Research on Parameter Regionalization of Distributed Hydrological Model Based on Machine Learning

In the past decade, more than 300 people have died per year on average due to mountain torrents in China. Mountain mostly occur ungauged small and medium-sized catchments, so it is difficult maintain high accuracy of flood prediction. order solve problem low simulation areas, this paper studies influence different methods parameter regionalization distributed hydrological model parameters hilly areas Hunan Province. According terrain, landform, soil land use characteristics each catchment, we Shortest Distance, Attribute Similarity, Support Vector Regression, Generative Adversarial Networks, Classification Regression Tree Random Forest create schemes. total, 426 floods 25 catchments are selected calibrate parameters, 136 8 used for verification. The results showed that values Nash–Sutcliffe coefficients scheme were 0.58, 0.64, 0.60, 0.66, 0.61 0.68, worst 0.27, 0.31, 0.25, 0.43, 0.35 0.59. random forest most stable solution significantly outperforms other methods. Using regionalize can improve which great significance flash forecasting early warning.