Current Trends and Future Scenarios: Modeling Maximum River Discharge in the Zhaiyk–Caspian Basin (Kazakhstan) Under a Changing Climate

Abstract

In the context of intensifying climate change, it is particularly important to assess the transformation of spring floods as a key phase of the hydrological regime of rivers. This study provides a comprehensive analysis of the characteristics of maximum runoff in the Zhaiyk–Caspian basin for the modern period and projected changes for 2030, 2040, and 2050 based on CMIP6 climate scenarios (SSP3-7.0 and SSP5-8.5). Analysis of observations at 34 hydrological stations showed a reduction in spring runoff by up to 35%, a decrease in the duration of high water and a reduction in maximum water discharge on some rivers by up to 45%. It has been established that those rising temperatures, more frequent thaws, and reduced autumn moisture lead to lower maximum water discharge and a redistribution of the seasonal flow regime. Scenario projections revealed significant spatial heterogeneity: some rivers are expected to experience an increase in maximum discharge of up to 72%, while others will see a steady decline in maximum discharge of up to 35%. The results obtained indicate the need to transition to an adaptive water management system focused on the regional characteristics of river basins and the sensitivity of small- and medium-sized watercourses to climate change.

1. Introduction

Assessing current changes in maximum river discharge and forecasting them under conditions of climate uncertainty is one of the key tasks of modern hydrology. In semi-arid regions such as the Zhaiyk–Caspian basin, maximum discharge accounts for the bulk (up to 80–90%) of annual discharge and occurs during the relatively short spring period [1,2,3,4]. Seasonal unevenness of runoff increases the region’s vulnerability to climate change and must be taken into account in water resource management and flood risk management.

Research on the impact of climate change on the hydrological regime of rivers, especially on changes in maximum runoff, is being conducted by scientists around the world [5,6,7,8,9,10]. Studies conducted in North America confirm that anthropogenic climate change contributes to increased precipitation intensity, which increases the likelihood of flooding [11]. Another study published in the journal Hydrology and Earth System Sciences analyzes the impact of climate change on 100-year floods in Bavaria, Germany. The results show an increase in the frequency and intensity of extreme floods, which is associated with rising temperatures and changes in precipitation patterns [12]. A global analysis conducted as part of the RheinBlick2050 project [13] assesses the impact of climate change on the flow of the Rhein River and its main tributaries. The results indicate an increase in winter water consumption and a decrease in summer consumption, which could affect water management and flood protection in the region.

Studies of river flow fluctuations caused by climate change in Russia [14,15,16,17] did not reveal any significant long-term trends in annual runoff, but the distribution of runoff within the year by season has changed significantly over the past decades. In particular, runoff during the winter low-water period has increased significantly. This increase was associated with a rise in the frequency of thaws, which led to frequent flooding during the winter low-water period, while spring runoff and floods from snowmelt decreased due to the depletion of water reserves in the snow before the spring season. According to scientific research [18], in recent decades, an increase in river runoff during the winter months has been observed in most catchment areas in the Baltic Sea basin, while pronounced negative trends have been identified in the spring and summer months. Similar seasonal changes in runoff are also observed in the Ural River basin in Russia. The results of an analysis of long-term variability confirm the trends toward a decrease in spring floods and a simultaneous increase in winter runoff, which is associated with an increase in the frequency of thaws and changes in precipitation patterns [19,20]. Scientific studies [21] emphasize that a modern trend of transformation of the intra-annual distribution of runoff is observed in the Ural River basin: the share of low-flow (especially winter) runoff is increasing, and the contribution of spring floods is decreasing [22]. This leads to a levelling of the annual hydrograph and a change in the water regime of Kazakh-type Rivers. These seasonal changes exacerbate hydrological instability and increase the likelihood of extreme floods. A particularly striking example was the flood situation in the spring of 2024 in the Zhaiyk River basin [23,24,25]. Early snowmelt combined with heavy precipitation led to large-scale flooding that affected the border areas of Russia and the western regions of Kazakhstan. According to ReliefWeb and ACAPS, tens of thousands of people were evacuated in Kazakhstan and Russia during the spring floods of 2024, hundreds of settlements were flooded, and a state of emergency was declared in 10 regions of Kazakhstan [26,27]. Scientific studies [28,29] confirm that extreme runoff values are intensifying due to changes in climatic conditions, including early snowmelt, increased winter precipitation, and rising air temperatures. These processes exacerbate the risks of both floods and droughts, making it necessary to transition to more accurate scenario-based water resource forecasting using new-generation climate models such as CMIP6.

Scenario-based forecasting of maximum runoff using CMIP6 ensembles allows for consideration not only of climate trends, but also of a wide range of uncertainties associated with socio-economic trajectories and greenhouse gas emissions [30,31,32]. CMIP6 provides extensive opportunities for analyzing climate parameters in various socio-economic development scenarios, ranging from moderate (SSP1) to high emissions (SSP5). However, GCM results need to be adapted to the regional scale through correction methods, as the models have relatively low spatial resolution and do not always reproduce local features of runoff formation [33]. To obtain high-resolution source data, statistical and dynamic downscaling methods or a hybrid of both are used [34,35].

In this regard, dynamic downscaling of global GCMs is a critically important tool for improving the accuracy of regional climate projections in conditions of complex orography and high spatial heterogeneity of climate processes characteristic of the region. This is especially true for the reproduction of extreme convective precipitation, which is sensitive to detailed orographic and thermodynamic contexts. Despite certain systematic errors in dynamic downscaling, it provides physically consistent regional estimates, enabling high-quality climate data describing the entire range of climate variables with a high degree of reliability. Among all these international climate databases, only the Inter-Sectoral Impact Model Intercomparison Project (ISIMIP) (https://www.isimip.org/) currently provides modeling results that have undergone dynamic downscaling based on CMIP6 scenarios for the Central Asian region [36]. As part of this initiative, CMIP6 climate models were further calibrated to account for regional orography and climate conditions, providing more reliable and applicable assessments for our region. This approach improves the accuracy of climate risk assessments, facilitates the development of adaptation measures, and provides a reliable basis for climate-informed decision-making at the regional level. These advantages are confirmed by the results of the analysis presented in the study [36], which shows that the application of dynamic resolution downscaling to ISIMIP data significantly improved the accuracy of precipitation reproduction compared to the original CMIP6 models. In addition, the study demonstrates that the adjusted data more accurately reflect key climate trends, such as temperature increases and changes in precipitation patterns in Kazakhstan and thus provide a reliable basis for climate projections in regions with complex orography.

To assess future climate change in the Kazakh part of the Zhaiyk–Caspian basin, high-resolution climate data with a spatial resolution of 0.25° × 0.25° and dynamically downscaled CMIP6 models from the third phase of ISIMIP were used within two socio-economic pathways (SSP3-7.0 and SSP5-8.5) [37].

In scenario-based forecasting of maximum runoff, changes in seasonal precipitation and temperature patterns play a key role, as climatic conditions in autumn, winter, and spring largely determine future runoff formation. The controlling factors can be grouped into two categories: relatively constant physiographic factors (basin size, relief, geological structure, soil composition, vegetation, and anthropogenic influences) and hydrometeorological factors (snow water reserves, snowmelt intensity, autumn soil moisture, soil freezing, evaporation, and precipitation). These variables determine the volume, timing, and peak discharge of spring floods, as well as the shape of the hydrograph [38,39,40].

The objective of this study is to analyze the current trends in maximum spring flood discharge across the rivers of the Zhaiyk–Caspian basin and to develop scenario-based forecasts of its potential changes under future climate change conditions. The research aims to improve the understanding of hydrological responses to climatic variability and to support the adaptation of water management practices in the region. The scenario forecast is carried out for three-time horizons: 2030 (2026–2035), 2040 (2036–2045), and 2050 (2046–2055). The period 1991–2020 was selected as the base period. The scientific novelty of this study lies in the comprehensive assessment of climatic factors influencing the formation of spring runoff using a scenario-based approach. The practical significance of the results is determined by their applicability for long-term water management planning, adaptation to climate risks, and the sustainable management of water resources in the studied region.

