Quality assessment and hydrochemistry of a coastal aquifer adjacent to a hypersaline lake: A case study of Western Asia


 The study is motivated by the quality degradation of groundwater with emphasis on salinity between 2012 and 2018 in an aquifer, namely Azarshahr aquifer, located in Western Asia (a case of north-west of Iran), adjacent to Lake Urmia. The groundwater of Azarshahr plain in the south-east, namely, the nutritional zone of the plain, has a low amount of salts (0.7–18.06 mg/L), but by moving to the north-west and west of the plain, which is the location of the outflow of water, the concentration of salts is significantly increased (35.42–87.5 mg/L). Ca–Cl has the cardinal influence on quality of water, making it different from the common type of fresh waters in Iran (Ca–HCO3). High concentrations of SO42−, Cl−, and Na+ detected in almost all the water samples, which indicates aquifer’s good quality for portable applications. Regarding agricultural applications, from 2012 to 2018, the percentage of samples in the good class of C2S1 has been reduced by 35%, which could also be a further indicator of increasing salinity in the aquifer. The spatial distribution of electrical conductivity (EC), cations, and anions tends to follow an increasing pattern toward western regions. Comparing the results of hydrochemical analysis of Lake Urmia with the corresponding results from the groundwater in Azarshahr plain, there is an acceptable correlation between the hydrochemical features of Lake Urmia with the saltwater of western parts of the Azarshahr plain. Hence, it is likely that another source of water salination in this plain, especially in western parts, is the influx from Urmia Lake. Due to the presence of clay mass of mountains between the Lake Urmia and Azarshahr plain as well as the lack of a hydraulic relationship, the probability of penetration of saline water from reinforced water reservoirs of Lake Urmia to the aquifer of Azarshahr plain is further strengthened.


Introduction
Industrial development accompanied by population growth has imposed heavy pollution loads to natural resources (Mehrdadi et al., 2006;Nasrabadi and Bidabadi, 2013;Rowshan et al., 2007). Water resource contamination is one of the major challenges in the way of sustainable development (Mehrdadi et al., 2006). From the total accessible fresh water all around the world more than 90% is contributed to groundwater resources (Petersen et al., 2017). Accordingly, sophisticated attention towards monitoring the quality and quantity of such resources would play a key role in achieving the global sustainable development in near future (Gerten et al., 2013).
During the past two decades, groundwater quality evaluations in different parts of the world have been studied by various researchers (Hu et al., 2019;Haidu and Nistor, 2019;Malov and Tokarev, 2019;Selvam et al., 2014;Pei-Yue et al., 2010;Melaku and Wang, 2019;Almeida et al., 2018). In this regard, Thirumalini and Joseph (2009) have examined various sampling wells to determine regression equation between electrical conductivity (EC) and total dissolved solids (TDS) for fresh water and saltwater in the Triuvallur district located on the northern border of Tamil Nadu, India. They reported that a linear correlation exists between these two parameters for fresh water, whereas there is a logarithmic correlation for saltwater. Shah et al. (2008) have compared groundwater quality in Gandhinagar Taluka in India with standard values described by World Health Organization (WHO, 2017) and have come up with a water quality index for that area.
Urmia Lake is one of the largest hyper-saline lakes in the world and the largest one in the Middle East (Gorgij et al., 2019). Once it was considered as the second largest hypersaline lake in the world, is suffering from sewer drought and saltwater intrusion. Ghale et al. (2017) reported that salty soil areas of the Lake Urmia have increased dramatically from 1995 to 2014 and more 4 than 5000 km² of Lake Urmia's water surface area was converted to salty soil bodies during recent years. Mardi et al. (2018) reported that the combined area of salt and salty soil bodies around Urmia Lake have increased by two orders of magnitude in the past two decades. Jeihouni et al. (2018) assessed the groundwater quantity over 11 years using a novel groundwater balance estimation method. The results indicated the negativity of the groundwater balance during this period which furthermore decreased the quality of the groundwater over the study period with the most severe condition in west and southwest of the study area. Shakerkhatibi et al. (2018) found that the total hardness of the groundwater in the region is highly correlated with the magnesium concentration than calcium. They reported that the dominant cations and anions in Lake Urmia were in the order of Mg 2+ > Na + > Ca 2+ > K + and HCO3 -> SO4 2-> Cl -> NO3 -> F -, respectively. They also reported that the major water types in the area were fresh (Ca-HCO3) and saline (Ca-Mg-Cl).
The need for water has produced an increasing withdrawal of groundwater in sensitive areas like deserts, where aquifers suffer from saline water intrusion, which consequently results in a deterioration of its quality (Arabameri et al., 2019). Groundwater salinization occurs in many aquifers around the world (Argamasilla et al., 2017;Moreau et al., 2019;Ferrer et al., 2019;Ponsadailakshmi et al., 2018;Vižintin et al., 2018;Bertrand et al., 2006). Understanding the origin and mechanisms of the salinization process is an important point for preventing further deterioration of groundwater resources. In the nature, elements such as radon and arsenic can leach into aquifers that have been drawn down (Amiri et al., 2014). Meanwhile, lowered levels of freshwater in the top layers of aquifers can also expose ponderous and non-buoyant saltwater to settle down which in longer time results in salinization of both ground water and any other adjacent water body. Temporal changes in the origin and constitution of the recharged water, hydrologic and human factors may cause periodic changes in groundwater quality (Reza and Singh, 2010).
Considering the relevant literature gap on the present subject, a hypothesis was defined to assess the suitability of the groundwater in Azarshahr aquifer for drinking and agricultural applications. Besides completing the information about groundwater quality in the Azarshahr aquifer, an assessment of the quality and hydrochemistry of an aquifer adjacent to a hypersaline lake was conducted. In this regard, the main objective of the present work was to provide a comprehensive assessment on the quality gradients of the groundwater in Azarshar aquifer in a period between 2012 and 2018.

