82]. In addition, the existence of a linear relationship between chloride and sodium, boron, fluoride and silica concentrations of waters also supports the mixing event. Considering the cold waters and R-1 thermal well water (in the KıGF field) as the two end members (Figure 7a and Figure 8), a strong positive correlation occurs between Cl⁻ and δ¹⁸O, Na⁺, B, F⁻ and SiO₂ with R² values ranging from 0.60 to 0.89. Although the B and F⁻ concentrations of cold waters are unknown (Table 2), the average concentrations of these ions were reported as 0.06 mg/L for B and 0.27 mg/L for F⁻ [22], which are significantly lower than the thermal waters. In conclusion, Figure 7 and Figure 8 show that there is an increasing cold groundwater contribution to the thermal waters from TeGF to TGF, GGF, YGF, KGF and PGF.

In the studied geothermal fields, the Cl⁻/SO₄²⁻ ratios increase whereas Ca²⁺/HCO₃⁻, Ca²⁺/SO₄²⁻and Ca²⁺/Na⁺ ratios decrease from PGF to KıGF (Figure 9). Considering the aforementioned assessments, the temperatures of thermal waters and their residence time (the extent of interaction with reservoir rocks) increase from PGF to KIGF, and the mixing ratios of cold groundwater decrease. Especially, the degree of silicate weathering increases from YGF to KıGF. The highest variations in these ionic ratios are observed in GGF thermal waters since the depths of thermal wells vary greatly (137 to 1500 m) and they receive water from different reservoirs (Figure 2).

Periodical analyses carried out in the Pamukkale, Karahayıt and Gölemezli thermal waters in the periods of 1991–1993 [65], 1993–1996 [62], 1995 [83] and 2023–2024 [84] indicated that their physicochemical compositions did not significantly change in time. Although there were substantial fluctuations in the chloride and nitrate levels of Pamukkale thermal waters due to anthropogenic pollution in the white travertine area (before 2000), the anomalous chloride concentrations disappeared with the demolition of hotels.

In the TeGF field, sometimes the thermal spring outlets become sealed due to CaCO₃ precipitation and the thermal water resurfaces from another nearby point forming a new spring. The decrease in thermal water pressure, the drying up of existing thermal springs or the emergence of new springs due to earthquakes are common to TeGF. It has been determined that variations in the temperature, electric potential and acoustic emission parameters of the thermal waters in this area are related to seismic activity [85]. For example, after the 12 April 2019 earthquake (Mw = 5.1) [86] with the epicenter very close to the İnaltı thermal spring, the flow rate of natural mud pools increased and their temperature rose to a value that would not allow people to enter. Another earthquake (Mw = 4.2) with an epicenter around the Pamukkale thermal spring area occurred on 28 September 2016. Unfortunately, there is no record documenting the possible changes in water chemistry before, during and after these earthquakes.

5.1.2. Mineral Saturation

The saturation states of studied thermal waters with respect to various minerals were calculated at discharge temperatures with the use of the PhreeqC program (version 3) [87]. Accordingly, all thermal waters are saturated with calcite, aragonite, dolomite, illite, K-feldspar, montmorillonite, muscovite and quartz, and undersaturated with fluorite and gypsum minerals. With the exception of Tosunlar thermal waters, all other samples are oversaturated with respect to gibbsite and kaolinite. Albite is undersaturated in waters PAM, KH-2 and KH-3 but it equilibrates with all other waters. Excluding the Göl-2 sample, all other waters are undersaturated with anhydrite. Thermal waters are at near saturation with respect to chalcedony.

In XRD analyses, calcite and aragonite were detected in the scaling samples from the T-1, Göl-1 and DG-3 wells. In the scaling material of the Göl-2 well, quartz and calcite, as well as gypsum, were recorded, considering that this well was drilled at a shallower depth (141 m) compared to Göl-1 (604 m) and was discharged from the evaporite-bearing units (e.g., gypsum and anhydrite). In YK-5 and KH-2 wells, only calcite and dolomite were detected as the scaling. In the Kızıldere field, calcite, aragonite, dolomite, strontianite and amorphous silica are the major scaling minerals [88,89,90]. In all studied fields, the scaling problem possibly creates a significant problem during energy production which might be solved by inhibitor injection.

5.2. Isotope Composition

5.2.1. Oxygen–Hydrogen Isotopes

For thermal waters, δ¹⁸O values vary from −9.28 (PAM) to −8.01‰ (YK-5) and δD values range from −60.82 (T-2) to −55.17‰ VSMOW (YK-5) (Table 3). In previous studies, δ¹⁸O and δD were, respectively, reported −10.0 to −8.84‰ and −61.0 to −57.0‰ (n = 18) [23,25,71,72,91,92,93] for PGF, −9.14 to −8.07‰ and −59.5 to −51.8‰ (n = 14) [22,25,52,71,72,83,91,92] for KGF, −8.54 to −8.13‰ and −61.09 to −56.2‰ (n = 11) [21,91,92] for GGF and −8.80 to −7.50‰ and −59.7 to −50.2‰ VSMOW (n = 18) [20,66,71,91,92,94,95] for YGF. These values are consistent with the oxygen and hydrogen isotope compositions of studied waters. δ¹⁸O and δD of TeGF and KıGF thermal waters are significantly higher than those of other fields, ranging from −7.40 to −4.10‰ and −57.60 to −51.30‰ for TeGF [52,71] and −8.50 to −4.30‰ and −61.81 to −34.10‰ for KıGF [26,52,71,94], respectively. No significant change was observed in the δ¹⁸O and δD values of the Pamukkale and Karahayıt thermal waters during the monitoring carried out in 1994–1995 [62]. However, stable isotope values of some thermal samples collected in the TeGF field during dry and rainy periods showed negligible seasonal differences [24]. Oxygen and hydrogen isotope values of cold waters are between −8.80 and −7.24‰ and from −56.0 to −50.01‰ (n = 14) [21,71,83], respectively.

In the δ¹⁸O–δD graphic (Figure 10a), cold waters and most low-temperature waters (<70 °C) plot between the local meteoric water line (LMWL) [96] and global meteoric water line (GMWL) [97]. One sample from the PGF thermal waters plots close to the Mediterranean meteoric water line (MMWL) [98]. It is concluded that studied cold and thermal waters have a meteoric origin. There is no significant ¹⁸O shift (between 0.4 and 1‰) for the PGF and KGF thermal waters. For samples GGF, YGF and TGF, δ¹⁸O deviation from LMWL is around 1.2–1.5‰. The highest deviations are seen in waters TeGF (2.7 to 3.3‰) and KıGF (3.2 to 4.6‰) (Figure 10a).

Table 3. Results of isotope analyses for the studied waters.

