Hydrochemical Characteristics and Water Quality Assessment for the Upper Reaches of Syr Darya River in Aral Sea Basin, Central Asia

: Based on water sampling of the upper reaches of the Syr River and its tributaries from the parts of Aral Sea Basin in Kyrgyzstan, the chemical compositions of river waters were systematically analyzed for revealing the hydrochemical characteristics and evaluating the water quality. Research indicates that there are some di ﬀ erences in ion concentration between the low-ﬂow season (LFS) and high-ﬂow season (HFS), but the hydrochemical classiﬁcation reﬂected that all water samples fall in the calcium bicarbonate category, except that only three samples fall in the not dominant category during the LFS. The water quality classiﬁcation shows that the water samples fall in the excellent to good categories for irrigation use. The analysis shows that the main ions of river waters come from the weathering of rocks, and the dissolution of carbonates is higher than that of silicates. Human activities have had an impact on the waterbody, especially inferred from the indicators of NH 4 -N and fecal coliform (FC). FC groups were detected in some rivers, in which the detection rate at the high-water level increased. The contents of potentially toxic elements are lower than international drinking water standards, but there are clustering di ﬀ erences between the LFS and HFS. There may be anthropogenic intrusions of Cu, Pb, and Zn during the LFS period and of Cu, Pb, Zn, and Cd during the HFS period. The results ﬁll the gaps in the study of the hydrochemical composition and water quality assessment in the Aral Sea Basin and will also provide a basis for water resource management and for the study of water quality evolution in the future.


Introduction
The shrinkage of Aral Sea [1], related to changes in water storage [2,3], water quality [4], regional climate [5,6], and environmental conditions [7][8][9], have been hotspots of research attention in the past few decades. The Syr Darya River, which originates in the Tian Shan Mountains in Kyrgyzstan and is the second largest river in the Aral Sea Basin, has also received more research focus. Former studies focused on the source of water evaporation in the Syr Darya River [10,11], the changes in water volume of the Syr Darya River [12][13][14][15], the coping strategy for water resource [16,17], and the climate change effects [18][19][20]. The deteriorated water quality of lakes and rivers threatens the livelihood of local people, the economic development, and biological diversity of riverside

Geographic Background
The Syr Darya River is mainly supplied from glaciers and snowmelt at a length of 2500 km [12]. The 75% of the annual flow of Syr Darya (36.57 km 3 ) waters is formed in Kyrgyzstan [30]. The study area refers to parts of the Aral Sea Basin in Kyrgyzstan ( Figure 1). Kyrgyzstan has a total population of about 6.3 million (2018). Tian Shan mountain ranges covering over 80% of its surface area. Of the total land area, 56.2% is classified as agricultural land, of which 87% is pasture and only 7.3% (equal to 1.4 million ha) is arable. Three quarters (1.1 million ha) of this arable land are irrigated [31]. The Syr Darya Rivers in Kyrgyzstan (Naryn River) flows through the Fergana Valley into Uzbekistan. The peak flow in the study area occurs in July and August, and the dry season is from November to April [27]. The bedrocks in the study region are mainly composed of sedimentary carbonate rocks, siliciclastic rocks, acid plutonic rocks, mixed sedimentary rocks and unconsolidated sediments [32]. According to the soil classification in the Harmonized World Soil Database (v 1.2) [33], the soils in the study area are composed of Calcisols, Kastanozem, and Leptosol. The study area has a continental climate caused by its position between the temperate and sub-tropical zones [34]. The summer temperatures in southwestern Fergana Valley can reach up to 40 • C [34]. The yearly precipitation in Kyrgyzstan varies between 100 and 1000 mm and is distributed unevenly [34]. From the perspective of the interannual temperature distribution, the monthly average temperature curve showed a single-peak distribution, and the highest temperature occurred in July [35]. However, the monthly total precipitation curve shows a bimodal distribution pattern, and the two peaks appear in April and November [36,37].  (Table S1).

Data Processing Methods
To evaluate the irrigation suitability, a classification diagram (USSL diagram) was plotted by the electrical conductance EC and sodium absorption ratio (SAR) (Equation (1) [47]). The SAR reflects the risk of an alkali or sodium hazard to crops [48]. A high Na + concentration in irrigation water will result in negative influences on the soil structure resulting in poor internal drainage and restricted circulation of air and water when Na + is displaced by Mg 2+ and Ca 2+ and is adsorbed by clay soil particles [48]. If not indicated below, the ionic concentrations are expressed in milliequivalents per liter (meq L −1 ).

