Spatial transformation of river runoff under hydroclimatic change and agricultural water use in the Shu River basin, Central Asia

Understanding the spatial differences in runoff response to hydroclimatic variability and anthropogenic water use is essential for sustainable water resource management in arid transboundary river basins. This study evaluates spatial differences in runoff–hydroclimatic relationships along the Shu River basin using long-term river runoff observations and key hydroclimatic indicators. Hydroclimatic changes were assessed using potential evapotranspiration (ET), climatic water balance, and the Standardized Precipitation Evapotranspiration Index (SPEI) for the upper and lower parts of the basin. Long-term trends were assessed using the Modified Mann–Kendall test (Yue and Wang modification) and Sen’s slope estimator, while relationships between runoff and hydroclimatic variables were examined using Spearman correlation analysis. The results reveal pronounced spatial differences in the mechanisms governing runoff variability across the basin. In the upper reaches, runoff remained strongly associated with long-term moisture availability (SPEI-12; ρ = 0.73, p < 0.01), while a significant drying trend was detected at the Kordai station (Modified Mann–Kendall, Z = −5.93, p = 3.07 × 10 −9). Climatic water deficit increased markedly downstream, from approximately 520 mm yr −1 at Kordai to −1130 mm yr −1 at Moiynkum. In contrast, runoff in the lower reaches exhibited weak and statistically insignificant relationships with hydroclimatic indicators, suggesting that intensive agricultural water withdrawals likely contribute to modifying the natural climate–runoff relationship. The results suggest that the Shu River basin is undergoing a transition from a hydrological system strongly associated with hydroclimatic conditions in the upper reaches to a human-modified water-limited system downstream.

Water resources in the arid and semi-arid regions of Central Asia are under increasing pressure due to climate change and intense agricultural use. Rising air temperatures, increased atmospheric evaporative demand, and increasing aridity are leading to a decline in water availability in river basins, especially in regions where water resources are already limited by natural conditions (Akhtar et al., 2021).

For many river systems in Central Asia, an additional factor transforming the hydrological regime is the high anthropogenic load, primarily associated with irrigated agriculture. In such basins, reduced river flow is the result of combined influence of climate change and internal redistribution of water resources along the channel system (Stephens et al., 2021).

The Shu River basin can be considered a typical example of such a system. The upper part of the basin develops in relatively more humid conditions, while the middle and lower reaches are located in an arid zone with a high moisture deficit. Moreover, the middle reaches experience significant pressure associated with agricultural water use, potentially exacerbating the reduction in available water flow in the lower reaches of the basin (Kaliyeva et al., 2021).

Recent studies have considerably improved understanding of hydrological change in Central Asian river basins. Research has documented increasing temperatures, altered runoff regimes, and growing pressure on water resources caused by climate change and expanding irrigation systems in transboundary basins (Didovets et al., 2021; Lobanova et al., 2021). Similar processes have been reported for the Amu Darya, Yellow River, and other semi-arid river systems, where river discharge is increasingly controlled by the combined effects of climatic variability and anthropogenic water regulation rather than by climatic forcing alone (Kong et al., 2025).

Recent studies have identified several regions of the world as particularly vulnerable to climate change, especially arid and semi-arid environments where increasing temperatures and changing precipitation regimes intensify water scarcity and environmental stress. The Eastern Mediterranean and Middle East region has been recognized as one of the global climate change hotspots. They are characterized by warming rates that exceed the global average as well as highly modified hydrological cycles. Observational records indicate increasing temperatures, more frequent and severe droughts, prolonged heatwaves, and greater variability in precipitation patterns, all of which contribute to increasing pressure on water resources and ecosystem stability (Zittis et al., 2022).

Climate change also poses substantial challenges for agricultural production and food security. Rising temperatures, altered precipitation regimes, and increasing frequency of extreme weather events directly affect soil moisture availability, crop productivity, livestock production, and aquatic ecosystems. Changes in drought occurrence and flood frequency can disrupt agricultural systems and reduce the resilience of food production networks. Consequently, climate adaptation and mitigation strategies are becoming increasingly important components of sustainable agricultural management (Gomez-Zavaglia et al., 2020).

Similar hydroclimatic changes have been documented in semi-arid regions of Africa. In the Sahel, long-term rainfall variability has been linked to large-scale atmospheric circulation patterns and ocean-atmosphere interactions. Although total rainfall has partially recovered in recent decades, precipitation has become more irregular, with increasing rainfall intensity and longer dry intervals between precipitation events. These changes highlight the complexity of hydroclimatic responses to global warming and emphasize the importance of evaluating not only total precipitation but also drought characteristics and extreme events (Biasutti, 2019).

In addition to climatic pressures, land degradation processes such as soil salinization represent a major environmental challenge in dryland regions. Both natural factors and anthropogenic activities, including irrigation practices and inadequate drainage systems, contribute to increasing soil salinity. Remote sensing studies have demonstrated the effectiveness of satellite-based approaches for monitoring salinity dynamics and identifying areas vulnerable to land degradation. Such assessments provide valuable information for sustainable land and water resource management in arid and semi-arid environments (Feng et al., 2025; Gorji et al., 2017).

Taken together, these studies demonstrate that climate change is altering hydrological processes, agricultural productivity, and ecosystem functioning across dryland regions worldwide. Understanding the interactions among temperature, evapotranspiration, water balance, drought conditions, and land degradation is therefore essential for developing effective adaptation strategies and ensuring sustainable management of natural resources under changing climatic conditions.

Evidence from the Yellow River Basin demonstrates that the combined effects of climate change and intensive water management can fundamentally alter river hydrology by reducing runoff, increasing evapotranspiration, and exacerbating water scarcity (Wang et al., 2025). Although numerous studies have investigated streamflow responses to climate change and human activities, recent systematic reviews indicate that considerable methodological challenges remain in separating the relative contributions of these drivers. Improving attribution approaches has therefore become an important research priority for sustainable water resources management (Ngwira and Damazio, 2025). Recent studies conducted in arid inland river basins of Northwest China have highlighted that both climate variability and human activities contribute to long-term changes in runoff and hydrological conditions. These findings emphasize the importance of integrated attribution approaches capable of separating climatic and anthropogenic influences on river systems (Jia and Niu, 2024). Mountain headwater catchments are essential sources of freshwater for downstream regions, but they are becoming increasingly vulnerable to the combined effects of climate variability and human-induced land-use change. Studies conducted in mountainous river basins have demonstrated that these interacting drivers substantially alter runoff dynamics, seasonal water availability, and water security, emphasizing the importance of integrated assessments for sustainable watershed management (Soomro et al., 2026).

Although similar approaches have been applied in several arid river basins, such integrated assessments remain limited for the Shu River basin, where the relative importance of hydroclimatic aridization and agricultural water withdrawals has not been comprehensively evaluated. Previous studies have considerably improved the understanding of hydrological processes in the Shu River basin. Kaliyeva et al. (2021) projected future changes in the hydrological regime under climate change scenarios, while Issanova et al. (2018) assessed regional water resource availability across Kazakhstan’s water-economic basins. More recently, Issaldayeva et al. (2023) demonstrated increasing hydroclimatic variability and runoff changes in southeastern Kazakhstan, highlighting the growing influence of climate change on mountain-fed river systems. Similar studies conducted in Central Asia and other semi-arid regions have shown that river runoff is increasingly affected by the combined influence of climatic aridization and anthropogenic water use (Rakhimova et al., 2020; Saedi et al., 2022; Gu et al., 2024). However, previous studies have generally considered climatic impacts or water resource use separately and have not explicitly evaluated how the relative importance of climatic and anthropogenic drivers changes along the longitudinal profile of the Shu River basin. This knowledge gap limits understanding of the mechanisms responsible for downstream runoff transformation and constrains the development of effective adaptation strategies.