2. Materials and Methods

2.1. Study Area

The distribution of the river network in the Zhaiyk–Caspian basin is determined by the presence of the Caspian Sea in the southwest and the mountainous formations of the Southern Urals in the northeast, which is why the rivers flow in a general direction from northeast to southwest. There are more than a hundred rivers (temporary watercourses) in the basin under consideration, including 12 rivers with a length of more than 200 km. The main river is the Ural (Zhaiyk), with a total length of 2534 km. The river is formed in the Russian Federation, originates in the Southern Urals, and flows into the Caspian Sea. The length of the river within the Republic of Kazakhstan is 1084 km (Figure 1) [2].

The main focus of this study is on the assessment of changes in the spring flood and its hydrological characteristics. The spring flood represents the dominant phase and a distinctive feature of the river flow regime within the Zhaiyk–Caspian basin, accounting for 70–90% of the total annual runoff. According to the classification proposed by B.D. Zaikov [41], the rivers of the basin mainly belong to the Kazakhstan type of hydrological regime, which is characterized by a single, sharp, and short flood wave caused by the melting of the seasonal snow cover. After the flood peak, during the summer–autumn period, the river discharge significantly decreases, and in some cases, smaller rivers may temporarily dry up. The peak of the spring flood generally occurs in April–May, consistent with the natural and climatic conditions of the region.

The location of the Zhaiyk–Caspian basin inland, far from oceans and high mountain systems, causes a continental, arid climate on its territory, characterized by a high amplitude of diurnal and seasonal temperatures, a small amount of unevenly distributed precipitation, high solar radiation, and relatively low humidity. The climatic conditions of the basin largely determine the features of river flow formation, including the dominance of snowmelt-fed spring floods typical of rivers of the Kazakhstan type. The distribution of air temperature over the territory of the Zhaiyk–Caspian basin as a whole has a zonal character and changes from north to south. The average annual air temperature varies from 2.8 °C in the north-east to 12 °C in the south of the territory under consideration. The intra-annual course of air temperature is characterized by stable frosts in winter, intense heat build-up in short periods and hot summers. The diversity of landforms and atmospheric processes occurring over the territory under consideration cause significant variability in precipitation. The average annual precipitation varies from 150 to over 400 mm, with the least precipitation falling on the Ustirt Plateau and the southern edges of the Mangystau Peninsula [3].

2.2. Input Data

In this study, official hydrological data from 34 hydrological observation posts and meteorological data on air temperature and atmospheric precipitation from 17 meteorological stations were used to analyze contemporary changes in the maximum flow of rivers in the Zhaiyk–Caspian basin (Figure 2,

Table 1. List of hydrological posts (HP) of the Zhaiyk–Caspian basin.

№	Gauging Station	Observation Period	Watershed Area, km2	Multi-Year Average Discharge, m3/s
I district—left-bank tributaries of the Zhaiyk River
1	Shyngyrlau–Lubenka	1963–present time	641	0.40
2	Shyngyrlau–Kentubek	1940–present time	4660	3.42
3	Elek–Shelek	1940–present time	37,300	32.6
4	Karakobda–Alpaisai	1962–present time	2240	2.32
5	Ulken Kobda–Kobda	1940–present time	8110	4.84
6	Ulken Kobda–Kogaly	1980–present time	14,200	7.84
7	Sarykobda–Bessarabsky	1956–1995	675	0.45
8	Terisakkan–Astrakhanskiy	1956–1995	446	0.38
9	Elek–Rzd № 47	1950–1990	1090	1.12
10	Karagandy–Kandagach	1948–1989	222	0.26
11	Kosestek–Kosestek	1956–present time	281	0.77
12	Aktasty–Belogorskiy	1946–present time	45	0.17
13	Terisbutak–Belogorskiy	1946–1990	19.8	0.09
14	Elek–Aktobe	1938–present time	11,000	14.8
15	Kargaly–Kargalinskoe	1940–present time	5000	9.54
16	Or–Bogetsai	1956–present time	7480	5.43
II district—rivers of the western part of the Common Szyrt
17	Derkul–Taskala	1963–present time	392	0.76
18	Derkul–Beles	1962–present time	1820	4.38
19	Shagan–Kamenny	1931–2010	4000	8.80
20	Shyzhyn 2–Chizha 2	1940–present time	509	1.61
21	Shizhyn 1–Chizha 1	1957–present time	456	1.43
IV district—rivers flowing from the southern part of the Pre-Urals plateau
22	Olenty–Zhympity	1940–present time	1290	1.25
23	Shiderty–Araltobe	1963–present time	750	0.81
24	Buldurty–Karaagach	1963–1990	457	0.43
25	Kaldygaity–Zhigerlen	1940–present time	2510	1.72
26	Shili–Akshy	1956–1992	318	0.32
27	Kopirankaty–Algabas	1940–present time	723	0.76
V district—southern rivers
28	Temir–Leninskiy	1932–present time	5310	3.84
29	Zhem–Zharkamys	1941–1991	26,000	10.8
30	Oyil–Taltogai	1936–1982	17,000	7.21
31	Shigyrlykumdy–aul № 10	1956–1992	1110	0.64
32	Kiyil–Novonadezhdinskiy	1956–1998	720	0.79
33	Ashyoyil–Maimak	1956–1992	4900	0.94
34	Sagyz–Sagyz	1949–1998	9930	1.00

and

Table 2. List of meteorological stations (MSs) of the Zhaiyk–Caspian basin.

№	Meteorological Station	Observation Period	Altitude Above Sea Level, m
I district—left-bank tributaries of the Zhaiyk River
1	Chingirlau	1928–present time	103
2	Aksai	1937–present time	63
3	Novoalexeevka	1937–present time	143
4	Novorossiyskoe	1925–present time	415
5	Ilinski	1958–present time	191
6	Kos-Istek	1956–present time	338
7	Rodnikovka	1930–present time	372
8	Aktobe	1898–present time	219
II district—rivers of the western part of the Common Szyrt
9	Kamenka	1954–present time	72
10	Uralsk	1888–present time	35
IV district—rivers flowing from the southern part of the Pre-Urals plateau
11	Dzhambeita	1897–present time	31
12	Karatobe	1958–present time	44
V district—southern rivers
13	Temir	1894–present time	234
14	Emba	1904–present time	252
15	Uil	1886–present time	95
16	Karauylkeldy	1937–present time	200
17	Sagyz	1954–present time	55

). The source data covers the entire period of instrumental measurements up to and including 2022. All data were obtained from open sources of the Kazhydromet State Agency, available on its official website: https://www.kazhydromet.kz/ru/gidrologiya (accessed on 5 May 2025).

To analyze the dynamics of spring floods, annual data on the following key parameters were used: spring runoff volume, maximum water discharge, start and end dates of the flood, and its duration. The data was processed for two time periods: from the beginning of instrumental observations until 1973 and the modern period from 1974 to 2022. This division is based on the fact that the 1970s are recognized as the onset of the modern stage of climate change, accompanied by pronounced warming trends and a transformation of the hydrological regime in the region [2].

To ensure comparability between these periods and to reveal patterns of long-term variability, data from key hydrological gauging stations with long observation series (starting from the 1940s) were used. These stations, located across different hydrological regions, are representative of the general trends in spring flood formation. In cases where observation series contained gaps or interruptions, missing peak discharge values were reconstructed using the hydrological analogy method. Gauging stations with shorter observation periods were mainly used for scenario-based forecasting, which allowed for a wider spatial coverage and consideration of the recent dynamics of river discharge under changing climatic conditions.

The study used data from natural or semi-natural regimes in the following hydrologically homogeneous areas:

–

I—left-bank tributaries of the Zhaiyk River (Or, Elek, Ulken Kobda, Shyngyrlau, Kargaly, Kosestek, Aktasty, etc.);

–

II—rivers in the western part of the Obshchy Syrt (Shagan, Derkul, Shyzhyn 1, Shyzhyn 2);

–

IV—rivers flowing from the southern part of the Podural Plateau (Olenty, Kupirankaty, Buldurty, Shiderty);

–

V—southern rivers (Oyil, Sagyz, Zhem).