Site	Sample ID	δ18O	δD	d-Excess	T	δ13C	δ34S(SO4)	δ18O(SO4)	14C	Age *	Age
		(VSMOW)	(VSMOW)	(VSMOW)	(TU)	(VPDB)	(VCDT)	(VSMOW)	(pmC)	(YBP)	(YBP)
This study
Pamukkale	PAM	−9.28	−59.75	14.49	2.16	5.68	16.82	13.94	na	na	na
Karahayıt	KH-3	−8.81	−57.69	12.79	2.13	5.45	17.11	13.73	1.104	35,954	36,202
	KH-4	−8.57	−55.63	12.93	na	na	na	na	na	na	na
Gölemezli	Göl-1	−8.54	−57.62	10.70	0.54	7.65	16.43	12.64	0.617	40,613	40,878
	Göl-2	−8.50	−60.56	7.44	0.02	3.79	20.04	15.71	0.676	39,909	40,143
Yenicekent	YK-5	−8.01	−55.17	8.91	0.05	6.32	14.10	10.65	0.930	37,408	37,662
Tosunlar	T-1	−8.43	−59.35	8.09	0.71	2.55	16.46	11.33	5.879	22,538	22,764
	T-2	−8.58	−60.82	7.82	na	na	na	na	na		na
	T-3	−8.16	−59.57	5.71	0.54	2.20	16.59	10.64	1.578	33,107	33,328
Rainwater	RW	na	na	na	5.64	na	na	na	na	na	na
Previous studies
Pamukkale	PAM a	−9.90	−58.70	20.50	3.50	na	14.30	16.74	na	na	na
	Beltes b	−8.30	−58.90	7.50	na	na	na	na	na	na	na
	Çukurbağ b	−9.00	−58.70	13.30	na	na	na	na	na	na	na
Karahayıt	KH-2 c	−8.45	−56.49	11.11	na	9.76	na	na	na	na	na
	KH-3 d	−8.24	−53.00	12.93	na	na	na	na	na	na	na
Gölemezli	DG-3 e	−8.33	−57.41	9.23	0.40	6.98	na	na	na	na	na
	Göl-1 e	−8.40	−57.80	9.40	0.23	6.18	na	na	na	na	na
	Göl-2 e	−8.26	−60.46	5.62	0.21	5.11	na	na	na	na	na
Yenicekent	YK-1 f	−7.98	−56.25	7.59	na	6.80	na	na	na	na	na
Tekkehamam	İnaltı a, g	−6.70	−57.50	−3.90	0.10	0.90	11.10	18.69	2.130	-	34,898
	W-3 h	−6.43	−53.32	−1.88	0.00	na	na	na	na	na	na
Kızıldere	R-1 i	−4.10	−51.30	−18.50	0.33	1.26	na	na	2.270		31,034
Cold waters	CS-1 c	−8.44	−54.70	12.82	na	−10.38	na	na	na	na	na
	CS-2 c	−8.51	−54.19	13.89	2.73	−9.71	na	na	na	na	na
	CS-3 c	−7.90	−54.13	9.07	na	−8.16	na	na	na	na	na

Note: ^a Yıldırım and Güner [71], ^b Kele et al. [25], ^c Alçiçek et al. [22], ^d Özkul et al. [72], ^e Alçiçek et al. [21], ^f Alçiçek et al. [20], ^g Güner and Yıldırım [99], ^h Avşar and Altuntaş [24], ⁱ Şimşek [52]. * Uncorrected radiocarbon ages; na: not analyzed.

Table 3. Results of isotope analyses for the studied waters.

Site	Sample ID	δ18O	δD	d-Excess	T	δ13C	δ34S(SO4)	δ18O(SO4)	14C	Age *	Age
		(VSMOW)	(VSMOW)	(VSMOW)	(TU)	(VPDB)	(VCDT)	(VSMOW)	(pmC)	(YBP)	(YBP)
This study
Pamukkale	PAM	−9.28	−59.75	14.49	2.16	5.68	16.82	13.94	na	na	na
Karahayıt	KH-3	−8.81	−57.69	12.79	2.13	5.45	17.11	13.73	1.104	35,954	36,202
	KH-4	−8.57	−55.63	12.93	na	na	na	na	na	na	na
Gölemezli	Göl-1	−8.54	−57.62	10.70	0.54	7.65	16.43	12.64	0.617	40,613	40,878
	Göl-2	−8.50	−60.56	7.44	0.02	3.79	20.04	15.71	0.676	39,909	40,143
Yenicekent	YK-5	−8.01	−55.17	8.91	0.05	6.32	14.10	10.65	0.930	37,408	37,662
Tosunlar	T-1	−8.43	−59.35	8.09	0.71	2.55	16.46	11.33	5.879	22,538	22,764
	T-2	−8.58	−60.82	7.82	na	na	na	na	na		na
	T-3	−8.16	−59.57	5.71	0.54	2.20	16.59	10.64	1.578	33,107	33,328
Rainwater	RW	na	na	na	5.64	na	na	na	na	na	na
Previous studies
Pamukkale	PAM a	−9.90	−58.70	20.50	3.50	na	14.30	16.74	na	na	na
	Beltes b	−8.30	−58.90	7.50	na	na	na	na	na	na	na
	Çukurbağ b	−9.00	−58.70	13.30	na	na	na	na	na	na	na
Karahayıt	KH-2 c	−8.45	−56.49	11.11	na	9.76	na	na	na	na	na
	KH-3 d	−8.24	−53.00	12.93	na	na	na	na	na	na	na
Gölemezli	DG-3 e	−8.33	−57.41	9.23	0.40	6.98	na	na	na	na	na
	Göl-1 e	−8.40	−57.80	9.40	0.23	6.18	na	na	na	na	na
	Göl-2 e	−8.26	−60.46	5.62	0.21	5.11	na	na	na	na	na
Yenicekent	YK-1 f	−7.98	−56.25	7.59	na	6.80	na	na	na	na	na
Tekkehamam	İnaltı a, g	−6.70	−57.50	−3.90	0.10	0.90	11.10	18.69	2.130	-	34,898
	W-3 h	−6.43	−53.32	−1.88	0.00	na	na	na	na	na	na
Kızıldere	R-1 i	−4.10	−51.30	−18.50	0.33	1.26	na	na	2.270		31,034
Cold waters	CS-1 c	−8.44	−54.70	12.82	na	−10.38	na	na	na	na	na
	CS-2 c	−8.51	−54.19	13.89	2.73	−9.71	na	na	na	na	na
	CS-3 c	−7.90	−54.13	9.07	na	−8.16	na	na	na	na	na

Note: ^a Yıldırım and Güner [71], ^b Kele et al. [25], ^c Alçiçek et al. [22], ^d Özkul et al. [72], ^e Alçiçek et al. [21], ^f Alçiçek et al. [20], ^g Güner and Yıldırım [99], ^h Avşar and Altuntaş [24], ⁱ Şimşek [52]. * Uncorrected radiocarbon ages; na: not analyzed.

