The Key Laboratory of Regional Numerical Weather Prediction, located at Guangzhou, is co-founded by the People’s Government of Guangdong Province and China Meteorological Administration. Its mission is to develop and operate the advanced regional numerical weather prediction (NWP) systems over South China area and East-Asia surroundings, which focus on the tropical severe weather warning and forecasting issues. The main features of the tropical NWP systems include full compressible and non-hydrostatic dynamic core, full package of physical parameterization, variational data assimilation, coupled with a regional ocean model, to further improve the marine weather prediction and customized meteorological services by applying forecast outputs. The lab’s vision is to be an Asian first class and trusted voice in the forecast of typhoons and be an operational center to develop and implement numerical weather prediction.
To meet its mission the lab is organized in several working groups. A model development group focuses on development and transition of new techniques into the GRAPES model to improve dynamical structure and physical parameterization. A data assimilation group synthesizes ensemble-based and variational method to meliorate data assimilation frame work which increases the application of local observations. An operation group, in conjunction with other departments of Guangdong Meteorological Service, maintains operational systems and issues analyses and reports in the form of text and graphics.
The primary objectives and activities of the lab are:
1) Design and develop advanced dynamical core and physical parameterization schemes of tropical mesoscale numerical weather prediction model for more accurate forecast of a typical weather system, such as typhoons, thunderstorms and squall lines.
2) Improve and explore exclusive and efficient techniques for data assimilation by ensemble-based and variational methods, which will be suitable to observations in tropical and marine environment.
3) Implement and maintain numerical weather prediction systems aiming to brace operational forecasts and warnings to diminish property loss and enhance economic efficiency.
Nowadays, the lab is responsible for multiple operational systems, which cover from synoptic scale to convective scale. The latest TRAMS model has 18km resolution and provides a 72 hours leading time forecast twice a day when a typhoon enters a critical area. The real time meso-scale forecast is generated by 9km resolution GRAPES model with several featured physical schemes. A 3km resolution model after assimilating local observation data generates convective scale forecast running with partial cycle scheme. In the future, the lab is planning to upgrade the model resolution to 1km.