The area of the transboundary Zhaiyk River (III), as well as Areas VI and VII classified as drainless territories, were not included in the analysis.

CMIP6 climate models were used in the study of long-term scenario projections of maximum runoff. The ISIMIP projection uses datasets based on five models from the Climate Model Intercomparison Project Phase 6 (CMIP6), identified as “priority” within the Inter-Sectoral Project on Climate Impacts [42] to cover the diversity of climate sensitivity and provide a diverse set of projections for impact modeling:

–

GFDL-ESM4 (NOAA Geophysical Hydrodynamics Laboratory, USA);

–

IPSL-CM6A-LR (Pierre-Simon Laplace Institute, France);

–

MPI-ESM1-2-HR (Max Planck Institute for Meteorology, Germany);

–

MRI-ESM2-0 (Meteorological Research Institute, Japan);

–

UKESM1-0-LL (Hadley Centre for Meteorological Forecasting, United Kingdom).

This limited number of models was selected to take into account the diversity of climate scenarios and uncertainties on the one hand, and to meet the limitations of computing resources and the volume of data to be processed on the other. The five selected models reliably reproduce the average climate response of the entire CMIP6 ensemble, making them a sufficient basis for assessing climate impacts without the need to use all models.

Three models (GFDL-ESM4, IPSL-CM6A-LR, and MPI-ESM1-2-LR) have low climate sensitivity, while two (UKESM1-0-LL and MRI-ESM2-0) have high climate sensitivity, providing a balanced representation of the CMIP6 ensemble. In addition, these models were selected for their structural independence in ocean and atmospheric components. Their representations of processes have been recognized as reliable by climate impact modeling experts. Furthermore, their selection is consistent with previous ISIMIP phases, ensuring consistency and comparability with previous studies.

2.3. Research Methods

Hydrometeorological calculations and statistical analyses were performed using the standard software packages MS Excel 2021 and Statistica 12.0. Observation series containing missing data were reconstructed using the hydrological analogy method, which is a widely accepted and established approach in hydrological research [43,44,45]. In cases of short observation periods or unrepresentative data, the time series were adjusted relative to a representative reference period.

Reconstruction of Maximum Discharge Series. The main focus of the study was on the analysis of changes in spring floods and their characteristics. To assess current changes in maximum spring discharge, quasi-natural series of maximum runoff were reconstructed for representative gauging stations, taking into account climatic factors. When filling data gaps, analogous rivers and nearby meteorological stations located within the flow formation zone were selected [43,44]. In cases where river basins were influenced by human activity, the river flow corresponding to the natural regime was restored in order to separately identify the contribution of climatic and anthropogenic factors.

Time series analysis. When studying the internal dynamics of time series, it is necessary to identify trends in changes in the maximum flow time series in the form of a linear or parabolic trend. Based on the type of trend with a certain degree of reliability, it is possible to forecast the time series for the long term [46]. To estimate a linear trend, it is proposed to use the least squares method for independent observations [47,48,49]. The regression equation in the case of linear dependence is

$$y_i=a\times x_i+b$$

(1)

where $a$ is the linear trend coefficient characterizing the rate of change in the studied variable; $x_i$ is time, year; $b$ is the level of the series at the initial moment of time.

The essence of the least squares method is to determine the calculated parameters and at which the sum of the squares of the deviations of the observed values from the calculated value using the formula will be minimal.

Methods for assessing scenario-based forecasting of maximum runoff based on statistical analysis. A statistical model—multiple linear regression—was used for scenario forecasting of maximum runoff volume. Statistical analysis methods allow us to investigate the relationships between climate variables (temperature, precipitation) and maximum runoff, as well as to predict possible changes in runoff in the future. Meteorological parameters selected on the basis of both physical and geographical representations of flood formation processes and the results of correlation analysis were used as predictors. Significant predictors were selected based on their degree of correlation with spring runoff volumes. The following meteorological factors were found to be the most significant:

–

total precipitation during the cold period (September–March), reflecting the degree of autumn soil moisture and snow accumulation ($P_{IX-III}$);

–

average air temperature in the period from June to October, determining the water absorption capacity and evaporation losses of the soil layer ($T_{VI-X}$);

–

average winter temperature, affecting the frequency of thaws and the intensity of early snowmelt ($T_{XI-III}$).

The calculation scheme for maximum runoff is shown in

Table 3. Calculation scheme of the maximum runoff of the rivers of the Zhaiyk–Caspian basin.

№	River–Point	Meteorological Station	Calculation Dependencies
I—left-bank tributaries of the Zhaiyk River
1	Shyngyrlau	Novoalexeevka	$Q_{max}=f\left(T_{VI-X};\sum P_{IX-III}; T_{XI-III}\right)$
2	Elek, Or	Aktobe, Kos-Istek, Rodnikovka
3	Kargaly, Kosestek, Terisbutak	Kos-Istek, Rodnikovka
4	Ulken Kobda, Karakobda, Sarykobda, Terisakkan	Novoalexeevka, Ilinski	$Q_{max}=f\left(\sum P_{IX-III}; T_{XI-III}\right)$
II—rivers of the western part of the Obshchy Syrt
1	Derkul, Shyzhyn 2, Shizhyn 1	Uralsk, Kamenka	$Q_{max}=f\left(T_{VI-X};\sum P_{IX-III}; T_{XI-III}\right)$
2	Shagan	Uralsk, Yanvartsevo
IV—rivers flowing from the southern part of the Podural Plateau
1	Olenty	Dzhambeita, Karatobe	$Q_{max}=f\left(T_{VI-X};\sum P_{IX-III}; T_{XI-III}\right)$
2	Kopirankaty	Dzhambeita, Aksai
3	Buldurty, Kaldygaity	Karatobe
4	Shiderty	Dzhambeita, Karatobe, Aksai	$Q_{max}=f\left(T_{VI-X};\sum P_{IX-X};\sum P_{XII-III};T_{XI-III}\right)$
V—southern rivers
1	Temir, Zhem	Emba, Temir	$Q_{max}=f\left(T_{VI-X};\sum P_{IX-III}; T_{XI-III}\right)$
2	Sagyz	Sagyz, Karauylkeldy
3	Oyil	Karauylkeldy, Uil, Temir
4	Ashyoyil	Karauylkeldy, Uil
5	Shigyrlykumdy	Karauylkeldy, Temir

.

3. Results

3.1. Contemporary Changes in the Characteristics of Spring Floods

Spring flood volume. Spring flood volume is the main indicator characterizing the total water discharge in a river during the flood period. Long-term changes in the spring flood discharge of rivers in the studied basin are characterized by certain spatial differences. During the modern period (1974–2022), a reduction in the volume of spring floods has been observed across virtually the entire territory. On average, the volume of floods in the basin’s rivers has decreased by 5% to 35%.

The left-bank tributaries of the Zhaiyk River (Or, Elek, Ulken Kobda, Shyngyrlau). Rivers in this hydrological region are characterized by the greatest reduction in runoff during the modern period (1974–2022) compared to the period before 1973, with an average reduction of 20–30%. The most obvious change in flood discharge is observed for the Ulken Kobda River (a decrease of 35%) and the Elek River (25–30%) (Figure 3).

A study of the long-term dynamics of changes in the volume of spring floods has shown that before 1973, the decrease in spring runoff was more intense, while in the period of 1974–2022, along with a sharp decrease in the volume of spring runoff, the rate of decrease and the amplitude also decreased. For example, for the Elek River in Aktobe, the rate of decrease was 85.1 million m³/10 years until 1973, and 4.9 million m³/10 years during the period 1974–2022—4.9 million m³/10 years, for the Ulken Kobda River—Kobda village, before 1973—35.6 million m³/10 years, for the period 1974–2022—1.8 million m³/10 years (Figure 4).