¹⁸O exchange between mineral and water depends on several parameters which include temperature, reservoir mineralogy, mineral dissolution–precipitation processes, mixing, mineral alteration, circulation rate and water/rock ratio [100]. The reservoir rocks in the study area are comprised of Cenozoic/Mesozoic limestone, Paleozoic marble, schist, and quartzite. For this reason, ¹⁸O exchange in the reservoir occurs by the reactions between thermal waters and calcite, quartz (or chalcedony) and feldspar minerals [100]:

CaCO₂¹⁸O + H₂O ↔ CaCO₃ + H₂¹⁸O

(5)

SiO¹⁸O + H₂O ↔ SiO₂ + H₂¹⁸O

(6)

CaAl₂Si₂O₇¹⁸O + H₂O ↔ CaAl₂Si₂O₈ + H₂¹⁸O

(7)

Steam separation from thermal fluids in the TeGF and KıGF fields might result in ¹⁸O enrichment of residual waters. The degree of isotope exchange rate is increased with increasing temperature and therefore ¹⁸O shift in thermal waters reflects the establishment of mineral–water equilibrium at reservoir temperatures [100]. This is more pronounced for the Kızıldere waters which have the highest bottom-hole temperatures. The oxygen isotope composition of Kızıldere waters (samples labeled KD and R-1 in Figure 10a) display a positive deviation up to 4‰ from the global meteoric water line indicating an intense water–rock interaction with carbonate and silicate rocks by the reactions shown in Equations (5)–(7).

Using the δD–elevation relation of cold water springs on the Yenice horst, Yıldırım and Güner [71] estimated the average recharge elevations of PGF, KGF, GGF and YGF thermal waters in the range of 1070 to 1370 m which correspond to area comprised by permeable lithologies such as limestones of Sazak formation and sandstones of Kolankaya formation. Results of the same study showed that TeGF and KıGF thermal waters are manifested through the gneiss and schists at elevations of 1300 to 1900 m. In a recent study by Alçiçek et al. [22,23], based on the δ¹⁸O–altitude relation, the recharge elevation of PGF and KGF waters is reported as 850–900 m. The reaction between high-temperature waters and oxygen-bearing minerals within the reservoir may cause δ¹⁸O enrichment in the thermal water. For this reason, in the estimation of recharge elevation, δD–altitude relation is expected to yield much more realistic results. We made an attempt here to determine the recharge elevations of sampled waters using δD values of cold spring waters that issue at different elevations of studied fields (Figure 10b). There is a strong positive correlation between δD contents of waters and elevations (R² = 0.99) and the elevation effect is found –4‰/100 m. As shown in Figure 10b, thermal waters sampled in this study have recharge elevations of 1030–1175 m, which agree with values proposed by Yıldırım and Güner [71].

Figure 10. (a) δ¹⁸O vs. δD diagram for the thermal and cold waters of the Denizli basin. Stable isotope data from previous studies are also included for comparison. In addition to studies cited in Table 3, the Tekkehamam data of Karakuş and Şimşek [19] and Kızıldere data of Şimşek [52] and Yaman and Özgür [101] were also used. (b) Recharge elevations estimated for the sampled waters. Data for Gürpınar, Sakızcılar and Sepet springs are from Önhon et al. [102] and Pınarbaşı spring data are from Yıldırım and Güner [71]. Shaded area represents the estimated range of recharge elevation.

Figure 10. (a) δ¹⁸O vs. δD diagram for the thermal and cold waters of the Denizli basin. Stable isotope data from previous studies are also included for comparison. In addition to studies cited in Table 3, the Tekkehamam data of Karakuş and Şimşek [19] and Kızıldere data of Şimşek [52] and Yaman and Özgür [101] were also used. (b) Recharge elevations estimated for the sampled waters. Data for Gürpınar, Sakızcılar and Sepet springs are from Önhon et al. [102] and Pınarbaşı spring data are from Yıldırım and Güner [71]. Shaded area represents the estimated range of recharge elevation.

5.2.2. Water Circulation

Tritium contents were measured for both thermal and local rainwater. Tritium ranges from 0.02 (Göl-2) to 2.16 TU (PAM) in thermal waters (Table 3). Tritium in local rainwater was measured at 5.64 TU. At the Yeşiloba station 50 km NE of the study area, the ³H content of 1988 precipitation was recorded in the range between 4 and 22 TU [102]. Tritium values in previous studies were reported at 0.10 to 4.50 TU for PGF thermal waters and <1.20 TU for waters in other fields [21,22,23,24,52,71,83,92,95]. These values are consistent with our tritium data. ³H content in cold waters is quite high varying from 0.39 to 17.4 TU (mean 12) [22,71,83].

Except for the Pamukkale thermal spring and Karahayıt KH-3 well, ³H values of thermal waters are close to zero indicating a deep circulation. In other words, these fluids are sub-modern waters recharged by precipitation of pre-1952 (<0.8 TU; [100]). In this study, tritium values of PAM and KH-3 thermal waters are very close (2.16 and 2.13 TU). In previous studies, ³H values of PGF and KGF waters were reported 0.10–4.50 TU [23,52,71,93] and 0.00–1.20 TU [22,52,71,91], respectively. It is likely that PGF thermal waters and, depending on the sampling period, KH-3 thermal water reflect the mixture of sub-modern and modern precipitation (0.8 > TU > ~4; [100]). The limited number of cold water samples indicate sub-modern recharge whilst the majority imply recharge by recent precipitation (5–15 TU, <5–10 years; [100]). Because chloride behaves in a conservative manner, it may yield qualitative information on the circulation time of waters. Comparison of tritium and chloride contents of studied waters yields that waters with low Cl⁻ concentrations (<20 mg/L) are correlated with high tritium values (>2 TU) whereas waters with high Cl⁻ concentrations have low tritium contents (<0.8 TU) (Table 2 and Table 3).

¹⁴C values measured in selected thermal waters are 0.617 (Göl-1) to 5.879 pmC (T-1) (Table 3). In previous studies, pmC values were reported at 2.13 [99] and 0.44 [24] for TeGF thermal waters and 94.13 for the cold water (T = 7.7 °C, EC = 118 μS/cm) [24]. The pmC in four thermal wells in KıGF was measured between 0.21 and 2.27 [99]. Cold spring at the recharge area of KıGF (T = 9.9 °C, EC = 118 μS/cm) has a pmC of 116 [99]. The pmC values of waters sampled in this study are slightly higher than thermal fluids in TGF and KıGF fields. The apparent ¹⁴C age of these samples was computed using the method outlined below.

