Introduction

Groundwater, especially in arid and semi-arid regions, has become the most important water source for drinking, irrigation, industry, and all other sectors (Kolsi et al. 2013; Singh et al. 2010; Solangi et al. 2019). Many areas in Central Asian, such as Iran (Dehbandi et al. 2018; Keshavarzi et al. 2011), Pakistan (Naseem et al. 2018; Solangi et al. 2019; Tabassum et al. 2019), Afghanistan (Ali et al. 2020; Hayat and Baba 2017), and northwest China (Huang and Pang 2012; Li et al. 2013; Lin et al. 2018) have been affected by degraded of groundwater quality, which has been preventing regional economic development. The quality of groundwater is deteriorating by the natural processes (e.g., evapotranspiration, mineral dissolution and recharge water quality) and anthropogenic activities (e.g., agricultural irrigation input, infiltration of domestic sewage and industrial wastewater and change of land-use and land-cover patterns) (Bhakar and Singh 2018; Bouzourra et al. 2015; Schwarzenbach et al. 2010). Groundwater quality of confined aquifer in the Yinchuan region, northwest China, was controlled mainly by the dissolution of minerals, mixing between the confined groundwater and polluted unconfined groundwater, and effects of ion exchange (Zhang et al. 2016). In addition, over-abstraction of groundwater for agricultural irrigation was causing the depletion and deterioration of aquifers around the world, and over-fertilizer was also degraded water quality (Pulido-Bosch et al. 2018). Therefore, the groundwater scientists have focused on assessing groundwater quality and its suitability for drinking, domestic, irrigation, and industrial uses (Adimalla et al. 2020; Bouteraa et al. 2019; Islam et al. 2017a, 2020b; Saha et al. 2020).

Groundwater quality evaluation is essential for managed groundwater development (Egbi et al. 2018). Water quality index (WQI) is an effective method to assess groundwater quality (Bouteraa et al. 2019; Rabeiy 2018; Solangi et al. 2019). The WQI integrated each parameter as well as many qualitative parameters into a single value using the drinking water quality standards proposed by the WHO (2011). A weight of each groundwater quality parameter was assigned according to their significance in the overall water quality. Based on their practical experience, many researchers (Alghamdi et al. 2020; Bhuiyan et al. 2016; Islam et al. 2020b; Reyes-Toscano et al. 2020; Sethy et al. 2017) assigned a weight between 1 and 5 to each indicator, which was used in calculating the relative weight using weighted arithmetic index method. However, weights were determined by experts’ experience might lead to over- or under-emphasizing some parameters, thus affecting the results (Maskooni et al. 2020). To reduce the errors while subjectively selecting the weights, WQI based on entropy weight (Adimalla et al. 2020; Islam et al. 2017a, 2020a; Maskooni et al. 2020) was applied for assessing the groundwater quality in the plain area of the Yarkant River Basin in Xinjiang, China.

The plain area of the Yarkant River Basin has an irrigated area of approximately 500,000 hm2, the largest irrigated areas in Xinjiang, northwest China. Groundwater is the main water source for agricultural and domestic purposes in the Yarkant River Basin. With rapid economic development, deteriorating groundwater quality has become a major constraint to sustainable socio-economic development and environmental protection in the region (Luan et al. 2017). It was very essential to deeply understand the factors influencing groundwater quality and its spatial variability for decision-making in any particular region (Islam et al. 2017a). However, the research on the assessment of groundwater quality for drinking purposes in the Yarkant River Basin was very limited. The characterization of groundwater quality and its spatial variation in the Yarkant River Basin using comprehensive evaluation and geostatistical approaches were yet to be conducted. In this study, WQI, statistics, geostatistics, ionic ratios, and geochemical equilibrium modelling were used to assess the groundwater quality for drinking purposes and analyze the spatial distribution characteristics and its factors of groundwater quality in this region.

Materials and methods

Study area

The Yarkant River Basin is located in the southern part of Xinjiang, northwest China, and can be divided into two units: southern mountainous and northern plain areas. The plain area (37° 22′ to 40° 29′ N, 76° 38′ to 80° 45′ E) is located in the middle of northern piedmont of the Kunlun Mountains, with Jiashi and Yopurga counties in the Kashi Prefecture and Taklimakan Desert situated in the west and east, respectively. The study area is shown in Fig. 1. The plain area of the Yarkant River Basin is a typical dry continental climate (Chang et al. 2016). The average annual temperature is 11.9 °C, annual precipitation is 52.7 mm, and annual potential evaporation is 2454 mm in this area.

