Search

Controlling weather is an outstanding and pioneering challenge for researchers around the world, due to the chaotic features of the complex atmosphere. A control simulation experiment (CSE) on the Lorenz-63 model, which consists of positive and negative regimes represented by the states of variable x, demonstrated that the variables can be controlled to stay in the target regime by adding perturbations with a constant magnitude to an independent model run (Miyoshi and Sun, 2022). The current study tries to reduce the input manipulation of the CSE, including the total control times and magnitudes of perturbations, by investigating how controls affect the instability of systems. For that purpose, we first explored the instability properties of Lorenz-63 models without and under control. Experiments show that the maximum growth rate of the singular vector (SV) reduces when the variable x was controlled in the target regime. Subsequently, this research proposes to update the magnitude of perturbations adaptively based on the maximum growth rate of SV; consequently, the times to control will also change. The proposed method successfully reduces around 40 % of total control times and around 20 % of total magnitudes of perturbations compared to the case with a constant magnitude. Results of this research suggest that investigating the impacts of control on instability would be beneficial for designing methods to control the complex atmosphere with feasible manipulations.

BSD 3-Clause License Copyright (c) 2022, Shunji Kotsuki All rights reserved. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

A flood forecasting system (FFS) is widely recognized as essential to protect people’s lives and prosperities. Developing an FFS with high accuracy, longer lead time, and high resolution is the ideal goal, but there are lots of obstacles to achieving this challenge. Here, we would like to introduce our progress in the development of 5-km resolution FFS system in Japan by Today’s Earth (TE) system (Ma et al., 2021). TE was developed by the collaboration between JAXA and The University of Tokyo and is routinely run at https://www.eorc.jaxa.jp/water/index.html. Among various events, we focus on a case study for forecasting Typhoon Hagibis by assessing its forecasting performance. The results showed that this method was accurate in predicting floods at 130 locations, approximately 91.6% of the total of 142 flooded locations, with a lead time of approximately 32.75 h. In terms of precision, these successfully predicted locations accounted for 24.0% of the total of 542 locations under a flood warning. On average, the predicted flood time was approximately 8.53 h earlier than a given dike-break time. Further, we would like to present our current work for developing an FFS with much higher resolution (1 km), with a probabilistic approach by the ensemble method using NEXRA (NICAM-LETKF JAXA Research Analysis, Kotsuki et al. 2017, https://www.eorc.jaxa.jp/theme/NEXRA/) data, and other developing versions of Today’s Earth system of Global scale (https://www.eorc.jaxa.jp/water/). Ma, W., Ishitsuka, Y., Takeshima, A. et al. (2021). Applicability of a nationwide flood forecasting system for Typhoon Hagibis 2019. Sci Rep 11, 10213. https://doi.org/10.1038/s41598-021-89522-8.Kotsuki S, Miyoshi T, Terasaki K. Lien GY, Kalnay E (2017) Assimilating the global satellite mapping of precipitation data with the Nonhydrostatic Icosahedral Atmospheric Model (NICAM), J. Geophys. Res. Atmos., 122, 631–650. doi: 10.1002/2016JD025355.

豪雪地帯に位置するダムでは降雪期に流入量が減少するため，流入量が増加する融雪期に利水容量を回復することが求められている．本研究では，深層強化学習を用いたダム操作モデルに集水域の融雪量や気温データを入力することで，豪雪地帯のダムにおける融雪洪水の抑制と利水容量を回復・確保するための操作の両立を目的とした学習を行った．学習の結果，AIダム操作モデルは6カ月以上の長期的なダムの操作であっても，期間中に発生した洪水の抑制操作と，融雪期における利水容量回復の操作を両立できることが明らかになった．また，気温や融雪量などの条件に応じて適切に放流量を制御できるという結果が得られた．, Dams located in heavy snowfall regions are required to restore their water utilization capacity during snow melting period because the inflow volume into reservoir decreases during the snowfall period. In this study, a dam operation model using deep reinforcement learning was trained by inputting snowmelt and temperature data of the catchment area to achieve both control of snowmelt floods and operation to restore the water utilization capacity of dams in heavy snowfall areas. From the results, this model can be adopted to both flood control and restoreing water utilization capacity during snowmelt period. We also obtain the results that this model can control discharge rate according to mass of snowmelt or weather conditions.

形態: カラー図版あり, Physical characteristics: Original contains color illustrations, Accepted: 2021-05-05, 資料番号: PA2210052000

Additional file 1. Supporting information.