Hydrogeochemical and isotopic investigations on the origins of groundwater salinization in Çarşamba coastal aquifer (North Turkey)

The aim of this study is to determine the origins of salinization and the main hydrogeochemical process that controls the chemistry of Çarşamba coastal aquifer in Turkey. Therefore, a total of 33 groundwater samples and three seawater samples were analyzed in the coastal region of Çarşamba Plain in July 2019, and for these samples’ physical parameters, major ions and environmental isotopes (δ18O, δ2H, and 3H) values were determined. Piper, Chadha, Gibbs diagrams and Stuyfzand Classification Systems were used to determine the origins of salinization and the key hydrogeochemical process controlling the groundwater chemistry. According to Stuyfzand classification system, the study showed that the freshwater and fresh-brackish water main types are the most widespread in the study area. Six water subtypes were observed in the study area that include CaHCO3, CaMix, NaMix, NaCl, NaHCO3, and MgHCO3. In addition, the subtypes CaMix indicated the locations of the transition zone, where the groundwater rich in Ca and HCO3 and gradually enriched in Na changes from CaMix with HCO3 as dominant anion to eventually CaCl and NaCl subtypes. Furthermore, the subtypes NaMix, NaHCO3, and MgHCO3 showed the locations of the transition zone where the flushing of the saline aquifer by freshwater takes place. All groundwater samples from study area had a positive cation exchange code, and show that four hydrogeochemical facies composed of CaHCO3, Ca–Mg–Cl, NaCl, and NaHCO3. Besides, groundwater samples have been influenced by two main mechanisms: the water–rock interaction and evaporation–crystallization. According to δ18O, δ2H, and 3H analysis, the water samples have meteoric origin, shallow circulation, and a short residence time.


Introduction
Although many factors, such as excessive pumping of groundwater, pollution, uncontrolled agricultural practices, and industrialization, lead to the depletion of the world water resources, salinization, especially in coastal aquifers, is the most important factors which pollutes groundwater resources. There are various potential sources of salinization of groundwater, such as natural saline groundwater, seawater intrusion, dissolution of halite, domestic, industrial, and agricultural wastes. In the coastal area, seawater intrusion is the principal source of salinization of groundwater, and very often requires the abandonment of wells in these zones (Al Farrah et al. 2011). Many research has been realized through the world to determine the sources of salinization of groundwater. The work carried out in Algeria by Djabri et al. (2008) and in Ghana by Kortatsi (2006) based on the physico-chemical analysis results of the water samples highlighted the influence of seawater intrusion and the dissolution of evaporate formations on the salinity of the coastal aquifers in their study area. In Syria, Abou Zakhem and Hafez (2007), using the results of physico-chemical and isotopic analysis of the water samples taken from the north of Latakia and south of Tartous, demonstrated that salinization of groundwater in these coastal areas is mainly due to the mixing between freshwater and seawater; and that evaporation process has very limited effects. In France, Veronique de Montety et al. (2008) carried out research to evaluate different salinity sources of the confined aquifer of the Rhône delta using chemical and isotopic groundwater compositions. According to the results of their study, seawater intrusion is the origin of increasing of Na, thus promoting the process of exchange between Ca 2+ and Na + . In addition, this seawater intrusion also led to the reduction of SO 4 2− which is confirmed by the strongly depleted of 13 C TDİC values.
In Turkey, while the average annual rain precipitation in the world is around 1000 mm, the average precipitation is around 646 mm. This situation shows that Turkey in terms of water resources is not among the rich countries. In addition to the scarcity and uneven distribution of water resources, rapid population growth and pollution of water resources conduct to a further worsening of the water deficit (Toklu 1999). On the other hand, unfavorable factors such as poor land use, urbanization and industrial organizations, wild storage areas, and the discharge of untreated liquid waste into basins cause rapid pollution of surface and groundwater. Among the sources of water pollution in most coastal areas of Turkey, the salinization process is one of the most worrying factors, which mainly threatens groundwater resources. In general, in coastal areas, the groundwater level decreases due to excessive pumping of groundwater. This excessive groundwater pumping lowers the level of the freshwater table and consequently reduces the pressure exerted by the freshwater column, which then allows the denser saltwater to move inland lateral (Barlow 2003). Arslan et al (2012), Fırat Ersoy et al. (2020), and Fırat Ersoy et al. (2021) carried out research to determine the influence of seawater on groundwater in the right bank of the Bafra Plain, using the results of hydrochemical and isotopic analysis. According to the results of their study, some wells have very high Electrical Conductivity (EC), Total Dissolved Solids (TDS), Cl − and Na + values, and the isotope distributions of δ 18 O and δ 2 H showed that all water samples were grouped around the freshwater-seawater mixing line. The work carried out in Van by Özler (2001) and in Mersin by Demirel (2004) based on the results of physicochemical analysis of groundwater samples taken in coastal area highlighted the influence of seawater intrusion due to excessive pumping of groundwater. According to the previous research that has been done in Çarşamba coastal aquifer, excessive increases of chloride, sodium, EC, and TDS values were observed. Between 1990 and 2012, water quality in this zone was severely deteriorated due to the excessive pumping of groundwater and the resulting seawater intrusion (Arslan 2017).