Study area
The study area is in the north-west of Iran, between the eastern longitude of 44°, 20′ and 45°, 20′ and northern latitude of 37°, 05′ and 38°, 05′ ( Fig. 1) with approximately 4268 km 2 in area and altitude in range of 1280 and 3608 above mean sea level. The Urmia Lake (UL) has a large storage capacity, regulating the inflow and outflow from a significant drainage area. The water depth varies from a few meters to near 130 m in northern and southern parts of aquifer, respectively. The groundwater flow direction is from west to east (UL). The major sources for recharging the Urmia Lake are four recurrent rivers, including Nazlou-chai, Rowzeh-chai, Shahrchai, and Barandouz-chai, which originate from the western mountains of the area. The mean input to the aquifer from these rivers and return flows from irrigated lands is about 290 million cubic meters (MCM) per year (Amiri et al., 2016a). In addition, infiltration from precipitation is near 37.7 MCM per year (Hamidi-razi et al., 2019). In this area, the irrigation is 6 mainly from groundwater sources (Khatami Mashhadi, 2013). The oldest rock units of the Pre-Cambrian are formed by meta volcanic series, acidic tuff, and diorite in the surrounding mountains of Urmia Lake, as well as metamorphous amphibolites and gneiss (Hamzehpour et al., 2018). Tertiary rocks in this plain are represented by limestone, conglomerate, sandstone, and shale (Hossein et al., 2018).

