Testing of the Multi-Physics Ensemble Kalman filter to Simulate the Track and Intensity of the Typhoon HaiYan (2013)

Abstract

This paper presents some experimental results of the multi-physics ensemble Kalman filter method which assimilating the satellite wind data in WRF model to simulate the track and intensity of the typhoon HaiYan (2013). The study conducted two experiments.: (1) Ensemble forecast with multiphysics ensemble Kalman filter assimilates the satellite wind data (CIMSS); (2) multi-physics Ensemble forecast (MPH). The results of atmospheric circulation analysis in the simulation started at 12 UTC (international time) until November 7, 2013, showing that the trend and intensity of the general circulation in the CIMSS test are similar to the real development. So the results of the storm trajectory forecast of CIMSS test at the 48-hour forecast limit onwards are better than the MPH test.
Moreover, experimental results based on the results of the 6 simulations for simulation error, the track error in the CIMSS test decreased 14% and 14.3% respectively in the 48-hour forecast period and 72 hours compared with the MPH test, and decreased 14% and 23.9% respectively compared to the global GFS forecast. For storm intensity (Pmin and Vmax), the CIMSS test also resulted in significantly improved errors in the 72-hour forecast period compared to MPH testing. These results confirmed that the assimilation of the satellite wind data into the input field of the model has a positive effect on Haiyan storm intensity and trajectory prediction skills. We may used this research to applying the multi-physics ensemble Kalman filter for forecasting storms affecting Vietnam.