Comparative Analysis of Regression Models for Approximating and Forecasting Annual Maximum Water Volumes of the Toktogul Reservoir in the Kyrgyz Republic

This work presents a comparative analysis of various regression models, including linear, polynomial, exponential, sinusoidal, and spline regression, for approximating the annual maximum water volumes of the Toktogul Reservoir in the Kyrgyz Republic. The aim of the study is to determine the most adequate model for describing the dynamics of reservoir filling, which is crucial for planning water management activities and forecasting water resources in the region. A comparative analysis of the predictive capabilities of the models was conducted based on the available data. The results demonstrate varying degrees of model fit to the observed data, highlighting the necessity of careful selection of regression analysis methods in hydrological studies. The analysis showed that simple models, such as linear regression, were insufficient, while spline regression demonstrated significantly better results.