Despite these advances, no previous study has comprehensively evaluated how the relative influence of hydroclimatic conditions and agricultural water withdrawals changes from the runoff-generation zone in the upper Shu River basin to the water-limited downstream reaches. Addressing this knowledge gap constitutes the primary motivation for the present study.

Previous studies on the Shu River basin have primarily focused on projected climate change impacts (Kaliyeva et al., 2021), regional water resource availability (Issanova et al., 2018), and recent runoff variability in mountain-fed river systems of southeastern Kazakhstan (Issaldayeva et al., 2023). Although these studies have substantially improved understanding of hydroclimatic processes in the region, they have largely examined climate change impacts, hydrological trends, or water resources separately. Comparatively little attention has been paid to the combined effects of hydroclimatic aridization and intensive agricultural water abstraction on the spatial transformation of river runoff along the river continuum. In particular, it remains unclear how the relative importance of climatic and anthropogenic controls changes from the mountainous runoff-generation areas to the downstream water-limited sections of the basin.

The main contribution of the present study is to evaluate spatial differences in runoff–hydroclimatic relationships along the Shu River basin under contrasting hydroclimatic and agricultural water-use conditions. By integrating long-term runoff observations with ET 0, climatic water balance, SPEI, and statistical analyses, the study provides a comprehensive regional assessment of runoff transformation under contrasting climatic and water-use conditions. By linking hydroclimatic variability with observed runoff dynamics, this study extends previous regional observations of the progressive transition from a hydrological system strongly associated with hydroclimatic conditions in the upper basin to a human-modified water-limited system downstream. This integrated assessment contributes to a better understanding of hydrological transformation in transboundary arid river basins and provides a scientific basis for adaptive water resources management under increasing climatic and anthropogenic pressures.

A recent study by Bakbergenov et al. (2026) investigated hydrological variability and water stress dynamics in the Shu River basin using recent hydrological observations (2011–2024), with emphasis on temporal changes in river discharge, water balance components, and water stress indicators (WSI and EFR). In contrast, the present study addresses a different research question by examining the long-term spatial transformation of runoff controls along the river continuum. Specifically, it integrates multi-decadal runoff observations with hydroclimatic indicators (ET 0, climatic water balance, and SPEI) to evaluate how the relative influence of hydroclimatic variability and agricultural water use changes between the upstream runoff-generation zone and downstream irrigated reaches. Therefore, the two studies should be regarded as complementary rather than overlapping. Unlike previous studies focusing on hydrological variability or water stress assessment, the present study specifically investigates the spatial evolution of runoff controls along the river continuum under contrasting hydroclimatic and water-use conditions.

The objective of this study is to assess the spatial transformation of the hydrological regime of the Shu River basin under conditions of increasing aridization and anthropogenic water use. To achieve this goal, we analyzed long-term river runoff dynamics in the upper and lower reaches, calculated indicators of potential evapotranspiration, climatic water balance, and aridity indices, and conducted a correlation analysis of the relationship between hydrological and hydroclimatic variables.

The study aims to answer these questions: (1) What long-term changes have occurred in river runoff in the upper and lower reaches of the Shu River basin under conditions of increasing aridization and water use pressure? (2) How have key hydroclimatic indicators, including potential evapotranspiration (ET0), climatic water balance, and aridity indices, changed over the study period? (3) What are the spatial differences in hydrological and hydroclimatic conditions between the upper and lower sections of the Shu River basin? (4) What relationships exist between river runoff and hydroclimatic variables, including evapotranspiration, climatic water balance, and aridity indicators?

Research area

Physical-geographical and climatic characteristics of the Shu River basin

The Shu flows through Kyrgyzstan and Kazakhstan. The river is 1,186 km long, 800 km of which are within Kazakhstan (Report of the MEGNR, 2023). Its drainage basin covers 67,500 km 2 (Kotlyakov et al., 2006). Its main tributaries are the Chon-Kemin, Yrgaity, and Kakpatas on the right and the Alamedin, Aksu, and Kuragaty on the left (Prokhorov, 1978).

The Shu flows into the Issyk-Kul Basin through the Upper and Lower Ortotokoy Gorges. Until the early 1950s, during floods, part of the Shu’s flow flowed through the Kutemaldy branch into Lake Issyk-Kul. Currently, 5–6 km before reaching Lake Issyk-Kul, the river turns northwest near the town of Balykchy, passing the Kapchigay tract, and through the Boom gorge enters the Shuya Valley. In its middle reaches, the river serves as the border between Kyrgyzstan and Kazakhstan. In its lower reaches, the river flows through Kazakh territory, where the river valley widens to 3–5 km, becomes swampy and, forming the northern border of the Moiynkum Desert, dries up in the sands, only during the flood flowing into the endorheic salt lake Akzhaykyn (Filatov, 2011) among the vast salt marshes of the Ashchykol Depression (Filatov, 2011).

The Shu River has an average annual discharge of approximately 130 m 3 /s at the outlet of the mountainous catchment, while the average sediment load reaches about 60 kg/s (Bakbergenov et al., 2026). The river is predominantly supplied by snowmelt, glacier meltwater, and groundwater contributions, resulting in a distinct seasonal hydrological regime. High-flow conditions typically occur between May and September, whereas river discharge decreases substantially during the late summer and autumn months. In the downstream reaches within Kazakhstan, the river frequently dries up from early August until late November due to the combined effects of natural aridity and intensive water abstraction for irrigation. A major hydraulic structure influencing the basin’s water distribution is the Big Shu Canal (BChC), constructed in 1958 in Kyrgyzstan. The canal diverts water from the Shu River to irrigate agricultural lands in the Shu region, substantially modifying the downstream flow regime.

The Shu River is one of the largest transboundary rivers in Central Asia, crucial for the water supply of southern Kazakhstan and northern Kyrgyzstan (Figure 1). The river originates on the northern slopes of the Kyrgyz Range and is formed primarily by snow and glacier feeding, with additional contributions from rain and groundwater. As it flows from its upper to lower reaches, the natural conditions of the basin change significantly: from foothill runoff formation areas to arid lowland areas with limited water resources (Issanova, G. et al., 2018; Issaldayeva S. et al., 2023; Yi Luo et al., 2018).

The hydrological regime of the river also varies along the basin. The upper reaches are characterized by comparatively higher water content due to mountain feeding, while the middle and lower reaches are characterized by intense economic use and a significantly drier climate. A significant portion of the water resources within the middle reaches is used for agricultural purposes, which further impacts the availability of surface runoff in the lower reaches of the basin (Gerlitz et al., 2020; Shahgedanova, M. et al., 2020).

The climate of the Shu River basin is characterized by pronounced continentality and significant spatial heterogeneity. According to long-term climate analysis for the upper and lower parts of the basin, natural conditions vary significantly along the river: from the relatively more humid foothill regions of the upper reaches to the sharply arid lowlands of the lower basin (Lyu, Y. et al., 2021; Vilhar et al., 2022).

