Downstream hydrochemistry and irrigation water quality of the Syr Darya, Aral Sea Basin, South Kazakhstan Open Access

Investigation of water chemistry can provide information about mechanisms that control the composition of surface water. Furthermore, major ions and their elemental (ionic) ratios data can be used to characterize the suitability of water sources for domestic, irrigation, or industrial purposes (Prasanna et al. 2010; Logeshkumaran et al. 2015; Qian et al. 2016). There are multiple processes and sources that contribute salts to large river chemistry, including rainwater, rock weathering, agricultural effluent, domestic, industrial, and mining wastes (Gibbs 1970; Meybeck 1987; Gaillardet et al. 1999; Mahamat Nour et al. 2020).

The Syr Darya annually supplies the Aral Sea with almost half of its freshwater, but its salinity and water quality are greatly affected by these processes often with significant consequences to its suitability for irrigation, fisheries, and other uses. Former Soviet programs to increase agricultural productivity have resulted in a serious decrease in inflow to the Aral Sea leading to a reduction in its size along with an increase in salinity. Because of this overuse, the river and the Aral Sea are considered one of the most serious ecological disasters of modern times.

Currently, reports on the occurrence and geochemical behavior of major ions in the Syr Darya basin are scarce. Among them, a reconnaissance study on the upper reaches (Ma et al. 2019) and lower stretches in Kazakhstan have provided some new information (Zhang et al. 2019). The recent study by Zhang et al. (2019) reported concentrations of dissolved ions but did not evaluate processes affecting downstream changes of ionic composition in the river. Our recent studies describing the occurrence of pesticides showed that residues of lindane were determined among the highest concentrations reported for rivers globally (Snow et al. 2020). These residues and presence of elevated metal concentrations suggest potential risks to fishery health (Uralbekov et al. 2021; Allen et al. 2023).

A recent review (Qadir et al. 2018) considered the Aral Sea water resource as one of the main hotspots of high magnesium worldwide. The issue of elevated magnesium concentrations in irrigation water has been raised recently and is now considered an emerging example of water quality deterioration and land degradation (Rengasamy & Marchuk 2011; Qadir et al. 2018). This is because long-term irrigation with excess magnesium may affect soil hydraulic properties over time (Smith et al. 2015) rendering soils unsuitable for crop production.

Evaluation of the mineral sources and equilibria contributing to surface water quality is of great interest, especially in transboundary rivers, and may allow the implementation of measures for controlling contaminants to improve water quality. The objectives of this study are to (1) evaluate the seasonal occurrence and distribution of major ions downstream the Syr Darya River from the Shardara reservoir to the Aral Sea; (2) determine mechanisms controlling river water composition; and (3) evaluate long-term suitability for irrigation. New data are presented here describing major dissolved salts based on samples collected over a single growing season in June and August 2021.

Flow is supplied from mixed sources at the upper reach, mainly from meltwater from mountain glaciers and snow, where groundwater contributes to baseflow in the lower reaches. In the Syr Darya basin, there are a number of reservoirs to provide storage for hydropower and irrigation. The recorded long-term discharge varies from location to location, but representative values range from 263 to 646 m³/s at the upper headwaters at Kal below the confluence of the Naryn and the Karadarya, 387–649 m³/s at Chilmokhrom, 216–940 m³/s at KzylKyshlak, 371–631 m³/s at KokBulakt (Yakubov et al. 2011), and 28–637 m³/s at the lowest stretch of the river (Supplementary material, Figure S2). Groundwater is estimated to supply the river in the amount of approximately 1.6 km³ year⁻¹ (Asarin et al. 2010). The average annual withdrawal of water for irrigation is estimated at ca. 10,000 million m³ or about 27 m³ per thousand hectares (Yakubov et al. 2011). Long-term weather stations record extreme variations in air temperature ranging from +45 °С in summer to −38 °С in winter. The annual amount of precipitation as rainfall ranges from 80 to 200 mm in the lower reaches to 300–400 mm in the upper reaches (Yakubov et al. 2011).