Geological conditions
Azarshahr plain is located in Azerbaijan Province, northwest of Iran. The plain in UTM coordinates is in the longitudinal axis between 518,200-and 575,700-meters east and between 4,197,000-and 4,163,500-meters north. This area is in the west of East Azarbaijan Province and on the eastern shore of Urmia Lake, which is limited to the north of Tabriz plain and south of Azarshahr Travertine hills. The plain is bordered from the west by the Urmia Lake salty flat plains. Sahand volcanic altitudes are in the east of this plain, and the coasts and salt marsh of Lake Urmia have limited the western part of this plain (Jeihouni et al., 2018).
The Azarshahr plain is comprised of igneous, sedimentary, and volcanic rocks. While igneous rocks extend in the southeast of the Azarshahr plain, volcanic sedimentary rocks mainly cover the east of the area, particularly from northeast to southeastern (Fig. 1). Even though central and western parts of Azarshahr plain are covered by sedimentary rock consisting of quaternary sedimentary rock such as travertine and young traces, it is considered the main source of potable and agricultural groundwater supply (ATWA, 2009). The plain includes layers with ages of Jurassic to Quaternary which are under the strong domination of movements with Alpine origin. The Sahand alluvial tuff conformably overlies Pliocene marls, sandstones, and fish-bed layers. The southwestern part of study area includes Jurassic and Cretaceous limestone with Pliocene travertine, which is believed to be connected to the thermal mineral issuing from the Cretaceous limestone as well as from alluvial tuff. The alluvial water course and plain deposits of the study area are derived from the erosion of Sahand pyroclastic materials, which have transported by water and other agents in to the Azarshahr Plain.
The total area of Azarshahr is around 457 km 2 while 124 km 2 of that is covered by Azarshahr plain. The common border of the aquifer with UL is approximately 92 km, which is the longest border with UL in compare to other aquifers in the region. However, this long mutual border and vicinity to UL makes the Azarshahr aquifer more vulnerable to the salty water invasion, which is more or less detectable in the samples taken from the outflow of the aquifer (Hossein et al., 2018). East and central parts of the Urmia region consist of wide plains where northern, western and southern partitions consist of mountainous regions (Alizadeh, 2013).
Geologically, the study area consists of Sahand volcanic rocks from Late Miocene to Pleistocene from southeast of plain. Quaternary units consist of alluvial terraces with alluvial plain sediments. Quaternary travertines also have outcrops in the south and southwestern parts of the plain, which are of great interest in leveraging the groundwater quality.
As illustrated in Fig. 1, geological formations that are adjacent to or in the direction of groundwater movement, depending on the lithology of the formation and the physicochemical properties of the water, affect the groundwater quality (Daw et al., 2018). Also, because of tectonics in relation to hydrogen units and groundwater flow, it is necessary to study the lithology of the formations and tectonics of the area and their role in changing water quality and groundwater movement. In terms of topography, the highest elevation is the western slopes of Sahand Mountain and the lowest points around the Lake Urmia, which have an altitude of 3100 meters and 1282 meters above sea level, respectively (Gorgij and Moghaddam, 2016).

8
Azarshahr plain only has a river called Gombrakhay River, which is also considered the main river. This river originates from the western slopes of Sahand Mountain and flows eastward to the west and after passing Azarshahr plain ends at Urumieh Lake. The existence of suitable agricultural land and flat plain has led to agricultural prevalence in this area and the main income of the majority of the villagers around Azarshahr is agricultural career. In the lower reaches of the hands it is dried, therefore, the groundwater is used for drinking, agriculture and sanitation in this area, but it is of great importance that increasing the harvesting of groundwater resources, in a rate higher than the recharge rate has led to exploitation of the aquifer. The maximum monthly mean precipitation falls in May (53.96 mm) and the minimum in August (4.16 mm) (Fatollahzadeh et al., 2016). Over the last few years, the average annual precipitation in the region has been significantly reduced. Therefore, it can be concluded that decreasing rainfall in recent years reduces feeding and increases hydraulic load loss in groundwater aquifers and can be effective in both degradation of water quality and infiltration of saline water. The climate of Azarshahr plain is under the influence of Mediterranean and cold weather streams (Delju et al., 2013;Motevalli et al., 2018).

Sample collection and analysis
To appropriately cover the study area, different sites in the plain were randomly selected for groundwater sampling and their respective location was recorded by a portable GIS device, (GRS-1, Skipper Technologies India Private Limited). Using Student's t-distribution, the number of samples within the 95% confidence level is determined as follows: where t is the t-statistic value selected for a given confidence level (2.23 for confidence level of 95%), s is the overall standard deviation, and e is the acceptable level of error or uncertainty. The degrees of freedom (df) which determines t is first selected randomly and then modified by successive iterations. For df = 10 with s = 3.86 and e = ±1.5 µg/g uncertainty, a total number (n) of 33 samples would be necessary.
Judgmental sampling strategy using prior knowledge of spatial and temporal variation of the pollutants was employed to identify the locations for sampling, particularly the location of the present pollutants (in our case around the outfall point). Moreover, a composite sampling method containing several separate samples taken at ten random time points has been employed.
If the pollutant of interest was detected, then the individual samples have been analyzed individually and the average was recorded.
Sampling was carried out using a set of white polythene bottles. To maintain the sampling accuracy up to a level, the wells were pumped for 5 min and the sampling bottles were washed thoroughly with groundwater to be collected. To determine the suitability of groundwater for different uses such as agricultural, industrial and domestic uses, the data obtained from the sampling wells, monitored during June 2012 and June 2018 by the Ministry of Energy, which were analyzed for physiochemical parameters such as pH, TDS, EC, and hardness values (Table 1 and 2). The experimental values were compared with standard values recommended by World Health Organization (WHO, 2017) for drinking water and public health (Table 2). According to these results, in most parts of this basin, these values exceeded the prescribed limit of WHO.

