Impact and sensitivity of Doppler radial velocity data assimilation on short-term simulations of typhoon Aere (2004)

Typhoons often inflict heavy damage on Taiwan due to floods, mud slides, strong winds, and storm surges. Accurate prediction is particularly important for hazard mitigation and prevention. The strong interaction between the typhoon circulation and the complex terrain of Taiwan island not only affect the track and intensity of the typhoon but also induce unique changes in its structure and the observed precipitation. In this study, we explore the utilization of radar observations to improve short-term prediction for typhoons approaching Taiwan. Doppler radar has the capability of remotly sensing the typhoon circulation with high spatial resolution, especially over the ocean. Given this advantage, Doppler velocity observations could capture the charactistics of typhoon circulation leading to improved typhoon prediction.

We assimulate the Doppler radial velocities obtained from a selected Taiwan radar using the Penn State/NCAR mesoscale model version 5 (MM5) three-dimensional variational data assimilation (3D-VAR) system. A series of assimilating experiments are carried out to simulate Typhoon Aere (2004). The results show that the Doppler radial velocity data may enable us to relocate the typhoon center, and increase cyclonic circulation and lowlevel inflow. Although there was slight improvement in the sea level pressure by the assimilation of Doppler velocity, the typhoon intensity was strengthened by model adjustment after simulating for one hour. The improved intensity, location, and circulation lead to better rainfall prediction. Further sensitivity experiments show that assimilating Doppler radial velocity has a more positive impact on Aere's predictions than that of assimilating sounding and surface station data.

Keywords: Doppler radial velocity, data assimilation, 3D-VAR