Regionalization of climate teleconnections across Central Asian mountains improves the predictability of seasonal precipitation

Abstract

Mountains play a critical role in water cycles in semiarid regions by providing for the majority of the total runoff. However, hydroclimatic conditions in mountainous regions vary considerably in space and time, with high interannual fluctuations driven by large-scale climate oscillations. Here, we investigated teleconnections between global climate oscillations and the peak precipitation season from February to June in the Tian-Shan and Pamir Mountains of Central Asia. Using hierarchical climate regionalization, we identified seven subregions with distinct precipitation patterns, and assessed correlations with selected climate oscillations at different time lags. We then simulated the seasonal precipitation in each subregion from 1979 to 2020 using the most prevalent teleconnections as predictors with support vector regression (SVR). Our findings indicate that the El Niño–Southern Oscillation, the Pacific Decadal Oscillation, and the Eastern Atlantic/West Russia pattern are among the major determinants of the seasonal precipitation. The dominant lead-lag times of these oscillations make them reliable predictors ahead of the season. We detected notable teleconnections with the North Atlantic Oscillation and Scandinavian Pattern, with their strongest associations emerging after onset of the season. While the SVR-based models exhibit robust prediction skills, they tend to underestimate precipitation in extremely wet seasons. Overall, our study highlights the value of appropriate spatial and temporal aggregations for exploring the impacts of climate teleconnections on precipitation in complex terrains.

Research on Water Resource Modeling Based on Machine Learning Technologies

“…In addition, the statistical-based model (Support Vector Regression (SVR)) has been used to predict seasonal precipitation. Umirbekov et al used SVR to investigate the long-term correlation between global climate change from February to June and the peak precipitation periods in the Tian Shan and Pamir Plateau of Central Asia [44]. Traditional machine learning methods can handle complex precipitation data.…”