In the upper part of the basin, climatic conditions are comparatively more favorable in terms of atmospheric moisture. The average annual air temperature is approximately 10 °C, and annual precipitation reaches approximately 500 mm. Temperatures are markedly seasonal: the average January temperature is approximately −4 °C, while in July it is approximately 24 °C. Most precipitation occurs in the spring, peaking in April, when average monthly precipitation reaches approximately 80 mm. The summer season is relatively dry, but maintains moderate atmospheric moisture, with total precipitation from June to August reaching approximately 110 mm.

The lower part of the basin is characterized by significantly more arid conditions. The average annual air temperature here is around 12 °C. In July, the hottest month, the average temperature reaches 28 °C–29 °C, and in January, the coldest month, the average temperature remains below freezing, at around −5 °C, reflecting the distinctly continental climate. Annual precipitation constitutes less than 300 mm–only almost half of annual precipitation in the upper reaches of the basin. Even during the wettest period, spring, the average monthly precipitation does not exceed 35–40 mm (Kaliyeva, K. et al., 2021; Gu et al., 2024; Zou et al., 2019).

The most pronounced climatic differences between the upper and lower parts of the basin occur during the warmer months. While the upper reaches receive approximately 110 mm of precipitation during the summer months, in the lower part of the basin, this figure drops to less than 40 mm. Combined with high temperatures, this creates extremely tense water supply conditions in the lower reaches, where atmospheric moisture is at its lowest precisely during the period of maximum evaporation (Rahimi, D. et al., 2025; Saedi J. et al., 2022).

Thus, the Shu River basin is characterized by a distinct increase in aridity downstream. The combination of higher temperatures, reduced atmospheric moisture, and pronounced summer aridity creates fundamentally different natural conditions for the formation and transformation of water resources within a single river system.

Hydrological and climatic data

To assess the spatial transformation of water resources in the Shu River basin, hydrological data from long-term observations and hydroclimatic reanalysis data were used. Hydrological data were provided by the Kazakh Scientific Research Institute of Water Economy (KazSRIWE) and included long-term river flow observations at two hydrological stations located in different parts of the Shu River basin (Table 1). Data from the Tashutkul hydrological station, located in the main river flow formation zone downstream of the upper basin regulation system, were used to characterize the upper reaches. Data from the Ulanbel hydrological station, reflecting the river’s water regime after passing through areas of intense economic use in the middle reaches, were used to assess the lower reaches. This selection of hydrological stations allowed us to compare the water content of the upper and lower parts of the basin and assess spatial variations in river flow along the channel system. Annual river flow data since 1950 were used to analyze long-term dynamics (Issanova, G. et al., 2018; Tursunova, A. et al., 2025; Rakhimova, M. et al., 2020).

To analyze climatic conditions, data from the ERA5 global atmospheric reanalysis developed by the European Centre for Medium-Range Weather Forecasts (ECMWF) were used. ERA5 was chosen due to its high spatiotemporal consistency, long time coverage, and widespread applicability in hydroclimatic studies, particularly for areas with limited availability of homogeneous, long-term meteorological observations.

Since the Shu River basin is characterized by pronounced spatial heterogeneity in climatic conditions, three representative meteorological points were selected for analysis, corresponding to the upper, middle, and lower parts of the basin: Kordai, Tole Bi, and Moiynkum (Table 1). This choice was motivated by the need to cover the basin’s main natural and climatic zones and compare hydroclimatic conditions in areas of water formation, transformation, and water shortage.

Since the ERA5 data are presented in a regular spatial grid, climate parameter extraction was performed using the nearest grid cell approach. For each selected meteorological station, the closest ERA5 cell with coordinates that best matched the actual location of the station was used. This approach allowed for spatial referencing of climate data to the actual conditions of the studied areas of the basin while maintaining the consistency of the original dataset.

The Kordai station, located at an altitude of 1,141 m above sea level, was used to characterize the climatic conditions of the upper part of the basin, where the main river runoff is generated and the influence of foothill conditions is more pronounced. The Tole Bi station, located in the middle reaches at an altitude of 455 m, represents the transitional part of the basin, characterized by active agricultural use of water resources. The Moiynkum station, located at an altitude of 351 m, reflects the climatic conditions of the arid lower part of the basin, where moisture deficits are most pronounced.

Monthly values of key meteorological parameters, including air temperature, precipitation, relative humidity, wind speed, atmospheric pressure, and solar radiation, were used from ERA5. These data were used to calculate potential evapotranspiration, climate water balance, and aridity indices (Figure 2).

Limitations

The use of various data sources was dictated by the availability and temporal coverage of the source information. Long-term hydrological observations of river runoff were available in the form of official archival data from the Kazakh Scientific Research Institute of Water Economy, enabling the use of real instrumental observations to analyze long-term river runoff dynamics. However, comparable continuous long-term meteorological observation series for all selected sections of the basin were either unavailable or publicly available, or were incomplete and heterogeneous. Therefore, ERA5 reanalysis data, which provides a spatially and temporally consistent set of meteorological parameters for the entire study basin, were used for the climate analysis.

Statistical analysis of the datasets

Assessment of climatic water deficit and aridity analysis

The analysis was conducted for three meteorological stations located in different physiographic conditions within the basin: Kordai (the foothill zone of the upper basin), Tole Bi (the middle part of the basin), and Moiynkum (the arid lowland of the central basin). To ensure comparability of the time series, the analysis was conducted over a single time interval, 1992–2026.

The following monthly values were used as input climate data: average air temperature; relative humidity; wind speed; solar radiation; atmospheric precipitation.

Based on these data, the reference evapotranspiration (ET 0) was calculated using the FAO Penman–Monteith method, recommended by the FAO as the standard approach for estimating atmospheric evaporative demand (Gocic and Trajkovic, 2010; Zomer R. et al., 2022): where: ET 0 - is the reference evapotranspiration (mm/day), Δ - is the slope of the saturated vapor pressure curve, R n - is the net radiation at the crop surface, G - is the density of heat flux in the soil, γ - is the psychrometric constant, T - is the daily average air temperature at 2 m height, u 2 - is the wind speed at 2 m height, e s − e a - is the water vapor pressure deficit.

Based on the calculated reference evapotranspiration, a climatic water balance (WB) was determined, characterizing the relationship between moisture input and atmospheric moisture loss. This indicator serves as the baseline variable for calculating the SPEI index and allows one to assess the degree of climatic water deficit in each month of the study period: where: P - is the monthly precipitation (mm), ET 0 - is the monthly reference evapotranspiration (mm).

The Standardized Precipitation Evapotranspiration Index (SPEI) was used to assess the spatiotemporal variability of climatic water deficit in the Shu River basin. It takes into account the combined influence of precipitation and atmospheric evaporative demand on the development of arid conditions. The SPEI index was calculated using the log-logistic distribution recommended in the original methodology (Jin, C. et al., 2024; Gorugantula and Kambhammettu, 2026).

Two time scales were chosen for the analysis: SPEI-3 - seasonal variability of aridity; SPEI-12 - long-term hydroclimatic changes.

Using two time scales allowed us to simultaneously assess short-term seasonal humidity fluctuations and long-term trends in climatic aridity in the study area.