Fig. 1
figure 1

Location map of groundwater sampling points situated in the study area

Full size image

The hydrogeological conditions of the study area are shown in Fig. 2. Groundwater recharge is mainly produced from the seepage of rivers and reservoirs, and irrigation water infiltration through permeable exposures (Zhang et al. 2019). Groundwater flows mainly along rivers from south to north. Evaporation and transpiration, spring drainage, and pumping are the main groundwater outputs. Piedmont plain area is mainly distributed in the north of Yecheng to Shache county, which is a single structure phreatic aquifer (SSPA). This area has good permeability and water conductivity, large hydraulic gradient, and smooth groundwater runoff, with highly thick medium coarse, medium fine, and fine sands (Kang et al. 2016). Alluvial plain is located in the middle of the study area, including Markit county and Bachu county, and the aquifer is mainly composed of multilayered structure phreatic and confined aquifers. This area has poor permeability and water conductivity, gentle hydraulic slope, and slow groundwater runoff (Wu et al. 2008). Clay layers situated in the middle of the study area are usually found at the depths of approximately 10 to 20 m below the land surface and separate the above multilayered structure phreatic aquifer (MSPA) from below multilayered structure confined aquifer (MSCA). Deposits in the confined aquifer consist of a sequence of fine, silty, and clayey sands up to 20 m thick, consisting of a fine-grained silty-clayey matrix.

Fig. 2
figure 2

Regional hydrogeological map of the plain area in the Yarkant River Basin

Full size image

Sample collection and analysis

In August–September 2018, 97 shallow groundwater samples (at the sampling depths of less than 100m) were collected from the study area (Fig. 1), including 14 from SSPA, 38 from MSPA, and 45 from MSCA. The total dissolved solids (TDS), total hardness (TH), dissolved oxygen (DO), pH, K+, Na+, Ca2+, Mg2+, Cl−, SO42−, HCO3−, NO3−, and F− concentrations were measured at the water quality laboratory of the NO.2 Hydrogeological and Engineering Geological Team of Xinjiang Bureau of Geology and Mineral Resources Exploration and Development. Concentrations of K+ and Na+ were determined using flame atomic absorption spectrophotometry. TH, Ca2+, Mg2+, and HCO3− concentrations were determined using ethylenediamine disodium tetraacetic acid titration method. TDS, Cl−, SO42−, NO3− and F− concentrations were measured by 105 °C drying-gravimetric method, silver nitrate volumetric method, barium sulfate turbidimetry, spectrophotometric method, and ion-selective electrodes method (PHS-3D), respectively. Concentrations of NO3− and other chemical constituents mentioned above were determined with the detection limits of 0.05 and 0.01 mg·L−1. For all samples, the ionic balance error was considered to be within the acceptable limit of ± 5%.

Water quality index and entropy weight method

To compute the WQI based on entropy weight, three steps have been followed (Adimalla et al. 2020; Maskooni et al. 2020). In the first step, entropy weight is assigned to each parameter. Entropy weight can be calculated by constructing a matrix. Considering that m (i = 1, 2, … , m) samples were monitored for water quality, with n parameters (j = 1, 2, … , n) being analyzed for each sample, then, xij represents the measured index value of parameter j in sample i. Matrix X is subsequently prepared as follows:

$$ \mathrm{X}=\left[\begin{array}{c}{x}_{11}\kern0.5em {x}_{12}\cdots \kern0.5em {x}_{1n}\\ {}{x}_{21}\kern0.5em {x}_{22}\cdots \kern0.5em {x}_{2n}\\ {}\vdots \kern1.5em \vdots \kern1.75em \vdots \kern1.25em \vdots \\ {}{x}_{m1}\kern0.5em {x}_{m2}\cdots {x}_{mn}\end{array}\right] $$
(1)

Owing to different units of different parameters and quantity grades, matrix X is normalized as matrix R (rij). The ratio of the index value of parameter j in sample i is calculated as follows:

$$ {f}_{ij}=\frac{r_{ij}}{\sum_{i=1}^m{r}_{ij}} $$
(2)

The information entropy of parameter j is expressed as follows:

$$ {h}_j=-\frac{1}{\ln m}{\sum}_{i=1}^m{f}_{ij}\ln {f}_{ij} $$
(3)

where hj is the entropy value of parameter j. The entropy weight of each parameter can then be calculated as follows:

$$ {w}_j=\frac{1-{h}_j}{n-{\sum}_{j=1}^n{h}_j} $$
(4)

where wj is the entropy weight of parameter j.

