講演情報
[PS-01-14]Detection of Flood-Related Large-Scale Climate Patterns Using Self-Organizing Maps and Ocean–Atmosphere Regime Analysis
*Wilmat Delankage Sineth Madusanka Priyasiri1, Sunmin Kim1, Yasuto Tachikawa1 (1. Graduate School of Engineering, Kyoto University)
キーワード：
Self Organizing Maps、Flood、Climate Regimes
This study investigates whether major Thailand flood events, especially those affecting the Chao Phraya River Basin, are linked to recurring large-scale climate regimes. Monthly ERA5 SST, MSLP, and WVF anomalies were classified using Self-Organizing Maps with 3×3, 4×4, and 5×5 grids. SOM node sequences were extracted for nine major flood events and compared across variables and resolutions. Four matched event-pair cases were then examined using raw/anomaly maps, spatial correlation, and RMSE. Results showed clearer repeated SST node patterns, while MSLP and WVF similarities were more variable. Future work will test node frequency, flood specificity, and combined-variable SOMs.