The Radiocarbon dating technique is based on the measurement of carbon lost by decay of parent radionuclide (¹⁴C). Application of this method prerequisites two major assumptions: (1) the initial activity of parent radionuclide is known which remained constant in the past and (2) the system is closed to the loss and gain of parent radionuclide except for radioactive decay. When these conditions are ensured, the age of water can be computed from the decay equation:

$${\mathsf{\alpha }}_{\mathrm{t}}={\mathsf{\alpha }}_0\times {\mathrm{e}}^{-\mathsf{\lambda }\mathrm{t}}$$

(8)

• ${\mathsf{\alpha }}_0$: Initial activity of parent radionuclide
• ${\mathsf{\alpha }}_{\mathrm{t}}$: Activity of parent radionuclide at time t (measured value)
• λ: Decay constant (for ¹⁴C:5730 years).

If it is solved from the following equation:

$$\mathrm{t}=-8267\times \ln \left({\displaystyle \frac{{\mathsf{\alpha }}_{\mathrm{t}}·{}^{14}\mathrm{C}}{{\mathsf{\alpha }}_0·{}^{14}\mathrm{C}}}\right)$$

(9)

In order to compute ¹⁴C ages from the above equation, initial ¹⁴C activity (${\mathsf{\alpha }}_0$) is required to be known. Assuming that initial ¹⁴C activity is 100 pmC [100], uncorrected ¹⁴C ages of samples are found 35,954 for KH-3; 40,613 for Göl-1; 39,909 for Göl-2; 37,408 for YK-5; 22,538 and for T-1 and 33,107 YBP for T-3 (Table 3).

Because ¹⁴C fractionates during organic and inorganic phase transformations, ¹⁴C activities measured should be normalized to a common δ¹³C value of −25‰ [100].

Correction factor = (δ¹³C_Sample − (−25))‰

(10)

Using the correction factor estimated for each sample, the corrected ¹⁴C activities for samples KH-3, Göl-1, Göl-2, YK-5, T-1 and T-3 are found 1.104, 0.617, 0.676, 0.93, 5.879 and 1.578 pmC, which compute the corrected radiocarbon ages of 36,202; 40,878; 40,143; 37,662; 22,764 and 33,328 YBP, respectively. In previous studies, the age of waters in the Kızıldere field is reported 31,034 for well R-1 and 22,500–27,500 YBP for the waters of other wells [99]. Waters in the Tekkehamam field are found in the range of 34,898 [99] to >43,500 YBP [24] which agrees with our radiocarbon ages measured in Kızıldere and Tekkehamam geothermal fields.

Güner and Yıldırım [99] measured tritium values up to 1.20 TU in the Karahayıt field (PAM Hotel well, 56 °C). In spite of the absence of systematic data, there is no doubt that KGF thermal waters contain tritium. It is interesting that KH-3 thermal water in this field with pmC of 1.104 (radiocarbon age of 36,202 YBP) contains tritium (2.13 TU) (Table 3). A similar case is also reported for CO₂-rich Melgaço mineral water in the north of Portugal (³H: 2.2 TU, pmC: 2.33 ± 0.07; [103]). Although the tritium content of the PAM thermal spring (2.16 TU) is consistent with modern waters, using the NETPATH XL 1.5 software Güner and Yıldırım [104] estimated the ¹⁴C age of Pamukkale springs to be 12,388 years. ¹⁴C dating study conducted on cold groundwaters in the Denizli basin yielded 5945, 4091 and 9631 YBP for three samples with zero tritium content [101]. As a conclusion, the ¹⁴C ages of PGF and KGF thermal waters are not correlated with their tritium contents. As mentioned in Section 5.2.3, CO₂ (free of ¹⁴C) originating from deep sources (upper mantle) or expelling from the metamorphism of carbonate rocks might dilute the studied waters masking the original pmC values. For this reason, ¹⁴C ages of waters should be evaluated with caution.

Except for two samples (T-1 and T-3) from the Tosunlar field, HCO₃⁻ and ¹⁴C contents of the remaining Denizli waters are negatively correlated (Figure 11). The diagram indicates that waters with relatively older ages have experienced prolonged water–rock interaction (e.g., Yıldırım et al. [105], Wei et al. [106]). Indeed, samples Göl-1 and Göl-2 with the lowest ¹⁴C values (0.617 and 0.676 pmC, respectively) have the highest EC HCO₃ values among the studied waters (Table 2 and Table 3). Not only HCO₃ but also sulfate concentrations of these samples (particularly for Göl-2) are markedly high, indicating that these waters have interacted with carbonate rocks and evaporite units as well.

5.2.3. δ¹³C

δ¹³C values measured on dissolved inorganic carbon (DIC) for thermal waters yielded 2.55 (T-3) to 7.65‰ VPDB (Göl-1) (Table 3). Carbon isotope compositions of thermal waters in the Denizli basin were reported in several studies. For example, according to Alçiçek et al. [20,21,22,23], δ¹³C values of PGF, KGF, GGF and YGF thermal waters are 7 to 10‰ VPDB (Figure 12a). Carbon isotope compositions in Tekkehamam and Kızıldere waters are much lower, ranging from −1.5 to 6.5‰ [24,99] and 0.32 to 1.26‰ VPDB [99] for the respective fields.

Carbon isotope composition of CO₂ (δ¹³C_CO2) is reported in the range of −2.5 to −4‰ in Pamukkale, Karahayıt and Yenice thermal fluids [91,107], −1.49 to 0.17‰ in Kızıldere [26,91,108], and −0.95 to 1.30‰ in Tekkehamam [107,109], which are about 4‰ lower (¹²C-rich) than DIC in studied thermal waters.

In previous studies, δ¹³C values of Pamukkale, Çukurbağ, Karahayıt, Gölemezli, Akköy and Yenicekent travertines are reported between +2.9 and +11.7‰ [38,39,40,110]. According to Kele et al. [25], such positive values are due to CO₂ released by thermometamorphic processes related to magmatic activity. δ¹³C_CO2 can be estimated from δ¹³C of studied travertines using the experimental calibration by Panichi and Tongiorgi [111] [δ¹³C_CO2 = 1.2 × δ¹³C_Trav. − 10.5]; results yielded δ¹³C_CO2 values varying from −7 to +3.54%. In a recent study by Rizzo et al. [112], the carbon isotope composition of Çukurbağ fissure ridge travertines in PGF was reported as 5.02‰ < δ¹³C < 7.22‰, which corresponds to δ¹³C_CO2 values of −4.5 to −1.8‰. It is concluded that δ¹³C values of studied thermal waters and travertines and δ¹³C_CO2 values estimated by the equation of Panichi and Tongiorgi [111] are consistent with δ¹³C_CO2 of Pamukkale, Karahayıt and Yenice thermal waters.