Data Processing Methods
To evaluate the irrigation suitability, a classification diagram (USSL diagram) was plotted by the electrical conductance EC and sodium absorption ratio (SAR) (Equation (1) [47]). The SAR reflects the risk of an alkali or sodium hazard to crops [48]. A high Na + concentration in irrigation water will result in negative influences on the soil structure resulting in poor internal drainage and restricted circulation of air and water when Na + is displaced by mg 2+ and Ca 2+ and is adsorbed by clay soil particles [48]. If not indicated below, the ionic concentrations are expressed in milliequivalents per liter (meq L −1 ).
A Wilcox diagram [49] was plotted by the EC and sodium percentage (Na%). The sodium percentage Na% was calculated as follows: A Piper diagram [50] is usually plotted to classify the major ion composition and to separate the hydrogeochemical facies [51][52][53][54]. The Gibbs diagram [23] reflects the ratios of cations [Na + /(Na + + Ca 2+ )] and anions [Cl -/(Cl -+ HCO 3 -)] against the TDS. Gibbs diagram can reveal three natural mechanisms (rock dominance, evaporation dominance, and precipitation dominance) that influence the hydrochemistry [55]. Mixing diagrams [56,57] can be used to reveal details about the processes of rock weathering. Hierarchical cluster analysis [58] and two-way cluster analysis [59] was used to examine the characteristics of ions and potentially toxic elements and to determine the similarity of the samples and the hydrochemical compositions. The colored cells were coded based on scales of Z-scores for the concentrations of ions and potentially toxic elements. The Z-score formula is: where, x is the original data, µ is the mean of all data, and σ is the standard deviation.

Results
As shown in Figure   μg L −1 during the HFS. The ΣCr value ranges from 0.01 to 0.08 mg L −1 with a median value of 0.03 mg L −1 during the LFS, and the ΣCr value ranges from 0.01 to 0.08 mg L −1 with a median value of 0.035 mg/L during the HFS. Cr 6+ ranges from 0.01 to 0.03 mg L −1 during the LFS and from 0.006 to 0.038 mg L −1 during the HFS.

