A Statistical Model for Tropical Cyclone Intensity Forecast in Northwest Pacific Ocean Considering the Land Effect

Tropical Cyclone (TC) is a destructive weather system. In order to alleviate enormous loss of lives and properties due to TCs, accurately and timely forecasting the TC's intensity and track is vital and essential. This study proposes a statistical regression method to forecast TCs' intensity at 12, 24, 36 and 48 hours over Northwest Pacific ocean. Besides the conventional factors of climatology and persistence, this study pays special attention to the land effect on the TC's intensity change by adding a new factor of ratio of land to sea into the statistical regression model. Three TC samples: ocean basin samples, offshore samples and the total TC samples from 2000 to 2014, are considered in this study to develop their intensity forecasting models respectively. 1°×1° NCEP/NCAR (National Centers for Environmental Prediction/National Centre for Atmospheric Research) and FNL (final) global reanalysis data are used as the predictors for considering the environmental effects on the TCs intensity change. Two methods of Stepwise Regression and Principal Component Analysis are employed to develop TC intensity forecasting models. Results show that the intensity forecasting skill for offshore TCs is significantly improved as the intensity change due to land decaying effect is specially considered. Therefore, the proposed models are valuable and practical to operational forecasters.