Progress Report
Artificial generation of upstream maritime heavy rains to govern intense-rain-induced disasters over land (AMAGOI)[4] Data Assimilation
Progress until FY2024
1. Outline of the project
To realize weather control, it is necessary to predict how the weather will respond when control inputs are given and to estimate when, where, and how we should add signals to the atmosphere to achieve desirable weather. However, there is currently no system capable of handling these tasks. In Item 4, we aim to develop a weather control computation system that can simulate the impact of control inputs on weather based on existing numerical weather prediction systems and estimate optimal control inputs.
In addition, the acceleration of prediction and control computations. To achieve weather control, as mentioned above, these computations must be completed before a disaster occurs or before the appropriate timing for control is missed. However, running numerical models used for weather prediction generally requires significant computational cost. With current computational resources and mainstream computational algorithms, it is anticipated that it would be difficult to complete these computations in a realistic timeframe. Therefore, we will introduce surrogate models and latent space representation techniques derived from mathematical research and deep learning into the prediction and control computation system. This will enable the calculation of control inputs in a realistic timeframe while evaluating effective elemental technologies for weather control computation. We also aim to utilize quantum computers, which are a cutting-edge computational technology, to optimize model predictive control and data assimilation calculations by mapping them to the Ising model and accelerating computations through quantum annealing.
2. Outcome so far
Since the project began in December 2023, we have focused on developing a system capable of predicting the effects of control inputs on the atmosphere. In the first year, we selected offshore structures as the control input and implemented actuator functionality within a regional weather model so that the effects of such structures could be simulated. We used the SCALE model, developed by RIKEN (Nishizawa et al., 2015; Sato et al., 2015), and introduced the effects of offshore structures by adding resistance forces on the model's surface layer. We prepared an environment to run simulations on the supercomputer Fugaku and conducted experiments with and without control to assess the impact of interventions by comparing simulation results. In collaboration with Item 5, we selected a case study and focused on the heavy rainfall event that affected the Kanto and Tohoku regions in September 2015. Our experiments confirmed that the installation of offshore structures generated a rainfall area of approximately 20 mm (24-hour accumulated total) on the leeward side of the structure. Sensitivity analysis on the size of the structure showed that larger structures resulted in increased rainfall directly downwind, but also a reduction in rainfall further downwind (Fig.1).
We are also developing methods to estimate when, where, and how to apply weather control interventions. In fiscal year 2024, we implemented Model Predictive Control (MPC) in the simplified Lorenz-96 weather model to build a system that estimates optimal control inputs. We investigated the feasibility of controlling extreme weather events and found that our approach achieved high success rates with smaller control inputs than in previous studies (Fig.2: Kawasaki et al., submitted to NPG).
However, MPC is known for its high computational cost, making direct application to full-scale weather prediction difficult. To address this, we also developed a method to accelerate MPC using ensemble approximations. We tested its effectiveness using the Lorenz-63 model and obtained promising results in terms of control success rates (Fig. 3: Kurosawa et al., submitted to NPG).
3. Future plans
We will continue to advance the development of methods for estimating when, where, and how weather control interventions should be applied. In fiscal year 2024, we used simplified, low-dimensional weather models for method development, but from the next year onward, we will incorporate the insights gained so far into real atmospheric ｓmodels and further develop the weather control computation system.