δ¹³C values of marbles in the Denizli region and marbles of the Menderes massif are −1.44 to 3.52‰ (average 2.05‰; [113]) and 0.21 to 3.94‰ (average 2.92‰; [114]), respectively. Metamorphic rocks of the basement (+0.7 to +2.9‰, [25]; +0.20 to +2.98‰, [115]) and carbonate rocks of the Neogene sequence (−2.80 to 5.04‰; [36]) yielded similar values for δ¹³C; δ¹³C is −3 to +3‰ for marine limestones, −40 to −20‰ for organic sediments and −9 to −4‰ (mean −6.5‰) for the mantle [100,116,117]. Our assessment implies that the CO₂ composition of thermal waters might be explained by several mechanisms which include the decomposition of carbonates within the metamorphic reservoir at a temperature of at least 350 °C [91], dissolution of Neogene carbonate rocks, CO₂ escape during ascend of thermal waters and release of mantle CO₂ (Figure 12a). The fact that δ¹³C values of travertines are higher than δ¹³C_DIC of thermal waters is attributed to the enrichment of heavy carbon isotope (¹³C) in the residual water by surface evaporation of thermal fluid and the escape of CO₂ gas to the atmosphere. The distribution of δ¹³C of DIC and CO₂ in studied thermal and cold waters is shown in Figure 12b. The diagram indicates that δ¹³C (DIC) of PGF, KGF, GGF, YGF and TGF thermal waters are consistent with marine limestone and metamorphic CO₂. δ¹³C_CO2 values, however, show mantle origin for PGF, KGF and YGF and marine limestone for the samples TeGF and KıGF. δ¹³C of cold waters in the basin is much lower, ranging from −10.5 to −4‰, probably contributed by an organic source [23].

5.2.4. Sulfur and Oxygen Isotopes on Sulfate

δ³⁴S values of sulfate in the studied thermal waters are in the range of 14.1 to 20.04‰ VCDT and δ¹⁸O‰ are from 10.65 to 15.71‰ VSMOW (Table 3). Yıldırım and Güner [71] reported δ³⁴S and δ¹⁸O in TeGF thermal waters as 16.72 to 18.69‰ VCDT and 13.5 to 11.1‰ VSMOW, respectively. Sulfur and oxygen isotope systematics of waters of the KıGF field are 18.96 to 20.00‰ VCDT and 0.90 to 6.12‰ VSMOW [26,71,94].

δ³⁴S and δ¹⁸O of gypsum deposits from the Sazak formation exposing 4 km SW of Tosunlar field are reported 10.65–10.92‰ (average 10.82‰) and 10.9–15.5‰ (average 13.12‰) for lenticular gypsum and 3.09–14.3‰ (average 8.69‰) and 7.6–17.6‰ (average 13.53‰) for the vein type gypsums [36], respectively. Yeşilyurt evaporites 10 km SW of Sarayköy yielded the following δ³⁴S–δ¹⁸O values: 21.84–26.45‰ (average 25.03‰) and 12.4–16.2‰ (average 14.75‰) for selenites, 9.93–13.73‰ (average 12.24‰) and 10.9–11.1‰ (average 10.96‰) for gypsumarenites and 24.7–26.28‰ (average 25.77‰) and 13.5–18.2‰ (average 15.7‰) for the vein type gypsums [36].

In the δ³⁴S(SO₄) versus δ¹⁸O(SO₄) diagram, studied waters and data from previous works were plotted. In this graphic, δ³⁴S (SO₄) values vary in a narrow range from 14 to 20‰ and δ¹⁸O (SO₄) values fall in a wide range from 1 to 16‰. Kızıldere samples with δ¹⁸O about 10‰ lower than other waters plot between the marine and terrestrial evaporite fields. During the dissolution of gypsum by meteoric waters, sulfate is likely to equilibrate with H₂O, which can explain the relatively low δ¹⁸O values of the Kızıldere thermal waters. On the contrary, waters sampled from the Pamukkale, Karahayıt, Gölemezli, Yenicekent and Tosunlar fields fall in the area of Cenozoic evaporites (Figure 13).

5.3. Rare Earth Element Fractionation

In this section, we compared REE + Y compositions of scale samples collected from the study areas with REY contents of thermal waters and various rocks exposed in the Denizli basin [118] (

Table 4. REY contents of scale samples from some thermal wells in the study area (ppm).

Field	Well	La	Ce	Pr	Nd	Sm	Eu	Gd	Tb	Dy	Ho	Er	Tm	Yb	Lu	Y
KGF	KH-2	2.0	2.8	0.42	1.2	0.17	0.07	0.3	0.05	0.19	0.05	0.16	0.03	0.13	0.02	1.8
GGF	Göl-1	2.4	2.7	0.45	2.1	0.49	0.18	0.96	0.17	1.27	0.33	1.00	0.14	0.90	0.14	12.4
	Göl-2	1.9	2.6	0.35	1.3	0.29	0.10	0.48	0.09	0.44	0.10	0.31	0.05	0.41	0.06	4.7
	DG-3	1.5	2.0	0.29	1.3	0.36	0.11	0.67	0.13	0.74	0.18	0.54	0.09	0.46	0.08	7.5
YGF	YK-5	1.8	3.7	0.60	3.2	1.05	0.32	1.82	0.31	2.13	0.45	1.36	0.18	1.08	0.14	17.3
TGF	T-1	3.0	5.6	0.67	3.0	0.68	0.18	0.87	0.15	1.04	0.16	0.49	0.06	0.43	0.06	6.5

, Figure 14). In the comparison, REY concentrations of samples are normalized with chondrite values of McDonough and Sun [119].

The rare earth elements (REEs), which represent the so-called lanthanides, show similar geochemical behavior through the geological processes and, with the exception of cerium (Ce) and europium (Eu), usually occur in the trivalent form [120]. Cerium can exist as Ce³⁺ like the majority of lanthanides, or as Ce⁴⁺ in oxidizing conditions. Eu³⁺ (with an ionic radius of 1.07 Å) occurs together with other REE³⁺, either on crystallographic sites dominated by REE or replaces for cations such as Ca²⁺ (with an ionic radius of 1.12 Å). On the other hand, Eu²⁺ has a larger ionic radius (1.25 Å, similar to Sr²⁺, 1.26 Å) and, therefore, Eu²⁺ and Eu³⁺ have different affinities for a specific crystallographic site [120]. Considering the order of atomic number, REEs are divided into three subgroups: LREE (light REE), HREE (heavy REE) and MREE (middle REE).

REY compositions of rock samples are nearly 1 to 2 orders of magnitude higher than the chondrite values (Figure 14). Neogene marl and Neogene limestone trends with a distinct positive thulium anomaly are lower than other lithologies. REY concentrations of quartzite, Mesozoic limestone, shale and schist-marble steeply descend from La to Sm and then continue with a sharp negative Eu anomaly and show a flat-like trend across the HREEs. REE + Y contents of travertine samples from the Pamukkale and Kızıldere fields are 1-fold lower than rock samples although their REY patterns are almost similar (Figure 14).