In the second step, a quality rating scale (qij) is assigned for each parameter as follows:

$$ {q}_{ij=\frac{x_{ij}}{s_j}\times 100} $$
(5)

where Sj is the standard value of parameter j as per the WHO (2011) guideline (Table 1).

Table 1 Statistics of hydrochemistry parameters of groundwater in single structure phreatic aquifer (SSPA), multilayered structure phreatic aquifer (MSPA), multilayered structure confined aquifer (MSCA), and all samples (All) of the Yarkant River Basin, relative weight of the parameters, and World Health Organization (WHO) water quality standards
Full size table

In the third step, the WQI of sample i is calculated as follows:

$$ {\mathrm{WQI}}_i={\sum}_{j=1}^n{w}_j\times {q}_{ij} $$
(6)

The WQI values can be categorized into five classes (Alghamdi et al. 2020), as shown in Table 2.

Table 2 Water quality index (WQI) classification
Full size table

Geostatistical analysis

The semivariogram models and ordinary kriging are applied for spatial distribution of groundwater quality parameter by ArcGIS software. The semivariogram was the main tool in geostatistics that expresses spatial correlation between adjacent observations (Arslan 2012; Rakib et al. 2020). Semivariogram model was important for structural analysis and spatial interpolation. Assuming the normality of groundwater quality variables is crucial for obtaining reliable results in parametric statistical tests (Machiwal et al. 2018). In this study, Kolmogorov-Smirnov normality test was used to examine the normal distribution of variables. Afterwards, semivariogram model was used to identify the best predictive model. The semivariogram is calculated as follows:

$$ {\gamma}_h=\frac{1}{2n}{\sum}_{i=1}^n{\left[z\left({x}_i\right)-z\left({x}_i+h\right)\right]}^2 $$
(7)

where represents one-half of the variance of difference between spatially distributed data points separated by distance h; z(xi) represents the value of variable at location xi; z(xi+h) represents the value of other points separated from xi by distance h; and n represents the number of sampled points used, separated by distance h.

Because of its easy calculation and prediction accuracy compared to the other kriging methods (Bhuiyan et al. 2016), ordinary kriging was considered to be an exact interpolator with minimum mean error for finding the best linear unbiased estimate. Interpolation acceptability criteria to ensure unbiased estimates were evaluated by cross-validation. The mean standardized error (MSE) and root mean square standardized error (RMSSE) values of the model approach to 0 and 1 respectively, which indicates that the model performance is fit (Kaur and Rishi 2018).

Hydrogeochemistry

Saturation index (SI) can be used to identify the geochemical reactions that control water-rock interaction (Ferchichi et al. 2017). SI was calculated as follows:

$$ SI=\log \left(\frac{IAP}{KT}\right) $$
(8)

where IAP represents the ionic activity product of the water samples and KT represents the equilibrium constant at the sample temperature. SI = 0 denotes the solubility equilibrium of mineral phase in the groundwater, SI < 0 indicates the undersaturation of groundwater with this mineral, and SI > 0 indicates the supersaturated of groundwater with the mineral and its inability to dissolve other minerals (Bouderbala and Gharbi 2017).

Sodium adsorption ratio (SAR) can reflect the cation exchange between Na+ in groundwater and Ca2+ and Mg2+ in aquifers (Ferchichi et al. 2017). The larger the SAR, the more obvious the cation exchange. SAR was calculated as follows:

$$ SAR=\frac{K^{+}+{Na}^{+}}{\sqrt{\frac{Mg^{2+}+{Ca}^{2+}}{2}}} $$
(9)

where Na+, K+, Ca2+, and Mg2+ are in meq·L−1.

