Transformation of seasonal distribution of river flow in the Zhaiyk–Caspian water basin under changing climate conditions

Studying the transformation of intra-annual runoff under conditions of climatic uncertainty is a critical issue for several reasons. First, changes in the seasonal distribution of runoff can significantly affect water availability at different times of the year, which is important for planning water abstraction, irrigation, hydropower, and other uses of water resources. Second, such changes affect the ecological systems of rivers and riparian zones. Understanding these changes is necessary to conserve biodiversity, maintain natural cycles, and keep ecological systems healthy.

Third, climate uncertainty leads to unpredictable changes in weather patterns. Studying runoff transformations helps to develop adaptation strategies that can reduce the negative impacts of climate change. Fourth, changes in river flow regimes can lead to increased risk of floods or droughts. Understanding these changes allows the development of measures to prevent and mitigate such natural extremes.

Fifth, water resources have a direct impact on industry, agriculture, drinking-water supply, and other key sectors of the economy. Changes in the distribution of runoff can affect the well-being of the population and the economic stability of the region under consideration. Sixth, flow transformation data are an important basis for the development of national and regional water strategies and policies, which helps to inform decisions aimed at the sustainable use of water resources.

Studying the intra-annual distribution of river flow is an important scientific and practical task in the field of organizing economically efficient and environmentally safe water use. The range of issues related to this topic is of interest from the point of view of scientific knowledge and for solving applied tasks. This statement finds much evidence both in the form of fundamental hydrological works that summarize research data and create a theoretical basis for further work in this area (Andreyanov 1960; Yevstigneyev 1990; Frolova et al. 2015), and in regulatory documentation that sets strict algorithms for calculating the intra-annual flow distribution for applying the obtained information in various areas of the economy (Set of Rules 2004; Frolova et al. 2015). Based on calculations of the intra-annual flow distribution, water management parameters such as the following are determined: guaranteed water release, electricity generation, and regulating reservoir capacity.

The problem of taking into account intra-annual variability of water flow often arises at the stages of planning measures to improve the ecological condition of small- and medium-sized rivers when rationalizing the use of their water resources. This need is also dictated by the need to adapt the complex water management system to seasonal changes in flow under climate influence. Comprehensive studies were carried out in the study basin in the 1970s during the period of publication of ‘Surface Water Resources of the USSR’ and in 2010–2012 during the period of publication of the 30-volume monograph ‘Water Resources of Kazakhstan: Assessment, Forecast, Management’ (Surface Water Resources of the USSR 1970; Galperin 2012). In the modern period, the intra-annual flow regime of the rivers of the Zhaiyk–Caspian water economy basin has undergone significant changes due to climate change, because the temperature regime of Kazakhstan, based on observed historical data, shows values consistently above normal (since 2000). Every year for 20 years since 2001 has been at least 0.4 °C warmer than the average for the base period 1961–1990 (Report 2022).

Changes in the flow regime are necessary to take into account within the framework of international cooperation for joint water resources management, as the Zhaiyk–Caspian water basin is a transboundary basin. The main river, the Zhaiyk, is formed in the territory of the Russian Federation, the downstream of the river is concentrated within the Republic of Kazakhstan, and the catchment areas of its left-bank tributaries Or and Elek are transboundary, flowing to the territory of Russia. The rivers flowing from the western part of the Common Syrt, including the Karaozen and Saryozen, are also transboundary. Under conditions of climatic uncertainty, studying the transformation of intra-annual runoff is critical to ensure sustainable water resources management, protect ecological systems, improve human security and welfare, and develop effective strategies for adaptation to changing climatic conditions.

The purpose of this research is to study the transformation of the seasonal distribution of river flow in the Zhaiyk–Caspian water basin under changing climate conditions, determine the estimated intra-annual distribution for the different time-periods, and prepare appropriate cartographic generalizations.

For the calculations of the intra-annual river flow distribution, we used official data from the RSE Kazhydromet (2021) (Long-term data on the regime and resources of land surface waters (LDL)). Data processing was carried out for two periods: from the beginning of instrumental observations until 1973 and the modern period from 1974 to 2021. The selection of the limits of the calculation periods is based on the results of analysis of long-term fluctuations in the monthly flow of the Zhaiyk River Basin, previously performed by JSC (Institute of Geography and Water Security). Hydrological calculations and statistical analysis were carried out using standard packages MS Excel and Statistica. Trends in intra-annual flow distribution were assessed using standard statistical analysis methods. The non-parametric Mann–Kendall test was used to determine and identify the trends in the time series (Mann 1945; Kendall 1975). The Mann–Kendall test is a statistical non-parametric test widely used to analyze trends in time series. There are two advantages to using this test. First, it is a non-parametric test and does not require normally distributed data. Secondly, the test has low sensitivity to sharp breaks due to the heterogeneity of time series. According to this test, the null hypothesis H₀ assumes the absence of trend (the data are independent and randomly ordered) and this is tested against the alternative hypothesis H₁, which assumes the trend presence. The Z-stat test statistic is used as a measure of trend significance. In fact, this test statistic is used to test the null hypothesis H₀. A positive Z-stat value indicates an increasing trend, while a negative Z-stat value indicates a decreasing trend. A significant trend at the 0.1, 0.05, 0.01, and 0.001 significance levels exists when |Z-stat| > 1.64, |Z-stat| > 1.96, |Z-stat| > 2.58, and |Z-stat| > 3.29 at confidence levels of 90%, 95%, 99%, and 99.9%, respectively. Therefore, this method is usually used to detect monotonic trends in a series of environmental data, climate data, and hydrological data (the Z-stat value can be used to determine the trend direction; if the value is ‘ + ’, then there is an increasing trend and vice versa) (Skaugen et al. 2012; Blinov 2022). The maps were created using the software package ArcGIS 10.8.

Calculation of hydrological characteristics from non-homogeneous data. The non-homogeneous character of multi-year series of river flow can be caused by the impact of the following: economic activities at catchments and in river channels and changing climatic conditions. At the initial stage, before determining the calculated values of hydrological characteristics, hydrological observation data are subjected to verification and careful analysis of their completeness and quality. The homogeneity of river flow series was assessed on the basis of genetic and statistical analyses.

Genetic analysis is undertaken to study the pattern of multi-year flow fluctuations and identify the physical reasons for the heterogeneity of the raw hydrological observation data. The main techniques that allowed the possible heterogeneity of flow characteristics to be revealed before the application of statistical methods are as follows: construction of difference and total integral curves and construction of complex plots of hydrometeorological elements. These techniques allowed us to analyze the structure of multi-year fluctuations in river flow and to determine the change in mean values and variability of the series over time for further analysis using statistical methods.

