Establishing a macroscopic-scale rainfall climate and water resources estimation model by machine learning method

Abstract

With the impact of climate change and the main rainfall seasons in Taiwan are concentrated in the plum rain season from May to June and the typhoon season from July to September each year.There are significant differences in rainfall and spatial and temporal distribution between the wet season and the dry season,the droughts will occur and even lead to severe water shortages, such as the worst drought in half a century in 2021.From a macroscopic spatial scale, for example, the El Niño phenomenon and solar activity may have a certain impact on the overall climate and water resources of the earth.Therefore, this study analyzes the correlation between rainfall and large-scale influencing factors such as sunspots, El Niño-Southern Oscillation,and uses machine learning models to predict and classify rainfall under different conditions,the prediction accuracy rate through historical data can reach 89.9% , with sunspots as the most significant factor. It is hoped that relevant units can provide reference for water resources management and planning.