Descriptive statistical analysis (range, median and mean) and comparison with WHO standard of measured values were done by using Excel. The PHREEQC was used for calculating SI. The ion ratio diagram was drawn by Origin software. The data of land use in 2018 was obtained from the Data Center of Resources and Environmental Sciences, Chinese Academy of Sciences (http://www.resdc.cn), and the described in detail were shown in Ning et al. (2018).

Results and discussions

Hydrochemical characteristics of groundwater

Detailed statistics of hydrochemical parameters of the groundwater in the plain area of the Yarkant River Basin are shown in Table 1. The pH of the groundwater was slightly alkaline to neutral, ranging from 6.48 to 8.60 with a mean value of 7.60 in all groundwater samples (All). TDS in All varied from 468.00 to 9250.00 mg·L−1, with a mean value of 2821.87 mg·L−1. TH ranged from 314.85 to 3840.00 mg·L−1, with a mean value of 1259.56 mg·L−1 in All. The mean values of TDS and TH shown the trend of MSCA > SSPA > MSPA. The order followed by the mean value for cation in all aquifers was Na+ > Ca2+ > Mg2+ > K+ and for anions was SO42− > Cl− > HCO3− > NO3− > F−. Variances were observed in the groundwater chemistry types in different aquifers. Piper trilinear diagram (Fig. 3) shows that the dominant water types were SO4•HCO3-Ca•Na, SO4•HCO3-Ca•Mg, and SO4•Cl-Na•Ca in SSPA, SO4•Cl-Na•Mg, SO4•Cl•HCO3-Na•Ca, and SO4•HCO3•Cl-Ca•Na in MSPA, and of SO4•Cl-Na•Ca, Cl•SO4-Na•Ca, and HCO3•SO4-Mg•Ca in MSCA.

Fig. 3
figure 3

Piper trilinear diagram in single structure phreatic aquifer (SSPA), multilayered structure phreatic aquifer (MSPA), and multilayered structure confined aquifer (MSCA)

Full size image

Table 3 shows the number and proportion of parameters greater than the acceptable values for drinking water as per WHO standards. The largest exceedance was observed for HCO3− in SSPA (85.71%), followed by TH and SO42− (57.14% proportion for both), while Mg2+, F−, pH, and DO were not found to exceed the limit. The largest exceedance was observed for HCO3− (100.00%) in MSPA, followed by SO42− (94.74%). The smallest exceedance rate was observed for pH (2.63%). Furthermore, the largest exceedance was observed for HCO3− (95.11%) in MSCA, followed by SO42− (91.11%), with pH and NO3− exhibiting the smallest exceedances (2.22%). The main parameter exceeding the acceptable value of drinking water in WHO standard were HCO3− , SO42− , TH, TDS, and Cl− (Table 3). Luan et al. (2017) showed that the groundwater parameters exceeding the acceptable value of the standards for drinking water quality of China were TH, TDS, Cl−, and SO42− in rural areas of Yarkant River Basin. Different from this study, HCO3− and cation (K+, Mg2+, Na+ and Ca2+) were not considered in Luan et al. (2017) according to the standards for drinking water quality of China.

Table 3 Number and rate of exceeding standard in single structure phreatic aquifer (SSPA), multilayered structure phreatic aquifer (MSPA), and multilayered structure confined aquifer (MSCA)
Full size table

Groundwater quality assessment

Entropy weight does not rely on subjective judgement for assigning weights, which can improve the WQI. The physico-chemical parameters with the highest entropy weight have the greatest impact on overall groundwater quality (Islam et al. 2017a). The entropy weights of NO3− was the largest, which had the highest effect on overall groundwater quality of the study area (Table 1). The effects of other parameters on overall groundwater quality decreased in the following order: DO > K+ > Cl− > Na+ > TH > TDS > SO42− > Ca2+ > Mg2+ > F− > HCO3− > pH. The pH has a minimal impact on overall groundwater quality of the study region. The calculation results indicated that the ranges of WQI of all samples was 31.79 to 549.37 with an average of 158.50. The groundwater quality varied from excellent to extremely poor. In total, 31 samples were classified as good quality, constituting the largest proportion (31.96%) (Fig. 4), and 22 (22.68%), 18 (18.56%) and 14 (14.43%) samples were classified as poor, medium and excellent quality, respectively. Extremely poor quality constituted the least proportion of 12.37% (12 samples) in All. The ranges of WQI of SSPA were 31.79 to 291.43, with an average of 89.74. Excellent quality constituted the largest proportion (50.00%) in SSPA, followed by good quality (35.72%). The WQI of MSPA ranged from 36.69 to 321.17, with an average of 163.01. The proportion of good water in MSPA was the highest (34.21%), followed by poor quality (28.95%). The ranges of WQI of MSCA was 36.63 to 549.37, with an average of 179.83. Good quality constituted the highest proportion in MSCA (28.89%), with a non-negligible proportion (20.00%) of extremely poor quality.