Statistical homogeneity means that the elements of hydrological characteristics and their parameters (mean value, dispersion, coefficients of variation, asymmetry, and autocorrelation of individual parts of the series) belong to the same population. Assessment of the homogeneity of hydrological series of observations by the statistical criteria of Student (for mean values) and Fisher (for variance) consists in comparing the calculated value of the statistic of the criterion for homogeneous consecutive parts of the series, obtained from empirical data, with its critically generalized value at a given significance level, sample size, autocorrelation, and skewness coefficients. As a rule, the significance level is set at 5%, which corresponds to the acceptance of the null hypothesis of time-series homogeneity with a probability of 95%.

Assessment of homogeneity (stationarity) by Student's criterion is carried out by comparing the calculated values of t and critical values of the statistic.

Student's and Fisher's tests are traditional methods for analyzing the significance of differences in samples and testing the null hypothesis. However, their application assumes that the data are independent, i.e., there is no autocorrelation between successive values of time series. Of course, this limitation is critical when analyzing river flows, which tend to be temporally structured and show short-term dependence between values. In conditions where river discharge exhibits autocorrelation, the use of Student's and Fisher's tests may lead to overestimation of the significance level. To overcome this limitation, statistical methods that take into account the temporal dependence of the data have been developed. For example, the modified Mann–Kendall test proposed by Hamed (2008), which takes into account short-term dependence and autocorrelation, making it most suitable for analyzing trends in hydrologic data. In addition, the methods described by Dimitriadis et al. (2021) can be applied to analyze time-dependent variability. This scientific study examines statistical tests that account for structural dependencies in the data and allows the variability of key hydrological-cycle processes to be analyzed at a global level. The approaches described in this scientific study provide more robust results for river flow time-series because they account for short-term dependencies and can provide a more accurate representation of the nature and magnitude of variability and trends. Nevertheless, the application of Student's and Fisher's tests in hydrological studies, including streamflow analysis, has a number of advantages. These tests are widely used because of their simplicity, reliability, and easy interpretation of results. The use of these methods can be justified because for many hydrological time-series, the dependence between successive values is of short term and does not critically affect the results of the analysis, which makes the use of these tests quite acceptable.

To determine the homogeneity or heterogeneity of the river flow time-series, a threshold based on statistical methods of trend and variation analysis was used. In this case, the indicators of average discharges of low-water seasons (summer–autumn, winter low-water periods) are analyzed on the basis of multi-year dynamics. Methods aimed at identifying changes in the intra-annual distribution of runoff (method of difference integral curves and method of total integral curves) are used to identify two different periods. Threshold values, at which part of the series is considered heterogeneous, were established by statistical analysis at the confidence level where significant differences between the mean values of seasonal flow indicate structural changes in the water regime.

The method of determining estimated flow values from composite distribution curves is applicable to a heterogeneous total series of observations including two homogeneous samples of different lengths. In case of heterogeneity of initial data of hydrological observations, empirical and analytical distribution curves are determined separately for each homogeneous part of the series.

At present, the main method is based on the construction of probability distributions of water discharge (mean annual, maximum, and minimum) based on available observations of flow and then extrapolation of distribution curves to the area of low or high availability. It should be noted that extrapolation of empirical distribution curves obtained from a short series of observations to the area of small and large supplies leads to large errors. The magnitude of these errors depends on the type of theoretical distribution curves, the method of determining their parameters, and the length of available flow-observation series. Risk reduction in determining the maximum design discharge is achieved by introducing guarantee corrections (Code of Rules 2004); calculated values can be increased by a correction of up to 20%.

To reduce the element of subjectivity when extrapolating the availability curves of the characteristics under consideration (maximum discharge and maximum level), theoretical curves are used, but when processing series of maximum water discharges, the upper points corresponding to the highest discharges deviate upwards from the theoretical availability curve. Increase of the asymmetry coefficient (selection of Cs) does not always correct the situation; as a rule, the curve obtained as a result of such actions already deviates from the main mass of points, the reason being that the empirical security of the upper points significantly deviates from the theoretical curve. Due to the different conditions of high and low flood formation, the series of water discharges are often heterogeneous. That is, two parts of the ranked series obey different distribution laws. In these cases, it is doubtful whether a single supply curve can be successfully selected for the whole series, regardless of whether the distribution law is log-normal or stepped.

There is another way of ‘fitting’ the theoretical curve to empirical data. These are truncated distribution curves and when applying these curves, we achieve correspondence of empirical points to the theoretical curve only for the part of the distribution we are interested in. For high discharges and water levels, this is the upper part of the ranked series. The possibility of using truncated distributions was envisaged in Chen (2014) and in Code of Standards and Rules (1983) for a heterogeneous series, although no recommendations for its application are given there. In the new Russian Code of rules for design and construction (Galperin 1999; Code of Rules 2004), the use of truncated distributions is recommended for heterogeneous series of maximum flow. However, the proposed methodology is far from controversial. In particular, the following are suggested: fixed truncation point and use of normal and gamma distributions only.

It should be noted that the Pareto–Burr–Feller family of distributions presents significant advantages (Dimitriadis et al. 2021) when analyzing river flow processes (accounting for rare, extreme events; accounting for highly variable, unbiased river flow data; adapting this distribution to different hydrological conditions). Nevertheless, while the Pareto–Burr–Feller distribution has a number of advantages, it also has a number of disadvantages. First, the high flexibility of parameters requires complex calibration, which may lead to statistically uncertain results. Second, the computational complexity requires significant computational resources, which can be a limitation of large datasets.

Under the same conditions of climate and other elements of the geographical landscape within several closely spaced river basins, variations in annual flow usually correspond. However, it can be expected that due to the instability of moisture transport pathways in different seasons of the year and throughout the annual time-interval, the correspondence of flow changes of different rivers will be disturbed. A characteristic feature in the distribution of average perennial river flow over the territory is its zonality, which is most clearly expressed in the plain parts. In addition to zonality, another characteristic feature in the distribution of mean perennial river flow over the territory is the distinctly pronounced influence of the relief on this distribution.

A characteristic feature of the region is the decreasing water availability of rivers downstream. The rivers of the region under the conditions of water regime formation belong to the Kazakhstan-type with sharply expressed predominance of spring flow. They are fed mainly by snow melt water. Rivers are fed by groundwater to a small extent.