This river crosses four Central Asian republics and is heavily used for power generation chiefly in the upper reach, for irrigation in the entire stretch, and is intended to support fisheries habitat in reservoirs and the Aral Sea. Despite the fact that the average annual flow has remained relatively constant, the inflow at the delta began to decrease starting in 1960, as the river became heavily used for irrigation and crop production (Asarin et al. 2010). Former Soviet programs aimed at maximizing agricultural productivity in this river basin increased the diversion of water through the development of an extensive canal system drastically affecting its chemical composition today. In addition, water quality is likely affected by industrial plants for polymetallic ores processing (U, Au, and Ag), oil production (UNECE website 2023) as well as municipal wastewater discharge (Klimaszyk et al. 2022).

In the laboratory, grab samples were filtered using 0.45-μm pore size membrane filters to separate suspended and dissolved fractions. Major ions were measured using standard hydrochemical methods (State Standard 26449.1-89). Sodium and potassium concentrations were measured using flame photometry. Calcium, magnesium, chloride, and alkalinity were determined by titration with standardized solutions. Sulfate ions and total dissolved solids (TDS) were determined by gravimetric analysis. A strong positive linear correlation (r² = 0.99, p < 0.01) had been found between TDS and specific electrical conductance (EC), while an average TDS/EC ratio was at 0.63 with the coefficient of variation (C_v) equal to 8%. concentrations were determined by standard colorimetric methods. The quality of chemical analyses was checked by calculating the charge balance, which was within ±10%, as well as the observed strong correlation between the total cation (Tz⁺) and anion (Tz⁻) (r² = 0.92, p < 0.01). Quality controls included laboratory duplicates, method blanks, fortified blanks, and matrices to check accuracy and precision. Certified reference materials (Ecohydrocontrol, Kazakhstan) were used to check calibrations and measured element concentrations deviated by less than 10% from the reference values. The precision of the methods used was assessed through analysis of duplicates and the relative standard deviation (%RSD) was <10%.

The U.S. Geological Survey chemical thermodynamics program PHREEQC (version 3) (Parkhurst & Appelo 2013) was used for equilibrium modeling of mineral phases and aqueous species. The program PHREEQC provides a computational method for predicting geochemical equilibrium. Chemical equilibrium is achieved by the anion-association aqueous model and can predict speciation saturation indices (SIs), which can be implemented in a wide variety of reversible and irreversible reactions. For data input to PHREEQC, log K (a temperature-dependent equilibrium constant), the chemical equation for mole-balance and mass-action expressions, and the activity coefficients for each aqueous species are defined (model expressions and theory are presented in the supplement file). Furthermore, the model is based on Debye–Hückel expressions to account for the non-ideality of aqueous solutions and performs satisfactorily at low ionic strength but is not suitable at higher ionic strengths in the range of seawater and above.

PHREEQC can predict aqueous species and reactions by using equilibrium constants from the ‘WATEQ4F’ database (Ball 1992), and model output includes the activity of dissolved ionic species as well as SIs of mineral species. The calculated product of ionic activity (IA) at equilibrium reaction is compared with the solubility product (KT). When log(IA) is greater than log(KT), the mineral is considered saturated and will precipitate out of the system; when log(IA) is lower, it is undersaturated and there is potential for mineral dissolution in the aqueous system; and when these two values are equal, mineral dissolution and precipitation in the aqueous system is at equilibrium (Chidambaram et al. 2012). Mineral precipitation and dissolution are important in riverine systems to understand how the concentration varied based on mineral occurrence in the system. In the present study, predicted aqueous species and SIs of key minerals were calculated. The model uses a nitrite/nitrate couple to calculate redox-sensitive species (Kölling 2000) and predicted equilibrium SIs of different minerals occurring at different sampling locations and events. Model inputs included measured temperature, pH, dissolved oxygen, alkalinity as bicarbonate, other anions (fluoride, chloride, sulfate, nitrite, nitrate, and silicate), and cations (sodium, potassium, calcium, magnesium, and ammonium) (Nordstrom et al. 1979; Malakar et al. 2020, 2022).

The Piper diagram provided a means to classify water based on major ion concentrations and identified downstream changes in water composition. Ion concentrations, expressed in meq L⁻¹, were plotted on a Piper diagram using the GWChart (Winston 2000). Gibbs diagram was used to interpret the effect of hydrogeochemical processes such as rock weathering, precipitation and evaporation, on surface water chemistry. Mixing diagrams were used to identify the contribution of various minerals weathering to the major ion composition of the surface water. Gibbs diagram plots of TDS versus the weight ratios and ⁠, and mixing diagram plots of Ca/Na molar ratio versus Mg/Na and HCO₃/Na molar ratios were plotted using OriginPro, Version 8.5 (OriginLab Corporation, Northampton, MA, USA). End-member coordinates for the Gibbs plots were taken from Gibbs 1970, while those for the mixing diagrams were taken from Gaillardet et al. (1999).

