Graphical Abstract
Liang J., K. Terasaki, T. Miyoshi, 2023: A Machine Learning Approach to the Observation Operator for Satellite Radiance Data Assimilation. J. Meteor. Soc. Japan, 101, 79-95.
Editor's Highlight
https://doi.org/10.2151/jmsj.2023-005.
Graphical Abstract
Plain Language Summary: Numerical weather prediction becomes less accurate if we forecast longer, but it can be improved by the effective use of observations such as satellite radiance observations. To use observations in numerical weather prediction, we need to simulate observations. For satellite radiances, we usually compute complex radiative transfer processes. This study explored a potential simplification using machine learning models. The proposed method simulates satellite radiance data using machine learning and obtained promising results. The method would be useful to accelerate the system development to use new satellite observations as quickly as possible after the satellite launch.
Highlights:
- Model forecast and satellite microwave radiance observations are used to train machine learning models to obtain the observation operator for satellite data assimilation.
- Data assimilation experiments using the machine learning-based observation operator show promising results without a separate bias correction procedure.
- The machine learning-based observation operator can potentially accelerate the development of using new satellite observations in numerical weather prediction.