In general, the rivers of the region are in synphase, i.e., only phases of the water cycle (high water or low water) with a shift in the beginning and end within a few years. All rivers in the region show an increase in flow in the 1980s, followed by a decrease since 2005; this change in flow corresponds to a sharp increase in temperature since the 1980s. The fluctuation of flow is also consistent with the trend of anomalies of annual precipitation totals in the territory of the Zhaiyk–Caspian water basin: in the 1980s in atmospheric precipitation there was an increase in the trend until the 2000s, and after that there was a decrease in the trend. The flow fluctuations of the rivers of the eastern part of the Caspian Lowland and the southern rivers are synchronous. There is an increase in flow until the early 1990s, then a decrease in flow after the 1990s, followed by an increase in flow until the 2000s.

The main flow of the region is in the spring period. The main climatic factors determining the magnitude of spring and, consequently, annual river flow are snow accumulation in the river basin at the beginning of melting and rainfall during the flood period. In spring, with rapid snowmelt, high floods can be formed at medium and even low snow reserves. Autumn soil moistening and soil freezing also play a significant role in this process. Thus, with equal winter–spring precipitation, the flood flow due to different autumn moistening of soils can differ twice.

The choice of the design period in a changing climate is a very difficult task. First, flow is influenced by economic activity, which has different scales in different parts of the territory and varies considerably over time. Second, information on river water use is almost always insufficient.

Regarding the assessment of water resources, the following should be noted:

• - It is reasonable to calculate river flow resources in two variants: for a multi-year period and for the period characterizing the current climate phase and the current level of anthropogenic influences on flow (1940–2021, 1974–2021).

• - The beginning of the multi-year calculation period is inevitably determined by the hydrological study of the basin.

• - The present-day climate phase is mainly characterized by the period since 1974.

The study of the water regime of rivers is of fundamental and applied importance for the Zhaiyk–Caspian water economy basin, which is characterized by low water availability, especially in dry years. The spatiotemporal variability of the water regime complicates the organization of cost-effective and environmentally safe use of the rivers’ water resources in the considered region.

The main share of the river flow in the study basin occurs in the spring season, which is a characteristic feature of rivers with the Kazakhstan-type of water regime. Long-term dynamics of river flow for certain seasons of the year demonstrate that the water regime of the rivers of the considered basin is characterized by certain seasonal transformations, primarily reflected in reduction in the share of spring flow and in increase in the share of runoff in low-water seasons (summer–autumn and winter low water).

To understand the changes taking place, first, it is required to analyze the factors influencing the flow formation. All factors can be divided into two groups. The first group includes physical and geographical factors that are fairly constant. The second group includes fluctuating hydrometeorological factors that directly depend on the particular season conditions. For example, the main hydroclimatic factors that determine the magnitude of the spring flow of the rivers of the considered basin are as follows: transition of air temperature toward positive values, amount of precipitation, snow reserves in the river basin at the beginning of the high water, intensity of snowmelt, rainfall during the high-water period, autumn moisture, and soils’ freezing depth in the watershed. Let us consider each factor separately.

Climate change theory is based on extensive scientific research linking changes in climate conditions to various factors, including human activities. The main aspects of this theory are global warming, greenhouse gas emissions, and their impact on atmospheric composition. Scientific research points to the possible influences of solar activity, volcanic activity, and geological processes on climate. However, despite their importance, the anthropogenic factors such as greenhouse gas emissions remain dominant in modern climate changes.

Therefore, there is no doubt that the climate of our planet is variable and various hypotheses, combined into three main categories, represent the main theories about climate change (cosmic factors (change in the solar constant); astronomical factors (change in the eccentricity of the Earth's orbit, inclination of the ecliptic plane); and geological factors (changes in the poles of the Earth's axis, land relief, and movement of continents)) (Schwarzbach 1955).

Climatic variability represents some temporary fluctuations, which to varying degrees affect the processes occurring in various spheres of the Earth. In climate variability, there is an undoubted direction of the process, on which is superimposed a series of cycles of different durations unstable in time. Climate change is associated with the evolution of the Earth as a planet and, in particular, the fact of alternation of long warm (non-glacial) and cold (glacial) periods has been confirmed. Ice ages have occurred six times over a billion years, the last of which, the Cenozoic, continues today (Raaben 1976; Yasamanov 1985). Information on changes in fluctuations in the average temperature of Europe during the Cenozoic era over the past 60 million years is given in the studies of Sergin & Sergin (1978). At the beginning of the period, the average temperature of the Earth's surface was approximately 10 °C higher than at the beginning of the Pleistocene. In the Pleistocene, the range of temperature fluctuations increased, conditions for continental glaciation appeared, and self-sustained oscillations arose in the ocean–atmosphere–continental ice-sheets system, expressed in the alternation of glacial and interglacial periods. Significant climate changes have begun to occur over relatively short periods of time. In the current post-glacial period, changes in climatic characteristics have been oscillatory; there have been two climatic optima – the ‘Iron Age cooling’ and the ‘Little Ice Age’ (Gribbin & Lem 1980), the latter ending in the middle of the 19th century.

Within relatively short periods, an oscillatory process predominates – secular and intrasecular cycles. Consequently, in principle, the climate is never stationary; we are always in some phase of its fluctuations. Only in limited periods of time can it be conditionally considered stationary (‘piecewise stationary model’) (Leonov 1989). Consideration of changes in modern climate from the point of view of analysis of the dynamics of globally averaged air temperature at the Earth's surface shows the following (Kislov 2001): first, these are changes from year to year (inter-annual fluctuations); second, fluctuations with a characteristic rhythm of about ten (or several tens of) years (‘decadal fluctuations’), a typical example of which is the warming of the 1940s and cooling in the 1960s; third, the global temperature trend.

Currently, from immediately after the ‘Little Ice Age’, a warming phase has begun. In the 1930s and 1940s, the warming was quite pronounced, especially in high latitudes; it was sometimes called ‘Arctic warming’. In the 1950–1960s, the warming there seemed to have stopped (and, importantly, with a delay from high to low latitudes); there was even talk about changing the phases of the cycle, but then in the 1970s and later, the warming became obvious and sharp.

In the early 1980s, the warming was interrupted by the consequences of the El Chichon eruption (Mexico, 1982), then continued with some years with those decades even being called ‘record warm’ for the entire period of instrumental observations. At the same time, the prevailing opinion has been that ‘global’ warming has an anthropogenic cause – the accumulation of greenhouse gases in the atmosphere and the resulting greenhouse effect. Drozdov (1992) indicated the exact date of the start of a new warming wave – from 1973, which is confirmed by some hydrological data from Kazakhstan (Galperin & Moldakhmetov 2003).