Rivers in the eastern part of the Caspian Depression that do not flow into the Zhaiyk River (Olenty, Kopirankaty, Buldurty, Shiderty). This hydrologically homogeneous region is also characterized by a significant decrease in the volume of spring floods during the period 1974–2022 compared to the period before 1973. The greatest reduction is observed on the Olenty River, amounting to 30%, which indicates a significant decrease in water inflow in the spring period. A reduction in the volume of spring floods has been recorded on the Kopirankaty and Kaldygaity rivers, but it is less pronounced (Figure 5).

The long-term dynamics of changes in the flood volume of rivers flowing from the eastern part of the Caspian Depression show a mixed trend for the periods under review for the Kaldygaity River—Zhigerlen village, and a slight negative trend for the Kopirankaty River—Algabas village (Figure 6).

Right-bank tributaries of the Zhaiyk River, rivers in the western part of the Obshchy Syrt that do not flow into the Zhaiyk River (Shagan, Derkul, Shyzhyn I, Shyzhyn II). In the basin of the rivers in the western part of the Obshchy Syrt, which form on the territory of the Republic of Kazakhstan, the volume of spring runoff for the period 1974–2022 decreased significantly compared to data before 1973, from 15% to 33% (Figure 7). The long-term dynamics of flood volume changes also confirm a decrease in spring flood volume. The decrease in flood volume on the Shagan River—Kamenny village occurs at a rate of 43.1 million m³/10 years until 1973, 4.9 million m³/10 years for the current period.

For the Shyzhyn 2–Chizha 2 village, a negative trend is observed, 13.3 million m³/10 years until 1973 and a slight decrease of 0.6 million m³/10 years in the modern period, indicating a less pronounced trend of reduction in the volume of spring floods on the right-bank tributaries of the Zhaiyk River (Figure 8).

Southern rivers (Temir, Oyil, Sagyz, Zhem). The following picture can be observed in the southern rivers: the Temir River in the village of Leninskiy and the Zhem River in the village of Zharkamys, where the flood volume decreased by 24%; the Oyil River in the village of Taltogai, where the flood volume decreased by 22% (Figure 9). The long-term dynamics of changes in the flood volume of southern rivers are shown in Figure 10.

An analysis of Figure 10 demonstrates the trends in changes in the flood volumes of southern rivers. Over the long term, the flow volume of the Temir River in the village of Leninskiy until 1973 showed a negative trend of −15.1 million m³/10 years, while in the modern period from 1974 to 2022, a positive trend of 2.2 million m³/10 years was observed. For the Oyil River in the village of Taltogai, a slight negative trend was observed in both periods.

Maximum water discharge. In addition to changes in the volume of runoff during the spring flood, the maximum discharge also decreased in the region under consideration. Maximum water discharge in the rivers of the study area reached significant values in the period before 1973, on the Elek River—about 2500 m³/s, on the Ulken Kobda River—1300 m³/s, and on the Temir River—about 800 m³/s. However, after 1974, a significant decrease in this hydrological characteristic was observed. On the Elek River, the maximum water discharge decreased to 1500 m³/s, on the Ulken Kobda River—to 1000 m³/s, and on the Temir River, the maximum water discharge decreased to 500 m³/s (Figure 11).

An analysis of trends in maximum water discharge shows that a decrease is observed on all rivers in the basin under study. However, the significance of these trends varies depending on the river. The most pronounced trends are observed on the Elek and Ulken Kobda rivers, where the decrease reaches about 5.3 m³/s per year. Less pronounced but still significant trends are observed on the Shagan and Oyil rivers, where the rates of decline are 1.22 m³/s and 2.20 m³/s per year, respectively. The least significant trends are observed on the Temir and Kopirankaty rivers.

Start and end dates of the flood season and duration of the flood. The main elements of the spring flood are the start and end dates of the flood and its duration. The start and end dates of the flood were determined based on hydrographs of the runoff. The beginning of the flood is taken as the first day with a noticeable increase in water discharge, usually preceding a sharp rise in level and discharge, and the end of the flood is taken as the day at the end of the decline curve, when the intensity of the decline has already sharply decreased as a result of the end of the discharge of the main volume of meltwater. Table 3 shows data for the main rivers of the Zhaiyk–Caspian basin on the dates of the beginning and end of the spring flood, its duration for two time periods—from the beginning of instrumental observations to 1973 and for the period 1974–2022.

The results in

Table 4. Characteristics of spring flooding of the rivers of the Zhaiyk–Caspian basin.

River–Point	Period	Flood Start Date, day	Flood End Date, days	Duration of Flood, day	Qmax, m3/s	Volume of Runoff During the Flood, mln. m3
I district—left-bank tributaries of the Zhaiyk River
Ulken Kobda–Kobda	Before 1973	05/IV	08/V	34	411	176
	1974–2022	28/III	03/V	37	175	115
	Δ	−7	−5	+3	−57%	−35%
Elek–Aktobe	Before 1973	01/IV	18/V	49	736	537
	1974–2022	31/III	02/V	33	525	410
	Δ	−1	−16	−16	−29%	−24%
Aktasty–Belogorskiy	Before 1973	02/IV	29/IV	28	15	5.8
	1974–2022	30/III	23/IV	25	10	3.9
	Δ	−3	−6	−3	−35%	−32%
Kargaly–Kargalinskoe	Before 1973	06/IV	14/V	39	505	343
	1974–2022	30/III	29/IV	32	386	266
	Δ	−7	−15	−7	−24%	−23%
II district—rivers of the western part of the Obshchy Syrt
Shagan–Kamenny	Before 1973	02/IV	06/V	35	408	207
	1974–2022	28/III	01/V	35	385	175
	Δ	−5	−5	0	−6%	−15%
Shyzhyn 2–Chizha 2	Before 1973	30/III	01/V	33	79	34
	1974–2022	26/III	15/IV	20	58	23
	Δ	−4	−15	−12	−26%	−33%
IV district—rivers flowing from the southern part of the Podural Plateau
Kaldygaity–Zhigerlen	Before 1973	01/IV	03/V	33	123	43
	1974–2022	25/III	18/IV	25	85	39
	Δ	−7	−14	−8	−31%	−8%
Kopirankaty–Algabas	Before 1973	31/III	30/IV	31	71	23
	1974–2022	26/III	18/IV	24	57	22
	Δ	−5	−12	−7	−20%	−5%
V district—southern rivers
Temir–Leninskiy	Before 1973	01/IV	04/V	34	338	142
	1974–2022	29/III	28/IV	32	269	109
	Δ	−3	−6	−2	−21%	−24%
Oyil–Taltogai	Before 1973	01/IV	12/V	42	320	235
	1974–2022	01/IV	08/V	38	209	182
	Δ	0	−4	−5	−35%	−23%
Zhem–Zharkamys	Before 1973	27/III	14/V	49	454	402
	1974–2022	25/III	02/V	38	335	305
	Δ	−2	−13	−10	−26%	−24%

Note: Δ—indicates the difference between the two periods (shown in bold).

show that most rivers in the Zhaiyk–Caspian basin tend to experience an earlier start and earlier end to spring floods, corresponding to a reduction in the duration of spring floods. On average, the flood season in modern conditions begins 2–7 days earlier and ends 5–16 days earlier than in the period before 1973. The most significant shift in the start dates of the spring flood season is observed on the Ulken Kobda, Kargaly, and Kaldygaity rivers. On most rivers in the study region, the flood ends 5–16 days earlier (the Elek, Kargala, Shyzhyn 2 and Zhem rivers), so the duration of the flood has decreased.

The downward trend in maximum runoff remains stable in all hydrologically homogeneous areas, reflecting current climate change. These changes in spring flood characteristics are due to a number of factors, including higher air temperatures in winter and spring, more frequent thaws, and shallower soil freezing. Together, these processes lead to a reduction in moisture reserves by the start of the flood season. Rapid snowmelt during sharp increases in air temperature often prevents meltwater from reaching the main river channels, while reduced soil freezing depth contributes to additional water losses due to seepage into underground horizons. Such complex and ambiguous variability in runoff is probably explained by the high degree of diversity and multidirectional influence of factors that shape spring floods.