Next, the accumulated values of the climatic water balance were fitted to a log-logistic distribution. Then, the cumulative probability obtained from the fitted distribution was transformed into SPEI values: where: W = − 2 ⁡ ln ⁡ P, P - is the exceedance probability obtained from the log-logistic cumulative distribution function, C 0, C 1, C 2, d 1, d 2, d 3 - are the empirical coefficients of the inverse normal approximation.

The obtained SPEI values were interpreted according to the generally accepted aridity classification. Positive values indicate humid conditions, while negative values indicate moisture deficit. Values below −1.0 characterize moderate drought, below −1.5 indicate severe drought, and below −2.0 indicate extreme drought.

Trend detection and magnitude estimation

To assess the direction and statistical significance of long-term changes, the Modified Mann–Kendall test proposed by Yue and Wang (2004) was applied (Gadedjisso-Tossou et al., 2021; Yue and Wang, 2004). This modification corrects the variance of the Mann–Kendall statistic for the effects of serial correlation, thereby reducing the likelihood of detecting spurious trends in autocorrelated hydroclimatic time series. where: x j – successive values of a time series, x k – length of the time series.

The application of the Modified Mann–Kendall test allowed us to determine the presence of long-term changes in the SPEI time series and to assess the direction of climate trends in the study region.

Trend strength was determined using the Sen’s slope estimator, which reflects the rate of change in the index over time (Danandeh and Vaheddoost, 2020; Fernandes, F. et al., 2024; Bello N. et al., 2026): where: x j − x k - difference between values of the time series at time points j and k, j > k, Q i - individual slope calculated from two observations.

The final trend value was determined as the median of all calculated slopes: where: Q med - median rate of change of the rime series per time unit.

The results were interpreted as follows: Q med > 0 - an upward trend (a tendency toward increasing humidity); Q med < 0 - a downward trend (a tendency toward increasing aridity); the absolute value of Q med characterizes the intensity of changes over time.

In this study, the Sen’s slope estimator was used only for the SPEI-12 time series, as this index reflects long-term hydroclimatic changes and interannual variability in water deficit. Using the Sen’s slope estimator for SPEI-12 allowed us to quantitatively assess the rate of climatic aridization or humidification of the study area.

The combined use of the Modified Mann–Kendall test and the Sen’s slope estimator provided a comprehensive analysis of long-term trends. The Modified Mann–Kendall test was used to determine the statistical significance of trends while accounting for serial correlation, whereas the Sen’s slope estimator was used to quantify their magnitude.

Correlation analysis

To assess the relationship between river runoff and the hydroclimatic characteristics of the basin, a correlation analysis was performed using Spearman’s rank correlation coefficient. This method was chosen because hydrological and climatic time series typically do not conform to the assumption of normal distribution, may contain outliers, and are characterized by pronounced interannual variability, limiting the application of parametric correlation analysis methods.

The Spearman coefficient (ρ) was calculated using the formula:

Where: ρ - is the Spearman rank correlation coefficient, d i - is the difference between the ranks of the two compared variables for the i th observation, n - is the number of observations.

Values of the ρ coefficient range from −1 to +1, where negative values indicate an inverse relationship between the variables, positive values indicate a direct relationship, and values close to zero indicate the absence of a pronounced monotonic relationship.

Correlation analysis was conducted between annual river runoff values and corresponding hydroclimatic indicators, including potential evapotranspiration (ET 0), climatic water balance (WB), and the SPEI-3 and SPEI-12 aridity indices. Since the hydroclimatic parameters were calculated based on meteorological station data, the analysis was spatially referenced to the corresponding parts of the basin: data from the Kordai station were used for the upper reaches (Tashutkul hydrological station), and data from the Moiynkum station were used for the lower reaches (Ulanbel hydrological station).

To ensure comparability of the time series, monthly hydroclimatic data were aggregated to an annual time scale. Annual ET0 and climate water balance values were calculated as the sum of monthly values, while annual average index values were used for SPEI-3 and SPEI-12. Correlation analysis was performed only for overlapping observation years for which both river runoff and relevant climate indicator data were available. Statistical significance was assessed at significance levels of p < 0.05 and p < 0.01.

Temporal variability of river runoff in the Shu River basin

Long-term river flow dynamics in the Shu River basin indicate marked spatial heterogeneity in the hydrological regime and significant differences between the upper and lower reaches of the river (Figure 3a). Time series analysis indicates both changes in water content over time and significant flow transformations as water moves downstream.

In the upper reaches, at the Tashutkul gauge, flow was characterized by relatively high values during the first half of the observation period. In the 1950s and 1960s, annual flow volumes generally ranged from 1,800 to 2,200 million m 3, indicating a relatively stable hydrological regime during this period. In subsequent decades, the structure of the time series became more heterogeneous. In particular, the noticeable decline in flow in the 1970s should be considered with account for the construction of Tashutkul Reservoir, which significantly influenced the redistribution of river flow and altered the natural water regime in the upper reaches.

Beginning in the late 1980s, interannual variability became significantly more pronounced. Periods of relatively low flow alternate with high-flow phases with sharp peaks. Maximum flow volume was observed in the mid-2000s, when it exceeded 3,500 million m 3 in the upper reaches. However, this increase was not sustained. After 2005, a decline was recorded again, and in the final years of the period under review, values decreased to 1,400–1,600 million m 3. Thus, the temporal structure of the series allows us to distinguish three main hydrological phases: a period of flow decline, a subsequent recovery in flow, and the current stage of repeated water resource decline.

In the lower reaches, at the Ulanbel gauge, the hydrological regime is characterized by significantly lower flow throughout the entire observation period. At the beginning of the time series, annual runoff volume was approximately 600–1,600 million m 3, but, in subsequent years, there was a significant decline, which was particularly noticeable in the 1970s and 1980s, when values dropped to extremely low levels in some years. Even during high-flow periods, increased runoff in the upper reaches was not accompanied by a comparable increase in water content in the downstream section. For example, in the mid-2000s, while runoff volume in Tashutkul exceeded 3,500 million m 3, Ulanbel recorded around 2,200 million m 3, indicating a significant reduction in available runoff as water moved downstream.

The current water use pattern in the basin allows us to consider anthropogenic factors as one of the key mechanisms behind this transformation. The Shu River basin is characterized by intensive irrigated agriculture, where significant volumes of river water are withdrawn for agricultural purposes (Figure 3b). According to available data, the total volume of agricultural water consumption in some years exceeds 1,500–2,000 million m 3, which is comparable to a significant portion of the annual river flow. Given this level of water withdrawal, the reduction in water availability in the lower reaches appears to be a natural consequence of the active redistribution of water resources within the basin.

However, the observed flow transformation, particularly pronounced in recent years, likely cannot be explained solely by agricultural water use. For arid river basins, climatic factors are additionally significant, including increased atmospheric evaporative demand and increasing moisture deficits, which can significantly limit water availability even while inflow in the upper reaches remains unchanged. Therefore, further analysis of potential evapotranspiration, climatic water balance, and aridity indices is necessary to assess the contribution of climate-driven processes to the transformation of the hydrological regime of the Shu River.

Spatiotemporal variability of reference evapotranspiration (ET 0) and changes in climatic water balance (WB)

A comparative analysis of hydroclimatic characteristics at the three meteorological stations (Kordai, Tole Bi, and Moiynkum) suggests a pronounced spatial gradient in aridity and significant differences in the water balance structure (Figure 4).