One of the important characteristics of river flow is its long-term dependence, known as the Hurst phenomenon (Hurst 1951). This phenomenon plays a key role in the long-term variability of flow and requires special attention when studying the trends of river systems, especially in a changing climate.

It should be noted that fractional-integrated models and methods help improve forecasts by effectively removing seasonal and long-term components (Pizarro et al. 2022). This study emphasizes the analysis of structures and long-term relationships, which is a promising approach for examining temporal and spatial scales of river flow (this approach allows for the investigation of multi-decadal climatic and hydrological changes). In principle, new approaches to water management are possible when studying the interaction of climate variability, taking into account the models proposed in the study (Guo et al. 2024).

Analyses based on the Hurst phenomenon have shown the presence of long-term trends in river flow changes that are related to global climatic changes, and the inclusion of fractionally integrated models will make it possible to clarify the temporal characteristics of flow and identify significant deviations from the traditional forecast values of seasonal flow. In particular, the increase in the proportion of winter flow is possibly related to long-term climatic fluctuations, as supported by similar studies (Pizarro et al. 2022).

Studying the transformation of seasonal river flow in the Zhaiyk–Caspian water basin, taking into account all factors influencing its change, is a key element for understanding modern hydroclimatic processes. Under conditions of climatic uncertainty, a comprehensive study is needed, which includes analysis of temperature anomalies, changes in precipitation, changes in the nature of snow cover, and other hydrometeorological parameters. An important aspect is the impact of anthropogenic activities such as regulation of rivers and changes in land use. Given the interrelationship of all these factors, the study helps to further understand how the hydrological regime of rivers is changing (in particular, the decreasing proportion of spring flow and increasing proportion of low-water flow). An integrated study is especially important for adaptation of water management in the region under consideration to modern climatic conditions and ensuring sustainable water use in the Western region of Kazakhstan.

It should be noted that for the period before 1973, at the meteorological stations of the Zhaiyk–Caspian water economy basin, no significant increase in air temperature was observed in February–March, with the exception of MS Dzhanybek (March) and MS Novyi Ushtogan (March).

Regarding atmospheric precipitation in the continents in extratropical latitudes, a positive trend is noted (10% of the annual amount). Reductions in precipitation occurred in tropical North Africa (10%–25% of the annual amount), this phenomenon is known as the Sahel and Southeast Asian drought (10% of the annual amount). Less pronounced, but clearly detectable, is the increase in precipitation in the southwestern part of the subtropics of the North American continent. The western part of Australia is experiencing an increase in precipitation, while the eastern part is experiencing a decrease from 50% to 100%. In the territory of Russia (Bulygina et al. 2000), the situation is the following: in winter, precipitation increased in Western Siberia and decreased along the entire Arctic coast. In spring, in Western Siberia, the Urals, and the east of the European part of Russia a tendency of increasing precipitation was detected.

The results of the Mann–-Kendall Z-stat test for time series of atmospheric precipitation for the modern period after 1974 revealed the following picture:

• - No significant trends were identified in the West Kazakhstan region in January. In February, significant positive trends were observed at three meteorological stations – Aksai (Z-stat = 2.52), Urda (Z-stat = 1.85), and Uralsk (Z-stat = 1.64). In March, a significant positive trend was observed at almost all meteorological stations in the region; no negative significant trends were identified for March. In April, significant positive trends were observed at two meteorological stations – MS Kamenka and MS Uralsk. From May to July inclusive, no significant trends were identified – neither positive nor negative. In August, significant negative trends were identified at three meteorological stations (MS Uralsk, MS Urda, and MS Dzhambeity). No significant trends were identified for September, October, and November (unlike the period before 1973, for this period in November, significant positive trends in atmospheric precipitation were identified at almost all meteorological stations in the region). A significant negative trend in atmospheric precipitation in December was detected at one meteorological station, Chapayevo (Z-stat = −2.20).

• - In the Atyrau region, a significant positive trend was identified only at one station – Atyrau; in February, a significant positive trend was also identified at one meteorological station – Inderborskiy. In March, significant positive trends were identified at almost all meteorological stations in the region; the Z-stat value varied from 1.86 to 3.08. In May, positive trends were also noted at almost all meteorological stations in the region; the Z-stat value varied from 1.64 to 2.12.

• - In the Aktobe region in January–February, no significant trends were identified, with the exception of MS Emba (January – negative significant trend) and MS Karaulkeldy (February – positive significant trend). In March, a significant positive trend was identified for all meteorological stations, and Z-stat values varied from 1.70 to 3.96. No significant trends were identified in April. In May, a significant positive trend was identified at MS Karaulkeldy; no significant trends were identified at other meteorological stations. In June, significant negative trends were identified at four meteorological stations (Karaulkeldy, Rodnikovka, Irgiz, and Aktobe). In July, no significant trends were identified with the exception of MS Nura (a significant positive trend was identified). In August, a significant negative trend was revealed at five meteorological stations (Novorossiysk, Kos-Istek, Aktobe, Temir, and Rodnikovka). From September to December, no significant trends were identified with the exception of MS Ayakkum (a significant negative trend was identified in October).

The amount of autumn precipitation plays a key role in the formation of water flow during the first month of winter low water, as it determines the water content of the previous period. Climate warming has led to an increased role of thaws in the formation of winter flow. Thus, since 1974, there has been a significant change in a number of flow-forming climatic indicators (increase in the sum of positive air temperatures (February–March, August, October) and increase in the amount of precipitation (March)), which has naturally led to intra-annual unevenness in river flow and to increased low-water flow. Consequently, changes in atmospheric precipitation affected the river flow regime, groundwater regime, and soil moisture reserves. Starting from the second half of the 1970s of the 20th century, the following changes occurred in the intra-annual distribution of river flow: an increase in water content in low-water months and a decrease in the flow of spring high water.

The territory of the Zhaiyk–Caspian water economy basin belongs to an area with stable snow cover (from 2.5 to five months a year). The exception is the Atyrau region, where stable snow cover exists in less than 50% of winters. Snow cover in the considered basin plays a very significant role in moistening the soil, contributing to the river flow and lakes and replenishing groundwater reserves. The territory of Western Kazakhstan comprises areas of insufficient moisture, which is reflected in the formation of surface runoff. The main share of the river flow in the study basin occurs in the spring season, which is a characteristic feature of rivers with the Kazakhstan-type of water regime. Surface runoff within the study area is formed almost only by melted snow waters. Rains in hot summer conditions and very dry soils are largely lost to evaporation and have no practical significance in the river and temporary watercourse flow in the area. The groundwater supply of watercourses is extremely small and often completely absent, which is associated with the deep occurrence of groundwater and slight incision of river valleys.