3.2. Contemporary Changes in Climatic Factors Affecting Maximum Runoff

Naturally, the main factors shaping flood levels are a group of climatic factors, among which precipitation plays a leading role, both in solid and liquid form. An increase in solid precipitation contributes to the accumulation of maximum snow reserves, which, in turn, leads to an increase in the volume of spring runoff. Preparation of the catchment area for the formation of spring floods begins in the fall, when the river basin is moistened (the sum of precipitation for the period September–October). During this period, an important factor is the average soil moisture content before the onset of the winter season. If the autumn was dry and the soil moisture content was low, water losses due to infiltration increase in the spring when the snow melts, leading to a decrease in runoff. Conversely, if soil moisture is high in autumn, infiltration losses of surface runoff in spring will be minimal, which contributes to an increase in the runoff coefficient and the formation of higher flood levels.

Another important climatic factor is air temperature. While snow reserves determine the total volume of high water and its peak values, the temperature regime affects the rate of snowmelt. The higher the sum of positive temperatures during the snowmelt period, the more intense the melting process, which leads to a faster rise in water levels and a shorter duration of high water.

Table 5. Coefficients of linear trend of climatic characteristics.

№	Meteorological Station	Period	ΣP for Months IX–X, mm/10 years	ΣP for Months XI–III, mm/10 years	T av for Months VI–X °C/10 years	T av for Months XI–III °C/10 years	ΣT—(Negative Temperatures) for Months XII–III
1	Sagyz	1961–2022	−2.5	1.4	0.4	0.3	48.5
2	Karatobe	1959–2022	−1.5	3.2	0.3	0.5	61.6
3	Ilinski	1960–2022	−0.2	−0.9	0.3	0.4	50.8
4	Kos-Istek	1958–2022	−3.1	7.4	0.2	0.3	32.9
5	Kamenka	1956–2022	−0.3	−3.8	0.3	0.4	51.0
6	Novorossiyskoe	1961–2022	−4.1	−1.4	0.3	0.3	32.9
7	Karauylkeldy	1940–1973	2.6	19.0	0.2	0.2	−8.3
		1974–2022	−0.1	10.4	0.4	0.4	53.5
		1940–2022	−1.1	6.7	0.3	0.4	47.1
8	Emba	1940–1973	4.2	24.0	0.0	0.2	−7.3
		1974–2022	−1.7	−2.4	0.2	0.3	37.9
		1940–2022	−1.6	3.8	0.2	0.3	37.6
9	Uil	1940–1973	2.2	24.2	0.0	0.2	3.1
		1974–2022	0.5	5.2	0.4	0.5	65.7
		1940–2022	−1.4	3.9	0.2	0.4	49.5
10	Temir	1951–1973	−0.5	16.9	−0.2	0.5	−5.9
		1974–2022	−3.2	−1.5	0.3	0.4	55.2
		1951–2022	−2.9	3.9	0.2	0.5	57.1
11	Novoalexeevka	1940–1973	−0.5	12.8	0.1	0.1	−24.8
		1974–2022	−2.5	−2.2	0.4	0.6	51.2
		1940–2022	−1.7	7.2	0.2	0.4	43.7
12	Dzhambeita	1940–1973	−5.7	19.7	0.2	0.2	−21.7
		1974–2022	1.2	−3.9	0.4	0.5	58.9
		1940–2022	−2.3	1.2	0.2	0.5	56.4
13	Aktobe	1940–1973	−0.5	19.7	0.2	0.5	29.1
		1974–2022	−1.6	5.1	0.4	0.5	53.1
		1940–2022	−1.2	11.8	0.2	0.5	57.6
14	Rodnikovka	1940–1973	8.0	40.5	0.0	0.3	3.8
		1974–2022	−4.6	−4.3	0.4	0.4	48.6
		1940–2022	−4.3	5.5	0.3	0.4	44.3
15	Aksai	1940–1973	−2.7	13.6	0.2	0.5	32.5
		1974–2022	−0.04	9.3	0.4	0.5	62.0
		1940–2022	−1.5	7.8	0.2	0.5	60.4
16	Uralsk	1940–1973	−2.9	20.5	0.2	0.4	−2.1
		1974–2022	−0.7	2.1	0.5	0.5	52.1
		1940–2022	0.05	5.6	0.2	0.5	55.6
17	Chingirlau	1940–1973	−3.6	15.9	0.2	0.3	11.3
		1974–2022	−0.9	−1.9	0.5	0.4	47.0
		1940–2022	−2.4	4.3	0.2	0.4	50.2

Note: Statistically significant trends are shown in blue (precipitation) and orange (air temperature).

presents the main climatic factors that shape spring floods, as well as their linear trend coefficients for three time periods: before 1973, 1974–2022, and for the entire observation period. For meteorological stations with less than 20 years of data available before 1973 (Sagyz, Karatobe, Ilinski, Kos-Istek, Kamenka, and Novorossiyskoe), the changes in meteorological parameters were calculated for the entire available observation period.

Precipitation. An analysis of linear trends in total precipitation during the winter period (November–March) showed predominantly positive changes at most meteorological stations in the Zhayik–Caspian basin. The first period (1940–1973) saw the most pronounced increase in winter precipitation, reaching 40.5 mm/10 years at the Rodnikovka MS, 24.2 mm/10 years at the Uil MS, 24.0 mm/10 years at the Emba MS and 20.5 mm/10 years at the Uralsk MS. These values reflect the relatively wet climate phase of the mid-20th century and the increase in moisture accumulation during the cold season.

In the second period (1974–2022), the positive trend continues but becomes less pronounced. The highest values of winter precipitation increase were recorded at the Karaulkeldy MS (10.4 mm/10 years), Aksai MS (9.3 mm/10 years) and Aktobe MS (5.1 mm/10 years). This indicates a continuing but more moderate increase in winter moisture under current climatic conditions. At the same time, some stations recorded a slight decrease in winter precipitation: the Dzhambeita MS (3.9 mm/10 years), Rodnikovka MS (4.3 mm/10 years), Emba MS (2.4 mm/10 years). These differences reflect local climatic features and the influence of orographic factors that determine the uneven distribution of precipitation across the basin.

With regard to precipitation during the autumn months (IX-X), a significant decrease has been recorded since 1974. For example, at the Rodnikovka MS (−4.6 mm/10 years) and at the Temir MS (−3.2 mm/10 years). The decrease in autumn precipitation reduces soil moisture before the onset of winter, which negatively affects the formation of a stable snow cover. In spring, during snowmelt, this leads to increased water losses due to infiltration into the soil layer, resulting in a reduction in spring runoff in almost all rivers in the region.

Air temperature. The analysis showed that the temperature regime of the study area has also undergone tremendous changes—the average air temperature during the winter period (November–March) has increased significantly in the modern period after 1974. For example, before 1973, the air temperature at the Novoalexeevka MS was 0.1 °C/10 years, and after 1974, it rose to 0.6 °C/10 years. At the Uil MS, a similar increase was observed, from 0.2 °C in the period before 1973 to 0.5 °C/10 years after 1974, confirming the general warming trend across the entire study region.

Similarly to the winter season, the average air temperature during June–October (VI–X), which reflects the soil water absorption capacity and evaporation losses, also exhibits a steady upward trend. For 1974–2022, a positive trend was observed at all stations in the Zhaiyk–Caspian basin: from +0.2 to +0.5 °C/10 years. The most intense warming was observed at the Uralsk and Chingirlau meteorological stations, +0.5 °C/10 years. In other cases, the values fluctuate between +0.2 and +0.4 °C/10 years. Higher temperatures during the warm season increase evaporation losses, reduce soil moisture, and limit the moisture reserves that form before winter. This, in turn, can reduce the water content of spring floods.

