Classification of Frontal and Typhoon Weather Fields Using Self-Organizing Maps and Its Application to Disaster Prevention Information Provision

2023 Fiscal Year Final Research Report

• Project/Area Number: 21K04275
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 22040:Hydroengineering-related
• Research Institution: Yamaguchi University
• Principal Investigator: Asai Koji 山口大学, 大学院創成科学研究科, 教授 (70202570)
• Co-Investigator(Kenkyū-buntansha): 西山 浩司 九州大学, 工学研究院, 助教 (20264070) ; 白水 元 東海大学, 建築都市学部, 特任助教 (60808210)
• Project Period (FY): 2021-04-01 – 2024-03-31
• Keywords: 自己組織化マップ / 豪雨災害 / 気象場分類 / 梅雨前線 / 台風 / 診断モデル

Outline of Final Research Achievements

This study aims to classify past weather patterns using SOM and to elucidate the characteristics of weather patterns and heavy rain. Additionally, it seeks to develop a system that alerts residents by identifying similarities between numerical weather prediction results and the weather patterns associated with past severe rain disasters.
Comparing the heavy rainfall in Hitoyoshi, Kumamoto Prefecture in July 2020 with that in July 1965, it was found that the most intense rainfall periods in both events occurred under the same weather pattern. By analyzing the trends in the occurrence times of heavy rainfall in the East China Sea region, it was determined that the frequency of heavy rainfall is higher in the morning. The study identified weather patterns with high heavy rain risk for the basins of major rivers in Kyushu. A system was developed to automatically acquire numerical weather prediction data and display the heavy rain risk for the next 84 hours automatically.

Free Research Field

水工学

Academic Significance and Societal Importance of the Research Achievements

本研究では豪雨になる前に住民避難を促進するシステムの開発を主目的としている．これまで類似の気象場の判断は人の感覚によるものであったが，SOMを用いることで客観的に類似気象場を判断することが可能となった．これを応用して住民に危機感を与え避難行動のスイッチを入れるシステムを開発した．具体的な社会実装が行われれば，豪雨災害の犠牲者を減少させることができ防災に大きく寄与するものと考えられ，社会的意義は大きい．