Consequently, due to the exclusive role of snow in the process of surface runoff formation, the volume of annual river and temporary watercourse flow in the considered basin is almost completely determined by the volume of spring flow. The main factor in the formation of spring floods in the rivers of the considered basin, as well as in other areas of lowland Kazakhstan, is the snow cover of the river basin. The considered territory includes the following natural landscape zones: Mugodzhar Mountains, Subural Plateau, and Caspian Lowland.

As a result of the assessment and analysis of the initial information, the main characteristics of the snow cover were obtained. The maximum values of snow-cover depth exceed the long-term average values by approximately two times (averaged value for the territory). Averaged minimum values of snow depth differ from maximum values by seven times. The greatest differences in the maximum values of snow-cover depth from the long-term average values are observed at the following meteorological stations: MS Aksai, MS Kamenka, MS Dzhambeity, MS Ilyinskiy – three times, and MS Chapayevo – four times. The depth of snow cover varies significantly in certain years, during the winter, and also depending on the location of the meteorological station.

For a number of meteorological stations, a slight decrease in the values of the maximum snow-cover depth is observed, such as at MS Novorossiysk, MS Ilyinskiy, MS Uralsk, and MS Martuk. For a number of meteorological stations, there is a slight increase in the values of the maximum snow-cover depth, such as at MS Rodnikovka, MS Aksai, MS Kamenka, and MS Dzhambeity.

A comparative assessment of the obtained data was made with the data presented in the scientific and applied reference book on climate of the USSR (Climate Reference Book 1989). Comparative analysis of changes in the timing of formation and destruction of stable snow cover and the duration of its occurrence was carried out for two periods: the first period is from the beginning of instrumental observations until 1973 and the second is the modern period after 1974.

In the considered region, the average date of the first appearance of snow cover (the first ten days of November) in the area of Martuk and Chingirlau meteorological stations has remained practically unchanged (shifting toward earlier dates by one day). In the area of meteorological stations Aktobe, Rodnikovka, Chapayevo, and Uralsk, the average date of the first appearance of snow cover has shifted toward earlier dates by seven, six, four, and two days, respectively. According to the Dzhambeity meteorological station, the appearance of temporary snow cover has shifted toward later dates, from seven to nine days.

The earliest date of appearance of the first snow cover according to the meteorological stations Novorossiysk (11 days), Martuk (eight days), and Chingirlau (two days) has shifted toward later dates. At the meteorological stations Aktobe (six days) and Chapayevo (three days), the situation is the opposite; the earliest date of appearance of the first snow cover has shifted toward earlier dates.

The latest date of appearance of the very first snow cover over the territory has undergone significant changes. According to the meteorological stations, these dates have moved earlier by 29 days (MS Uralsk and MS Chingirlau). In the area of Martuk and Novorossiysk meteorological stations, the most extreme date has shifted, on the contrary, to later dates by 12 and 11 days, respectively.

Regarding the average date of stable snow-cover formation, a comparative assessment showed that for almost all meteorological stations, the average date has moved toward an earlier date from two to eight days. In the area of Aktobe and Novorossiysk meteorological stations, the average dates of stable snow-cover formation remained unchanged.

The earliest date of stable snow-cover formation, according to the vast majority of meteorological stations, has shifted toward earlier dates from three to 17 days. The latest date of the stable snow cover formation has moved toward a later date at the following meteorological stations: Aktobe (two days), Chapayevo (four days), Rodnikovka (six days), Martuk (eight days), Uralsk (eight days), and Chingirlau (11 days). In the area of meteorological stations Novorossiysk (nine days) and Dzhambeity (25 days), the latest date of stable snow-cover formation moved toward an earlier date.

In the territory of the Zhaiyk–Caspian water basin, the destruction of stable snow cover occurs from the south to the north direction. This is due to the following reasons: physical and geographical features, the intensity of spring solar radiation, advection of warm air masses, and the value of snow water equivalent. The process of stable snow-cover destruction in the considered territory occurs on average in the last ten days of March and the first ten days of April. In recent decades, the start of stable snow-cover destruction has begun to occur earlier. The average date of stable snow-cover destruction at most meteorological stations has shifted to earlier dates from two to ten days, with the exception of the area of Chingirlau meteorological station (the average date of stable snow-cover destruction has shifted to later dates, by four days).

Regarding the stable snow-cover destruction, all dates have shifted to earlier in comparison with climatic norms: the average date, the earliest, and the latest. For example, according to the Chapayevo meteorological station, the average date is by ten days, the earliest is by nine days, and the latest is by eight days. According to the Dzhambeity meteorological station, a similar situation is observed: the shift in the average date is by five days, in the earliest is by ten days, and in the latest is by seven days.

The complete snow-cover melting over the considered territory occurs very unevenly, both in terms of thermal conditions and relief conditions. The picture of complete snow-cover melting in the area of Uralsk meteorological station has changed as follows: the average date has shifted earlier by four days, the earliest date has shifted earlier by 21 days, and the latest date has shifted earlier by one day.

The average number of days with snow cover in the considered territory for the modern period ranged from 60 days (MS Atyrau) to 162 days (MS Kos-Istek). According to observations of the Martuk, Novorossiysk, and Chapayevo meteorological stations, the number of days with snow cover increased by one day. In the area of Aktobe meteorological station, the duration of stable snow cover increased by five days, and in the areas of Rodnikovka and Dzhambeity meteorological stations, on the contrary, it decreased by one and two days, respectively. The more significant reduction in the duration of snow cover is observed in the area of Uralsk and Chingirlau meteorological stations; the number of days with snow cover decreased by four days. At the same time, all characteristics of snow-cover dynamics indicate a stable trend toward their change (Figure 4(b)).

Snow cover, which is the main source of runoff formation in the considered territory, is characterized by its relatively small thickness. The average snow water equivalent value by the beginning of its melting within most of the study area ranges from 10 to 180 mm. In the snowiest winters, the snow-water equivalent value reaches from 40 to 420 mm, and in winters with little snow, it reaches up to 65 mm.

