応用数理問題―計算機科学から社会科学までA

[S05A-01] A Hybrid Scheme of Kinematics and Koopman Operator Analysis for Short-time Precipitation Forecast

*Takashi MIYAMOTO¹, Shitao ZHENG¹, Masato ABE², Shingo SHIMIZU³, Ryohei KATO³, Koyuru IWANAMI³ (1. University of Yamanashi, 2. BMC Corporation, 3. National Research Institute for Earth Science and Disaster Resilience)

Keywords:data-driven analysis, Koopman operator analysis, nonlinear dynamics, short-time precipitation forecast

This study proposes a new precipitation forecast method based on Koopman operator analysis, which is a more recent type of data-driven method. We develop a model to decompose the temporal evolution of weather states into global spatial movement and the evolution or attenuation of state, and to predict the former by kinematic manipulation and the latter by Koopman operator analysis. A numerical experiment was performed wherein the precipitation was predicted for a certain lead time based on the observation records in XRAIN. From the experimental results, it was verified that the proposed model shows high accuracy as compared to the persistent prediction and simple kinematic extrapolation models.