Fig. 4
figure 4

Proportion of different levels of groundwater quality in single structure phreatic aquifer (SSPA), multilayered structure phreatic aquifer (MSPA), multilayered structure confined aquifer (MSCA), and all samples (All)

Full size image

The WQI values of phreatic aquifer (PA, including SSPA and MSPA) and MSCA followed normal distribution after logarithmic-transformation by the Kolmogorov-Smirnov test. Exponential and Gaussian models were applied for the WQI of PA and MSCA, respectively. The nugget, sill, and the range values of the model are shown in Table 4. The R2 of the model was 0.737 and 0.791 (approximately 1) for PA and MSCA, and RSS was 0.002 and 0.005 (approximately 0), respectively. The results indicated that the fit semivariogram represents very well the spatial structure of these variables in the groundwater. According to nugget to sill ratio, the spatial dependence of groundwater quality parameters can be classified as strong spatial dependence with ratio less than 0.25, moderate spatial dependence with the ratio of 0.25 to 0.75, and weak spatial dependence with the ratio of more than 0.75 (Arslan 2012; Islam et al. 2017b). Nugget to sill ratio of the WQI values in PA and MSCA were 0.078 and 0.198, respectively, which indicated strong spatial dependence (Table 4), which might be due to the low-flow conditions in major rivers, prevailing aquifer geology, and long-term geogenic processes, such as groundwater source rock, rainfall, and infiltration processes.

Table 4 Semi-variation of water quality index (WQI) of phreatic aquifer (PA, including single structure phreatic aquifer, SSPA and multilayered structure phreatic aquifer, MSPA) and multilayered structure confined aquifer (MSCA)
Full size table

During cross-validation of ordinary kriging interpolation, the MSE was found to be −0.062 and −0.053, and RMSSE was found to be 1.098 and 1.054 in the PA and MSCA, respectively, indicating the accuracy of predictions. The spatial distribution of WQI in southern of PA was mainly good and excellent water, and medium water in the north was the largest, followed by poor water (Fig. 5a). As groundwater flows from the relatively high south areas toward the north direction, the quality of groundwater degraded from south to north in the PA. The distribution of WQI exhibited excellent, good, and an alternation of medium and poor quality (including extremely poor quality) from south to north in the MSCA (Fig. 5b). The MSCA towards north was limited by the surface water supply, slow groundwater runoff, and strong evaporation and concentration, which resulted in a larger WQI. The relatively lower WQI in the southern part of the MSCA was related to its recharge by the PA. The spatial distribution characteristics of WQI in PA and MSCA were consistent with those of Cl−, SO42− and TDS (Zhang et al. 2019).

Fig. 5
figure 5

Spatial distribution of water quality index (WQI) in a phreatic aquifer (including single structure phreatic aquifer, SSPA and multilayered structure phreatic aquifer, MSPA) and b multilayered structure confined aquifer (MSCA)

Full size image

Evolution process of groundwater quality

Pearson correlation analysis of 13 hydrochemical parameters was used to explain the influencing factors of groundwater quality. The absolute values of correlation coefficient were 0.30–0.50, 0.50–0.75, and > 0.75 which denoted weak, moderate, strong correlation, respectively (Islam et al. 2017a). A positive significant correlation indicated the same source, either natural or anthropogenic sources. The correlation coefficient was found to be the highest between Cl− and Na+, and a highly positive correlation was observed between SO42− and Mg2+, Ca2+, and Na+ (Fig. 6), indicating that Na+, Ca2+, Mg2+, Cl−, and SO42− may have common sources. The pH exhibited a negative correlation with HCO3− and F−, indicating the favorability of acid environment for HCO3− and F− enrichment. No obvious correlation was observed between the chemical constituents mentioned above and NO3−, indicating that NO3− may have originated from anthropogenic activities.