Measured parameters and analytic procedure
Parameters such as pH, EC, TDS (total dissolved solids), TH (total hardness), major cations (Na + , K + , Ca 2+ , and Mg 2+ ) and major anions (CO3 2− , HCO3 -, SO4 2− , and Cl − ) were taken into consideration. pH and EC of each sample were measured in situ by a digital pH and EC meter, respectively. TDS was determined gravimetrically at 105-110 °C (Kazi et al., 2009). In laboratory, the duplicate aqueous samples of 1000 mL of each batch collected were filtered through polycarbonate filter (0.45 mm pore size), and the samples were divided in two parts. One part was used for analysis of anions, while the second part treated with 2 mL of concentrated HNO3 solution was used for metal analysis.
The acid-treated water samples were analyzed for the determination of major cations by further 20-time dilution with ultra-pure water. Ca 2+ , Na + , and K + were measured by flame photometry (Model 410 Sherwood Incorporation, USA), while Mg 2+ was determined by the flame atomic absorption spectrometer (FAAS), HACH Incorporation, USA. In case of anion concentrations, sulfate has been measured by HACH DR/2000 (direct reading spectrophotometer) using the method number 8051, chloride has been measured through argentometric course using the method number 2330 and bicarbonate has been measured by titration using the method number 4500, respectively (APHA, 1998).

Statistical analysis
The relationship between different dissolved species is used to reveal the origin of solutes as well as the main hydrogeochemical processes (such as dissolution, ion exchange, and enrichment or depletion of major ions) (Kumar et al., 2009 Excel 2016 were used for this purpose.

Principal components analysis
Principal components analysis (PCA) is one of the best ordination methods that consist of an eigen analysis of a covariance or correlation matrix calculated on the original measurement data. Graphically, it can be described as a rotation of a swarm of data points in multidimensional space so that the longest axis (the axis with the greatest variance) is the first PCA axis, the second longest axis perpendicular to the first is the second PCA axis, and so forth. Assuming the sample parameters as the original set of variables, and the Euclidean distance matrix as an estimate of the correlation matrix explaining the correlations between each pair of samples, the PCA framework for grouping the samples into separate independent clusters was automatically assigned and formed. In the PCA method, the initial clusters are extracted out by the eigenvalueeigenvector analysis of the similarity matrix as presented in Eq. (1): where S is a P × P Euclidean distance matrix, I is the identity matrix, λi are the characteristic roots (eigenvalues), and Yi are the corresponding eigenvectors. Eq. (3) is an eigenvalueeigenvector equation, λ1 ≥ λ2 ≥ … ≥ λp are the real, nonnegative roots of the determinant polynomial of degree P given as For the values obtained from Eq. (2), Yi can be calculated. According to the PCA method, each of the P independent principal components (factors) can be written as a linear combination of the original variables (water samples), with the elements of the P eigenvectors as the coefficients of these linear combinations. There should be low similarities among samples that are associated with different clusters and high similarities among samples strongly associated with the same cluster.

Hierarchical cluster analysis
Nine variables (Ca 2+ , Mg 2+ , Na + , K + , HCO3 -, CO3 2-, Cl-, SO4 2-, and TDS) and 33 water samples were considered for the cluster analysis. The Q-mode clustering analysis resulted in two major water groups (Groups A and B) and seven subgroups ( Table 3. The average values for each of the composition of the subgroups produced by the HCA analysis reveal trends between them and they are the basis for the distinction of the subgroups. The HCA 13 result is consistent with the analysis made on the basis of Piper plots. Most important trends are increase of Na and TDS towards the rift and the reverse in Ca and Mg as one goes down from SB-1 to SB-7. The relationship between the statistically defined clusters of samples and geographic location was prepared by plotting subgroup values for each sample. The seven subgroups are separated geographically, as well as physiographically with good correspondence between spatial locations and the HCA results. Samples that belong to the same subgroup are located in close proximity to one another suggesting more or less the same hydrogeochemical processes (evolution) and/or flow paths.