Thermal waters with REY contents 5 to 9 orders of magnitude lower than the rock samples show varying source rock lithologies. Positive Y anomaly (not pronounced for the Kızıldere samples) and negative Eu and Ce (only for Pamukkale samples) anomalies are the common features of the Denizli thermal waters (Figure 14). For example, Kızıldere and Gölemezli waters with negative Eu pattern are in agreement with shale and limestone. Such REY patterns depend on local infiltration of meteoric water and are greatly controlled by the degree of water–rock interaction [118].

Scales sampled in this study from the Kızıldere, Gölemezli, Yenice and Tosunlar fields are represented by negative Eu and positive Y anomalies and their REY compositions are quite consistent with the Neogene limestone. Scale YK-5 (ΣREY: 18.1 ppm) from the Yenicekent field and scale KH-2 (ΣREY: 7.6 ppm) from the Karahayıt field have the highest and lowest REE contents, respectively. REY concentrations of scales sharply descend from La to Sm and then flow a horizontal pattern with a saw-like shape including a negative Eu anomaly and positive Y anomaly. Eu anomaly is probably inherited from plagioclase removal from the source rock whilst yttrium anomaly possibly implies the replacement of calcium by this element (e.g., Rankama and Sahama [121]).

Although REEs are more compatible in calcite (partition coefficient of log DREE = 4.4 [122]) than in silicate minerals (log DREE = 0.9–1.6 [123]), REE concentrations of scale samples are noticeably low (Figure 14). Available REE data may imply a rapid rise of CO₂-rich waters without sufficient interaction with the host rocks [105]. The retardation of water–rock interaction by CO₂-dominated fluids or retention of REE by the clays [124,125] may well explain the low rare earth element concentrations of thermal waters.

5.4. Geothermometry

Silica, cation and isotope geothermometers applied to the studied waters and estimated reservoir temperatures are shown in

Table 5. Chemical and isotope geothermometers applied to the sampled thermal waters (°C).

Geothermometer	Calibration	PAM	KH-2	KH-3	KH-4	Göl-1	Göl-2	YK-5	T-1	T-2	T-3
Tmeasured		33.9	56.0	60.7	53.4	63.3	69.0	61.5	97.0	98.0	84.1
SiO2 1–Chalcedony	$\mathrm{t}°\mathrm{C}={\displaystyle \frac{1032}{4.69-{\mathrm{logSiO}}_2}}-273.15$	68	62	75	48	91	130	79	131	143	118
SiO2 1—Quartz No steam loss	$\mathrm{t}°\mathrm{C}={\displaystyle \frac{1309}{5.19-{\mathrm{logSiO}}_2}}-273.15$	99	92	104	79	119	155	109	156	166	144
SiO2 1—Quartz Max. steam loss at 100 °C	$\mathrm{t}°\mathrm{C}={\displaystyle \frac{1522}{5.75-{\mathrm{logSiO}}_2}}-273.15$	99	94	105	83	118	148	109	148	157	139
Na-K 2	$\mathrm{t}°\mathrm{C}={\displaystyle \frac{855.6}{0.8573+\log \left(\mathrm{Na}/\mathrm{K}\right)}}-273.15$	215	254	246	287	251	201	244	247	270	233
Na-K 3	$\mathrm{t}°\mathrm{C}={\displaystyle \frac{1217}{1.483+\log \left(\mathrm{Na}/\mathrm{K}\right)}}-273.15$	239	268	262	292	266	228	260	263	280	253
Mg-Li 4	$\mathrm{t}°\mathrm{C}={\displaystyle \frac{2200}{5.47+\log \left(\sqrt{\mathrm{Mg}}/\mathrm{Li}\right)}}-273.15$	27	38	41	43	44	74	69	68	66	95
Na-Li 4	$\mathrm{t}°\mathrm{C}={\displaystyle \frac{1590}{0.779+\log \left(\mathrm{Na}/\mathrm{Li}\right)}}-273.15$	213	183	197	185	145	210	168	136	142	173
K-Mg 5	$\mathrm{t}°\mathrm{C}={\displaystyle \frac{4410}{14.00-\log \left(\mathrm{K}/\sqrt{\mathrm{Mg}}\right)}}-273.15$	30	58	55	68	81	77	99	116	115	124
Na-K-Ca 6,7	$\mathrm{t}°\mathrm{C}={\displaystyle \frac{1647}{\log \left(\mathrm{Na}/\mathrm{K}\right)+\mathsf{\beta }\left[\log \left(\sqrt{\mathrm{Ca}}/\mathrm{Na}\right)+2.06\right]+2.47}}-273.15$	16	61	64	73	55 *	52 *	107 *	121 *	104 *	141 *
18O(SO4-H2O) 8	$1000\ln \mathsf{\alpha }{\mathrm{SO}}_4-{\mathrm{H}}_2\mathrm{O}=3.25\left({10}^6/{\mathrm{T}}^2\right)-5.6$	63	-	67	-	75	57	93	85	-	-

Note: ¹ Fournier [126], ² Truesdell [127], ³ Fournier [128], ⁴ Kharaka and Mariner [129], ⁵ Giggenbach et al. [130],⁶ Fournier and Truesdell [131], ⁷ Fournier and Potter [132], ⁸ Lloyd [133]; * Mg correction applied to Na-K-Ca temperatures.

. The lowest reservoir temperatures are calculated by the chalcedony geothermometer. For samples PGF, KGF, GGF and YGF, which have a maximum discharge temperature of 69 °C, the chalcedony geothermometer yielded reservoir temperatures varying from 48 to 130 °C. TGF waters with relatively higher discharge temperatures have chalcedony temperatures falling in the range between 118 and 143 °C. The quartz geothermometers with temperatures about 30 °C higher than the chalcedony geothermometer gave reservoir temperatures of 139 to 166 °C for the Tosunlar field and 79 to 155 °C for the other fields.

The use of ratios, rather than absolute abundance of cations, makes the Na-K geothermometer less sensitive to the secondary processes (e.g., mixing and boiling) in the reservoir. Therefore, Na-K geothermometer might produce inaccurate results when applied to low-temperature waters because at low temperatures Na/K ratio of waters is controlled by leaching rather than chemical equilibrium [10]. In fact, most of the anomalously high reservoir temperatures in Table 5 (201–292 °C) were obtained by the Na-K geothermometers for the waters with low discharge temperatures. Temperatures estimated by the Mg-Li geothermometer are generally lower than the measured values. For the Tosunlar thermal well waters, Na–Li geothermometer yielded temperatures of 136–173 °C that are close to bottom-hole temperatures (107–149 °C) and the same geothermometer gave higher reservoir temperatures (145–213 °C) for the other fields. Temperatures estimated by the K-Mg geothermometer are lower than the measured in PAM and KH-3 thermal waters but slightly higher for other water samples. Na-K-Ca temperatures are lower than the measured in PGF and GGF waters and are close to bottom-hole temperatures of samples from the Tosunlar field.