Statistical parameters used in this study (variation, С_v, and correlation, r coefficients) were assessed by OriginPro, Version 8.5 (OriginLab Corporation, Northampton, MA, USA). A Generalized Extreme Studentized Deviate test (ESD-test) was used to evaluate the difference, where p < 0.05 was considered as statistically significant.

The results of physico-chemical analyses along the Syr Darya are given in Tables 1 and 2 and Supplementary material, Tables S2 and S3. Field measurements indicate that river water was well-oxygenated with dissolved oxygen concentrations ranging from 7.0 to 10.7 mg L⁻¹. In all cases, the pH was slightly alkaline, ranging from 7.7 to 8.4. Specific EC increased downstream, changing from 1,341 μS cm⁻¹ in the upper points to 2,220 μS cm⁻¹ in the lower reaches. TDS (and EC) values decrease in the Shardara reservoir from ∼1,110 mg L⁻¹ (1,894 μS cm⁻¹ ) to ∼800 mg L⁻¹ (1,341 μS cm⁻¹ ) for samples collected in June 2021 and from ∼1,500 mg L⁻¹ (2,184 μS cm⁻¹ ) to ∼1,250 mg L⁻¹ (1,808 μS cm⁻¹ ) for samples collected in August 2021, suggesting the occurrence of some dilution in the reservoir. The C_v of measurements is generally within 15%, indicating no significant spatial variability in composition along the river. Concentrations of major ions are presented in Tables 1 and 2. The sequences of average major ions concentrations (in meq L⁻¹) in descending order for the Syr Darya, were: for cations, for anions. For sulfate and magnesium, concentrations in all water samples exceed Kazakhstan's national maximum permissible concentrations (MPC) (MINRYBKHOZ USSR 1990) for the protection of fishery water bodies, and for sodium, in most cases, exceed the corresponding MPC of 100 mg L⁻¹.

Table 1

Major ions and TDS (in mg L⁻¹) collected for Syr Darya waters in June 2021

Site ID .	TDS .	.	.	.	.	.	.	.	.
1	1,104	207	ND	87	688	140	90	235	2.27
2 (Shardara)	802	158	3.0	65	455	110	63	114	4.2
3	818	171	3.0	68	457	113	61	114	4.3
4	841	180	ND	76	481	112	67	114	0.20
5	840	165	ND	75	474	113	68	142	2.27
6	848	168	3.0	78	485	113	67	135	4.7
7	869	159	3.0	79	483	114	66	142	1.23
8	1,117	168	ND	113	682	138	99	253	8.5
Mean	905	172	–	80	526	119	72	156	3.4
Maximum	1,117	207	3.0	113	688	140	99	253	8.5
Minimum	802	159	ND	65	455	110	61	114	0.20
9	10,497	128	ND	3,040	4,276	681	730	3,554	85
Kazakhstan MPCs (MINRYBKHOZ USSR 1990)				300	100	180	50	120	50

Site ID .	TDS .	.	.	.	.	.	.	.	.
1	1,104	207	ND	87	688	140	90	235	2.27
2 (Shardara)	802	158	3.0	65	455	110	63	114	4.2
3	818	171	3.0	68	457	113	61	114	4.3
4	841	180	ND	76	481	112	67	114	0.20
5	840	165	ND	75	474	113	68	142	2.27
6	848	168	3.0	78	485	113	67	135	4.7
7	869	159	3.0	79	483	114	66	142	1.23
8	1,117	168	ND	113	682	138	99	253	8.5
Mean	905	172	–	80	526	119	72	156	3.4
Maximum	1,117	207	3.0	113	688	140	99	253	8.5
Minimum	802	159	ND	65	455	110	61	114	0.20
9	10,497	128	ND	3,040	4,276	681	730	3,554	85
Kazakhstan MPCs (MINRYBKHOZ USSR 1990)				300	100	180	50	120	50

ND, non-detectable at a detection limit of 0.7 mg L⁻¹ for ⁠.