Sources and Influencing Factors for the Major Ions and Potentially Toxic Elements
Natural processes, including precipitation, evaporation, and water-rock interactions, control the ionic composition of the river water samples [57]. Most of the river samples were plotted in the water-rock interaction array (Figure 3). The anions and cations in the river water samples from the upper reaches of Syr Darya River and its tributaries are mainly controlled by water-rock interactions. All water samples plot the arrays of carbonates and silicates (Figure 3), which indicates that the ionic compositions are mainly controlled by carbonate dissolution and silicate weathering. However, Figure 3 Water 2019, 11, 1893 7 of 16 shows that the water samples from the upper riches of the Syr Darya River had higher Ca/Na ratios, which suggests that the dissolution of carbonates is more notable than that of silicates [61].
Natural processes, including precipitation, evaporation, and water-rock interactions, control the ionic composition of the river water samples [57]. Most of the river samples were plotted in the water-rock interaction array (Figure 3). The anions and cations in the river water samples from the upper reaches of Syr Darya River and its tributaries are mainly controlled by water-rock interactions. All water samples plot the arrays of carbonates and silicates ( Figure  3), which indicates that the ionic compositions are mainly controlled by carbonate dissolution and silicate weathering. However, Figure 3 shows that the water samples from the upper riches of the Syr Darya River had higher Ca/Na ratios, which suggests that the dissolution of carbonates is more notable than that of silicates [61]. Almost all water samples fall in the calcium bicarbonate array (Figure 4). During the LFS, only three samples fall in the not dominant category of the lower parts of the ternary diagrams. Generally, there were no significant differences in hydrochemical facies classification between the water samples collected during the LFS and HFS (Figure 4). Because the water samples from the upper reaches of Syr Darya River and its tributaries mainly belong to the calcium bicarbonate category, carbonate dissolution may dominate the hydrochemical composition. If only carbonate dissolution occurs, the values of Ca 2+ /HCO 3 − and (Ca 2+ + mg 2+ )/HCO 3 − will be close to 0.5 [62]. In this area, the values of Ca 2+ /HCO 3 − and (Ca 2+ + mg 2+ )/HCO 3 − were higher than 0.5 ( Figure 5), which suggests that the excess Ca 2+ and mg 2+ were not only the result of carbonate dissolution but were also influenced by other natural sources, e.g., gypsum and the silicate mineral of anorthite [62].   If the Ca 2+ originates from gypsum dissolution, the ratio of Ca 2+ /SO 4 2− will be nearly equal to 1 [63]. However, in this study, the ratio of Ca 2+ /SO 4 2− is much greater than 1 ( Figure 5), suggesting that the excess Ca 2+ may be derived from the dissolution of carbonates or silicate minerals [64]. In addition, if the ion ratio of Ca 2+ /Mg 2+ was higher than 2, the natural process is the weathering of silicates [63]. Similarly, in this region, the value of the Ca 2+ /Mg 2+ ratio is much higher than 2, which suggests that the major Ca 2+ ion should also be influenced by the process of silicate weathering in addition to carbonate dissolution. If the Na + in the water samples come from halite dissolution, then the ratio of Na + /Cl − is about 1 [65]. However, in this study, the ratio of Na + /Cl − is much greater than 1, indicating that the excess Na + comes from the weathering of silicates due to the ion exchange process. The results are similar to studies in Kanding, southeastern China [63], the Yellow River source region, the northeastern Qinghai-Tibetan Plateau [56], and the results are different from studies in southern Germany [62], India [66] and China [67]. NH 4 -N was detected at each sampling point in different sampling periods, which reflected the influences of human activities. The concentration during the HFS period is lower than that during the LFS period, which reflects the dilution effect to a certain extent. Kyrgyzstan has large grazing livestock, which is the traditional vocation of the Kyrgyz people. The fecal contamination caused by livestock and humans will lead to increased risks of disease outbreaks, reflected by the fecal contamination [68]. In this study, FC bacteria were detected in four river samples (Isfairam-Sai (2), Jazy (2) Kok-Art (2) and Naryn (1) in the Supplementary Materials (Table S1)) among the 38 samples, and the FC values were the same at 2 MPN 100 mL −1 (most probable number per 100 mL) for the four river samples during the LFS. However, there are clear differences with those during the HFS period, and FC bacteria were detected in 16 river samples (Ala-Buga (1), Ak-Buura (1), Ak-Buura (2), Ak-Suu (1), Ak-Suu (2), At-Bashy (1), At-Bashy (2), Chon Naryn, Isfairam-Sai (1), Jumgal (2), Kichi-Naryn, Kurshab (1), Mailuu-Suu (2), Naryn (1), Soh (1), and Soh (2) in the Supplementary Materials (Table S1)), which indicates that a large amount of untreated feces flowed into rivers from pastures with surface runoff during the HFS period.  From the dendrogram of hierarchical clustering analysis (Figure 6a,b), during the HFS period (Figure 6b), Cd was close to the Zn, Cu, and Pb groups, which may reflect differences with that during the LFS (Figure 6a). The similarity between heavy metals did not reflect sources of heavy metals and their influencing factors. The dendrogram created by two-way analysis illustrates the apparent characteristics of the analytes and the spatial sample locations (Figure 6c,d). The Euclidean distance was used to reflect the measure of similarity between the monitoring points or the analytes. From the point of view of the clustering dendrogram between the heavy metals and major ions, there are also significant differences between the different periods. During the LFS period (Figure 6c), Zn, Cu, Pb exhibited significant differences from the main ions and the other potentially toxic elements. During the HFS period (Figure 6d), Zn, Cu, Pb and Cd had significant differences from the main ions. However, the other potentially toxic elements are close to the major ions. There may be anthropogenic intrusions of Cu, Pb, and Zn during the LFS period and of Cu, Pb, Zn, and Cd during the HFS period. The affinities between samples in the two-way clustering diagram were mainly determined by the relationships between ions and heavy metals. From the view of the sample clustering, the sampling points are quite different between the HFS and LFS, which reflects the complexity of the geographical background and hydrological conditions. From the dendrogram of hierarchical clustering analysis (Figure 6a,b), during the HFS period (Figure 6b), Cd was close to the Zn, Cu, and Pb groups, which may reflect differences with that during the LFS (Figure 6a). The similarity between heavy metals did not reflect sources of heavy metals and their influencing factors. The dendrogram created by two-way analysis illustrates the apparent characteristics of the analytes and the spatial sample locations (Figure 6c,d). The Euclidean distance was used to reflect the measure of similarity between the monitoring points or the analytes. From the point of view of the clustering dendrogram between the heavy metals and major ions, there are also significant differences between the different periods. During the LFS period (Figure 6c), Zn, Cu, Pb exhibited significant differences from the main ions and the other potentially toxic elements. During the HFS period (Figure 6d), Zn, Cu, Pb and Cd had significant differences from the main ions. However, the other potentially toxic elements are close to the major ions. There may be anthropogenic intrusions of Cu, Pb, and Zn during the LFS period and of Cu, Pb, Zn, and Cd during the HFS period. The affinities between samples in the two-way clustering diagram were mainly determined by the relationships between ions and heavy metals. From the view of the sample clustering, the sampling points are quite different between the HFS and LFS, which reflects the complexity of the geographical background and hydrological conditions.