This study seeks to determine the salinization origins and the main hydrogeochemical process controlling groundwater chemistry of Çarşamba coastal aquifer based on the results of physico-chemical and isotopic analyses (δ 2 H, δ 18 O and 3 H) of groundwater samples collected in this region. Due to its position on the Black Sea coast, Çarşamba plain is a region where intense agricultural activities take place. It is assumed that the excessive pumping of groundwater in this region for agricultural activities leads to contamination of the aquifer by the seawater intrusion. In addition, the excessive use of fertilizers, pesticides in agricultural areas as well as industrial wastes from the industrial areas leads to pollution of groundwater in this region (Maman Hassan 2021). This study therefore made it possible to highlight the influence of seawater intrusion caused by the agricultural and industrial activities taking place in the study area on the groundwater chemistry in the Çarşamba aquifer.

General properties and geology
The study area is located on the left bank of the Yeşilırmak river and the western part of the Çarşamba plain, Turkey, lying between the latitudes 41° 11′ 30'' and 41° 16′ 30'' and the longitudes 36° 22′ 00'' and 36° 36′ 00''; the plain of Çarşamba is situated in the Middle of Black Sea Region (Fig. 1). The climate of this region is semihumid and the average annual precipitation is about 700 mm and the average annual temperature is about 17 °C (Arslan 2014). The wet period is from October to the end of December. There are three formation types in the study area, which are the Tertiary age Tekkeköy formation, Eocene age Sarıyurt formation, and Quaternary age Alluvium formation (Fig. 2). The Sarıyurt formation consists of sandstones, siltstones, marls, and conglomerates. Its upper part is composed of interstratified siltstones, sandstones, and marls, while an alternation with sandstones is observed in the middle part. In the lower part of this formation, conglomerates and sandstones are observed. The Tekkeköy formation consists of sandstones, mudstones, basalts, marls, and tuffites alternation and agglomerates. The Quaternary Alluvium, which covers almost the entire study area, consists of sands, gravels, clays, sandstones, muds, and silts (Yoldaş et al. 1985). The geology of the study area indicates that only the Quaternary Alluvium can be considered as important groundwater reservoirs. The volcanic rocks of the Çarşamba plain are geological formations which have the characteristics of the secondary aquifer after the alluvial deposits in terms of groundwater. In these formations, the weathered basalts and in particular the interconnected cracks and the fracture systems developed in the agglomerate levels, facilitate transportation and accumulation of groundwater (DSI 1993). The geological cross-section of study area is given in Fig. 3. In terms of hydrogeological characteristics, the Çarşamba plain can be divided into two parts: the left bank and the right bank of the Yeşilırmak River. The left bank is generally composed of unconfined and semi-confined aquifers, while on the right bank, semi-confined and confined aquifers are the most frequent (DSI 1993). On the left bank of Yeşilırmak river where the study area is located, the aquifer begins generally at a depth varying between 4 and 40 m, and its thickness values ranged from 1 and 20 m. The majority of aquifers in the left bank of Yeşilırmak River consist of various layers of gravels, clayey-sands, clayey gravels, and silty sands within the Quaternary alluvium. The aquifer of Çarşamba plain has available discharge between 2 and 32 l/s, and a variable hydraulic gradient with a maximum value of 0.0025 (DSI 1993;Arslan 2014).