The Na-K-Mg diagram proposed by Giggenbach [134] is very useful for estimating the reservoir temperatures of thermal waters and evaluating the equilibrium state of waters with respect to certain silicate minerals they interact. This diagram also allows a quick assessment of reservoir temperatures and tests the applicability of cation geothermometers (Figure 15). In this diagram, R-1 well water from the Kızıldere field plots very close to the full equilibrium line and the bottom-hole temperature (242 °C) is almost the same as Na-K temperature. The remaining waters from other fields fall in the immature waters field extending along a line from very immature PGF waters to the partially equilibrated Tekkehamam waters. According to Giggenbach [134], for waters plotting in the immature waters field cation geothermometers should be assessed cautiously. With the exception of Kızıldere R-1 well water, studied waters do not attain equilibrium and reflect mixing with cold groundwater, therefore results of cation geothermometers are regarded as uncertain. For such (immature) waters, temperatures calculated by the quartz geothermometers are much more realistic. The average quartz temperature estimated with SiO₂ contents of Tekkehamam thermal waters (İnaltı and W-3, Table 2) is 170 °C, which is consistent with the bottom-hole temperature of Tekkehamam TH-2 well (168 °C; Table 1). The average of quartz and Na-K temperatures estimated for Kızıldere R-1 well water (243 °C) is consistent with temperatures of geothermal fluids in this field. It is concluded that temperatures estimated for PGF, KGF and YGF (79–109 °C), GGF (118–155 °C) and Tosunlar field (144–166 °C) may be accepted as reservoir temperatures.

Reservoir temperatures of studied geothermal fields were evaluated with saturation states of waters with respect to several minerals that were estimated by the PhreeqC program at temperatures of 25 to 200 °C [87]. When a mineral group is close to the equilibrium line (SI = 0) at a certain temperature, it is assumed that the water is in equilibrium with these minerals and the temperature reflects the reservoir temperature [135]. From the mineral equilibrium diagrams given in Figure 16, the reservoir temperatures determined for PAM, KH-4 and Göl-1 are 96, 97 and 119 °C, respectively, from 99 to 108 °C for sample YK-5 and from 121 to 140 °C for sample T-1.

We used a ¹⁸O(SO₄-H₂O) isotope geothermometer as another alternative to estimate the reservoir temperatures of thermal waters. Calculations were first made for KıGF, which is home to deep geothermal wells. ¹⁸O(SO₄) and ¹⁸O(H₂O) values of thermal waters in KıGF are 2.7, 1.3, 0.9‰ and −5.52, −5.12, −5.91‰, respectively [26]. For the waters of the same field, Yıldırım and Güner [71] reported ¹⁸O(SO₄) values between 2.83 and 5.29‰ and ¹⁸O(H₂O) values from −6.6 to −5.1‰. Using these values and the equation proposed by Lloyd [133] reservoir temperatures for KıGF were estimated in the range of 175 to 244 °C which are quite consistent with both quartz and Na-K temperatures and discharge temperatures of thermal waters. However, for the low-temperature waters, ¹⁸O(SO₄-H₂O) temperatures computed from the available data (this study and Yıldırım and Güner [71]) are not realistic (Table 5, Figure 17) which might be attributed to the fact that equilibrium is not established between sulfate and water because of bacterial sulfate reduction and/or slow exchange of ¹⁸O between sulfate and water [136,137,138].

Chemical and ¹⁸O isotope geothermometers applied to the studied waters yielded a wide range of reservoir temperatures. Discharge and borehole temperatures of waters from deep wells are generally consistent with Na-K and SiO₂ reservoir temperatures for the Kızıldere field and SiO₂ reservoir temperature for the Tekkehamam and Tosunlar fields. The depth of wells in Karahayıt (452 to 650 m) and Gölemezli (except for DG-1 137 to 750 m, DG-1 well was shut down because of inefficiency) is relatively low. The shallowest wells with depth of 54–300 m are in the Yenicekent field where the reservoir is overlain by a thin Neogene cover. Deep wells were not drilled since the waters in the Karahayıt and Yenicekent fields were intended to be used for thermal spas and greenhouse heating, respectively. Because the Pamukkale thermal spring site is a conservation area, geothermal drilling is not allowed. It is highly probable that deep wells opened in the PGF, KGF, YGF and GGF fields might yield water temperatures much higher than the current discharge temperatures.

5.5. Fluid Characterization

In previous studies, fossil fluid temperature in the Pamukkale geothermal field was studied in detail using different techniques [112]. Homogenization temperatures of the inclusions trapped in the Pamukkale travertines were measured in the range of 132 to 145 °C (with an average of 140 °C). Salinities of inclusions are found between 0.4 and 7.0 wt% NaCl equivalent indicating the existence of moderately saline solutions. The oxygen isotope composition of thermal waters (δ¹⁸O) that precipitated the Pamukkale travertines was estimated using the oxygen isotope fractionation between water and calcite (Δ¹⁸O_{calcite-water}). Results of calculations that were carried out at a discharge temperature of 56 °C yielded −8.3 to −3.2‰ (VSMOW) indicating a meteoric origin [112]. δ¹⁸O values of paleofluids in the Pamukkale geothermal field are consistent with modern waters that fall in the range of −9.28 to −5.59‰ VSMOW.

Helium isotope variation (³He/⁴He) can be used as a tracer to investigate the origin of crustal or mantle-derived volatiles in different tectonic and volcanic environments. Helium has two isotopes; ⁴He is produced by the decay of radioactive elements (e.g., U and Th), while ³He is mostly of primordial origin and therefore was trapped within the earth during the accretion [139]. Due to the significant difference in isotopic composition between the mantle and crustal helium, helium isotope studies in natural waters and minerals provide valuable information on the extent of mantle-derived contributions to crustal fluids and the rate of mantle degassing through the crust.

The ³He/⁴He ratios measured in fluid inclusions of Pamukkale travertines yielded 0.5–1.3 Ra [112] which are notably higher than values characteristic of crustal lithologies (average Ra = 0.05 [139]. Helium isotope composition is much higher up to 3.68 Ra in Pamukkale thermal fluids [140], 0.95 to 1.67 Ra in Kızıldere thermal water [108,141] and 2.36 to 2.86 Ra in Tekkehamam waters [141]. These findings indicate a significant He contribution from the mantle (8 Ra). Assuming a simple binary mixing between the mantle and crustal helium components, the proportion of mantle-derived helium in Pamukkale, Kızıldere and Tekkehamam thermal waters is found to range from 12 to 46%.

