JOURNAL OF COASTAL RESEARCH, v.91, no.SP1, pp.166 - 170
Abstract
To predict the typhoon tracks and intensity in Northwestern Pacific (NWP), an operational typhoon prediction system has been established. Real-time 5-day prediction is conducted two times per day using the WRF model forced with results from NCEP-GFS. To improve the accuracy of typhoon track prediction the spectral nudging is adopted for the bias reduction in the large-scale environmental field. As a way of improving the prediction accuracy near the initial stage of tropical cyclone (TC) development, the dynamic initialization scheme is employed which helps to obtain a stable initial condition without numerical shocks between different models in the system. The real-time typhoon simulations for the typhoon NEPARTAK using WRF forced with NCEP-GFS prediction data is compared with results from major typhoon prediction agencies. The best results on the track prediction are obtained with the multi-nested case of WRF model. It is however noted that underestimation of the maximum wind speed of WRF prediction still exists, obviously requiring better understanding of RI-related processes to improve the model prediction.