Results and discussion
In this study, scree plot (introduced by Raymond B. Cattell in 1966) was used to select the number of the clusters. It can be observed from the scree plots ( Fig. 2) for two periods that only two clusters are needed to group the water samples. These two groups contribute 85.62% in 2012 and 88.22% in 2018, respectively. Therefore, it is clear that the water sample data for two seasons can be clustered into two groups.
[ Fig. 2, here] pH According to WHO guidelines (WHO, 2017), the range of pH value prescribed for drinking purposes is 6.5-8.5. The pH values of groundwater in the study area varied between 7.68 and 8.35, indicating slightly acidic to slightly basic water. These pH values were all in the desirable ranges. According to WHO, pH less than 6.5 or greater than 9.2 would markedly impair the potability of drinking water. pH usually has no direct impact on human health; however, higher value of pH can increase the scale formation in water pipes and also reduce disinfection potential of chloride (WHO, 2017). More alkaline water requires a longer contact time or a higher free residual chlorine level at the end of the contact time for adequate disinfection (WHO, 2017). For example, at pH 6.0-8.0 the free residual chlorine must be 0.4-0.5 mg/L, at pH 8.0-9.0, it is rising to 0.6 mg/L, chlorination become ineffective above pH 9.0 (WHO, 2017).

Electrical conductivity
Electrical conductivity (EC) shows the concentration of ionized substances in water (WHO, 2017). The maximum permissible concentration of EC for drinking water is 1400 μs/cm (WHO, 2017). In this basin the EC values in the samples were in the range of 1553-12,300 μs/cm, indicating the higher EC values than that of prescribed limit for drinking water. The EC distribution is shown in Fig. 3a and 3b, which indicates high concentrations of total salts, especially in the east of this basin. Chebotarev (1955) studied on the chemical evolution of groundwater in the length of flow and stated that a bicarbonate-sulfate-chloride water type exists from recharge to discharge. In addition, the most important reason for EC increase in the Azarshahr aquifer most probably would be the salt water intrusion from the Urmia Lake (Amiri et al., 2014a).

[Fig. 3, here]
In general, according to Fig. 3a and 3b, it is possible to receive changes in the direction of increase in electrical conductivity in the Azarshahr plain from the eastern part and highlands to the western regions. In 2012, the lowest electrical conductivity in the southeast and northern part of the plain (stations No. 29 and No. 19) was reported to be 473 and 638 μS/cm, respectively.
However, in 2018, the lowest electrical conductivity was observed in the eastern and southeastern parts only at 567 μS/cm. The northern parts of Azarshahr plain show more values of electrical conductivity than in 2012.
As Fig. 3a and 3b indicate, the electrical conductivity in the northern and western parts of the plain is increasing. The process of increase is such that, for example, in well No. 4, the electrical conductivity has reached 12,650 µS/cm. Also, the increase in the central parts of the plain can be noticed. For example, at station No. 11, the electrical conductivity is reached to 10,340 µS/cm in 2018. Considering the relatively high distance of these areas from Urmia Lake, it can be expected that, in addition to the effect of water penetration, the sharp increase in fossil fuels has a significant effect on salinity increase in the central parts of the plain. In general, by comparing the results, the amount of salts that can be increased from west to east in 2018 could be considered as one of the possible reasons for the influx of Urmia lake water into Azarshahr plain aquifer.

Chloride
The high concentrations of chloride can give a salty taste to drinking water (WHO, 2017).
It can increase the rate of corrosion in water pipes (WHO, 2017). According to ( water. About 90% of water samples had higher chloride values than that of prescribed limit for drinking water. The chloride distribution is shown in Fig. 3c and 3d, which indicates a high concentration of total chloride. According to Table 3

Sodium
Since sodium ion is often associated with ionic chloride due to its chemical properties, it is expected that in areas with a relatively high chloride ion, the sodium ion concentration would also be high. Accordingly, in areas where groundwater has low salinity and of relatively good quality, the concentration of sodium is similar to that of chloride and its amount gradually increases in direction of groundwater movement.
Generally, according to Fig Generally, in all studied stations, the trend of increasing sodium ion is visible. Based on this, the reasons for higher concentration of sodium ion in this period can be considered as the reasons for higher Chloride concentration. In other words, the excessive increase in water extraction and the effect of saline water penetration on the coastal beds of Urmia Lake, especially in western areas, are due to the increase of salinity.