In a recent study by Süer et al. [109], isotopic compositions of various gases (CO₂, CH₄, and N₂) from the Kızıldere and Tekkehamam geothermal fields were investigated. The stable isotopic composition of carbon (δ¹³C)CO₂ varied in a narrow range from −0.95 to +1.3‰ indicating that most part of the carbon (CO₂) derives from limestone sources. δ¹³C and δ²H of methane from these fields showed distinct values ranging from −34.4 to −20.8‰ and −143.3 to −126.7‰, respectively. CO₂-CH₄ isotope geothermometry estimations yielded temperatures significantly higher than the bottom–hole temperatures recorded in Kızıldere and Tekkehamam. Finally, the N₂/³⁶Ar ratios of the majority of samples indicate the presence of a non-atmospheric nitrogen component, which is likely to be a mixture of sedimentary and mantle nitrogen [109].

Chromatographic determination of gas compositions of samples collected from the Denizli region yielded that CO₂ is the primary gas phase followed by N₂ and O₂. The volume of CO₂ is 69–86% in Pamukkale, 66–90% in Karahayıt and 98% in Gölemezli fields [142]. High nitrogen abundance (up to 24%) in these fluids accompanied by O₂ with a content of 3 to 8% is attributed to mixing with shallow waters.

Synthesis of our findings with the published water chemistry and gas isotope data reveals that Pamukkale, Tekkehamam and Kızıldere geothermal fluids interact with carbonate-type lithologies at a wide range of temperatures from about 35 to 242 °C. Regarding the volatiles in these fluids, some portion of helium and nitrogen is derived from the mantle whilst methane and carbon dioxide originate entirely from crustal sources.

5.6. Limitations of the Study

This study chiefly depends on data collected at a particular point in time potentially excluding seasonal or long-term variations in thermal water chemistry. For this reason, future studies should incorporate temporal data on the chemical and isotopic compositions of the examined water sources across different seasons to assess the temporal variability. The main limitation of the study is the small number of water samples (n = 10), which limits the use of further methodological approaches. The disadvantage of our limited data was partially overcome by the use of water and scale chemistry data available in the literature.

It is important to note that here anthropogenic factors could be of great importance to the sustainability and environmental impact of geothermal sources. For this, physicochemical parameters and ionic concentrations including trace elements could be monthly monitored to ensure observation of natural or anthropogenic variations in these parameters. Special attention should be given to practicing proper management to discharge waters from spas or greenhouses, as they are not considered trade waste.

Pamukkale thermal waters are used in agricultural irrigation after the bleaching process, and Karahayıt thermal waters are purified after use in thermal baths. All the wastewater in other fields and the power plant waters that cannot be reinjected are discharged into the receiving environment (usually the Büyük Menderes River). This might negatively affect the ecological balance in the receiving environment due to temperature differences and boron, arsenic and heavy metal input. On the other hand, during the reinjection of thermal waters in TeGF, in some locations water at boiling temperature is temporarily manifested at the surface, probably due to the use of unsuitable well equipment or the low depth of reinjection wells that only reach the Neogene units. This may affect the physicochemical properties of existing thermal springs. However, such an impact is not in question since the İnaltı thermal spring was sampled (analysis is given in Table 2) at the time there were no geothermal power plants in this area [71]. Reinjection wells that could contaminate the existing freshwater aquifers might be considered another problem. Due to financial reasons, the lack of periodic or seasonal monitoring of the chemical and isotopic composition of wastewater and receiving water bodies prevented a more comprehensive assessment.

6. Conclusions

Studied geothermal fields have more than one reservoir. For all the fields, limestones of the Neogene Sazak formation comprise the shallow (first) reservoir whilst schist, marble and quartzite of the Paleozoic Menderes massif represent the deep (second) reservoir. Apart from these, limestones and dolomitic limestones of the Mesozoic Çökelez formation in PGF form the medium reservoir between the shallow and deep reservoirs, while gneiss and quartzites comprise the deepest reservoir in the Kızıldere field. Cover rocks are represented by the impermeable units of the Neogene formations. The high geothermal gradient resulting from neotectonic activity is the heat source of geothermal systems. All thermal waters are of meteoric origin and tritium values of PGF and KGF waters reflect the mixture of sub-modern and modern precipitation, while the tritium contents of thermal waters in other fields reveal the pre−1952 recharge. Using the altitude versus δD relationship, recharge elevations for PGF, KGF, GGF, YGF and TGF thermal waters are estimated in the range of 1030 to 1175 m.

Different temperatures, the extent of the water–rock interaction process, dissolution-precipitation, mixing, cation exchange and microbial processes in geothermal fields have resulted in different hydrochemical types of thermal waters. Dissolution of carbonate rocks (e.g., limestones and marble) exerts a great control on the formation of Pamukkale and Karahayıt thermal waters with the composition of Ca-Mg-HCO₃. The waters of Gölemezli, Yenicekent and Tosunlar geothermal fields are of Ca-Na-HCO₃ type and probably interacted with claystone and marl units. On the other hand, high sulfate content in Tekkehamam geothermal fluids is attributed to derivation from organic matter-rich sediments or bacterial reduction of sulfate. Finally, the Na-HCO₃ composition of the Kızıldere field implies interaction with silicate-type reservoir rocks. Temperature, circulation depth, residence time of water and the degree of silicate weathering increase from PGF to KıGF but the mixing ratio of cold groundwater decreases. The noticeable shift in the δ¹⁸O values of the thermal waters from TeGF and KıGF is explained by water–rock interaction at high temperatures.

δ¹³C data indicate that Mesozoic and Neogene limestones, high-temperature decomposition of carbonates within the metamorphic reservoir and mantle-derived CO₂ are the major sources of carbon in the studied thermal waters and δ³⁴S values showed that the sulfate is originated from the Neogene gypsum.

Scale samples collected from the Kızıldere, Gölemezli, Yenice and Tosunlar fields are represented by negative Eu and positive Y anomalies and their REY trends show that Neogene limestone is the major source rock.

Silica and mineral equilibrium geothermometers and Na-K-Mg diagram applied to geothermal fields other than TeGF and KıGF gave realistic reservoir temperatures in the range of 79 to 166 °C, with the highest belonging to the TGF field.

In order to determine the natural or anthropogenic effects on the physical and chemical composition of thermal waters in the study areas and to make an assessment of the protection of the receiving environment and fresh groundwater aquifers, temperature, electrical conductivity, pH, ion, element and isotope analyses should be carried out at selected sampling points at monthly intervals (at least 12 months). For Pamukkale, which is on the World Heritage List, this monitoring work should be indefinite.