Since 1974, there has been a steady increase in average winter air temperatures, leading to an increase in the sum of positive temperatures, indicating more frequent winter thaws and more intense snowmelt. The number of thaws (total number of days with a positive mean daily air temperature during the period of stable frost conditions) lasting 16–20 and 20–30 days in the modern period has almost doubled compared to the period before 1973 (

Table 6. Number of thaw events with different durations.

Meteorological Station	Periods	Number of Thaw Events with Duration
		0–5 Days	6–10 Days	11–15 Days	16–20 Days	20–30 Days
Aksai	1940–1973	15	9	7	4	0
	1974–2022	8	16	13	8	1
Uralsk	1940–1973	11	8	7	4	4
	1974–2022	8	7	9	11	11
Chingirlau	1940–1973	13	7	7	1	4
	1974–2022	10	17	7	3	8
Chapaevo	1940–1973	11	5	7	4	7
	1974–2022	6	7	14	7	13
Aktobe	1940–1973	15	7	5	4	2
	1974–2022	11	15	11	8	3
Martuk	1940–1973	15	7	9	1	2
	1974–2022	13	14	10	6	4
Novoalexeevka	1940–1973	11	9	5	3	5
	1974–2022	12	8	11	8	8
Uil	1936–1973	16	9	9	3	0
	1974–2022	16	14	11	3	3
Karauylkeldy	1940–1973	18	9	5	3	0
	1974–2022	12	19	12	4	0

). For example, at the Uralsk MS, the number of thaws lasting 16–20 and 20–30 days in the period 1940–1973 were observed four times, while in the modern period their number increased to 11 cases; at the Chapaevo meteorological station, the number of thaws lasting 16–20 days during the period 1940–1973 was observed four times, while during the period 1974–2021—increased to 7 cases, thaws lasting 20–30 days during the period 1940–1973 were recorded 7 times, during the period 1974–2021 their number increased to 13 cases.

The sum of negative temperatures during the winter period (December–March) has decreased significantly in recent years, indicating a general warming of the winter season and a decrease in frostiness. Until 1973, the sum of negative temperatures in most regions was high, indicating stable cold winters with a prolonged period of low temperatures. A decrease in negative temperatures leads to a reduction in the depth of soil freezing, which in turn increases the infiltration of meltwater into the soil horizons. Most of the meltwater flowed into river beds, forming floods (period before 1973), while in the modern period after 1974, a significant part of the water goes into the soil, reducing the volume of surface runoff.

Analysis of the data obtained shows that until 1973, weather conditions remained relatively stable: precipitation and air temperature did not change significantly. However, in the modern period, there has been a decrease in precipitation in the autumn-winter period and an accelerated rise in temperature, which has led to a decrease in snow reserves, an earlier start to spring flooding, and a reduction in its duration. All this together confirms the increase in climate risks for the region’s water resources and highlights the need to adapt water management to new climatic conditions.

3.3. Scenario-Based Forecasting of Maximum River Discharge

In this study, scenario-based forecasting was performed exclusively for maximum discharge, as this value demonstrates the strongest correlation with seasonal meteorological predictors. Forecasting other flood parameters, such as maximum water discharge, flood onset and end dates, and flood duration, requires the use of daily meteorological data. Despite the availability of daily meteorological data for CMIP6, their use in this study proved difficult. This is because working with such data requires prior reduction in spatial resolution and correction of daily data, without which the original values can significantly distort the amplitude, frequency, and temporal characteristics of precipitation and temperature. In this regard, the scenario forecast was limited to seasonal characteristics, where modeling errors are significantly lower.

The multiple correlation coefficients between maximum runoff volumes and the above predictors vary between R = 0.65 and 0.87.

Table 7. Forecast of air temperature changes for the period June–October (T_VI-X) according to climate scenarios SSP3-7.0 and SSP5-8.5.

Hydrological Districts	Meteorological Station	SSP Scenarios	Periods				Change in °C from the 1991–2020 Baseline Period
			1991–2020	2030 (2026–2035)	2040 (2036–2045)	2050 (2046–2055)
I district	Novorossiyskoe	3–7.0	15.7	17.6	18.3	18.7	1.9	2.6	3.1
		5–8.5		17.8	18.6	18.8	2.1	2.9	3.1
	Aktobe	3–7.0	17.1	18.7	19.3	19.9	1.6	2.2	2.8
		5–8.5		18.8	19.5	20.0	1.7	2.4	2.9
	Kos-Istek	3–7.0	14.9	18.1	18.8	19.3	3.2	3.8	4.4
		5–8.5		18.2	18.9	19.4	3.3	4.0	4.4
	Rodnikovka	3–7.0	15.5	18.0	18.7	19.2	2.5	3.2	3.6
		5–8.5		18.2	18.9	19.4	2.6	3.3	3.8
	Novoalexeevka	3–7.0	17.6	19.4	20.1	20.6	1.8	2.4	3.0
		5–8.5		19.6	20.3	20.8	2.0	2.6	3.1
	Aksai	3–7.0	17.3	19.3	19.9	20.4	2.0	2.7	3.1
		5–8.5		19.5	20.1	20.6	2.2	2.9	3.4
	Chingirlau	3–7.0	17.5	19.1	19.8	20.2	1.6	2.3	2.7
		5–8.5		19.3	19.9	20.4	1.8	2.5	3.0
II district	Kamenka	3–7.0	16.9	19.2	19.9	20.4	2.3	3.0	3.5
		5–8.5		19.4	20.1	20.6	2.5	3.2	3.7
	Yanvartsevo	3–7.0	17.2	19.3	20.0	20.5	2.2	2.9	3.4
		5–8.5		19.6	20.2	20.6	2.4	3.1	3.4
	Uralsk	3–7.0	17.4	19.2	19.9	20.4	1.8	2.5	3.0
		5–8.5		19.5	20.1	20.6	2.1	2.7	3.2
IV district	Dzhambeita	3–7.0	18.3	20.2	20.9	21.4	1.9	2.6	3.1
		5–8.5		20.4	21.1	21.6	2.1	2.7	3.3
	Karatobe	3–7.0	18.4	20.9	21.6	22.1	2.5	3.1	3.7
		5–8.5		21.1	21.7	22.3	2.7	3.3	3.9
V district	Emba	3–7.0	18.2	18.8	19.5	20.0	0.7	1.3	1.9
		5–8.5		19.0	19.6	20.1	0.8	1.5	1.9
	Temir	3–7.0	17.9	19.2	19.9	20.5	1.4	2.0	2.6
		5–8.5		19.4	20.0	20.5	1.5	2.2	2.6
	Karauylkeldy	3–7.0	19.9	20.4	21.0	21.6	0.6	1.2	1.8
		5–8.5		20.6	21.2	21.8	0.7	1.4	1.9
	Sagyz	3–7.0	20.2	20.8	21.4	22.0	0.6	1.2	1.8
		5–8.5		21.0	21.6	22.2	0.7	1.4	1.9
	Uil	3–7.0	19.2	20.5	21.1	21.7	1.4	2.0	2.5
		5–8.5		20.7	21.3	21.9	1.6	2.2	2.7

Note: Air temperature values are given in °C, and bold text indicates the most significant changes at particular meteorological stations.

,

Table 8. Forecast of changes in total precipitation for the period September–March (X_IX-III) according to climate scenarios SSP3-7.0 and SSP5-8.5.