View Large

Table 2

Major ions and TDS (in mg L⁻¹) collected for Syr Darya waters in August 2021

Site ID .	TDS .	.	.	.	.	.	.	.	.	.
1	1,477	247	ND	118	795	170		109	213	3.8
2 (Shardara)	1,257	98	6,0	113	688	116		91	158	4.3
3	1,316	128	ND	127	720	120		85	163	4.5
4	1,331	134	ND	129	749	129		86	168	4.1
5	1,335	122	ND	122	661	119		96	183	4.8
6	1,345	125	ND	125	747	120		94	168	4.3
7	1,308	110	9,0	119	729	119		91	158	2.40
8	1,502	156	9,0	147	804	144		109	189	4.8
Mean	1,359	140	–	125	737	130		95	175	4.1
Maximum	1,502	247	9,0	147	804	170		109	213	4.8
Minimum	1,257	98	ND	113	661	116		85	158	2.40
9	10,262	134	24	3,057	4,252	561		675	1,588	62
Kazakhstan MPCs (MINRYBKHOZ USSR 1990)				300	100	180		50	120	50

Site ID .	TDS .	.	.	.	.	.	.	.	.	.
1	1,477	247	ND	118	795	170		109	213	3.8
2 (Shardara)	1,257	98	6,0	113	688	116		91	158	4.3
3	1,316	128	ND	127	720	120		85	163	4.5
4	1,331	134	ND	129	749	129		86	168	4.1
5	1,335	122	ND	122	661	119		96	183	4.8
6	1,345	125	ND	125	747	120		94	168	4.3
7	1,308	110	9,0	119	729	119		91	158	2.40
8	1,502	156	9,0	147	804	144		109	189	4.8
Mean	1,359	140	–	125	737	130		95	175	4.1
Maximum	1,502	247	9,0	147	804	170		109	213	4.8
Minimum	1,257	98	ND	113	661	116		85	158	2.40
9	10,262	134	24	3,057	4,252	561		675	1,588	62
Kazakhstan MPCs (MINRYBKHOZ USSR 1990)				300	100	180		50	120	50

ND, non-detectable at a detection limit of 0.7 mg L⁻¹ for ⁠.

View Large

A comparison of June and August composition indicates that the Ca–Mg–SO₄–Cl composition type remains unchanged. However, there is an upward deviation observed for water collected in August 2021 to calcium chloride from samples collected in June 2021. Increasing contributions of magnesium and sulfate ions to the river dissolved load were observed as illustrated in the triangle parts of the diagram. Water composition differs substantially from that of the upper reaches of the Syr Darya in Kyrgyzstan reported by Ma et al. (2019), which falls in the calcium bicarbonate region (green star, Figure 3).

This difference may be due to the return of a large amount of irrigation collector-drainage waters back to the mainstream in the upper and middle reaches of Fergana Valley, Mirzachoʻl Steppe, and Tashkent irrigation regions. In the upper reaches, there are no natural depressions to collect waters formed after irrigation. On average, the amount of salts entering the river along with drainage water with TDS ranging from 1.2 to 4.4 g L⁻¹ is ca. 21 million t year⁻¹ and the flowrates of large collectors vary from 7 to 52 m³/s, which are formed in the upper and middle reaches of the river (Yakubov et al. 2011). The majority of collector-drainage samples belong to the Ca–Mg–SO₄–Cl chemical facies (Chembarisov 1988).

This suggests that expected evaporation–crystallization processes contribute to downstream river dissolved load for samples collected in June, and to a lesser extent for samples in August. Recorded flow rates downstream the Syr Darya River further support this interpretation, with gradually decreasing flow from the Shardara reservoir to the river delta and slightly higher flow rates recorded for August (see Supplementary material, Figures S2–S5). The observed ∼40% increase of TDS downstream in June can be explained by the evaporation concentration effect, which can be seen from the increasing proportion of and relative to and ⁠, respectively. Possible sources for these ions include groundwater inflow, and runoff or irrigation return waters following leaching of secondary salts. The 14% downstream increase in TDS in August 2021 together with constant values of ⁠) and ⁠) ratios is probably due to chemical weathering or added irrigation drainage water return flow. The lower stretch of the Syr Darya in South Kazakhstan and the Kyzyl Orda region has typically been characterized as a Ca–Mg–SO₄–Cl chemical facies (Chembarisov 1988).

