Improving Wind and Rain Simulations for Tropical Cyclones with the Assimilation of Doppler Radar Data

Abstract

More accurate prediction of the strong winds and heavy rain associated with tropical cyclones using numerical weather prediction (NWP) models would be helpful in the provision of weather services for the public. In this paper, the impact of assimilating radar data in the simulation of Typhoon Neoguri and Severe Tropical Storm Kammuri in 2008 is studied using Weather Research and Forecasting (WRF) version 2.2 and WRF VAR version 2.1. Only the data from the radar at Tate's Cairn in Hong Kong are considered. Four experiments are conducted, namely, (a) simulation without radar data, (b) simulation with radar data assimilated at the initial time, (c) cycling simulation with the assimilation of radar data (Doppler velocity and reflectivity) directly assimilated, and (d) cycling simulation with the assimilation of 2D wind field retrieved from the Doppler velocity data from the radar. By comparing with actual observations of the surface wind distribution in Hong Kong and the actual radar reflectivity data, it turns out that both (c) and (d) outperform (a) and (b), and (c) and (d) show comparable skills. As a result, cycling simulation with the assimilation of weather radar data (even for a single radar) could improve the prediction of winds and rain bands associated with tropical cyclones.