Hydrological Districts	Meteorological Station	SSP Scenarios	Periods				Change in mm from the 1991–2020 Baseline Period
			1991–2020	2030 (2026–2035)	2040 (2036–2045)	2050 (2046–2055)
I district	Novorossiyskoe	3–7.0	195	256	279	283	31	43	45
		5–8.5		264	285	288	35	46	47
	Aktobe	3–7.0	180	198	226	234	10	26	30
		5–8.5		215	232	230	20	29	28
	Kos-Istek	3–7.0	213	237	275	284	11	29	33
		5–8.5		254	279	266	19	31	25
	Rodnikovka	3–7.0	198	268	288	282	36	46	43
		5–8.5		275	296	288	39	50	46
	Novoalexeevka	3–7.0	165	199	222	222	21	35	34
		5–8.5		224	226	222	36	37	35
	Aksai	3–7.0	159	188	191	182	18	20	14
		5–8.5		179	206	193	13	30	21
	Chingirlau	3–7.0	156	190	204	200	22	31	28
		5–8.5		196	215	212	26	38	36
II district	Kamenka	3–7.0	170	162	163	157	−5	−4	−7
		5–8.5		161	161	155	−5	−5	−9
	Yanvartsevo	3–7.0	200	233	246	249	17	23	25
		5–8.5		250	254	245	25	27	23
	Uralsk	3–7.0	189	234	241	227	24	27	20
		5–8.5		230	237	226	21	25	20
IV district	Dzhambeita	3–7.0	135	178	194	179	32	44	33
		5–8.5		180	193	186	33	43	38
	Karatobe	3–7.0	137	159	172	170	16	26	24
		5–8.5		166	172	173	21	26	26
V district	Emba	3–7.0	118	140	170	190	19	44	61
		5–8.5		160	175	172	36	49	46
	Temir	3–7.0	163	179	212	233	10	30	43
		5–8.5		207	223	229	27	37	40
	Karauylkeldy	3–7.0	123	144	169	182	17	38	48
		5–8.5		174	179	188	42	46	53
	Sagyz	3–7.0	92	122	138	145	32	49	57
		5–8.5		140	143	147	52	55	59
	Uil	3–7.0	145	173	185	201	19	28	39
		5–8.5		191	191	195	32	32	34

Note: Precipitation values are given in millimeters (mm).

and

Table 9. Forecast of air temperature changes for the period November–March (T_XI–III) according to climate scenarios SSP3-7.0 and SSP5-8.5.

Hydrological Districts	Meteorological Station	SSP Scenarios	Periods				Change in °C from the 1991–2020 Baseline Period
			1991–2020	2030 (2026–2035)	2040 (2036–2045)	2050 (2046–2055)
I district	Novorossiyskoe	3–7.0	−9.8	−8.1	−7.4	−6.6	1.8	2.5	3.3
		5–8.5		−8.1	−7.3	−7.0	1.8	2.5	2.8
	Aktobe	3–7.0	−8.3	−6.8	−5.9	−5.3	1.5	2.3	2.9
		5–8.5		−6.8	−5.9	−5.4	1.4	2.4	2.8
	Kos-Istek	3–7.0	−9.8	−7.7	−6.9	−6.3	2.0	2.8	3.4
		5–8.5		−7.8	−6.9	−6.4	1.9	2.9	3.4
	Rodnikovka	3–7.0	−9.2	−7.2	−6.4	−5.8	2.0	2.8	3.4
		5–8.5		−7.3	−6.3	−5.9	1.9	2.8	3.2
	Novoalexeevka	3–7.0	−7.7	−5.8	−5.0	−4.4	1.9	2.7	3.4
		5–8.5		−5.9	−5.0	−4.5	1.8	2.7	3.3
	Aksai	3–7.0	−7.1	−5.4	−4.7	−4.0	1.7	2.4	3.2
		5–8.5		−5.6	−4.6	−4.1	1.5	2.5	3.0
	Chingirlau	3–7.0	−7.5	−5.7	−5.0	−4.3	1.8	2.5	3.2
		5–8.5		−5.9	−4.9	−4.4	1.6	2.5	3.1
II district	Kamenka	3–7.0	−7.0	−4.7	−4.1	−3.4	2.3	3.0	3.6
		5–8.5		−5.0	−3.9	−3.4	2.0	3.1	3.7
	Yanvartsevo	3–7.0	−7.0	−6.0	−5.2	−4.5	1.0	1.8	2.5
		5–8.5		−6.0	−5.2	−4.8	1.0	1.8	2.2
	Uralsk	3–7.0	−6.7	−5.0	−4.4	−3.6	1.6	2.3	3.0
		5–8.5		−5.3	−4.3	−3.7	1.4	2.4	2.9
IV district	Dzhambeita	3–7.0	−6.5	−4.5	−3.8	−3.1	2.0	2.7	3.4
		5–8.5		−4.7	−3.8	−3.1	1.8	2.8	3.4
	Karatobe	3–7.0	−6.3	−4.2	−3.4	−2.9	2.1	2.9	3.4
		5–8.5		−4.3	−3.3	−2.9	2.0	3.0	3.4
V district	Emba	3–7.0	−8.1	−7.0	−6.1	−5.6	1.2	2.0	2.6
		5–8.5		−7.0	−6.1	−5.7	1.1	2.1	2.5
	Temir	3–7.0	−7.9	−6.3	−5.5	−5.0	1.5	2.4	2.9
		5–8.5		−6.3	−5.4	−5.0	1.5	2.5	2.8
	Karauylkeldy	3–7.0	−6.8	−4.7	−3.8	−3.3	2.1	3.0	3.5
		5–8.5		−4.6	−3.7	−3.3	2.2	3.1	3.5
	Sagyz	3–7.0	−5.7	−4.3	−3.3	−2.9	1.4	2.4	2.8
		5–8.5		−4.2	−3.3	−2.9	1.5	2.4	2.8
	Uil	3–7.0	−6.3	−4.7	−3.9	−3.4	1.6	2.3	2.9
		5–8.5		−4.8	−3.8	−3.4	1.5	2.4	2.9

Note: Air temperature values are given in °C, and bold text indicates the most significant changes at particular meteorological stations.

show scenario-based changes in key meteorological parameters used as predictors in forecasting spring runoff volumes.

Table 7 shows a steady increase in average air temperature during the warm period (June–October) in all hydrological regions. According to scenarios SSP3-7.0 and SSP5-8.5, by 2050, the temperature increase could reach +4.4 °C compared to the baseline period of 1991–2020. The most intense warming is expected in the northeastern part of the basin (Kos-Istek, Karatobe MS), while in the southern regions the temperature increase is less pronounced.

Under both climate scenarios (Table 8), most areas of the Zhaiyk–Caspian basin are expected to experience an increase in precipitation during the cold season (September-March), with the exception of the Kamenka MS, where a slight decrease of up to 10% is forecast compared to the baseline period. A particularly significant increase is expected in the southern part of the basin, at weather stations such as Emba, Temir, Karauylkeldy, and Sagyz, with an increase of up to 60% by 2050 compared to the baseline period of 1991–2020.

The results obtained (Table 9) indicate that during the cold period (November–March), when air temperature strongly influences the frequency of thaws and snowmelt intensity, a steady increase in average air temperature is projected throughout the Zhaiyk–Caspian basin. By 2050, winter temperatures are expected to rise by +1.5 to +3.7 °C relative to the baseline period of 1991–2020.

The most intense warming is predicted in the northwestern and central parts of the basin—at the meteorological stations of Kamenka, Karauylkeldy, Kos-Istek, Chingirlau, and Dzhambeita. In the southern regions (Emba, Sagyz, Temir, Uil MS), the warming rates are more moderate, ranging from +1.5 to +2.9 °C.

Analysis of the projected maximum runoff for the 2030, 2040 and 2050 under climate scenarios SSP3-7.0 and SSP5-8.5 in the Zhaiyk–Caspian basin revealed pronounced spatial heterogeneity (Figure 12, Figure 13 and Figure 14).

In the area of the left-bank tributaries of the Zhaiyk River, changes in maximum runoff under scenarios SSP3-7.0 and SSP5-8.5 are ambiguous. Under the SSP5-8.5 scenario, most major rivers, such as the Sarykobda, Or, Ulken Kobda, and Terisbutak, are projected to experience an increase in maximum flow ranging from 6% to 48% compared to the baseline period of 1991–2020. At the same time, a number of rivers, including Kos-Istek, Karakobda, Terisakkan, and Elek, are expected to experience a steady decline in runoff of up to 35%.