All stations showed a negative average annual climatic water balance (the difference between precipitation and potential evapotranspiration, P − ET 0), indicating a persistent moisture deficit throughout the study area. The degree of deficit varies significantly between stations. The lowest water deficit is observed at Kordai, where the average climatic water balance is approximately −520 mm/year. At Tole Bi, the deficit nearly doubles, reaching approximately −920 mm/year, while the most extreme conditions are observed at Moiynkum, where the average annual moisture deficit is approximately −1130 mm/year (Figure 4a).

This pattern reflects a distinct transition from relatively more humid foothill areas to extremely arid lowland desert conditions. The difference between Kordai and Moiynkum exceeds 600 mm, demonstrating extremely high spatial heterogeneity of the climatic water regime even within a relatively limited geographic area.

Analysis of the annual precipitation distribution shows that Kordai station is characterized by the most favorable moisture conditions. The median annual precipitation here is approximately 500–520 mm, with interannual variability ranging from approximately 330–730 mm (Figure 4b). This indicates a relatively high supply of atmospheric moisture, although the range of values remains significant, reflecting the instability of the precipitation regime.

For Tole Bi station, the median precipitation decreases to 370–390 mm/year, and the interannual variability ranges from approximately 210–530 mm. Despite its proximity to Kordai on a regional scale, precipitation has decreased by approximately 25%–30%, significantly contributing to the development of a moisture deficit.

Minimum precipitation is observed at Moiynkum, where the median is only 280–300 mm/year, with an overall range of approximately 160–410 mm. Thus, compared to Kordai, precipitation here is almost 1.8 times lower, which is a key factor in the development of extremely arid conditions.

However, precipitation amounts were not the only reason for the differences in water balance, because the intensity of atmospheric evaporative demand also contributed to the change. Analysis of potential evapotranspiration (ET 0) indicates the opposite spatial trend. Minimum ET 0 values are recorded at Kordai, where median values are approximately 1050–1070 mm/year, with an interannual range of approximately 950–1150 mm.

At the Tole Bi station, a significant increase in evaporative demand is observed: the median ET 0 value reaches approximately 1,300 mm/year, with a variability range of 1,150–1,420 mm. This represents an increase in potential moisture loss of approximately 20%–25% relative to Kordai.

The highest ET 0 values are found at the Moiynkum station, where the median level reaches 1,410–1,430 mm/year, and in some years exceeds 1,500 mm. Compared to Kordai, this increase is nearly 35%–40%, confirming the significantly harsher thermal and radiation conditions of this area.

Thus, the increasing water deficit from Kordai to Moiynkum is explained by the dual impact of two climatic factors: a simultaneous decrease in precipitation and an increase in atmospheric evaporative demand. This is a typical sign of the transition from subhumid or semi-arid conditions to a pronounced arid climate.

Further confirmation of spatial heterogeneity is provided by an analysis of the distribution of the annual climatic water balance (Figure 4c). For the Kordai station, values range from approximately −800 to −200 mm, indicating the presence of years with relatively less pronounced moisture deficits. The median value is approximately −530 mm.

For Tole Bi, the range shifts toward more negative values (approximately −1180 to −630 mm), with the median value approaching −940 mm. This indicates not only an increase in water deficit but also its persistent nature.

At the Moiynkum station, the distribution is even more shifted toward extremely negative values—from −1370 to −870 mm, with a median of approximately −1150 mm. Notably, even the most favorable years at this station remain significantly drier than the average conditions at Kordai. This suggests that the region’s aridity is structural rather than temporary.

A seasonal analysis of the climatic water balance (Figure 4d) allows us to clarify the mechanisms underlying the annual moisture deficit. At all stations, summer is the main period of water deficit. For Kordai, the average summer deficit is approximately −135 mm, for Tole Bi it is approximately −205 mm, and for Moiynkum it reaches −225 mm. This confirms the dominant role of high temperatures, intense solar radiation, and maximum evapotranspiration during the warm season.

A negative water balance is also observed in spring: approximately −25 mm for Kordai, −75 mm for Tole Bi, and almost −100 mm for Moiynkum. autumn is characterized by a moderate deficit, which intensifies toward more arid areas: from −30 mm in Kordai to −65 mm in Moiynkum.

Interestingly, the winter season is the only period with a positive water balance at all the stations studied. For Kordai, the surplus is approximately +10 mm, for Tole Bi it is +20 mm, and for Moiynkum it is +8–10 mm. This is due to minimal evaporation during the cold period while atmospheric moisture inflow remains constant.

Overall, the results indicate the presence of a stable spatial aridity gradient in the study area. Kordai can be characterized as a zone of moderate climatic moisture deficit, Tole Bi as a transitional semi-arid zone with a pronounced deficit, while Moiynkum belongs to a zone of persistently high aridity.

Drought characteristics based on SPEI-3 and SPEI-12

To assess the spatiotemporal variability of aridity in the Shu River basin, we used the Standardized Precipitation Evapotranspiration Index (SPEI) on seasonal (SPEI-3) and long-term (SPEI-12) time scales, calculated using data from three meteorological stations representing contrasting physical and geographical conditions in the basin - Kordai, Tole Bi, and Moiynkum (Figure 5).

Analysis of the SPEI-3 time series demonstrates pronounced intra-annual and interannual variability in hydroclimatic conditions at all stations studied. Index values regularly fluctuate between positive and negative values, reflecting the alternation of wet and dry seasons characteristic of the region’s sharply continental climate. The amplitude of seasonal fluctuations at all stations remains high, reaching ±2.0–2.5 SPEI units in some periods. At the Moiynkum station, negative anomalies in the most extreme years approach −3.0, corresponding to the extreme seasonal drought category.

The most pronounced short-term drought episodes were recorded in the mid-1990s, 2007–2009, and 2011–2015, when SPEI-3 values at some stations dropped below −2.0. These episodes are particularly pronounced at Moiynkum, where seasonal fluctuations are sharper and have a higher frequency of extreme negative deviations. Meanwhile, the Kordai and Tole Bi stations demonstrate a more consistent pattern of seasonal dynamics, which is likely due to the similar climatic conditions of the upper and middle parts of the basin. Thus, seasonal analysis indicates high sensitivity of the entire basin to short-term hydroclimatic anomalies, although the most pronounced seasonal instability is characteristic of the arid central part of the basin.

The SPEI-12 index, which reflects long-term changes in the climatic water balance, suggests more stable long-term phases of humidification and aridization. For all three stations, the beginning of the study period is characterized by relatively dry conditions, followed by a transition to a stable positive phase in the mid-1990s. The most pronounced period of increased humidification is recorded in the late 1990s and early 2000s, particularly at the Kordai station, where SPEI-12 values exceeded +2.0, corresponding to extremely humid conditions. Similar positive dynamics are observed at the Tole Bi and Moiynkum stations, indicating a synchronous regional nature of hydroclimatic changes.

Starting in the second half of the 2000s, the nature of the long-term dynamics has changed significantly. Following a period of relative climatic stability, a pronounced negative phase developed at all stations, becoming particularly pronounced after 2010. The most prolonged and intense drought period occurred between 2011 and 2015, when SPEI-12 values at all stations studied remained consistently negative. During this period, the index dropped below −2.0, and at the Tole Bi station, it reached values of −2.4 to −2.5, consistent with extreme long-term drought. Deep negative anomalies were also recorded at Kordai and Moiynkum, confirming the regional scale of this drought episode.