Based on these results, the primary mechanisms controlling Syr Darya water composition differ from those suggested by Zhang et al. (2019), where most samples reflected the effect of rock weathering. When compared with the upper reaches of the Syr Darya reported by Ma et al. (2019), an evolutionary path can be traced, starting from rock weathering dominance at the upper reaches (zone with Ca-rich and medium TDS) toward evaporation dominance at the lower reaches (zone with the Na-rich and high TDS values). Evaporation has an increasingly significant influence on distance downstream, while the upstream river water composition is mostly controlled by rock weathering. The maximum enrichment in Na was determined for the Aral Sea which is characterized by ⁠) and ⁠) ratios at 0.84 and 0.96, respectively, and a TDS ∼10,000 mg L⁻¹. This pattern of increasing Na and Cl contributions along with increasing salinity is typical for other rivers located in arid regions (Gibbs 1970).

These results show that Syr Darya composition is not significantly influenced by carbonate dissolution, supported by an average ratio of 4.3 and the lack of correlation for samples collected in June 2021 (see Table 3). The excess of Mg and Ca ions over bicarbonate ions should be balanced with other ions like sulfate (Meybeck 1987; Tsering et al. 2019). In addition, the calculated excess of over ⁠, indicated by their ratio ranging from 1.2 to 1.3 should be balanced by ⁠, which are in much high amounts than ions. Sulfate ions may be derived from the weathering of either evaporitic sulfate or oxidation of pyrite and sulfur containing organic matter (Meybeck 1987).

Table 3

Correlation coefficients (p < 0.01) for physico-chemical parapets, major ions for Syr Darya waters collected in June 2021 (left triangle) and August 2021 (right triangle)

.	Temp. .	pH .	EC .	DO .	TDS .	.	.	.	.	.	.	.
Temp.	1	–	−0.61a	–	–	–	–	–	–	–	–	–
pH	−0.61a	1	−0.61a	0.65a	–	−0.91	–	–	−0.79a	–	−0.78a	–
EC	–	0.74	1	–	0.98	0.80a	–	0.86	0.84	0.80a	0.80a	–
DO	–	–	–	1	–	−0.62	–	–	–	–	–	–
TDS	–	–	0.99	–	1	0.80a	–	0.82	0.86	0.90	0.86	–
	–	–	–	–	–	1	–	–	0.97	0.72a	0.92	–
	−0.75a	0.72a	0.87	–	0.88	–	1	–	–	–	–	–
	–	–	0.99	–	0.99	–	0.85	1	0.77a	–	–	–
	–	–	0.99	–	0.99	–	0.82	0.99	1	0.80a	0.91	–
	–	0.75a	0.98	–	0.98	–	0.92	0.98	0.96	1	0.86	–
	–	0.72a	0.98	–	0.99	–	0.90	0.98	0.98	0.98	1	–
	−0.71a	–	–	–	–	–	–	–	–	–	–	1

.	Temp. .	pH .	EC .	DO .	TDS .	.	.	.	.	.	.	.
Temp.	1	–	−0.61a	–	–	–	–	–	–	–	–	–
pH	−0.61a	1	−0.61a	0.65a	–	−0.91	–	–	−0.79a	–	−0.78a	–
EC	–	0.74	1	–	0.98	0.80a	–	0.86	0.84	0.80a	0.80a	–
DO	–	–	–	1	–	−0.62	–	–	–	–	–	–
TDS	–	–	0.99	–	1	0.80a	–	0.82	0.86	0.90	0.86	–
	–	–	–	–	–	1	–	–	0.97	0.72a	0.92	–
	−0.75a	0.72a	0.87	–	0.88	–	1	–	–	–	–	–
	–	–	0.99	–	0.99	–	0.85	1	0.77a	–	–	–
	–	–	0.99	–	0.99	–	0.82	0.99	1	0.80a	0.91	–
	–	0.75a	0.98	–	0.98	–	0.92	0.98	0.96	1	0.86	–
	–	0.72a	0.98	–	0.99	–	0.90	0.98	0.98	0.98	1	–
	−0.71a	–	–	–	–	–	–	–	–	–	–	1

^aAll values marked are significant at the 95% confidence limit.

