Statistical prediction of typhoon-induced total accumulated rainfall in the Western North Pacific using typhoon track similarity indices

Abstract

Typhoons can cause significant water problems if not effectively managed, but predicting typhoon-induced rainfall remains difficult due to the complexity of these storms and their interactions with land. In this study, we developed a new statistical prediction model for typhoon-induced accumulated rainfall (TAR) based on the principle of similarity in Asia. Our model utilizes track data from the Regional Specialized Meteorology Center, Tokyo, and precipitation data from the National Oceanic and Atmospheric Administration's Climate Prediction Center. This study focused on precipitation from 1979 to 2020 and developed the TAR calculation for typhoons, evaluated track similarity, and applied intensity correction to optimize the model output. After comparing the performance of the fuzzy C-means clustering and polygon methods, we used the triangle mesh area index to select the most similar typhoons. Further, the optimal ensemble number was selected based on the minimum root-mean-square error. Finally, we applied translation speed intensity correction to four selected cases, demonstrating significant improvements in TAR prediction compared to previous studies.