For the assessment of the degree of stability of the snow-cover period, G.D. Richter proposed a stability coefficient (C_sc), determined from the ratio of the difference between the greatest and shortest durations of snow cover to its average long-term duration in days (Uteshev 1959; Table 1). Conventionally, the maximum stability of snow cover according to this formula is taken to be a coefficient that is close to zero. The distribution of this coefficient reflects a natural decrease in stable snow cover from north to south direction across the territory of Kazakhstan.

Snow cover is characterized by the greatest stability when the stability coefficient is 0.40 (MC Martuk, C_sc = 0.44). The stability coefficient is 1.0 or more when winters are characterized by a large number of thaws and low snow-depth, which corresponds to extremely unstable snow cover (MS Kaztalovka, MS Chapayevo, MS Dzhambeity, MS Zhalpaktal, MS Makhambet, MS Peshnoy, and MS Atyrau).

Consequently, the instability of snow cover is one of the typical features of the winter landscape of the considered region. The main reason for this is the temperature regime of winters, characterized by frequent thaws, as well as the insignificant depth of snow cover.

One of the key factors forming the spring flood is the depth of soil freezing.

Time series of soil freezing depth anomalies at meteorological stations located in the Atyrau and Mangistau regions are not provided due to the insufficiency and lack of observational data. Anomalies are calculated relative to the modern period since 1974. The calculated trends indicate a significant trend toward a decrease in the depth of soil freezing in recent decades.

The maximum depth of soil freezing in the Aktobe region was observed at the Martuk meteorological station (120 cm, 1976–1977). The minimum depth of soil freezing in the Aktobe region was observed at the Rodnikovka meteorological station (9 cm, 1991–1992).

In the West Kazakhstan region, the maximum depth of soil freezing was observed at the Yanvartsevo meteorological station (123 cm, 1976–1977). The smallest depth of soil freezing in the West Kazakhstan region was observed at the Taipak meteorological station (5 cm, 2015–2016).

The development of hydrotechnical construction in Kazakhstan, which was accompanied by the construction of the largest reservoirs and cascades of hydroelectric power plants, led to fundamental changes in the natural hydrological regime of most river systems, formed under the influence of natural conditions. Problems of river surface-flow management are solved through the creation and operation of reservoirs of multi-year, seasonal, weekly, and daily regulation, which ensure redistribution of water resources in time. Reservoirs of multi-year and seasonal flow regulation are mainly created and operated on large rivers of the republic. The problem of changes in annual river flow under the influence of reservoirs is well enough studied (Galperin & Moldakhmetov 2003). However, the decisive influence in river flow regulation by reservoirs is manifested, first of all, in the intra-annual flow distribution in the closing gauging station. Here, the role of reservoirs is to eliminate natural flow irregularity, with an increase of flow volume in low-water periods due to reduction of flood flow (Makhmudova et al. 2023a, b).

The flow of rivers in the Zhaiyk–Caspian water basin is widely used in various sectors of the economy. Large reservoirs and many small ponds have been built on some rivers, under the influence of which the flow tends to reduce due to losses from filling the reservoirs and additional evaporation from the water surface, as well as water intake for the needs of various sectors of the economy. The main way to guarantee water supply to the economy and population in territories with uneven river flow is accumulation of water volumes in reservoirs. In the basin of the Ural River in the Russian Federation, the largest reservoirs are located in the upper reaches of the main river (Verkhneuralskoye, Magnitogorskoye, and Iriklinskoye), meeting the water needs of industrial enterprises and public utilities in the Southern Trans-Urals. A total of 12 reservoirs with a volume of more than 10 million m³ are located in the Ural River Basin in the Russian Federation, including Iriklinskoye (with a volume of 3,260 million m³), Verkhneuralskoye (with a volume of 601 million m³), and Magnitogorskoye (with a volume of 174.0 million m³) reservoirs. There are also 129 reservoirs with a volume of less than 10 million m³. Reservoirs allow the regulation of flow by cutting the peaks of spring floods and increasing the low-water flow (Table 2).

At present, 50 water reservoirs (with a capacity of more than 1 million m³) have been constructed in the basin of the Zhaiyk River in Kazakhstan for the use of river flow in economic sectors according to the Scheme of Integrated Use and Protection of Water Resources, with a total full volume of about 1.4 km³ and with a usable capacity of about 1.1 km³. Together, 20 reservoirs have a capacity of more than 10 million m³. Perennial flow regulation is carried out by the following four reservoirs: Aktobe (W_-use = 220 million m³), Kargaly (W_-use = 262 million m³), Magadjanovskiy (W_-use = 11.9 million m³), and Uysylka (W_-use = 21.2 million m³).

Almost all water reservoirs are used for agricultural needs (irrigation, watering), but Aktobe reservoir has a complex purpose. Information about operating reservoirs is given in Table 3.

Table 3 also includes additional information on 27 water reservoirs identified by the staff of JSC ‘Institute of Geography and Water Security’ during the 2024 field surveys. Water reservoirs are fundamental natural and man-made elements of hydrotechnical and water management systems at any level. They make it possible to regulate water resources of rivers and lakes to the extent necessary for sustainable development of the economy and population. In this connection, creation of water reservoirs has been widespread both in Kazakhstan and all over the world.

There are 297 hydraulic structures in the territory of the Zhaiyk–Caspian water basin (Table 4). Administratively, the water basin includes the whole of the West Kazakhstan, Atyrau, Mangistau provinces, and most of Aktobe province (without the territory of the Tobyl–Torgai water management basin). In terms of area, the Zhaiyk–Caspian water basin is the largest, occupying 23.5% of the territory of the Republic of Kazakhstan.

In the West Kazakhstan region, water reservoirs are mainly located on the Kushum canal and channels of the Zhaiyk River, hence do not directly affect the flow of the Zhaiyk River itself. On the Ural River (Zhaiyk River) in the territory of the Russian Federation since 1957, the large Iriklinskoye water reservoir of multi-year regulation with a usable capacity of 2,160 million m³ has been operating. Thus, the change in the annual flow of the Zhaiyk River at the site of the village of Kushum is due to the intake of water from the reservoir into the Kushum canal, as well as other water intake structures located along the length of the river, and the influence of the Iriklinskoye water reservoir.

There are five large reservoirs of perennial regulation in Aktobe province: Aktobe on the Elek River, Kargaly on the Kargaly River, Magadjanovskoye on the Sugursu River, Uysylka on the Sarymyrza River, and Sazdinskoye on the river of the same name. The largest water reservoir, Kargaly, has been in operation since 1975, and Aktobe reservoir since 1988. The flow of the Elek River near Shelek village has been influenced by these two reservoirs since they started operation.