‘–’ corresponds to no significant correlation.

View Large

In June 2021, and ions show a strong positive correlation with ⁠, and a poor correlation of with all other ions, indicating that the main source of these ions in the dissolved load of the river are leaching of evaporite minerals such as gypsum and halite. Major ions such as ⁠, ⁠, ⁠, strongly correlated positively with electrical conductivity as well as with TDS (see Table 3). The entry of these ions is possible from irrigation return flow through the drainage-collector system draining through various saline soil horizons. In comparison, samples from August 2021 show a strong positive correlation between and and ⁠, and the poor correlation of with all other ions is consistent with rock weathering as the major contributor of dissolved load in the studied area. Correlation coefficients of magnesium, calcium, and sodium are found to be maximum from linear regression 0.90, 0.86, 0.86, respectively.

The chemical equilibrium model predicts mineral SIs over the course of the river for the months of June and August. The predicated SIs of the major mineral with saturation value greater than −1 are presented (Supplementary material, Table S6), the color code red indicates positive SI, white indicates zero or equilibrium SI and blue indicates negative SI. A positive SI value indicates probable precipitation of the mineral due to oversaturation and a negative value indicates undersaturation of the predicted mineral.

The major mineral SIs are found to be similar in the 2 months at different sampling locations, and predicted equilibria are with Ca, Mg, and Si minerals (Supplementary material, Table S6). Geochemical modeling suggests that the Syr Darya is generally saturated with respect to carbonate minerals (dolomite, calcite, aragonite, huntite, and magnesite) and some silicate minerals (chrysotile, talc, and tremolite). Chloride (halite), sulfate (gypsum, anhydrite), and most other minerals present near equilibrium are undersaturated. Mineral SIs are consistent with previous findings suggesting the contribution of secondary minerals leaching and silicate weathering to the river chemistry. However, there is a clear distinction in the SI values of some minerals, which varies spatially along the river stream and temporally from June to August. Chrysotile, an asbestos group mineral presents positive SI in the upstream and downstream in June, but in August, it has positive SI from site 2 to site 9 locations of the river stream. Similar temporal differences in SI are also observed in huntite mineral, a magnesium carbonate mineral reported in other evaporation-driven river basins (Akbulut & Kadir 2003), where SI is positive in August throughout the river stream, compared to June.

The spatial and temporal differences in mineral SI such as huntite and chrysotile can indicate evaporitic mineral precipitation (Kinsman 1967; Hovland et al. 2018). Huntitite, magnesite, and other Mg-rich carbonate minerals have been reported in related depositional settings in Central Asia (Novoselov et al. 2019). A clear difference in mineral composition is also observed at site 9; however, given the high ionic strength of the water sample at the Aral Sea, model predictions may be affected (Parkhurst & Appelo 2013).

The atmospheric contribution of oceanic salts to the river dissolved load is negligible due to the fact that the river is strongly influenced by agricultural activity, and because average rainfall is low in the study region (about 100 mm year^–1, see Supplementary material, Table S7). The effect of agricultural effluent and fertilizer use on the composition of water can be assessed by the presence of nitrogen-containing compounds (Table 4). For most samples collected in August 2021, concentrations of were below detection suggesting that agricultural inputs from runoff are low by the end of summer. By contrast, for water collected in June 2021, were found in the range of 1.90–7.45 and 0.017–0.424, respectively. These results support the contribution of irrigation return waters to the mainstream water collected in June 2021 as indicated by the presence of nitrates along with a strong correlation. In August, the absence of nitrogen compounds suggests less influence from irrigation return flow.

