The goal of this research project is to prevent heavy rainfall damage on land by generating torrential rain over the sea. To achieve this, we need to redirect weather scenarios from causing heavy rain on land to reducing atmospheric water vapor over the sea, preventing damage. Comprehensive meteorological data, such as temperature, humidity, and wind direction, result in ultra-high-dimensional data, making prediction and control challenging. However, if weather conditions lead to several typical patterns, with one being undesirable, we can effectively describe these abstracted meteorological conditions to improve control efficiency. In Research Item 3, we aim to extract essential low-dimensional degrees of freedom, that we call as latent space representation, crucial for these branching points. We investigate three approaches to acquire latent space representations beneficial for weather control: reservoir computing, Koopman mode decomposition, and landscape analysis.

For reservoir computing, we studied data dimensionality reduction using deep learning and nonlinear time series prediction methods. We explored dimensionality reduction for meteorological fields in predicting linear rainbands using deep learning, examining Autoencoders and Convolutional Networks, and listing key references for technology development. Based on this, we obtained preliminary results on predicting nonlinear convective phenomena in FY 2024 (Fig. 2).
For Koopman mode decomposition, we studied dimensionality reduction techniques using Koopman mode and dynamic mode decomposition. We confirmed the Prony method's effectiveness for periodic data and the Original and Exact methods for quasi-periodic data, listing effective techniques for meteorological applications (Fig. 3).