Impact of Tropical Cyclone-Associated Background Structure on Tropical Cyclone Assimilation and Prediction

In the past two decades, the Tropical Cyclone (TC) track forecast have been improved significantly. However, little progress has been made in intensity forecast. Previous studies have shown that the TC intensity forecast is sensitive to the TC structure. The lack of improvement in intensity forecast is partly due to the suboptimal representation of TC structure in the model initial condition. Being able to reasonably representing TC structure in the initial conditions through data assimilation becomes and important task for improving the follow-up TC forecast. In this study, the WRF-LETKF data assimilation and prediction system is used to simulate the typhoon Fanapi (2010). By comparing the assimilation experiments initialized from different TC structures, we investigate the impact of TC structure in the model background state on TC assimilation and prediction. Results shows that after assimilating the dropsonde observations, the TC track forecast of all experiments can be improved significantly. However, the differences in track prediction among experiments suggest that the track prediction is less sensitive to the background TC structure. The background TC structure have larger impact on TC intensity forecast. Results show that the experiment with a better background TC structure can lead to better intensity forecast.