2023 Fiscal Year Final Research Report
Research on hydrometeors assimilation methods to elucidate the mechanisms of torrential rain development
| Project/Area Number |
21K03669
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 17020:Atmospheric and hydrospheric sciences-related
|
|---|
| Research Institution | Japan, Meteorological Research Institute |
|---|
Principal Investigator |
Ikuta Yasutaka 気象庁気象研究所, 気象観測研究部, 主任研究官 (80878249)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
澤田 謙 気象庁気象研究所, 気象観測研究部, 主任研究官 (10847205)
堀田 大介 気象庁気象研究所, 気象観測研究部, 主任研究官 (60805365)
|
|---|
| Project Period (FY) |
2021-04-01 – 2024-03-31
|
| Keywords | データ同化 / 集中豪雨 / 深層学習 / 数値天気予報 |
|---|
| Outline of Final Research Achievements |
This study aims to estimate background errors according to various meteorological conditions using deep learning and utilize them in data assimilation to elucidate the mechanism of torrential rain development. Conventional methods for estimating background errors have been challenging difficulties in achieving high accuracy due to the diversity of cloud precipitation processes. Therefore, we developed a method to generate background error using a type of deep learning called conditional generative adversarial networks. By using the generated background errors to investigate the assimilation impact of rain and snow mixing ratios, we demonstrated superior analytical accuracy compared to conventional methods. Furthermore, to elucidate the reasons for errors in forecasts, we introduced new raindrop size distributions and non-spherical snow particles into the numerical model. As a result, an improvement in the predictive accuracy of hydrometeors within precipitation systems was achieved.
|
| Free Research Field |
気象学
|
| Academic Significance and Societal Importance of the Research Achievements |
この研究課題の成果は、気象レーダーや地球観測衛星データ同化技術の進展をもたらし、数値天気予報モデルによる集中豪雨の内部構造の精密な再現を可能とする。この数値モデルの現象再現性の向上は、顕著現象の原因となる積乱雲の発生や発達機構の高精度な分析を可能とする。豪雨の発達機構に関する理解の深化は、気象学の更なる進歩に必要不可欠である。また、数値天気予報モデルの初期値の精度向上は、日々の数値天気予報の精度も向上させる。特に、モデルの水物質の再現性能の向上は、水害や土砂災害の原因となる集中豪雨の予測精度向上に直結し、防災情報の高度化に寄与する。
|