Total dissolved solids
According to the results of the correlation analysis, there were correlations among the water quality parameters, which could reveal the possible sources of parameters or potential contributors of chemical components. Each pair of elements showed significant positive and negative correlations. A strong significant positive correlation was observed between EC and TDS and Cl − (0.804-0.889; p < 0.05), whereas TDS also demonstrated a strong significant positive correlation with Cl − (0.998) for both periods of the study. The spatial distributions of TDS are also shown in Fig. 6b and 6c, where red color denote high concentrations and blue color represents lower concentrations. The distributions of TDS and EC had significant positive correlations across the study area. The spatial distributions of calcium ion are shown in Fig. 6d and 6e for the two periods of the study. Table 3  The diagram shows the chemical properties of the water in terms of the relative concentration of its constituents, not in terms of its absolute concentration, so that the type of water can be concluded quite easily.
Figs. 8 and 9 demonstrate the Piper and Durov diagrams of the groundwater samples in the study area, respectively. As shown in left triangle part, the concentration of Ca 2+ is higher than the other cations, indicating the dominance of alkaline earths over alkali metals. However, In the right triangle, the concentration of Clis higher than that of HCO3 -, suggesting the dominance of strong acid over weak one, however, for some samples the concentration of HCO3is higher than that of Cland SO4 2-, highlighting the abundance of the weak acid over strong showing temporary to permanent hardness based on (Nagaraju et al., 2016). The water type ranges from the Ca-Cl type to mixed type (no cation-anion exceed 50%) type according to Sadashivaiah et al. (2008). Basically, the groundwater of the area resembles a non-carbonated typology. In the nutrition zone of the aquifer, the water flow change to sulfate type, and finally, it is converted to the chlorine type due to evaporation. These areas consist of 13 stations from 33 sampling stations (9)(10)(11)(18)(19)(20)22,24,28,30,32,and 33), mainly located in the eastern and southeastern parts of the plain. The sulfate type is striped after the bicarbonate type and maintains the adjacency to the eastern heights (including stations No. 13 to 15 and No. 23).
Possibly, the presence of gypsum compounds in these structures has caused sulfidation in these areas. The chlorine type consists of almost all western regions and parts of the central and southeast (stations No. 1 to 8,12,16,17,21,25,26 and 31). It should be noted that in 2014, due to increased concentration of chloride ions, more areas and stations contain chlorine type water.

Water quality for irrigation purposes
To evaluate suitability of groundwater for irrigation applications, TDS, EC, sodium adsorption ratio (SAR) were analyzed (Li et al., 2013). All determined groundwater concentrations used in these computations were in the unit of mg/L. The chemical quality of water is very important for agricultural use, and the quantity of agricultural products depends on the quality of water used in irrigation. From the agricultural perspective, the use of groundwater with high concentration of salts can cause soil salinity and increase the exchangeable sodium content. Particularly in arid and desert areas due to lack of rainfall and lack of proper soil leaching, this process is intensified. In the classification of irrigation water, in addition to determining the chemical properties of water, various factors such as soil gender, soil condition, irrigation water content, ambient temperature, chemical elements present in the soil and the type of cultivated plant should be studied.
One of the most common classifications of irrigation water is the classification based on the Wilcox diagram, presented by the US Department of Agriculture (Wilcox, 1955). The amount of sodium sucrose with calcium and magnesium can be estimated by the ratio of sodium absorption (SAR). The SAR value is calculated from the following equation : In this classification, two chemical agents (SAR) and electrical conductivity (salinity risk) are considered, each of which is divided into four classes and generates a total of 16 different classes. C1S1 water is the best water for agriculture and C4S4 is the most unfavorable water for agriculture applications. In this form, increasing the amount of electrical conductivity, the effect of sodium becomes more severe. The various groups listed in the Wilcox classification 23 are classified as follows (in this classification, C represents salinity and S represents the amount of sodium): (i) Very good waters with an EC of less than 250 micrometers per cm and placed in the C1S1 class; (ii) Good waters that belong to one of the C1S2, C2S1, or C2S2 classes; (iii) Moderate waters that belong to one of the C1S3, C2S3, C3S1, C3S2, and C3S3 classes and are suitable for irrigation of coarse texture and good drainage; and (iv) Inappropriate water, which is located in the C1S4, C2S4, C3S4, C4S4, C4S1, C4S2, and C4S3 classes, and the larger their coefficients, become inappropriate. For classification of groundwater in terms of agricultural uses during 2012 and 2018, the Wilcox diagram was used (Fig. 10).
[ Fig. 10, here] As shown in Fig. 10, many samples are placed in the C3S1, C4S1, and C4S2 classes for both periods. The number of very few samples in the C2S1 class, or the very good category, has been used for agriculture purposes. The percentage of each Wilcox classification for agricultural consumption in Azarshahr plain is given in Table 5. Based on the above tables, it can be concluded that, as time passes, the percentage of samples in the good class of C2S1 has been reduced, which could also be a further indicator of increasing salinity in the aquifer. Also, in Tables 6 and 7, the classification of each sampling station for June 2012 and 2018 from the perspective of agricultural application is presented, respectively.

