Radar Wind Retrieval, Ensemble Simulations, and Cluster Analysis of a Heavy Rainfall Event over Northern Taiwan during 1–2 June 2017

Abstract This study examines a heavy rainfall event over northern Taiwan from 1 to 2 June 2017, focusing on the interactions between the marine boundary layer jet (MBLJ), frontal cyclone, and barrier jet (BJ) in influencing rainfall distribution. Three mesoscale convective systems (MCSs) moved eastward, producing intense, prolonged rainfall and transforming the linear convective line from a trailing stratiform (TS)-type to a parallel stratiform (PS)-type structure. Ensemble simulations, divided into five clusters using k -means clustering with spatial adjustments, reveal distinct rainfall distribution patterns, each linked to specific wind dynamics and mesoscale topographic interactions. Further categorization identified two primary rainfall types: type A, closely matching the observed event, with substantial inland rainfall driven by a nearby low pressure center and southwesterly MBLJ; and type B, with offshore rainfall influenced by a more distant low pressure center and stronger southerly MBLJ that extends BJs further north. Findings underscore the role of MBLJ, BJ, and cyclone positioning in shaping rainfall distribution, offering insights for more accurate rainfall forecasts in Taiwan’s complex terrain by capturing atmospheric variabilities and informing future forecasting models. Significance Statement This study explores the factors behind intense rainfall during Taiwan’s mei-yu season, focusing on how weather systems and the island’s mountains shape rainfall patterns. By analyzing different rainfall scenarios, the study identified five main patterns influenced by changing winds and terrain. It also highlighted two key rainfall types: one with more rain inland and another over the ocean, each linked to different weather conditions. These findings help improve our understanding of heavy rainfall in Taiwan and offer valuable insights for better weather prediction.