In the western part of the Obshchy Syrt, a steady decrease in maximum runoff is predicted under both climate scenarios. On most rivers, such as Shyzhyn 1 and Derkul, runoff volumes will decrease by up to 35%. The exception is the Shagan River, which will see a steady increase in maximum runoff of up to 33%, regardless of the scenario selected.

The rivers in the southern part of the Podural Plateau show mixed changes in maximum flow. Under the SSP5-8.5 scenario, most watercourses show a steady increase in volume, which is particularly pronounced in rivers such as the Kopirankaty and Shiderty, where the increase reaches up to 72%. Southern rivers show the most stable positive dynamics under both scenarios. On average, under SSP3-7.0, maximum flow increases by up to 20%, and under SSP5-8.5, by up to 30%. The exception is the Kiyil River, which shows a steady decline in maximum flow to 34% of the baseline period.

An analysis of scenario projections of maximum runoff for rivers in the Zhaiyk–Caspian basin shows that small and medium-sized rivers in the region are highly sensitive to expected climate change. This is particularly evident in steppe and arid areas, where runoff fluctuations vary widely. Comparison of the projected changes in meteorological parameters with the results of peak flow modeling confirms the key role of air temperature and cold-season precipitation in shaping the spring flood regime. The increase in winter temperatures, observed throughout the Zhaiyk–Caspian Basin and particularly pronounced at the Kamenka, Kosestek, and Chingirlau meteorological stations, may lead to earlier and more intensive snowmelt, which in turn alters the temporal distribution of runoff and reduces peak flows in several northern catchments, including the Elek, Terisakkan, and Kosestek river basins. Conversely, the increase in total cold-season precipitation, most notable in the southern areas (Emba, Sagyz, Temir, and Karauylkeldy meteorological stations), creates conditions for the accumulation of larger snow reserves and, consequently, higher spring runoff.

In the western part of the Obshchy Syrt (Shyzyn and Derkul river basins), a substantial rise in winter temperatures is projected, while the increase in total cold-season precipitation is relatively small; at certain meteorological stations (e.g., Kamenka), a decline is even expected. This combination of factors-sharp warming accompanied by weak or negative precipitation trends-shortens the snow accumulation period, increases winter runoff, and ultimately reduces spring peak flows.

Thus, the scenario-based projections indicate that the combined effect of rising winter and autumn temperatures and changes in precipitation determines the spatially heterogeneous response of river systems: a substantial increase in snow reserves coupled with moderate warming leads to higher peak flows, whereas sharp warming with insufficient precipitation growth results in a consistent decline in their magnitude.

4. Discussion

The study of spring floods is a key area of river hydrology, with both scientific and practical significance. From a scientific point of view, spring flood parameters are an indicator of the stable characteristics of the seasonal regime of river flow, which is formed under the influence of climatic conditions and the natural and geographical characteristics of the catchment area. From a practical point of view, these data are crucial for hydraulic engineering design: knowledge of the timing of the onset of floods and the maximum discharge is necessary for calculating the strength of bridge crossings and determining the parameters of spillways, canals, and other elements of water management infrastructure [50,51,52,53].

The results obtained indicate significant changes in spring floods in the Zhaiyk–Caspian basin in the modern period: spring runoff volumes have decreased, the timing of its onset and end has shifted, and the duration of floods has decreased. These changes are consistent with trends observed in other regions of the world, where changes in precipitation and air temperature lead to a reduction in spring runoff and a redistribution of the seasonal structure of the river regime [18,54,55,56].

Similar processes of spring runoff transformation have been recorded in a number of domestic studies covering both the Kazakh and Russian parts of the Ural River basin. Such processes are explained by common climatic causes and confirm the existence of regional consistency in the response of hydrological systems to climate warming [57,58,59,60].

Climatic factors have had a significant impact on changes in maximum runoff. Analysis of data from 17 weather stations shows a steady increase in air temperature in both winter and summer-autumn periods, an increase in the number of thaws, and a decrease in the number of days with negative temperatures. Elevated temperatures in the winter-spring season contribute to earlier snowmelt, increase infiltration losses and, as a result, lead to a decrease in spring runoff. These results are consistent with the conclusions presented in the works of [40,61].

Intra-annual precipitation dynamics show mixed changes. In winter (December–March), most stations show a slight positive trend (up to +9.8 mm/10 years), similar to the trends recorded in the Volga and Ural basins [51,62,63]. At the same time, autumn precipitation (September-October), which plays an important role in soil moistening before freezing, shows a downward trend, averaging 5 mm/10 years. The lack of autumn moisture increases the filtration losses of meltwater in spring, which is also confirmed by the results of a number of studies on the peculiarities of runoff formation under climate change [50,64,65].

In conditions of climate instability and increasing interannual variability of meteorological factors, there is a growing need for a scenario-based approach to forecasting extreme runoff characteristics using global climate models such as CMIP6 [66]. This study used CMIP6 data for two scenarios—SSP3-7.0 and SSP5-8.5—covering the future horizons of the 2030s, 2040s, and 2050s. The forecast models indicate a possible increase in extreme floods in some regions under high emission scenarios. For example, calculations using the EC-Earth3 model show a significant increase in 100-year water discharge in European regions under the SSP5-8.5 scenario [67]. In the Nam Ou River basin (a tributary of the Mekong), calculations based on CMIP6 model data using the MIKE-NAM hydrological model show an increase in maximum discharge of 23–49% by 2080 [68]. These results are consistent with the findings of this study on the Zhaiyk–Caspian basin, where models also predict an increase in maximum flow of up to 48% under the SSP5-8.5 scenario, despite the presence of isolated areas with negative dynamics (up to −35% under SSP3-7.0).

With regard to annual runoff, global studies show that trends in annual runoff vary depending on the region—in some parts of Europe and Asia, both increases and decreases in annual runoff are predicted, depending on the ratio of precipitation to evaporation [69,70] found that 72% of the world’s land area could experience an increase in annual runoff of up to +16.1% by the end of the 21st century under the SSP5-8.5 scenario.

Regional scenario assessments for the Zhaiyk–Caspian basin are consistent with global trends: extreme floods are intensifying, the seasonal structure of runoff is being redistributed, and uncertainty is increasing due to the variability of climate models. This highlights the need to use model ensembles and adaptive approaches in water resource planning and management.

5. Conclusions

The analysis of data on spring floods in the Zhayik-Caspian basin (1974–2022) showed a steady trend of decreasing spring runoff volumes in all hydrologically homogeneous areas. The largest decrease (up to 35%) was observed in the left-bank tributaries of the Zhayik–Ulken Kobda and Elek, as well as in the rivers of the Caspian lowland (Olenty, Kopirankaty) and southern watercourses (Temir, Oiyl, Zhem).

Maximum water discharge decreased by 30–40%: on the Elek—from 2500 to 1500 m³/s, on the Ulken Kobda—from 1300 to 1000 m³/s, and on the Temir—from 800 to 500 m³/s. The flood season begins on average 2–7 days earlier and ends 5–16 days earlier than before 1973, which leads to a reduction in its duration.

The main reason for the transformation of the flow is climate warming: the average winter temperature increases by +0.5, …, +0.6 °C per decade, the frequency of thaws increases, the depth of soil freezing decreases, and filtration losses of meltwater increase. Despite the increase in winter precipitation, the decrease in autumn precipitation reduces the water-holding capacity of the soil.

Scenario modelling (CMIP6, SSP3-7.0 and SSP5-8.5) until 2050 shows a further increase in winter temperatures of 1.5–4.4 °C and an increase in precipitation of 30–60%. At the same time, the spatial response of runoff is heterogeneous: on the left-bank tributaries of the Zhaiyk, both an increase (up to +48%) and a decrease (up to –35%) in flood levels are possible; in southern rivers, an increase of up to +72% is expected.

Small- and medium-sized rivers are most sensitive to climate change. The transformation of flood levels requires the introduction of adaptive approaches to water resource management, including scenario planning, climate modelling and a consideration of local basin characteristics. The results obtained are consistent with global trends, confirming their significance for regional hydrological policy.