Spring high water is the main phase and distinctive feature of the water regime of the Zhaiyk–Caspian river basin. In the spring season, most of the annual river flow is formed. To study the characteristics of the high water of the main rivers of the Zhaiyk–Caspian water economy basin, an analysis of the annual high-water volumes, flood duration, and the long-term dynamics of maximum water flows was carried out (Table 5).

After the end of the spring high water, the summer–autumn low water begins. The flow share during the summer–autumn low-water period is 10%–20% of the annual flow. Against the background of a general increase in water content during low-water periods, changes in minimum monthly water flows are the most obvious and dramatic. The increase in summer–autumn minimum average monthly water flows is 5%–15%.

The obtained results indicate an increase in the minimum flow values (this trend was identified for most rivers in the basin, regardless of water content and the duration of the observation series). There is a stable tendency toward an increase in the values of the limiting season with maximum dynamics in the modern period.

The duration of the period of minimum flow is determined, first by the stability of the low water. Despite the identified tendency of increase in the share of winter flow, the characteristic of the river water regime of the Zhaiyk River Basin remains the excess of minimum summer–autumn discharges in comparison to winter discharges.

The maximum water discharge for the spring-flood period, minimum monthly discharge for the summer–autumn low-water period, and minimum monthly discharge for the winter low-water period were estimated for the present-day period of climatic changes, 1974–2021. Availability curves were constructed and statistical characteristics of availability curves were determined (mean value for the period under consideration, coefficient of variation, coefficient of asymmetry, and discharges of different availability). The parameters of the supply curves were selected during the construction of the supply curves. Statistical parameters of the maximum runoff distribution of the main rivers are given in Table 6. Statistical parameters of the minimum flow distribution for the summer–autumn low-water period of the main rivers are given in Table 7. Statistical parameters of the minimum flow distribution for the winter low-water period of the main rivers are given in Table 8.

Thus, the analysis of the long-term river flow dynamics for certain seasons of the year showed that the water regimes of the rivers of the considered basin are characterized by certain seasonal transformations, primarily in reduction in the share of spring flow and in increase in the flow share in low-water seasons (summer–autumn low water and winter low water).

There are two periods for this indicator. The first period (1936–1958) is distinguished by significant discharges in the summer–autumn season and is quite low in the winter season. In the second period (1959–2021), the amplitudes of inter-annual variations and differences in the flow of low-water seasons decreased significantly, which indicates a certain equalization of the shares of summer–autumn and winter discharge.

As noted earlier, the main reason for the increase in the share of winter flow is the positive dynamics of the frequency and duration of winter thaws, a characteristic of the territory of the studied basin. As a result of an increase in the proportion of liquid precipitation in winter, favorable conditions are appearing for the additional infiltration recharge of groundwater and natural increase in river flow (Dzhamalov et al. 2017).

Thus, spatial–temporal fluctuations in the river flow of the studied basin are reflected in the intra-annual aspect, in the formation of extreme hydrological phenomena – low or high water. In this regard, the study of current trends in the intra-annual distribution of lowland river flow is relevant for assessing the stability of the hydroecological situation and solving the problems of guaranteed water supply in limited seasons.

Western Kazakhstan experienced a severe flooding situation in 2024, which demonstrated the region's vulnerability to extreme hydrological events. This event emphasized the importance of studying river flow trends in the face of climatic uncertainty and increasing anthropogenic factors.

Floods are the result of natural causes and a variety of human activities and are therefore both a natural and a social phenomenon. The main natural causes of floods are hydrometeorological phenomena such as river surges, mudflows, prolonged rainfall, and downpours (Blöschl et al. 2019). A flood is defined as a significant increase in the water-carrying capacity of rivers that occurs annually in the same season and is accompanied by an increase in the water level of the river. The cause and timing of floods depend on the geographical location of the river and are related to the inflow of water into the river channel as a result of snowmelt in the plains, and snow and glacier melt in the mountains. During spring snowmelt, rivers overflow on the plains, with water levels rising up to 20 m and the width of the flooded area reaching many kilometres.

In Kazakhstan, floods occupy a significant place among natural hazards causing significant damage. There are about 800 rivers (over 50 km long), more than 570 water reservoirs of various capacities and purposes, and about 3,000 large lakes, the presence of which may be associated with floods. In recent decades, floods have become more frequent on the rivers of Kazakhstan, which is primarily due to anthropogenic causes: unsatisfactory flow regulation, development of flood-prone areas, and conflicts of interests of water users, including at the interstate level.

According to studies (CAREC 2022), the significance of flood risk in Kazakhstan is further supported by estimates of average annual human losses. More than 390 deaths per year are recorded, which is two-thirds higher than in any other Central Asia Regional Economic Cooperation Program (CAREC) member state, while floods are expected to affect on average more than 156,000 people each year. Flood risk in Kazakhstan is much more pronounced than earthquake risk, as heavy rainfall and snowmelt cause significant damage. The average annual flood losses are estimated at $419 million in the USA. The spatial pattern of flood risk is diverse, with damage exceeding US$ 30 million in many areas of northern, southern, and central Kazakhstan. Over a 100-year recurrence period, flood losses are modeled at US$ 1.8 billion, which is approximately 1% of the gross domestic product. In most of the reviewed works (Blöschl et al. 2019; Heinrich & Penning-Rowsell 2022; Akiyanova et al. 2023; Liu et al. 2023; Makhmudova et al. 2023a, b; Sanders et al. 2023), flood hazard is investigated as a function of hydrological factors, which leads to disruption of economic activity. In the process of floodplain development, flooding and waterlogging of floodplains automatically start to be categorized as floods of varying degrees of hazard, recurrence, and so on.

Floods caused by spring flooding occur when there is abundant snowmelt. The intensity of floods is largely determined by the nature of snow-cover melting, as well as the depth of soil freezing. The most dangerous situation is the rapid melting of the heavy snow cover after winter, which is accompanied by heavy precipitation in the form of rain. In this case, all melt and rainwater flow down the frozen ground directly into the river, and as a result, floods can be catastrophic. Given the vastness of river floodplains and their active multipurpose use, as well as the rapidity of spring-flood peaks, it is objectively difficult to adequately assess the damage caused by flooding.