Materials and methods
For this purpose, 33 groundwater samples and three seawater samples were taken in July 2019 in the study area. Groundwater samples were taken during the wet season from wells used in the study area for irrigation, industrial activities, and domestic use. Water sampling was taken according to the TS EN ISO 5667 Water Quality Sampling standard (ISO 1991) defined in Turkish Standards. According to this standard, water sampling is done after 15 min of pumping in deep wells. Water samples were taken into polyethylene bottles pre-washed with pure water. For each sampling point, 1 L bottle was used for the major ion analysis and the 100 ml bottles were used for the minor ion analysis.  ( 3 H) values were determined. The global coordinates of all the sampling points expressed in the sexagesimal system were taken and recorded with a Global Positioning System (GPS) device. The physical parameters of groundwater samples were determined in situ using HQD-Meter. The analysis and chemical parameters of groundwater samples were carried out at the Seventh Regional Directorate of Hydraulic works of the State (DSI) in accordance with TS EN ISO standards. The concentrations of HCO 3 − and CO 3 − were determined using titration method. Atomic absorption spectrometry TS EN ISO 14911 method was used to determine the concentrations of Na + , Ca 2+ , K + , and Mg 2+ , while NH 4 + . NO 2 − , NO 3 − SO 4 2− , and Cl − analyses were carried out using ion chromatography method. The environmental isotopes' (δ 18 O, δ 2 H, and 3 H) analyses were carried out at the Isotope Laboratories of the Technical Research-Quality Control Directorate of the State Hydraulic Works. The mass spectrometry method was used to determine the value of δ 2 H and δ 18 O; and the values of 3 H were determined using liquid scintillation method. For the analysis of 18 O and 2 H, the precision values are ± 0.15 ‰ and ± 2 ‰, respectively. In this study, all maps were made using ArcGIS software. The spatial distribution maps were produced using inverse distance weighted (IDW) interpolation method.

Hydrochemical interpretation methods
In this study, the determination of salinization origins and the hydrochemical process controlling the groundwater chemistry is mainly based on the classification of water types according to Stuyfzand and Chadha diagram (Chadha 1999). In addition, Spearman correlation analysis method was used to determine the relationships between the physico-chemical parameters of groundwater samples. This correlation method was used in this study, because the physico-chemical parameters of groundwater samples collected in the study area did not follow a normal distribution law. The Spearman correlation analysis method is usually used when the supposition of the bivariate normal distribution is not tenable (Artusi et al. 2002). The ratio of sodium to total sodium and chloride to total dissolved solids (TDS) was used to evaluate the dominant ion process that occurring in groundwater of study area. Hydrochemical facies of water samples was classified basis on the dominant ions using the Piper diagram (Piper 1944). The diagram of Gibbs was used for to determine the main mechanisms controlling the groundwater chemistry in the study area. Groundwater chemistry is in general controlled by various mechanisms such as weathering of rocks, precipitation, and evaporation. Thus, this diagram is constructed by plotting TDS on the Y-axis against Na/(Na + Ca) on the X-axis for cations; and TDS on the Y-axis against Cl/ (Cl + HCO 3 ) on the X-axis for anions (Gibbs 1970). Chadha diagram is used in general for the classification of natural waters as well as the identification of hydrochemical process. This diagram is constructed by plotting on the X-axis the difference between alkaline earths and alkali metals [(Ca 2++ Mg 2+ )-(Na ++ K + )] against weak acidic anions and strong acidic anions [(HCO 3 − + CO 3 2− )-(SO 4 2− + Cl − )] on the Y-axis, of water data expressed as milliequivalent percentage. The milliequivalent proportion differences from the axis X and Y are prolonged further into the principal study subfields of the proposed diagram, which defines the global character of water (Chadha 1999). The δ 18 O, δ 2 H, and 3 H analyses were carried out to determine the origin, the relative age of the groundwater, and whether the seawater intrusion into the aquifer is significant in the study area. The groundwater origin and the mixing process between freshwater and seawater were determined using the relationship between δ 2 H and δ 18 O, Sinop Meteoric Water Line (D = 8* δ18O + 16.4) (Dilaver et al. 2018) and Global Meteoric Water Line (δ D = 8* δ18O + 10) (Craig 1961). In addition, to assess the seawater influence on groundwater, the cross plots of Cl − vs δ 18 O, EC vs δ 18 O, EC vs 3 H, and Cl − vs 3 H were used. The Stuyfzand classification system was developed by Stuyfzand (1986) as a new system applied to aquifer systems by combining the characteristics of existing classification systems. This system had been improved for the first time specifically for groundwater from coastal aquifers limestone systems with cation exchange phenomena caused by the intrusion of fresh or salt water. After the system has been extended for non-calcareous aquifer systems (Stuyfzand 1989). Consequently, this classification can be applied for groundwater from all hydrogeological systems. This system, based on the determination of different types of water by subdividing the principal chemical characteristics of groundwater at 4 levels, is important in determining the principle of cation exchange in groundwater (Fırat Ersoy and Ersoy 2009).