Evaluation of the Irrigation Suitability
The concentrations of the potentially toxic elements were lower than the drinking water standards from the drinking water quality guidelines of the World Health Organization [69], the Council Directive 98/83/EC on the quality of water intended for human consumption, the European Union [70], the National Primary Drinking Water Regulations, the United States Environmental Protection Agency [71] and the China National Standards for drinking water quality (GB5749-2006) [72], which indicated that the water quality of the water from the parts of Aral Sea Basin in Kyrgyzstan is good (Table 1). The river water of the Syr Darya River is used extensively for irrigation [73], and the upstream water quality has an important impact on downstream irrigation. Moreover, the rivers in the study area flow into the Fergana Valley, which has one of the highest population densities in Central Asia and is among the conflict-prone areas for water resources [74,75], and the rivers in this study area are the main sources for irrigation. Wilcox and USSL diagrams were used to classify and assess the water quality for irrigation [76]. From the USSL diagram (Figure 7), the water samples fall into the C1-S1 to C2-S1 category of low sodium (alkali) hazards to low/medium salinity hazards, respectively, suggesting that the river water quality was satisfactory for irrigation use. The sodium percentage (%Na) vs. EC was plotted in a Wilcox diagram for assessing the irrigation water quality, which showed that all water samples fall in the excellent to good categories for irrigation use.

Evaluation of the Irrigation Suitability
The concentrations of the potentially toxic elements were lower than the drinking water standards from the drinking water quality guidelines of the World Health Organization [69], the Council Directive 98/83/EC on the quality of water intended for human consumption, the European Union [70], the National Primary Drinking Water Regulations, the United States Environmental Protection Agency [71] and the China National Standards for drinking water quality (GB5749-2006) [72], which indicated that the water quality of the water from the parts of Aral Sea Basin in Kyrgyzstan is good (Table 1). The river water of the Syr Darya River is used extensively for irrigation [73], and the upstream water quality has an important impact on downstream irrigation. Moreover, the rivers in the study area flow into the Fergana Valley, which has one of the highest population densities in Central Asia and is among the conflict-prone areas for water resources [74,75], and the rivers in this study area are the main sources for irrigation. Wilcox and USSL diagrams were used to classify and assess the water quality for irrigation [76]. From the USSL diagram (Figure 7), the water samples fall into the C1-S1 to C2-S1 category of low sodium (alkali) hazards to low/medium salinity hazards, respectively, suggesting that the river water quality was satisfactory for irrigation use. The sodium percentage (%Na) vs. EC was plotted in a Wilcox diagram for assessing the irrigation water quality, which showed that all water samples fall in the excellent to good categories for irrigation use.

Conclusions
With a total of 76 river water samples collected during the LFS and HFS periods, the hydrochemistry and potential influencing factors in the upper reaches of the Syr River and its tributaries were systematically analyzed. The results are as follows: 1.
There are some differences in ion concentration between the LFS and HFS periods, but the hydrochemical classification reflected that all water samples fall in the calcium bicarbonate category, except that only three samples fall in the not dominant category during the LFS.

2.
The analysis shows that the main ions in the water from the upper reaches of Syr Darya River and its tributaries in Kyrgyzstan, come from the bedrock, and the intensity of carbonate dissolution is higher than that of silicate weathering.

3.
Human activities have had an impact on the water body, which is especially inferred from the indicators of the NH 4 -N concentration, FC contamination, and levels of potentially toxic elements. FC bacteria were detected in four river samples from the main channel of Syr Darya River (Naryn) and its tributaries during the LFS. However, FC bacteria were detected in 16 river samples from the main channel of Syr Darya River (Naryn) and its tributaries during the HFS. There may be anthropogenic intrusions of Cu, Pb, and Zn during the LFS period and of Cu, Pb, Zn, and Cd during the HFS period.

4.
The contents of potentially toxic elements in the water from the parts of the Aral Sea Basin in Kyrgyzstan are lower than international drinking water standards, but there are clustering differences between the LFS and HFS periods. The water quality classification shows that the water samples fall in the excellent to good categories for irrigation use.