To spatially summarize the identified patterns, a heat map of the SPEI-12 temporal variability was constructed for all three meteorological stations in the Shu River basin (Figure 6). The resulting spatiotemporal distribution confirms the high consistency of long-term hydroclimatic fluctuations between the studied observation points. The heat map clearly indicates synchronous periods of increased moisture in 1993–1995 and 1998–2004, as well as spatially consistent dry periods in 2007–2009 and especially 2011–2015, occurring simultaneously at all stations.

At the same time, spatial comparison highlights the heterogeneity of the intensity of climate change within the basin. The most persistent accumulation of negative anomalies in recent years has been observed at the Kordai station, while Moiynkum is more characterized by frequent but less stable fluctuations between wet and dry phases. Tole Bi exhibits characteristics of both stations, combining pronounced long-term variability with the severity of individual dry episodes.

Statistical analysis of the SPEI-12 time series using the Modified Mann–Kendall test (Yue and Wang correction for serial correlation) confirms a trend toward increasing aridity, although its magnitude varies among stations. A pronounced negative trend was detected at the Kordai station (Modified Mann–Kendall Z = −5.93, p < 0.001), indicating a statistically significant increase in climatic water deficit in the upper part of the basin. The Sen’s slope estimator was −0.0032 SPEI/month, approximately −0.038 SPEI/year (Table 2).

A statistically significant negative trend (Modified Mann–Kendall Z = −2.53, p = 0.011) was also detected at the Tole Bi station, although its intensity was lower. The Sen’s slope estimator was −0.0017 SPEI/month (approximately −0.020 SPEI/year), indicating a gradual increase in climatic water deficit.

No statistically significant trend was detected at the Moiynkum station (Modified Mann–Kendall Z = −1.19, p = 0.233). Although the Sen’s slope estimator (−0.0008 SPEI/month) suggests a weak tendency toward increasing aridity, this trend was not statistically significant.

Overall, the obtained results indicate that aridization processes in the Shu River basin are spatially heterogeneous. The most pronounced and statistically confirmed increase in climatic water deficit is observed in the upper part of the basin (Kordai), while the middle part of the basin (Tole Bi) is characterized by the most severe and prolonged drought episodes. In the central arid zone (Moiynkum), the long-term trend is weaker, but the high frequency of short-term droughts indicates the ongoing climate vulnerability of the area. The obtained results are consistent with the general trend of increasing climatic water deficit in the arid and semi-arid regions of Central Asia and provide the basis for further analysis of the impact of climate change on the formation of river runoff in the Shu River basin.

Correlations between river runoff and hydroclimatic indicators

Correlation analysis revealed significant differences in the response of the upper and lower reaches of the Shu River to hydroclimatic conditions (Figure 7). Spearman’s rank correlation coefficient was used to assess these relationships, as the hydrological and climatic series do not exhibit a strictly normal distribution and contain significant interannual variability.

For the upper reaches, represented by the Tashutkul gauge, the relationship with climatic parameters is quite clear. The highest correlation was obtained with the SPEI-12 index (ρ = 0.73; p < 0.01), indicating a close relationship between river runoff and long-term moisture availability. This result is entirely expected, since runoff formation in the upper part of the basin is determined not by short-term weather fluctuations, but by the overall hydroclimatic conditions over an extended period.

The correlation with the SPEI-3 index also remains significant (ρ = 0.60; p < 0.05), but is somewhat weaker than for SPEI-12. This suggests that short-term humidity changes influence the river water content but are not determinant of the outcome. A similar correlation was found for the climatic water balance (WB), where the correlation coefficient was 0.61 (p < 0.05). In other words, in years with a more favorable balance of precipitation and atmospheric losses, river runoff volume consistently increases.

A negative correlation was found for potential evapotranspiration (ET 0) (ρ = −0.44), consistent with general hydroclimatic logic: an increase in evaporative demand is accompanied by a decrease in available water resources. However, this correlation is not statistically significant. A likely explanation is that runoff formation in the upper reaches is influenced not only by evaporative demand, but also by the overall water regime of the basin, which is better reflected by integrated indicators such as WB and SPEI.

The lower reaches present a fundamentally different picture. For the Ulanbel gauge, no significant correlations with climatic indicators are observed. The connection with the climatic water balance is practically nonexistent (ρ = 0.01), a similar situation is observed for SPEI-3 (ρ = 0.05). Even for SPEI-12, where the coefficient is somewhat higher (ρ = 0.39), statistical significance is not achieved.

These findings suggest that in the lower reaches no longer reflects the direct climatic signal as clearly as in the upper reaches. As water moves downstream in the basin, the climatically determined hydrological response undergoes significant transformation. Considering the previously identified significant reduction in flow between Tashutkul and Ulanbel, as well as intense agricultural water use in the middle reaches, this pattern is consistent with the hydrological setting of the basin. A significant amount of water is redistributed within the basin before reaching the downstream sections. As a result, the observed water flow does not reflect the influence of climatic conditions, but rather the combined effect of them with water abstraction, channel losses, and internal hydrological transformations.

Overall, correlation analysis shows that the mechanisms of flow formation in different parts of the Shu River basin vary significantly. The upper reaches retain a pronounced dependence on hydroclimatic conditions, while the lower reaches are characterized by a significantly more transformed regime, in which the direct influence of climatic factors is weakened.

The results of this study demonstrate a pronounced spatial transformation of the hydrological regime within the Shu River basin. River runoff in the upper reaches remains strongly associated with hydroclimatic conditions, whereas this relationship becomes substantially weaker downstream, where intensive agricultural water use increasingly modifies the natural flow regime. This spatial differentiation indicates that the dominant controls on runoff generation vary considerably along the river continuum, reflecting the combined influence of climatic forcing and anthropogenic water abstraction.

The present findings are generally consistent with previous studies conducted in other Kazakhstani river basins. Rakhimova et al. (2020) quantitatively demonstrated that climatic factors dominate runoff variability in the upper Buqtyrma River Basin, whereas anthropogenic activities exert a stronger influence downstream. Although our study does not provide a formal quantitative attribution, it identifies a similar spatial pattern using complementary hydroclimatic indicators. Unlike Rakhimova et al. (2020), whose objective was to quantify the relative contributions of climatic and anthropogenic drivers using the climate elasticity method, the present study focuses on evaluating spatial differences in runoff controls along the river continuum. Thus, the two studies address complementary rather than identical research questions.

A similar pattern has been reported in other transboundary river basins of Central Asia. For example, Didovets et al. (2021) showed that climate change is expected to substantially alter water availability across major Central Asian river basins, while increasing water demand further amplifies regional water stress. Likewise, Shahgedanova et al. (2020) demonstrated that glacier retreat and changing climatic conditions modify river discharge in the Northern Tien Shan, emphasizing the growing vulnerability of mountain-fed river systems. These studies are consistent with our finding that runoff in the upstream mountainous areas remains strongly associated with hydroclimatic conditions, where snowmelt and glacier melt continue to influence seasonal discharge.

