Hydrological modelling and river flow dynamics of small rivers under climate change in Kazakhstan

Climate warming is rapidly reshaping high-mountain environments worldwide. Negative mass balance of glaciers across most mountain regions leads to profound shifts in regional hydrology. This fact poses particular risks for Central Asia, where up to 90% of water resources originate in the mountains and support domestic, industrial, and agricultural needs across the arid lowlands. While major transboundary rivers such as the Amu Darya, Syr Darya and Irtysh have received considerable scientific attention, much less is known about the smaller rivers, such as Ulken Almaty, Kishi Almaty, and Kaskelen, that supply water to the largest city in Kazakhstan, Almaty. The rapid rise in air temperature in the Ile Alatau Mountains is accelerating glacier melt, diminishing a critical source of runoff for the studied rivers.

Therefore, understanding how climate change affects the runoff dynamics of these small but vital freshwater sources is essential for sustainable water management in southeastern Kazakhstan. Thus, this study aims to model long-term river flow dynamics considering climatic and anthropogenic factors.

The study is based on more than a century of observational records obtained from the state hydrometeorological monitoring network. Time series analysis was performed using linear regression, the parameters of which were estimated using the least squares method, and the degree of trend severity was determined by the coefficient of determination (D). Modelling was performed using the HBV conceptual semi-distributed hydrological model. The model was calibrated using the automated GAP optimization algorithm in combination with manual parameter adjustment. The quality of the modelling was assessed using the Nash–Sutcliffe efficiency (NSE), standard deviation and PBIAS criteria, and reliability was assessed by validation over an independent period.

An analysis of long-term flow dynamics since the 1920s-30s has revealed mixed trends. The Kishi Almaty experienced a steady decline in discharge, the Ulken Almaty showed an increase in flow, while the Kaskelen had a relatively stable regime without pronounced long-term trends. It is noteworthy that during the last two decades, there has been an overall increase in runoff compared to previous periods. In this regard, the period 2000-2015 was chosen for flow modelling, which allowed climate change to be taken into account and its impact to be adequately reflected in the model parameters.

HBV modelling showed high accuracy in reproducing runoff: NSE ranged from 0.80 to 0.93, PBIAS from –0.9 to –4.7%. The model correctly reproduced the key characteristics of spring-summer floods, including the start, peak, end and duration. A comparison of modelled and observed runoff volumes showed high consistency: for Ulken Almaty, 84.6 versus 82.3 million m³, for Kishi Almaty, 53.7 versus 51.9 million m³, for Kaskelen, 132 versus 127 million m³, with a ratio of 91–95%. The validation confirmed the high efficiency of the model for the Ulken Almaty and Kaskelen rivers and satisfactory efficiency for the Kishi Almaty.

The study confirms the possibility of accurately reproducing key characteristics of flow and flood levels for taking climate change into account when planning water use and managing water resources in the region.

Key words: hydrological modeling, Climate change, Water availability, Kazakhstan