Assessment of real-time Predictions and numerical simulation using WRF for Typhoon Neoguri (201408)

Abstract

The real-time predictions for typhoon track and strength are distributed in variable agencies. The real-time prediction data from KMA, RSMC and JTWC were compared the accuracy with the adjusted best-track data from Digital Typhoon (RSMC). The prediction periods are also variable and irregular from 3 days to 5 days. The sequential prediction data are summarized with 6 hour interval from 4th to 11th July 2014 for the typhoon Neoguri (201408). The JTWC prediction of the typhoontrack and the RSMC predictions of the maximum wind speed and t he centerpressure show the best predictions. Also the numerical simulation using WRF model forced NCEP GFS prediction data is compared. The errors on the track are shown similarly the differences of 100 km in 48 hour prediction and 200 km in 72 hour prediction. The better results on the track prediction are shown in WRF model. However the large errors on the maximum wind speed and the center pressure of WRF prediction near initial are shown. It is caused by instability in the numerical simulation, and can be improve by introducing the dynamic initialization (DI) method. The WRF prediction of the center pressure shows the best accuracy than the prediction agencies, and the WRF prediction of the maximum wind speed shows the similar accuracy of RSMC prediction.he prediction periods are also variable and irregular from 3 days to 5 days. The sequential prediction data are summarized with 6 hour interval from 4th to 11th July 2014 for the typhoon Neoguri (201408). The JTWC prediction of the typhoontrack and the RSMC predictions of the maximum wind speed and t he centerpressure show the best predictions. Also the numerical simulation using WRF model forced NCEP GFS prediction data is compared. The errors on the track are shown similarly the differences of 100 km in 48 hour prediction and 200 km in 72 hour prediction. The better results on the track prediction are shown in WRF model. However the large errors on the maximum wind speed and the center pressure of WRF prediction near initial are shown. It is caused by instability in the numerical simulation, and can be improve by introducing the dynamic initialization (DI) method. The WRF prediction of the center pressure shows the best accuracy than the prediction agencies, and the WRF prediction of the maximum wind speed shows the similar accuracy of RSMC prediction.