The Cutting Edge of Predictive Science: Integrating Theory and Data to Challenge Sudden Rainstorms at the Osaka–Kansai Expo

Abstract

We conduct research in predictive science centered on data assimilation techniques that integrate obser-vational data with numerical simulations. We developed the world’s first ultra–high-frequency weather forecasting system with 30-second update cycles. Following validation experiments conducted during the Tokyo 2020 Olympic and Paralympic Games, the system successfully achieved 30-minute lead-time pre-dictions of localized guerrilla downpours at the Osaka–Kansai Expo 2025. On the theoretical side, we proposed a unified mathematical framework for data assimilation and control in chaotic systems, demon-strating its potential applicability to weather control. Furthermore, we are developing novel methodologies that combine data assimilation with artificial intelligence and machine learning. Data assimilation tech-niques are applied not only to weather forecasting but also across a wide range of fields, including ocean environments, red tide prediction, forest ecosystems, estimation of the effective reproduction number of COVID-19, and inverse estimation of intercellular forces.