日本地球惑星科学連合2025年大会

講演情報

[J] 口頭発表

セッション記号 A (大気水圏科学) » A-AS 大気科学・気象学・大気環境

[A-AS09] 応用気象学

2025年5月27日(火) 10:45 〜 12:15 展示場特設会場 (4) (幕張メッセ国際展示場 7・8ホール)

コンビーナ:稲津 將(北海道大学大学院理学研究院)、竹見 哲也(京都大学防災研究所)、日下 博幸(筑波大学)、座長:丹治 星河(京都大学)

11:00 〜 11:15

[AAS09-02] 冬季豪雪に基づく日本の気象帯区分

*松岡 亮1稲津 將1 (1.北海道大学理学研究院地球惑星科学部門)

キーワード:豪雪、気象区分、自己組織化写像

In this study, we propose a weather zone classification for heavy snowfalls in Japan during winter using a Self-Organizing Map (SOM), which effectively maps recurrent weather patterns onto a two-dimensional latent space. This study examined data from 159 AMeDAS meteorological observation stations distributed across Japan. We utilized six-hourly sea level pressure data from the cold season (January, February, March, November, and December) over a 60-year period since 1960, covering the domain of [127.5°E, 157.5°E]×[27.5°N, 55°N]. A 10×10 SOM map was first generated to extract representative weather patterns. Heavy snowfall days were defined as days when the daily snowfall exceeded the 5th percentile of all snowfall days at each station. Stations where no snowfall was observed or where the threshold snowfall depth was 0 cm were excluded from the analysis. By linking the identified heavy snowfall days with SOM classifications, we extracted the characteristic synoptic patterns associated with heavy snowfall at each station. Furthermore, hierarchical clustering was applied to group similar weather patterns on SOM map.

Our results indicate that heavy snowfall events are generally associated with a west-high-east-low pressure pattern and the passage of extratropical cyclones. The classification revealed two major weather zones: the "Northern Japan" and the "Southern Japan," based on snowfall intensity. These were further subdivided into eight minor weather zones, reflecting distinct synoptic features. The spatial distribution of these weather zones exhibited a characteristic zonal pattern along the Japanese archipelago and its central mountain range, oriented perpendicular to the East Asian winter monsoon. This study demonstrates that organizing synoptic-scale meteorological fields responsible for heavy snowfall enables an objective characterization of Japan's snowfall weather, highlighting systematic pattern influenced by local topography.