Improving tropical cyclone forecasts with ensemble data assimilation of radar and geostationary satellite data
This work seeks to improve track and intensity forecast for tropical cyclones through advanced data assimilation and high-resolution radar and geostationary satellite data. Based on a three dimensional ensemble-variational (3DEnsVar) hybrid data assimilation system and regional WRF model, Shanghai Typhoon Institute has developed a Typhoon Ensemble Data assimilation And Prediction System (TEDAPS) which has been running in real-time since 2015. The evaluation for three-year performance indicates that TEDAPS produces promising TC track forecast with 24-h error around 65 km and 72-h error less than 200 km. Meanwhile, the system provides multiple uncertainty and probability information for TC track, wind and TC associated precipitation.
Recently this system has been extended to include high-resolution radar data and FY geostationary satellite data through which we expect to improve typhoon intensity forecast. Our preliminary results on convection-allowing assimilation shows that the radar data is able to improve both intensity and precipitation forecasts with little impact on track forecast, while the geostationary satellite data impact more on track as well as intensity forecasts.
Prof. Hong Li Shanghai Typhoon Institute of CMA