Weather forecasting method based on AI weather forecasting model and WRF model
By constructing a background field assimilation and multi-source observation data assimilation mechanism in the AI weather forecast model, the boundary conditions of the WRF model are dynamically driven, solving the problem that the boundary condition error of the WRF model cannot be corrected in the existing technology, thus improving the accuracy of weather forecasts and the forecasting ability of severe convective weather events.
Patent Information
- Application Number
- CN202610804822.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-06-05
- Publication Date
- 2026-07-03
- Estimated Expiration
- 2046-06-05
AI Technical Summary
In the current technology, the combination of AI weather forecasting models and WRF models has not yet formed an effective multi-source observation data assimilation and correction mechanism, which results in the boundary condition errors of WRF models being unable to be effectively corrected, affecting the accuracy of weather forecasts, especially in the forecasting of severe convective weather events.
By inputting ERA5 global atmospheric reanalysis gridded data into the AI weather forecast model for autoregressive iterative inference and bias correction, an assimilation background field is constructed. Multi-source observation vectors are then constructed by combining data from ground automatic weather stations, radiosondes, geostationary meteorological satellites, and Doppler radar VAD wind profiles. The optimal atmospheric analysis field is then solved, and the boundary conditions of the WRF model are dynamically driven to achieve hourly updates.
It effectively eliminated the systematic bias in AI model predictions, improved the WRF model's ability to forecast high-impact weather events such as severe convection, and enhanced the accuracy and timeliness of weather forecasts.