Anthropogenic causes of floods are related to human economic activities. They can be divided into direct and indirect activities. Indirect activities include those that take place in river basins, valleys, floodplains, and channels and can cause changes in their water regimes. Direct anthropogenic causes lead directly to large floods and are related to various hydraulic engineering activities and dam failures, as well as improper flood protection measures.

When analyzing hydrological observations over the entire period of instrumental observations, the most extreme values were observed in 1942, 1957, 1994, 2011, and 2024.

According to the multi-year course of average annual values of water discharge at gauging stations along the flow of the River Zhaiyk (Kazakhstan territory), there is a trend toward a decrease in the annual values of flow at hydrological stations on the River Zhaiyk, for the period from 2005 to 2021, compared with previous periods. The same trend is observed for the maximum flow for a multi-year period, i.e., since 1921, there is a decrease in the maximum values of water discharge at all hydrological stations of the River Zhaiyk. It should be noted that the share of flood flow in the annual volume according to the data for the modern period in the considered water management basin is as follows: on the River Zhaiyk, 65%–70% (decrease by 5%–10%); on the River Shagan, 70%–75% (decrease by 10%); on the River Elek, 50%–55% (decrease by 20%–40%); and on the River Or, 70%–75% (decrease by 10%–15%). Almost all rivers in the basin under consideration show a decrease in the volume of flow during the spring flood.

Consequently, there is a decrease in maximum spring-flood discharges. The reduction of maximum water discharges is as follows: River Zhaiyk, up to 40%; River Shyngyrlau, up to 65%; River Shagan, up to 45%; River Derkul, up to 15%; River Olenty, up to 25%; River Ulken Kobda, up to 40%; River Elek, up to 60%; River Or, up to 35%; River Zhem, up to 25%; and River Oyil, up to 40%. This trend is caused by the increase in winter air temperature, which is accompanied by an increase in the number and duration of thaws and, consequently, a decrease in pre-spring water reserves and maximum spring-flood discharges.

Taking into account the spring flood of 2024, the values of the highest levels for the period of instrumental observations on the River Zhaiyk have changed (Table 9). The analysis of Table 9 shows that, at the hydrological posts of Yanvartsevo, Makhambet, and Atyrau, the highest water levels for the whole multi-year period were observed exactly in 2024. At the hydrological posts of Uralsk, Koshim, and Taipak, the highest water levels are the levels of 1942.

In order to prevent the flood situation arising this year in the Zhaiyk–Caspian water basin, operational monitoring of water bodies during the pre-flood period (both in the territory of the Russian Federation and the Republic of Kazakhstan) is necessary. In addition, monitoring of the major tributaries of the Elek and Shyngyrlau Rivers is needed to forecast maximum discharges and their time to reach the Zhaiyk River. More detailed analyses require monitoring data on the tributaries, which are currently insufficient.

For the city of Uralsk, the issue of monitoring and forecasting the Shagan and Derkul Rivers is relevant, because the flooding period on these rivers is short term (up to 80% of the annual runoff occurs during the flooding period) and with a small catchment area, the city is flooded. Consequently, it is necessary to conduct stationary, permanent monitoring of the channel water balance from the border of the Republic of Kazakhstan and the Russian Federation to the Caspian Sea. In this regard, there is a need to carry out this kind of work throughout the entire plain of Kazakhstan, as comprehensive hydrological studies are needed to ensure water security of the country. It should be noted that the scientific research conducted by Wang et al. (2023) provides an opportunity to use distance data as a key indicator to assess the vulnerability of populations. Analysis of distances between settlements and areas at high flood risk is necessary in the development of adaptation strategies to minimize loss and damage. In addition, understanding the relationship between the level of flood protection and the behavior of the population allows risk factors to be taken into account when planning protective measures. In principle, the approach proposed in the study by Wang et al. (2023) can be useful for developing adaptation measures and flood-risk management.

As a result of analyzing the transformation of the seasonal distribution of river flow in the Zhaiyk–Caspian water basin, the following can be noted:

• - In recent decades, the intra-annual flow distribution of the rivers of the Zhaiyk–Caspian water basin has changed significantly (a statistically significant decrease in the unevenness of the intra-annual flow distribution for the period after 1974 was revealed); in the rivers of the basin under consideration, the change in the spring flood has reached the following scales: in the River Elek basin, the share of spring runoff has decreased to 40%; in the River Or basin, it has decreased to 15%; and in the River Shagan basin, it has decreased to 10% of the annual flow value.

• - Intra-annual redistribution of flow was expressed in significant degradation of floods as a phase of the water regime; the main cause of these changes is the increase in winter air temperatures and the increase in the number and duration of winter thaws, which leads to a decrease in pre-spring water reserves.

• - Maximum water discharges were also significantly changed (reduction of maximum water discharges of: River Zhaiyk to 40%, River Shyngyrlau to 65%, River Shagan to 45%, River Derkul to 15%, River Olenty to 25%, River Ulken Kobda to 40%, River Elek to 60%, River Or to 35%, River Zhem to 25%, and River Oyil to 40%), and these changes particularly emphasize the transformation of the water regime, which requires adaptation of water management systems;

• - Changes have affected not only the share of spring floods, but also the share of seasonal flow distribution as a whole: the share of flow of low-water seasons – summer–autumn and winter low-water periods – has increased, which indicates a significant redistribution of water during the year. This requires revision of approaches to water resources planning and management, taking into account new climatic conditions and changes in water regime,

• - Spatial and temporal fluctuations of river flow in the studied basin are manifested in the intra-annual aspect, in the formation of extreme hydrological phenomena – low water or high water. In this regard, the study of current trends in the intra-annual distribution of flow of plain rivers is relevant for assessing the stability of the hydroecological situation and solving the problems of guaranteed water supply in limiting seasons.

The results of the analysis show that under conditions of increasing climatic uncertainty, integrated water resources management becomes critical to ensure the sustainability of ecological systems in Western Kazakhstan and to prevent possible negative consequences for the socioeconomic and environmental situation.

We thank the Head of the Water Resources Laboratory, Professor Aisulu Tursunova, the Editor-in-Chief, Professor K. Srinivasa Raju, the Editorial Board of the journal, Ms Lucy Ibboston, the IWA publishing team, and the anonymous reviewers for their kind support during the preparation of the paper.

This research was funded by the Committee of Science of the Ministry of Science and Higher Education of the Republic of Kazakhstan Grant No. AP19678734 ‘Assessment of the current and predicted hydrological changes of Kazakhstan river basins based on modelling (ex. Buktyrma, Esil, Zhaiyk rivers)’.

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

The authors declare there is no conflict.