The first level of this classification of groundwater, which corresponds to the main type of water, is determined according to the chloride content. Therefore, depending on the chloride concentration, there are six different main types; fresh (F), fresh-brackish (Fb), brackish (B), brackish-salt (Bs), hyperhaline (H), and salt (S) (Stuyfzand 1986;Fırat Ersoy and Ersoy 2009). The classification of water main according to the Cl − is given in Table 1. The second level that represents the water type is determined according to the hardness values. The classification of water types according to the hardness values is given in Table 1. The third level of this classification that corresponds to the subtype of water is determined according to the dominant anions and cations. However, it should be noted that in this classification, the cation or anion with most meq/l does not necessarily determine the name. The strongest hydrogeochemical family is first determined for both cations and anions, e.g., the families [Na + K + NH 4 ] and [SO 4 + NO 3 + NO 2 ]. The stronger member of the two families is then chosen to form the combination. The theoretical maximum number of subtypes is 54. However, only 16 subtypes have been discovered, namely those listed in Table 2. The class of water is finally determined according to the sum of Mg, K, and Na in meq/l, adjusted for contribution of sea salt according to the following formula: (K + Mg + Na) corrected = (K + Mg + Na) measured -1.061*Cl − ; where: the symbol "-" often indicates the intrusion of salt water when (Mg + K + Na) corrected < − √ (0.5 Cl); the symbol " + " often indicates freshwater encroachment when (Mg + K + Na) corrected > √(0.5Cl); the symbol "ϕ" often indicates an equilibrium when − √ (0.5 Cl) ≤ (Mg + K + Na) c orrected ≤ √ (0.5 Cl).
The factor 1.061 of the previous formula corresponds to the ratio {(Mg + K + Na)/Cl} for the average seawater. It is also supposed that all chlorine ions come from the sea and the fractionation of the main components of seawater during spraying can be neglected and that the Cl concentration behaves conservatively (Stuyfzand 1986). The water classification according to Stuyfzand classification system is given in Table 2.

Hydrochemical parameters of groundwater samples
The pH values of groundwater samples are between 6.5 and 8, TDS value range is 2679.20-399.16 mg/L, EC range is 587-3940 µS/cm, and DO values are between 1.37 and 8.07 mg/ (Table 3). High values of EC and TDS were observed in wells located towards the center, in the northeast part, and locally in the west part of the study area, while the lower values are observed in the northern part, and towards the west and south part. Ca 2+ , Mg 2+ , Na + , and K + values are, respectively, between 52.91 and 167.2 mg/L, 16.65 and 114.21 mg/L, 20.01 and 609.04 mg/L, and 1.56 and 25.35 mg/L. The NH 4 + value range is 0-4.39 mg/L with a median value of 0.92 mg/L (Table 3). Higher Na + value was observed in well number 13 located, respectively, in the center of the study area. Anion's analysis results show that the Cl − value range is 18.08-1001.82 mg/L, HCO 3 − value range is 248.92-822.41 mg/L, and SO 4 2− value range is 19.69-504.80 mg/L. The NO 3 − value range is 0.25-101.68 mg/L with a median value of 16.06 mg/L. The concentration of nitrate in groundwater usually comes from the oxidation of ammonium (nitrification) (Rajmohan et al 2009). In the presence of oxygen,   (Mcclain et al. 1994;Abou Zakhem and Hafez 2015). In the study area, 87% of water samples have DO concentration greater than 2 mg/L. This situation favors the oxidation of NH 4 + to NO 3 − in the study area (Casciotti et al 2010; Abou Zakhem and Hafez 2015). However, it is important to clarify that DO concentration in groundwater and soils cannot be used to predict high nitrate levels, since there may not be a source of nitrogen. The nitrification process which leads to the oxidation of ammonium to nitrate is given in the following equations (Elisante and Muzuka 2016;Rajmohan et al. 2009): In the study area, high concentrations of NO 3 − are mainly due to discharges of industrial wastewater, urban domestic wastewater, and the use of nitrogen fertilizers. The work carried out by Arslan et al. (2017) in the coastal plain of Çarşamba to assess the spatial and seasonal changes in groundwater pollution by nitrates in the agricultural area showed that the leaching of nitrates to the groundwater due to excessive use of fertilizers could be the cause of the high concentrations of nitrates in groundwater (1) of the Çarşamba coastal aquifer. The high Cl − values were observed towards the center, the western part, and locally towards the northeast of the study area. It is observed that in the study area, the high values of Cl − , Na + , TDS, and EC have been observed in industrial and agricultural areas.