However, the downstream sections of the Shu River exhibit a fundamentally different response. The absence of significant correlations between runoff and hydroclimatic indicators suggests that anthropogenic regulation increasingly overrides the natural climatic signal. This observation agrees with recent findings from the Yellow River Basin, where decades of reservoir construction, ecological restoration, and intensive irrigation have profoundly altered natural runoff processes despite ongoing climatic change (Wang et al., 2025). In the Yellow River, increasing evapotranspiration together with expanding water withdrawals has substantially reduced downstream water availability. Although the Shu River basin differs considerably in size and hydrological characteristics, both systems illustrate how intensive water management can progressively weaken the direct relationship between runoff and hydroclimatic conditions in heavily managed arid basins.

Comparable evidence has also been reported for other inland river systems. Gu et al. (2024) found that runoff changes in the Manas River Basin of northwestern China resulted from the combined effects of climatic variability and human intervention, while recent research on the upper Shule River showed that nearly half of the observed runoff changes could be attributed to anthropogenic stressors, despite climate remaining the dominant driver of long-term hydrological evolution. These similarities indicate that the interaction between climate variability and human water use represents a common characteristic of many arid and semi-arid river basins rather than a phenomenon unique to the Shu River.

The present findings therefore contribute to a broader understanding of hydrological transformation in mountain-fed transboundary rivers. Unlike previous studies that primarily examined climate-induced runoff changes or future discharge projections, our results demonstrate that the relative importance of climatic and anthropogenic drivers changes considerably along the river continuum. The upper reaches preserve a hydrological system strongly associated with hydroclimatic conditions, whereas downstream sections increasingly reflect the cumulative effects of irrigation withdrawals, channel regulation, and water redistribution. This longitudinal shift in runoff controls provides additional regional evidence that sustainable management of the Shu River basin requires simultaneous consideration of both hydroclimatic processes and water management practices rather than treating these drivers independently.

The hydroclimatic analysis demonstrated a pronounced spatial gradient of increasing aridity from the upper mountainous areas toward the downstream lowlands of the Shu River basin. The progressive increase in reference evapotranspiration (ET 0), together with decreasing precipitation and an increasingly negative climatic water balance, indicates that atmospheric water demand has intensified throughout the basin. These findings are consistent with the general pattern of climatic aridization reported for Central Asia, where rising air temperatures and enhanced evaporative demand have become the dominant drivers of increasing water scarcity during recent decades (Sun et al., 2022; Lyu et al., 2021).

Our results are also in agreement with studies conducted in neighboring mountain-fed river basins. Gu et al. (2024) reported that increasing temperatures in the Manas River Basin substantially enhanced atmospheric evaporative demand, leading to reduced water availability despite relatively modest changes in precipitation. Similarly, Rahimi et al. (2025) observed that climate warming altered runoff generation in the Zagros Mountain region by increasing evapotranspiration and reducing effective water yield. Comparable responses have also been documented in the Kunhar River Basin of northern Pakistan, where climate change combined with land-use change significantly modified seasonal runoff patterns and reduced future water yield under different climate scenarios. These studies collectively demonstrate that increasing atmospheric evaporative demand has become one of the principal mechanisms controlling hydrological change in mountain-fed river basins.

An important finding of the present study is the considerable spatial variability of climatic water deficit within the Shu River basin. The transition from an average annual climatic water balance of approximately −520 mm in Kordai to nearly −1130 mm in Moiynkum illustrates the strong hydroclimatic contrast between the upper and lower parts of the basin. This pattern reflects not only decreasing precipitation but also progressively increasing evaporative demand under warmer and drier climatic conditions. Similar longitudinal increases in climatic aridity have been reported for several inland river basins in Central Asia and northwestern China, where downstream regions experience substantially greater atmospheric moisture deficits than mountain headwaters (Gu et al., 2024; Lyu et al., 2021).

The SPEI analysis further confirms that drought conditions have intensified across the basin during recent decades. Severe drought episodes identified during 2011–2015 were observed simultaneously at all meteorological stations, indicating that these events were driven primarily by regional-scale climatic forcing rather than by local hydrological anomalies. Similar drought episodes have been reported throughout Central Asia, where increasing temperatures and enhanced evaporative demand have intensified hydroclimatic drought despite relatively small long-term changes in annual precipitation (Sun et al., 2022). Comparable drying trends have also been documented in Turkey (Danandeh Mehr and Vaheddoost, 2020), Brazil (Fernandes et al., 2024), and Nigeria (Bello et al., 2026), suggesting that increasing drought severity represents a widespread characteristic of semi-arid environments under contemporary climate change.

Interestingly, the strongest statistically significant long-term drying trend was detected at the Kordai station, whereas the trend at the already arid Moiynkum station was weaker and not statistically significant. At first glance, this observation appears counterintuitive because the lower basin experiences much more severe climatic aridity. However, similar behavior has been reported in other arid regions, where humid or semi-humid mountain areas often exhibit stronger relative hydroclimatic changes than desert regions that already operate under persistently dry climatic conditions. In such environments, relatively small changes in precipitation or evapotranspiration can produce larger standardized anomalies in the upper basin than in chronically water-limited downstream areas. Consequently, the significant drying trend observed at Kordai likely reflects a rapid shift in climatic conditions within the runoff-generation zone rather than greater absolute aridity.

These findings have important implications for future water availability in the Shu River basin. Because the upper mountainous catchments provide the majority of river discharge, increasing hydroclimatic aridization in these areas may progressively reduce inflow to downstream regions even without further increases in irrigation demand. Consequently, future reductions in water availability are likely to result from the combined effects of decreasing climatic water supply and continued anthropogenic water withdrawals, emphasizing the need for integrated adaptation strategies that simultaneously address both climatic and water management challenges.

One of the most important findings of this study is the clear spatial shift in the dominant controls of river runoff along the Shu River. While runoff in the upper reaches remains strongly correlated with hydroclimatic indicators, particularly SPEI-12 and climatic water balance, these relationships almost disappear in the downstream section. This pattern indicates that the hydrological response of the basin changes considerably along the river continuum and that downstream runoff increasingly shows a weaker relationship with hydroclimatic variability and is consistent with the influence of human regulation and water use.

The strong relationship between runoff and long-term hydroclimatic indicators observed at the Tashutkul station suggests that runoff generation in the upper basin appears to be strongly associated with climatic water availability. Snow accumulation, glacier melt, and seasonal precipitation remain the principal sources of streamflow, making runoff highly sensitive to prolonged moisture deficits. In contrast, the weak correlations observed at the Ulanbel station indicate that climatic signals become progressively masked as water moves downstream through intensively irrigated agricultural areas. Under such conditions, river discharge reflects not only climatic forcing but also water withdrawals, reservoir operation, canal diversions, return flows, and transmission losses.

This interpretation agrees with recent systematic assessments of streamflow attribution. A comprehensive review of 101 studies by Ngwira and Damazio (2025) concluded that separating the relative contributions of climatic variability and anthropogenic activities remains one of the major challenges in hydrological research. Their review emphasized that integrated analytical approaches are required because climate and human activities interact simultaneously rather than independently. The methodology adopted in the present study, combining long-term hydrological observations with hydroclimatic indicators and statistical analyses, follows this integrated perspective and provides evidence of the spatial redistribution of climatic and anthropogenic controls within the Shu River basin.

