Home >

news Help

Publication Information


Title
Japanese: 
English:Data-driven non-deterministic forecasting of tropical cyclone rainfall 
Author
Japanese: HOKSON Jose Angelo Arocena, 鼎 信次郎.  
English: J.A. Hokson, S. Kanae.  
Language English 
Journal/Book name
Japanese: 
English: 
Volume, Number, Page        
Published date July 13, 2023 
Publisher
Japanese: 
English: 
Conference name
Japanese: 
English:XXVIII General Assembly of the International Union of Geodesy and Geophysics (IUGG 2023) 
Conference site
Japanese:ベルリン 
English:Berlin 
Abstract The Western Pacific Region is the most active tropical cyclone (TC) basin. On average, twenty-six tropical cyclones affect large populations in the region every year, including those of the Philippines, Vietnam, Taiwan, China, Japan, and Korea. This study presents a statistical methodology – involving fuzzy-c clustering – for predicting TC-induced rainfall in the region. The method relies on the idea that a TC’s rainfall can be predicted using the rainfall of past TCs. Track, mean sea level pressure, movement speed, etc. are utilized to determine which past TCs to consider. The results of prediction are represented as the probability of exceedance of selected rainfall thresholds. Such representation allows the inclusion of possible extreme events that a deterministic representation may miss.

©2007 Tokyo Institute of Technology All rights reserved.