These values indicate a seawater intrusion caused by an over-extraction of groundwater in these areas. The high NH 4 + values observed in some wells are influenced by industrial waste, the excessive use of fertilizers and pesticides in the study area (Maman Hassan 2021). In Sultanate of Oman, the work carried out by Rajmohan et al. (2009) showed that nitrogen fertilizers and farm manures, which are normally used for paddy and other crops, are the major ammonium sources in groundwater of agricultural areas.
Piper classification diagram showed four different water facies in the study area, which are CaHCO 3 , Ca-Mg-Cl, NaCl, and Ca-NaHCO 3 (Fig. 4). The CaHCO 3 water type, which is dominant in the study area, represents around 73% of analyzed samples. The abundance in the study area of this water type that is characterized by the dominance of Ca and HCO 3 ions reveals a strong carbonate rock-water interaction and strong recharge process. The Ca-Mg-Cl water type represents 18% of analyzed samples and were observed in 22, 19, 33, 5, 7, and 14 well number. The Ca-NaHCO 3 water types were observed in wells number 23 and 16, and represented around 6% of analyzed samples. NaCl water type is only observed in 13 well number; the Ca-NaHCO 3 water types of results from the cation exchange between seawater and fresh water during the freshening process. In addition, the presence of Ca-Mg-Cl and NaCl water facies also indicates the seawater intrusion in the study area.
The Piper diagram can also be used in determining of seawater intrusion into groundwater using Kelly (2005) classification. Therefore, on this diagram, the fresh groundwater samples will land near the area labeled "fresh water", while the water resulting from a conservative mixture (mixture without ion exchange reactions) between freshwater and seawater would be drawn along the line labeled "mixed water" (Kelly 2005).
According to this classification (Fig. 4), 75.76%, 21.21%, and 3.03% of the groundwater samples fall in freshwater water area, mixed water area, and seawater intrusion area, respectively. The water samples which fall in mixed and seawater intrusion areas indicated the influence of seawater on groundwater in the study area.
According to Chadha diagram, four water facies were observed in the study area which are CaHCO 3 , Ca-Mg-Cl, NaCl, and NaHCO 3 (Fig. 5). The CaHCO 3 water type that falls in field 5 of diagram is the most widespread in the study area and represents around 70% of analyzed samples. This dominant water type reveals as in the piper diagram that the weathering of host rocks and recharge process are most important in the study area. The Ca-Mg-Cl water type, which falls in field 6 of diagram, represents around 12% of analyzed samples and were observed in wells number 22, 7, 14, and 33. This water type indicates reverse ion exchange process that corresponds to the beginning of the salinization process in the study area. The Na-Cl water type, which falls in field 7 of diagram, represents around 9% of analyzed samples and were observed in wells number 13, 31, and 28. This water type indicates evaporation, mixing between freshwater, and seawater. The CaNaHCO 3 water type that falls in field 8 of diagram represents around 9% of analyzed samples and were observed in wells number 1, 16, 14, and 23. This water type shows the cation exchange reaction resulting from the flushing of the saline aquifer by freshwater. The ratio of sodium to total sodium and chloride to total dissolved solids (TDS) can also be used for to evaluate the dominant ion process (ion exchange and reverse ion exchange) in the aquifer (Ghezelsofloo et al. 2021). Thus, in this diagram, the ratio of sodium to total sodium and chloride is the main factor that determines the effects of ion exchange in the aquifer (Fig. 6). The values of (Na/Na+ Cl) less than 0.5 indicate that the ion exchange process is occurring in the aquifer, while those greater than 0.5 indicate that the reverse ion exchange is a dominant ionic process in the aquifer (Ghezelsofloo et al. 2021). According to Fig. 6, most of water samples from study area fall in ion exchange area. However, water samples from wells numbers 24, 14, and 13 fall in seawater or reverse softening area. Seawater intrusion into the aquifer is one of the most important factors which cause the reverse ion exchange process and hence the salinization of the groundwater water (Ghezelsofloo et al. 2021;Papazotos et al. 2019).