Fig. 6
figure 6

Correlation coefficients of parameters of the shallow groundwater

Full size image

With the extension of the river, the groundwater quality of PA within 10 km of the river gradually degenerated from excellent quality to good, medium, and poor quality (Fig. 7). The further away from the river, the groundwater quality of SSPA gradually degenerated from excellent quality to good and poor quality on western bank of the river and from excellent quality to good and medium quality on eastern. The groundwater quality of MSPA on western bank of the river gradually degenerated from good quality to poor quality at 120 km and from poor quality to extremely poor quality at 300 km. The WQI value in the south of PA was relatively small, especially SSPA, which might be related to the rapid flow of groundwater and the recharge of surface water with low concentration (Wu et al. 2008). The slow flow of groundwater in the northern low plain might increase the residence time of rock-water interactions and also increase the concentrations of groundwater ions (Kang et al. 2016). In addition to the hydraulic connection with SSPA in the south, the surface water supply to MSCA was limited. As the MSCA was situated in a more closed environment with longer durations of water rock action, the WQI in MSCA was found to be mostly higher than that in PA.

Fig. 7
figure 7

The groundwater quality of phreatic aquifer (including single structure phreatic aquifer, SSPA and multilayered structure phreatic aquifer, MSPA) along the Yarkant River

Full size image

In aquifers, the dissolution process of soluble mineral is main natural contributor to groundwater chemical composition (Abu-Alnaeem et al. 2018). The relationship between SI and WQI is shown in Fig. 8. The SI of dolomite and calcite were observed to be less than zero for excellent and good quality, indicating that the mineral dissolution of dolomite and calcite was the main hydrochemical process (Fig. 8a and b). In the medium, poor, and excellently poor quality, the SI of dolomite and calcite were greater than zero, indicating that dolomite and calcite had reached a saturated state, with a weak mineral dissolution. Therefore, the weak dissolution of calcite or dolomite could have likely occurred, which could be an important factor to control Ca2+ and Mg2+ concentrations in the groundwater with high WQI value. Furthermore, the concentration of Ca2+ (mean 269.15 mg·L-1) exceeds that of Mg2+ (mean 139.03 mg·L-1), indicating that the dissolution of calcite may be a dominant factor governing the groundwater chemistry (Islam et al. 2017a). The SI of halite, gypsum, anhydrite, and fluorite were less than zero and increased with increasing WQI (Fig. 8c, d, e, and f), indicating that Na+, Cl−, SO42-, and F− concentrations might have been controlled by the dissolution of evaporite minerals, such as halite, gypsum, and anhydrite.

Fig. 8
figure 8

Saturation index of calcite (a), dolomite (b), halite (c), gypsum (d), anhydrite (e), and fluorite (f) minerals

Full size image

Relationship between specific ions can indicate the main hydrochemical processes in groundwater. Most medium, poor, and extremely poor quality were located above the congruent dissolution line of salt rock (Fig. 9a), indicating that halite dissolution did not contribute to the dominance of Na+ over Cl− in the plot (El Alfy et al. 2018). Na+ might have come from cation exchange in addition to the dissolution of salt rock. Samples with end-members 0 (k = −1) in the scatter plot of (Ca2+ + Mg2+) − (HCO3− + SO42−) versus Na+ + K+ − Cl− represent ions originating from mineral dissolution (Argamasilla et al. 2017; Chegbeleh et al. 2020; Saha et al. 2020). The samples were plotted away from end-members 0, indicating that Ca2+ and/or Mg2+ were derived from processes not related with carbonate or gypsum dissolution (Fig. 9b). A plot with these two indexes was enriched in Na+ + K+ − Cl−, while (Ca2+ + Mg2+) − (HCO3− + SO42−) was observed to decrease, with a significantly deviating slope from mineral congruent dissolution line (y = −x), confirming the likelihood of ion exchange domination between Ca2+, Mg2+, and Na+ in the groundwater system.