Table 4

N-containing inorganic species (mg/L), F⁻ and SiO₂ for collected Syr Darya

.	June 2021 .					August 2021 .
Site ID .	.	.	.	F− .	SiO2 .	.	.	.	F− .	SiO2 .
1	ND	7.5	0.424	0.45	10.0	ND	8.9	0.14	0.68	10.5
2 (Shardara)	ND	4.2	0.201	0.45	5.0	ND	ND	ND	0.64	8.1
3	ND	5.7	0.056	0.44	3.8	ND	ND	ND	0.62	9.8
4	ND	5.6	0.046	0.44	5.1	ND	ND	ND	0.62	9.6
5	ND	5.2	0.054	0.44	4.0	ND	ND	ND	0.64	7.9
6	ND	5.6	0.039	0.46	4.2	ND	ND	ND	0.65	8.5
7	ND	4.2	0.017	0.47	4.3	ND	ND	ND	0.68	7.1
8	ND	1.9	0.052	0.54	4.0	ND	ND	ND	0.76	7.3
Mean	–	5.0	0.110	0.46	5.0	–	–	–	0.66	8.6
Maximum	–	7.5	0.424	0.54	10.0	–	–	–	0.76	10.5
Minimum	–	1.9	0.017	0.44	3.8	–	–	–	0.62	7.1
9	ND	ND	0.054	1.84	0.7	ND	ND	ND	2.37	ND
Kazakhstan MPCs	0.5	40	0.08			0.5	40	0.08

.	June 2021 .					August 2021 .
Site ID .	.	.	.	F− .	SiO2 .	.	.	.	F− .	SiO2 .
1	ND	7.5	0.424	0.45	10.0	ND	8.9	0.14	0.68	10.5
2 (Shardara)	ND	4.2	0.201	0.45	5.0	ND	ND	ND	0.64	8.1
3	ND	5.7	0.056	0.44	3.8	ND	ND	ND	0.62	9.8
4	ND	5.6	0.046	0.44	5.1	ND	ND	ND	0.62	9.6
5	ND	5.2	0.054	0.44	4.0	ND	ND	ND	0.64	7.9
6	ND	5.6	0.039	0.46	4.2	ND	ND	ND	0.65	8.5
7	ND	4.2	0.017	0.47	4.3	ND	ND	ND	0.68	7.1
8	ND	1.9	0.052	0.54	4.0	ND	ND	ND	0.76	7.3
Mean	–	5.0	0.110	0.46	5.0	–	–	–	0.66	8.6
Maximum	–	7.5	0.424	0.54	10.0	–	–	–	0.76	10.5
Minimum	–	1.9	0.017	0.44	3.8	–	–	–	0.62	7.1
9	ND	ND	0.054	1.84	0.7	ND	ND	ND	2.37	ND
Kazakhstan MPCs	0.5	40	0.08			0.5	40	0.08

ND, non-detectable at a detection limit of 0.01 mg L⁻¹ for ⁠, ⁠, 0.003 mg L⁻¹ for 0.05 mg L⁻¹ for SiO₂.

View Large

The average content of silica (8.6 mg/L, ∼143 μmole L⁻¹) in August is higher compared to its average concertation in the samples collected in June 2021 (5.0 mg/L, ∼80 μmole L⁻¹). Given that silica is essentially derived from silicate weathering (Meybeck 1987), a greater contribution of chemical weathering may contribute to the dissolved load in August 2021. The lowest silica concentrations occur at the mouth of the river at the Aral Sea, most probably due to removal through biological activity.

It is now estimated that water diverted for irrigation is approximately 84–86% of the total water intake (UNECE website 2023). It is difficult to assess withdrawal rates from the lower part of the Syr Darya River due to the presence of numerous irrigation distribution systems and canals in the study area. The average annual withdrawal of water from the lower part of the Syr Darya, excluding the Dostyk water canal, is estimated at ca. 5,000 million m³ (UNECE Website 2014), and water is almost entirely used for irrigation.

Numerous water quality measurements determine water suitability for irrigation. These water quality parameters assess the effects of dissolved salts on the physico-chemical properties of soils such as morphology, permeability and aeration. Traditional water parameters reported for the Syr Darya relate chiefly to the relative proportion of sodium to other cations (SAR, Na%, KI) as a measure of suitability for irrigation (Yakubov et al. 2011; Zhang et al. 2019). These studies showed that the Syr Darya is generally suitable for irrigation purposes due to its low fraction of sodium and bicarbonate ions. However, alkali hazard must be considered in conjunction with salinity hazard based only on TDS or EC values (US Salinity Laboratory 1954). Table 5 summarizes the irrigation water quality indices. Based on EC measurements alone, irrigation water salinity control is needed throughout the season, generally requiring the selection of crops with high salt tolerance. Other recent studies suggest a need to expand the water quality indices when assessing water suitability for irrigation by considering magnesium, hazardous trace elements, and bioavailability (Satybaldiyev et al. 2023; Malakar et al. 2019).