Data plotted on Gibbs diagram show that in the study area, groundwater are influenced by two main mechanisms that include water-rock interaction and evaporation-crystallization (Fig. 7). The groundwater samples that fall in rock dominance field are the most widespread and represents around 80% of analyzed samples. The rock-water interaction, which is dominant, indicates that in the study area, local geological context and hydrogeological conditions significantly influence the chemical composition of groundwater. The groundwater samples that fall in evaporation-crystallization dominance field are the least dominant and represent around 20% of analyzed samples. Gibbs diagram can be used to determine the influence of seawater on the aquifer (Seenipandi et al. 2019). Water samples (20%) that fall in evaporation-crystallization dominance field indicate predominantly evaporative conditions leading to higher amounts of total dissolved solids and increased salinity values by increasing ionic concentrations due to diffusion of saltwater through seepage movement (Seenipandi et al. 2019;Chidambaram et al. 2009). Similar results were observed in the work carried out by Umarani et al. (2019) where due to seawater mixture, most of water samples collected during the summer in coastal aquifers of Tamil Nadu fell into evaporative dominance field. In Malaysia, the work realized by Sefie et al. (2018) to assess the hydrogeochemistry and groundwater quality of lower Kelantan basin shows that the water samples falling in the evaporation dominance zone of the Gibbs diagram were influenced by fossil seawater, probably trapped during sedimentation in this area.
The results of the Kolmogorov-Smirnov and Shapiro-Wilk test performed to test the normality of variables are presented in Table 4. According to Table 4, almost all variables have no normal distribution. Thus, Spearman correlation method can be used to strength of the links between variables. Spearman correlation coefficient matrices for physical and hydrochemical parameters of groundwater samples are presented in Table 5. According to this table, Cl − was highly correlated with Na + (R = 0.816) and moderately correlated with K + (R = 0.551), TDS (R = 0.522),   − 0.627). The moderate positive correlation observed between Cl − and EC, and the strong positive correlation observed between Cl − and Na + indicate that these ions come from the common source which is seawater intrusion (Sae-Ju et al. 2020;El Moujabber et al. 2006). In the coastal area, the concentrations of Cl − and EC values are used as classic indicators of salinization of groundwater (El Moujabber et al. 2006). The bivariate diagram of Cl − /HCO 3 − ratio and Cl − is generally used to determine the mixing of groundwater and seawater (Seddique et al. 2019). According to this diagram, in the study area, most of water samples that have lower Cl − /HCO 3 − ratios can be characterized as fresh waters (Fig. 8). However, some water samples present high Cl − /HCO 3 − ration and high Cl values. These samples indicate so the mixing between groundwater and seawater in the study area.

Stuyfzand classification system
The spatial distribution of the first level of water classification according to Stuyfzand, which corresponds to the main type of water determined according to the chloride content, shows that the study area can be divided into three regions: the fresh, fresh-brackish, and brackish-salt water regions (Fig. 9). The freshwater, which represents 84.84% of the samples, occupies a very large portion of the study area. The fresh-brackish water represents around 9% of groundwater samples, and occupies in the south, the center, north, and northeast of the study area. This dominance of F and Fb can be explained by the significant recharge of the aquifer and the abundant infiltration of surface water, which contributes to the refreshing of aquifer and limits the mixing between salt and fresh water at least in the study area. The Bs water main type was only observed in one well located towards the center of the study area.
The spatial distribution of the second level of water classification, which corresponds to the type of water, determined according to the total hardness values, shows that the study area can be divided into three regions: hard, very hard, and extremely hard water regions (Fig. 10). The groundwater samples' hardness values are varying between 21 and 85.55 °F with an average of 44.22 °F. The very hard water region is the most widespread in the study area and represents around 67% of analyzed samples. The hard water region occupies locally the northeast, the Northwest, and a small part of the center and south of study area. The extremely hard water types were only observed in one well located towards the center of study area. The determination of the third level of this classification which corresponds to the water subtypes determined according to the dominant anions and cations shows that the water samples are distributed according to six types of facies: the CaHCO 3 facies, which represents around 60.60% of samples, the CaMix facies which represents around 18.19% of the samples, the NaMix facies which represents around 6.06% of the samples, the NaCl facies which represents 3.03% of samples, the CaNaHCO 3 facies, which represents 9.09% of samples, and, finally, the MgHCO 3 facies which represents 3.03% of samples (Fig. 11). The facies of water samples determined according to Stuyfzand classification system is given in Table 6.
The spatial distribution the water subtypes shows that the subtypes CaHCO 3 are widely distributed in the study area, the subtypes CaMix are found in the west and the center, the subtypes NaHCO 3 are found in the center and extreme northwest, the subtypes NaMix occupy locally the north and the east, the subtypes NaCl are found in the center, and finally, the subtypes MgHCO 3 are found in the extreme east part of study area (Fig. 12).
The last level of this classification, which corresponds to the water class, was determined according to the sum of K + , Na + , and Mg 2+ in meq/l, adjusted for the contribution of sea salt, shows that all groundwater samples have positive cation exchange code.