Fig. 9
figure 9

Biplot of a Na+ versus Cl- and b (Ca2+ + Mg2+) – (HCO3- + SO42-) versus Na+ + K+ – Cl- in the hydrochemistry of the study area

Full size image

The SAR of the all groundwater ranged from 0.29 to 23.91, with an average value of 5.27. Figure 10 shows that the SAR was positively correlated with WQI (R2 = 0.703), implying that the higher the WQI, the greater its SAR value and the more obvious the ion exchange effect. Most groundwater samples in the poor (95.83%) and extremely poor quality (100.00%) were located in MSPA or MSCA. The slow groundwater runoff, long water-rock interaction time, and fine sediment particles in the confined aquifer more obviously alternated the adsorption between Na+, Ca2+, and Mg2+. Ca2+ in the groundwater replaced a part of Na+ adsorbed in the water-bearing medium, thus increasing Na+ in groundwater and deteriorating the groundwater quality, which were the same as the groundwater in the Tarim River Basin (Xiao et al. 2014).

Fig. 10
figure 10

Relationship between sodium adsorption ratio (SAR) and water quality index (WQI)

Full size image

Natural processes play a dominant role in deteriorating groundwater quality, and the impact of human activities cannot be ignored. In recent decades, the increase of population and agricultural production had aggravated the environmental and water quality degradation of the region (Xiao et al. 2014). NO3− in some samples exceeded the standard value and their WQIs were found to be medium, poor, or extremely poor quality (Fig. 11a). The mean NO3− concentrations in SSPA and MSPA were higher than those in MSCA (Table 1), indicating the susceptibility of SSPA and MSPA to be anthropogenically contaminated. Poor and extremely poor groundwater quality were mainly located in the central and northern cultivated and built-up land (Fig. 11b). Fertilizer infiltration with irrigation water, untreated domestic sewage, and industrial effluents pollute groundwater along the groundwater channels, thus deteriorating groundwater quality (Zhang et al. 2016). As SSPA and MSPA is situated closer to the surface, it is more directly affected by man-made pollution than the MSCA. Although anthropogenic contamination has less impact on MSCA, some abandoned wells have damaged the aquifers, and contaminated of MSPA along the walls can deteriorate confined aquifer (Lin et al. 2017).

Fig. 11
figure 11

a Relationship between NO3- and water quality index (WQI) and b the distribution of groundwater quality in land use types

Full size image

Conclusions

Groundwater was found to be slightly alkaline with high TDS and TH in the plain area of the Yarkant River Basin. The dominant water types were found to be SO4•HCO3-Ca•Na, SO4•HCO3-Ca•Mg, and SO4•Cl-Na•Ca in SSPA, SO4•Cl-Na•Mg, SO4•Cl•HCO3-Na•Ca, and SO4•HCO3•Cl-Ca•Na in MSPA, and SO4•Cl-Na•Ca, Cl•SO4-Na•Ca, and HCO3•SO4-Mg•Ca in MSCA. In all, the proportion that concentration of HCO3− more than the standard values for drinking water as per the WHO guidelines was largest, followed by SO42−. Comprehensive evaluation value (WQI) (ranged between 31.79 and 549.37, with an average of 158.50) indicated that the groundwater quality varied from excellent to extremely poor. The largest proportion was of good quality (31.96%), followed by poor quality (22.68%). The proportions of medium, excellent, and extremely poor quality were 18.56%, 14.43%, and 12.37%, respectively. Nugget to sill ratio of the WQI values in PA (including SSPA and MSPA) and MSCA were 0.078 and 0.198, respectively, which indicated strong spatial dependence. The slow groundwater runoff, long water-rock interaction time, and fine sediment particles, the groundwater quality degraded from south to north in PA and MSCA. The deterioration of water quality may be controlled by the dissolution of evaporite minerals, such as halite, gypsum, and anhydrite and ion exchange process. In addition, the local effects of anthropogenic pollution on groundwater quality cannot be ignored. For sustainable development, the monitoring and management of groundwater quality for drinking and irrigation was needed in this area. The scale of groundwater exploitation should be controlled and the permission system of groundwater intake should be strictly implemented. The poor quality or abandoned wells should be treated in time by means of sealing, cementing, and backfilling to prevent the groundwater quality of MSPA from the MSCA.