[A-CG41] 衛星による地球環境観測

[ACG41-26] Advancement and AI integration of satellite data assimilation of clouds, precipitation, and the ocean

*三好 建正¹、大石 俊¹、Liang Jianyu¹、Goodliff Michael¹、大塚 成徳¹、富田 浩文¹、金丸 佳矢²、八代 尚³、寺崎 康児⁴、岡本 幸三⁴、小槻 峻司⁵、松岸 修平⁶、佐藤 正樹⁶、Bishop Craig⁷、久保田 拓志⁸、可知 美佐子⁸ (1.理化学研究所、2.情報通信研究機構、3.国立環境研究所、4.気象庁気象研究所、5.千葉大学、6.東京大学、7.メルボルン大学、8.宇宙航空研究開発機構)

キーワード：データ同化、人工知能、衛星、降水、雲、海洋

This presentation outlines the plan of a research project started in FY2025. This research aims to advance data assimilation, analysis, and prediction of clouds, precipitation, and the ocean, based on the achievements from the previous projects since 2013: “ensemble data assimilation of TRMM/GPM precipitation observations” (2013-2016); “advancing data assimilation of GPM observations” (2016-2019); “advancing precipitation prediction algorithm by data assimilation of GPM observations” (2019-2022); “development of a satellite ocean data assimilation system with the JAXA Supercomputer System Generation 2” (2017-2020); “satellite data assimilation using an ocean model” (2020-2022); and “advances and applications of satellite data assimilation of clouds, precipitation, and the ocean” (2022-2025). We developed the global atmospheric ensemble data assimilation system NICAM-LETKF with the local ensemble transform Kalman filter (LETKF) and the Nonhydrostatic ICosahedral Atmospheric Model (NICAM). We also developed a precipitation nowcasting system GSMaP RIKEN Nowcast (GSMaP_RNC) using the satellite-analyzed Global Mapping of Precipitation (GSMaP) dataset. By seamlessly merging data from the NICAM-LETKF numerical weather prediction and GSMaP_RNC, we established a real-time precipitation prediction system and have been continuously operating it for public data dissemination. Moreover, we have been deeply involved in the development of JAXA’s operational real-time atmospheric analysis NEXRA (NICAM-LETKF JAXA Research Analysis) which provides Level-4 products with proven data quality through validation for extreme weather cases such as typhoons and heavy rainfalls. For ocean data assimilation, we implemented the LETKF with an ocean model sbPOM and developed a daily-update ensemble analysis system using dense and frequent SST data from the Himawari-8 and -9 geostationary satellites. In the project starting in FY2025, we plan to enhance our data assimilation systems by integrating AI techniques. Through this research, we aim to deepen our integrated understanding of the earth system on clouds, precipitation, and the ocean and to advance the analysis, prediction, and real-life applications.