The subtype CaMix, which occupies the west, and the center of the study area shows the transition zone location, where the groundwater rich in Ca and HCO 3 and gradually enriched in Na, changes from CaMix with HCO 3 as dominant anion, to eventually CaCl and NaCl facies. This process corresponds to the beginning of the salinization process in this area. In the coastal areas, cation exchange reaction is the mainly process which controls hydrochemical composition of groundwater affected by seawater interference. This subtype CaMix was observed in wells number 2, 5, 7, 14, 22, and 33. The subtype NaCl observed in well number 13 also indicates the seawater intrusion in aquifer. Freshwater is in general considered as an infiltrating rain, which has dissolved calcite, and the water type in this situation is F-CaHCO 3 ϕ, while the seawater type is S-NaClϕ (Al Farrah et al. 2011). Therefore, in the study area, the salinization process is mainly caused by seawater intrusion which reduces the quality of groundwater. In addition, the excessive use of fertilizers, pesticides in agricultural areas, as well as industrial waste from the industrial area leads to pollution of groundwater in this region. During the salinization process, a cation exchange reaction takes place, resulting in a deficit of Na + , Mg 2+ , and K + and an excess of Ca 2+ according to the following chemical reaction:  During this cation exchange reaction, there is an adsorption of dominant Na + , K + and Mg 2+ ions and a release of [aCa] − X + bNa + + cK + + dMg 2+ ⇒ aCa 2+ + bNa, cK, dMg − X + aCa 2+ .
Ca 2+ ions, causing the move of water resulting from NaCl water type to the CaCl water type, which is characteristic of salinization process. This process is generally at the origin of formation of the (Na + K + Mg) deficit in groundwater. The process of salinization in groundwater can be represented as follows:  It is observed that the classes of samples number 2, 5, 7, 14, 22, and 33 affected by seawater intrusion present falsely positive cation exchange code. This false-positive cation exchange code is not due to cation exchange but due to Mg 2+ ions which dissolved from the carbonate rocks in the study area.
The subtypes NaMix, NaHCO 3 , and MgHCO 3 observed in the study area indicate the locations of the transition zone where the flushing of the saline aquifer by freshwater takes place. This process is observed when freshwater from coastal aquifers infiltrates into salt water, thus causing the following chemical reaction: During this cation exchange reaction resulting from the flushing of the saline aquifer by freshwater, there is an adsorption of Ca 2+ ions and a release of Na + , Mg 2+ , and K + ions, causing the formation of NaHCO 3 which is typical for freshening. The freshening process is generally at the origin of formation of the (Na + + K + + Mg 2+ ) surplus in groundwater. It is accepted that initial seawater conditions (SNaClϕ) dominated in the beginning of the process. Original seawater diluted gradually due to the recharge phenomena, and then, sea water conditions coming from S ended up F (Stuyfzand 1986). The cation exchanges that occur during the freshening process can also be schematized as follows (Stuyfzand 1986;Fırat and Ersoy 2009): NaClϕ ⇒ NaCl + ⇒ NaHCO 3 + ⇒ MgHCO 3 + ⇒ CaHCO 3 + ⇒ CaHCO 3 ϕ S ⇒ Bs ⇒ B⇒ Fb ⇒ F. It is observed also that the groundwater samples affected by freshening process in study area present a positive cation exchange code which is due essentially to cation exchange between fresh water and salt water.
The high Na + values observed in some samples could also come from ion exchange between the clay particles and groundwater. This exchange takes place between the Na + ions of the clay particles and the Mg 2+ and Ca 2+ ions of the groundwater. These ion exchanges therefore lead to an increase in Na + ions and a decrease in Ca 2+ and Mg 2+ ions in groundwater (Şahinci 1991).

Isotopic signatures
In hydrogeological studies, the analysis of natural variations in the content of heavy isotopes of oxygen ( 18 O) and hydrogen ( 2 H) present in the structure of water is one of the methods used in the study of the origin and dynamics of groundwater and surface water. Due to the differences in the heavy isotope content generally observed between groundwater and seawater, the isotope content of groundwater mixed with seawater can be easily separated from the isotope content of meteoric water. In the study area, δ 18 O and δ 2 H content in groundwater ranged from − 8.89 to − 7.21‰ and from Meteoric Water Line (δ D = 8* δ 18 O + 10) (Craig 1961). The relationship between δ 2 H and δ 18 O allows to understand the mixing process between freshwater and seawater (Nair et al  Figure 13 shows the relationship between stable isotopes δ 2 H and δ 18 O of groundwater and seawater. According to the isotope distributions of δ 18 O and δ 2 H, all groundwater samples have meteoric origins (Fig. 13). However, it is observed that the water samples taken from wells number 13, 14, 16, 19, and 26 have more positive δ 18 O and δ 2 H values than the other samples. More positive values of δ 18 O and δ 2 H indicate interference between seawater and groundwater in the study area.
