ENSEMBLE PREDICTION EXPERIMENTS OF TRACKS OF TROPICAL CYCLONES BY USING MULTIPLE CUMULUS PARAMETERIZATION SCHEMES

Abstract:

Ensemble prediction experiments of the tracks of eight tropical cyclones occurring between 2004 -2006 over the western Pacific have been performed by using MM5 with five cumulus parameterizations chemes. The results show that the predictions of the tracks of the tropical cyclones are sensitive to the selection of cumulus parameterization schemes. Each scheme has its own advantage and disadvantage, and the predications without cumulus parameterization schemes are not the worst, sometimes even better than the others. And all of the three ensemble methods improve the predictions of the tracks significantly, among which the ensemble method without parameterization schemes, the Grell, Betts-Miller and Kain-Fritsch schemes are the best.