Similar findings have been reported in several heavily regulated river systems worldwide. In the Yellow River Basin, Wang et al. (2025) demonstrated that reservoir construction, ecological restoration programmes, and rapidly increasing irrigation water use substantially altered natural runoff processes despite continuing climatic change. Likewise, studies in the Shule River Basin showed that although climate variability remained the dominant driver of long-term hydrological evolution, anthropogenic activities accounted for nearly half of the observed runoff changes during recent decades. Comparable conclusions have also been reported for the Buqtyrma River Basin in eastern Kazakhstan (Rakhimova et al., 2020), where climatic and anthropogenic influences jointly determined runoff variability. These studies indicate that interactions between climate and water management represent a common feature of arid river basins and should therefore be considered simultaneously rather than separately.

Although the present study did not explicitly quantify the percentage contribution of climate change and human activities to runoff reduction, the observed spatial differences strongly suggest that agricultural water abstraction plays a major role in modifying downstream hydrological conditions. This interpretation is supported by the substantial reduction in river discharge between the upstream and downstream gauging stations together with the documented volumes of irrigation water withdrawals. Future studies may further strengthen these conclusions by applying formal attribution methods such as the Budyko framework, elasticity analysis, or process-based hydrological modelling to quantify the relative contribution of individual driving factors.

Several limitations of the present study should also be acknowledged. The present study is based on statistical relationships between runoff and hydroclimatic indicators and therefore does not constitute a formal attribution analysis. While the observed spatial differences provide evidence of changing runoff–climate relationships within the basin, they do not allow the exact quantitative contribution of climatic variability and anthropogenic water use to be determined. Future studies should combine the present hydroclimatic framework with formal attribution approaches, such as the Budyko framework, climate elasticity analysis, isotope techniques, or process-based hydrological modelling, to quantify the contribution of individual driving factors.

From a water management perspective, the results emphasize that adaptation strategies should not focus exclusively on climate change. Increasing irrigation efficiency, improving canal conveyance systems, reducing transmission losses, and optimizing reservoir operation are likely to provide immediate benefits for downstream water availability. At the same time, continued climatic aridization in the upstream runoff-generation zone may progressively reduce natural inflows, suggesting that future water management should simultaneously address both climatic adaptation and sustainable allocation of available water resources.

Compared with our previous study (Bakbergenov et al., 2026), which focused on recent hydrological variability and water stress indicators, the present study extends the analysis in three important ways. First, it considerably expands the temporal perspective by analysing long-term runoff observations together with multi-decadal hydroclimatic variability. Second, it explicitly investigates spatial differences in runoff controls between the upstream and downstream parts of the basin. Third, it combines runoff observations with climatic water balance and drought indicators to interpret how hydroclimatic variability and agricultural water use jointly influence runoff dynamics. Consequently, the two studies address different scientific objectives and provide complementary insights into hydrological change within the Shu River basin.

Rather than proposing a fundamentally new hydrological mechanism, the present study confirms that spatial differences in the relative influence of hydroclimatic variability and anthropogenic water use, previously reported for other arid river basins, are also evident in the Shu River basin. The main contribution therefore lies in extending these observations to a transboundary Central Asian river system using a comprehensive hydroclimatic assessment and long-term observational data.

Overall, the results suggest that the hydrological evolution of the Shu River basin reflects the interaction of climatic aridization and anthropogenic water use rather than either factor alone. The upper basin continues to function as a runoff-generation area where runoff remains strongly associated with hydroclimatic conditions, whereas the lower basin increasingly reflects intensive human regulation of water resources. This transition from a predominantly natural to a strongly managed hydrological system represents one of the principal scientific contributions of the present study and provides a useful framework for understanding hydrological transformation in other transboundary arid river basins experiencing similar climatic and socio-economic pressures.

This study demonstrates significant spatial transformation of the hydrological regime within the Shu River basin under conditions of increasing aridization and intense water use. River runoff in the upper reaches remains strongly related to hydroclimatic conditions, particularly long-term moisture availability represented by SPEI-12 and climatic water balance.

A pronounced spatial gradient in climatic conditions was identified, with increasing evapotranspiration, decreasing precipitation, and increasingly negative climatic water balance from Kordai to Moiynkum. Drought analysis revealed severe basin-wide drought episodes, particularly during 2011–2015, indicating increasing hydroclimatic stress.

In contrast to the upper basin, runoff in the lower reaches showed weak relationships with hydroclimatic indicators, suggesting substantial modification of the natural flow regime. The results suggest that downstream water availability is associated with both increasing hydroclimatic aridity and intensive agricultural water withdrawals. However, the relative quantitative contribution of these drivers was beyond the scope of the present study and requires further investigation using formal attribution methods. Overall, the Shu River basin exhibits a transition from a runoff system strongly associated with hydroclimatic conditions in the upper reaches to a downstream water-limited system increasingly influenced by agricultural water use.

The results of this study have important implications for sustainable water resources management in the Shu River basin and other transboundary river systems in Central Asia. The observed transition from a hydrological system strongly associated with hydroclimatic conditions in the upper reaches to a downstream water-limited system is consistent with an increasing influence of agricultural water use, indicating that future water management strategies should address both climatic and anthropogenic drivers simultaneously. Adaptation measures focusing exclusively on climate change are unlikely to ensure long-term water security if increasing irrigation demand and inefficient water use continue to reduce downstream water availability.

The strong hydroclimatic control observed in the upper basin suggests that preserving runoff-generation processes in the mountainous headwaters should remain a management priority. Continuous monitoring of climatic variables, snow accumulation, glacier dynamics, and drought conditions is essential for improving seasonal water availability forecasts and supporting adaptive reservoir operation. At the same time, the downstream reaches would benefit from more efficient irrigation practices, modernization of water conveyance infrastructure, reduction of transmission losses, and optimization of agricultural water allocation. Such measures may substantially increase water availability without requiring additional water abstraction from the river.

Because the Shu River is a transboundary basin shared by Kazakhstan and Kyrgyzstan, effective adaptation also requires strengthened bilateral cooperation. Coordinated monitoring programmes, data sharing, and integrated water management policies would improve the ability of both countries to respond to increasing climatic variability and growing water demand. Similar recommendations have been proposed for other transboundary river basins in Central Asia, where sustainable water governance depends on coordinated management of both natural and human-induced hydrological changes.

Rather than assessing recent water stress dynamics, as presented in our previous work (Bakbergenov et al., 2026), the present study provides a long-term hydroclimatic perspective on spatial runoff transformation within the Shu River basin. Together, these studies provide complementary information for understanding both temporal water stress dynamics and spatial changes in runoff controls.

The methodological framework presented in this study may also be applied to other mountain-fed arid river basins experiencing increasing hydroclimatic stress and intensive agricultural water use. The combination of long-term runoff observations, hydroclimatic indicators (ET 0, climatic water balance, and SPEI), trend analysis, and correlation analysis provides a practical approach for identifying spatial differences in runoff controls and evaluating hydrological transformation under changing environmental conditions. The present study should therefore be regarded as a regional case study that expands the observational basis for understanding runoff transformation in arid transboundary river basins rather than as a formal quantitative attribution analysis.

Future research should expand the present analysis by integrating process-based hydrological models, remote sensing products, glacier and snowmelt observations, and detailed information on irrigation water withdrawals and reservoir operation. Such integrated approaches would enable quantitative attribution of runoff changes to individual climatic and anthropogenic drivers and improve the reliability of future projections under different climate and water management scenarios. Ultimately, these advances would contribute to more resilient and sustainable management of water resources in the Shu River basin and other arid transboundary river systems facing similar environmental challenges.