To determine the level of seawater intrusion into groundwater, important information can be obtained by examining the relationships between isotope analysis results and physico-chemical analysis results. There are close relationships between the values of 18 O, Cl − , SO 4 2− , and EC of groundwater. In general, seawater intrusion in aquifer results in an increase in the δ 18 O, Cl − , SO 4 2− , and EC values of the groundwater. Relatively high values of δ 18 O (0 ‰) as well as high concentrations of Cl and TDS indicate seawater origins, while low values of δ 18 O and low content of Cl and of TDS are characteristics of fresh groundwater origin (Wang and Jiao 2012). In this study to assess the seawater influence on groundwater, the cross plots of Cl − vs δ 18 O and EC vs δ 18 O were used. According to the Fig. 14 Because tritium element is a radioactive isotope that degrades over time, it is present in low value in water with long residence time and high value in water with a short residence time. Due to this situation, the concentration of tritium in young water is higher than in olds water. In the study area, groundwater values for 3 H ranged from 0.96 TU to 7.15 Tu. In this study to assess the seawater influence on groundwater, the cross plots of EC vs 3 H and Cl − vs 3 H were also used (Fig. 15). According to Fig. 15, the groundwater samples from wells number 15, 14, and 32 have lower tritium content (< 3 TU) than the water samples from other wells. Therefore, these wells were recharged by groundwater having a longer transit time than that of groundwater which recharged the other wells. In addition, although the water samples from wells 15 and 32 have low tritium values, its EC (< 1500 µS/cm) and Cl − (< 4 meq / l) values are also low. This situation indicates that these wells are not affected by seawater intrusion. High values of EC (> 1500 µS/cm) and Cl − (> 6 meq/l) were observed in wells 13, 28, and 22 due to seawater intrusion. The majority of the groundwater samples collected from study area have tritium values greater than 4 TU. In addition, it is found that most of these wells have EC values less than 1500 µS/ cm and Cl − values lower than 6 meq/l. The high values of 3 H indicate that these wells were recharged by water from recent precipitation.

Conclusion
This study seeks to determine the hydrochemical properties of the groundwater of Çarşamba coastal aquifer; one of the most important coastal aquifers in Turkey. According to the first level of Stuyfzand groundwater classification, the most widespread main types are freshwater and fresh-brackish water in the study area with, respectively, 84.84 and 9% of analyzed samples. However, brackish-salt water main type was observed in one water sample from well number 13. The second level of this classification showed that very hard water region is the most widespread in the study area and represents around 67% analyzed water samples. In addition, the third level of Stuyfzand groundwater classification showed that water samples were distributed into six types of facies, which are CaHCO 3 , CaMix, NaMix, NaCl, NaHCO 3 , and MgHCO 3 . Piper diagram indicated that in the study area, the hydrogeochemical facies of water samples are CaHCO 3 , Ca-Mg-Cl, NaCl, and Ca-NaHCO 3 . Furthermore, Chadha diagram revealed four chemically different groundwater facies in the study area. These facies include CaHCO 3 that reveals the weathering of host rocks and recharge process; Ca-Mg-Cl and NaCl that indicated reverse ion exchange process and NaHCO 3 , which indicated the cation exchange reaction, resulted from the flushing of salt water by freshwater. Gibbs diagram showed that the groundwater samples are influenced by two main mechanisms that include water-rock interaction and evaporation-crystallization. The rock-water interaction is dominant, and indicates that the local geological context and hydrogeological conditions significantly influence the chemical composition of groundwater. Groundwater values for δ 18 O, δ 2 H, and 3 H ranged from − 8.89 to − 7.21‰, from − 56.41 to − 46.64 ‰, and from 0.96 TU to 7.15 TU, respectively. The isotope distributions of δ 18 O and δ 2 H showed that all groundwater samples have meteoric origins. According to the cross plots of Cl − vs δ 18 O and EC vs δ 18 O, more positive δ 18 O values are observed in samples with higher Cl and EC values. High Cl − and EC values and more positive δ 18 O values are observed in well numbers 13, 14, 28, and 22. These wells were affected by seawater intrusion and the 3 H values indicated that more water samples have shallow circulation and a short residence time.