Previous investigations (Hwang et al. 2017; Anim-Gyampo et al. 2019) use a so-called MH or magnesium hardness (MH) index to evaluate suitability for irrigation. Based on the MH, all samples collected in August are unsuitable for irrigation (see Table 5) due to soil effects associated with elevated magnesium. Here, it is worth noting that this index shows only the ratio of magnesium-to-calcium, and thus, it is not possible to draw any conclusions regarding the suitability of water for irrigation. Elevated magnesium concentrations in irrigation water may lead to rapid soil degradation and impact crop yields negatively (Qadir et al. 2018). One of the earliest studies reporting the effects of irrigating with high magnesium-to-calcium ratio water was to exchangeable sodium (Rahman & Rowell 1979). A number of subsequent studies appeared that reported low productivity of some crops when irrigated with water containing high magnesium concentrations, even though infiltration may not be evident (Ayers & Westcot 1985; Rengasamy & Marchuk 2011; Qadir et al. 2018). It is thought that the antagonistic effect of high magnesium or competition for adsorption sites affects soil structure.

A recently proposed CROSS_opt water quality index evaluates sodicity hazard and the authors suggest substituting this for the commonly used SAR in evaluating irrigation water quality (Oster et al. 2016). Reasons for recommending the use of this parameter are as follows: (1) it includes the relative deleterious impact on soil hydraulic properties of four common cations (Na, K, Mg, Ca); (2) this parameter has been tested on many soils based on a comprehensive review of the literature; (3) a strong correlation was found between CROSS_opt and effects to the hydraulic conductivity of soils. In the present study, the predicted impact of water on soil permeability is not significant as shown using the CROSS_opt parameter (see Table 5), despite high levels of magnesium relative to calcium observed for waters collected in August 2021 (recorded magnesium-to-calcium ratios ranged from 1.1 to 2.2). We note here that the CROSS_opt water quality index of the Syr Darya waters is close to a ‘slight-to-moderate’ degree based on our hydrochemical analysis.

Previous investigations of soil properties in the Syr Darya basin (Bekbaev et al. 2005; Ramazanov et al. 2016) showed an increase in soil sodicity, with exchangeable Mg and Na percentages up to 66 and 17%, respectively. More than 30% of the irrigated territory in South Kazakhstan can be categorized as magnesium-affected soils, likely subject to long-term deterioration in soil properties, including decreased infiltration rates and degradation in soil structure (Bekbaev et al. 2005; Vyshpolsky et al. 2008, 2010). One cause for increased soil sodicity is the use of recycled drainage waters with high Mg concentrations for irrigation. Future evaluation of water quality may include the evaluation of the CROSS_opt index to better predict the long-term dispersive effect of Mg in cropland soils.

New hydrochemical analysis of the Syr Darya shows significant changes in the water type from Ca–Mg–HCO₃ to Ca–Mg–SO₄–Cl with distance downstream and contrasts with recent studies of upstream river water quality in this region. Our results show that evaporative concentration contributes to the TDS increasing downstream the Syr Darya in June, and to a lesser extent for samples in August. Element ratios and chemical equilibrium modeling indicate that evaporite dissolution contributed to dissolved load at the beginning of the irrigation season, while chemical rock weathering is a more dominant process toward the end of irrigation. Dissolved element ratios, together with correlation analysis and geochemical modeling, helped to identify predominant mineral dissolution controlling surface water chemistry. An increased proportion of magnesium relative to calcium (Mg/Ca: >1) in late summer has implications for the sustainable irrigation of soils in this region. Improved irrigation water management may be possible through controlling the elevated salinity and magnesium levels in South Kazakhstan croplands. Additional irrigation water quality studies will provide a better understanding of the continuous use of drainage waters for irrigation and the impacts of high magnesium waters on sustainable crop production in South Kazakhstan. Further study of spatial and temporal distributions of major dissolved ions with a focus on major uses will contribute to a better understanding of river chemistry in similar arid regions.

This research is funded by the Science Committee of the Ministry of Education and Science of the Republic of Kazakhstan (Grant No. AP08857167).

All authors contributed to the study conception and design. The first draft of the manuscript was written by B.U. and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.