EC vs Cl -/Cl -+ HCO 3molar ratio
There is less evidence that the anthropogenic activities significantly would influence solute concentrations in the aquifer since according to Fig. 11, most water samples fell within the Gibbs diagram (Kou et al., 2019). Rather, most hydrochemical parameters fell within the fields 24 defined by Na + / (Na + + Ca 2+ ) or Cl − / (Cl − + HCO3 − ) end members, which suggests that the aquifer chemistry is predominantly affected by rock weathering (Madhav et al., 2018).
[ Fig. 11, here] Na + /Cl − versus Cl − and salinization sources The relationship between Na + /Cl − molar ration versus Cl − is helpful in identifying the salinization sources in Aquifers. If the ration is approximately equal to unity then most probably the source of these ions is halite dissolution (Kumar et al., 2009). If the ratio is greater than the unity, it is typically interpreted as Na released from a silicate weathering reaction. A ratio less than unity indicates that the addition of Cl − , is due to water level rise which causes more salt dissolution from the soil or mixing of wastewater with ground water. As shown in Fig. 12a, while a portion of the wells fall below one in Na + /Cl − ratio, another portion of them fall over one Na + /Cl − ratio making it hard to clearly understand the source of the salinization. In comparison with these countries, Iran is less severely affected by groundwater salinity pollution (Table 8).

Degree of salinization
The relationship between HCO3 − /Cl − ratio with Cl − for the majority of wells (as a whole), exhibited a pronounced negative slope indicating the reverse relationship between carbonate (represents freshening) and Cl − (represents salinization) (Fig. 12b). Based on the inverse of Simpson's ratio, the influence of saline water can be classified into six classes. In general, a HCO3 − /Cl − ratio N2 indicates freshwater recharge whereas a ratio less than unity indicates higher degrees (or potential) of salinization (Chaudhuri and Srinivasulu, 2014;October et al., 2013).

Mg 2+ /Ca 2+ versus Cl − and salinization sources
In groundwater, Mg 2+ /Ca 2+ ratio is usually less than unity while in sea water the Mg 2+ /Ca 2+ ratio greater than one. Thus, higher ratios of Mg 2+ /Ca 2+ can be found in seawater intrusion zones and similarly in areas experiencing of dolomite dissolution. It should be noted that 81.3% of the wells exhibit Mg 2+ /Ca 2+ ratios N1, indicating that seawater intrusion, saltwater up-coning, reverse cation exchange, dolomite dissolution processes or sewage invasion may play a significant role on the groundwater chemistry signatures in the study area. Based on a study by Mayo and Loucks (1995), a distinction between the areas of calcite and dolomite dissolution can be identified by calculating the Mg 2+ /Ca 2+ molar ratios of these cations. If the molar ratio is less than one, the dissolution of calcite has greater contribution to the water chemistry, whereas greater prevalence of dolomite dissolution occurs, when the Mg 2+ /Ca 2+ molar ratio is equal or N1. Katz et al. (1997) describe that lower molar ratios (i.e. less than 0.5) indicate silicate weathering sources and/or processes are occurring. As shown in Fig. 13, only one well exhibited Mg 2+ /Ca 2+ a ratio lower than 0.5, which reveal that negligible silicate-weathering processes were occurring in the study area. The average of Mg 2+ /Ca 2+ molar ratio for wells within all clusters are N1, with the exception of cluster 6. This result reveals that calcite dissolution is occurring (and predominant) in cluster 6 (under the direct impact of rainfall), while the other clusters are impacted by seawater and saltwater intrusion (i.e. clusters 3 and 5) or influenced by dolomite dissolution (i.e. clusters 1, 2, 4, and 7). It can be concluded that dolomite dissolution is the predominant hydrogeochemical process in comparing with calcite dissolution in the study area.
Relative to the wells of cluster 5which located close to the shoreline, the wells of cluster 3 shows lower Mg 2+ /Ca 2+ molar ratio and lower Cl − concentration. This is a chemical signature that can be used to distinguish between seawater intrusion area and saltwater up-coning area. Meanwhile, the higher Mg 2+ /Ca 2+ ratio and lower Cl − concentration are characterized the wells of cluster 2 in the NE part of the study area when comparing them with the wells of cluster 1 in the SE part.
The higher Mg 2+ concentration in the north can be attributed to the dissolution of carbonate matrix of sandstone which is rich in Mg 2+ (Gavish and Friedman, 1969). This dissolution is a result of acidic medium of the nitrate pollution which is higher in the eastern side of the northern area. Moreover, among the seven clusters, the Mg 2+ /Ca 2+ ratio is recorded for the wells of cluster 2 then for clusters 1 and 5. This is a clear indication that the eastern clusters 1 and 2 were influenced by Eocene brackish water which is characterized by Na-Mg-Cl water type (Vengosh et al., 2005).

Conclusions
An assessment over the quality and hydrochemistry of an aquifer adjacent to a hypersaline lake has been presented. The north-west of Iran is the host for the largest hypersaline lake in 28 Western Asia, namely Urmia Lake. The results significantly highlighted the formulated hypothesis that the groundwater is suitable for irrigation and drinking applications as well as the physico-chemical analysis of groundwater samples revealed that the majority of water samples during 2012 were found to be within acceptable limits as prescribed by WHO. However, as per total hardness classification, groundwater samples of the majority (> 90%) of the sites fall in the category of absolutely hard water ranging from moderately hard to very hard. The results for chemical analysis of groundwater samples represented that the mean concentration for ionic dominance pattern is in the order of Ca 2+ > Na + > Mg 2+ > K + for cations and HCO3 − > SO4 2− > Cl − for anions during 2012, whereas the respective pattern for 2018 is calcium > Mg 2+ > Na + > K + for cations and Cl − > HCO3 − > SO4 2− for anions. Based on obtained values of EC, SAR, Na%, and the salinity diagram, it can be concluded that most of the sampling points in the study area are appropriate for irrigation purposes both in 2012 and 2018. According to the hydrochemical and statistical evaluations, the following points can be drawn: (1) The influence of saline water on Azarshahr plain, especially in western parts, is clearly visible. As can be seen in almost all stations in the western part (Qeshlagh region), a significant increase in electrical conductivity is observed.
(2) The thinner western areas of the plain compared to the eastern areas can increase groundwater retention in these areas and ultimately lead to increased solute dissolution and increased salinity. 29 Despite these promising results, it is recommended that some future research would be required to provide new insights in this area. Especially, in terms of new methods for aquifer vulnerability assessment, the use of methods such as Susceptibility Index-Contamination Degree (SICODE), soil contamination degree index (CD), combination of DRASTIC, Goldberg, PI, and Susceptibility Index with hydrological and meteorological data will provide deeper understanding of the leveraging parameters in aquifer and groundwater sustainability practices.
Although this study was conducted based on a comprehensive data gathered in the field, it took into consideration the most important elements active in aquifers which are limited to several compounds. In this regard, focusing on the more active elements and considering them in the analysis of water quality judgments shall be an interesting idea for upcoming future research on this topic. Some of the potential elements for consideration would be heavy metals, radioactive, and rare earth elements.

References
Groundwater quality degradation of an aquifer in Iran central desert.