An Empirical Model of Tropical Cyclone Intensity Forecast in the Western North Pacific

The relative impact of environmental parameters on tropical cyclone (TC) intensification rate (IR) was investigated through a box difference index (BDI) method, using TC best track data from Joint Typhoon Warning Center and environmental fields from the NCEP final analysis data over the western North Pacific (WNP) during 2000–2018. There are total 6307 TC samples with a 6-h interval, of which about 14% belong to rapid intensification (RI) category. The analysis shows that RI occurs more frequently with higher environmental sea surface temperature, higher oceanic heat content, and lower upper-tropospheric temperature. A moderate easterly shear is more favorable for TC intensification. TC intensification happens mostly equatorward of 20°N while TC weakening happens mostly when TCs are located in the northwest of the basin. Mid-tropospheric relative humidity and vertical velocity are good indicators separating the intensification and non-intensification groups. A statistical model for TC intensity prediction was constructed based on six environmental predictors, with or without initial TC intensity. Both models are skillful based on Brier skill score (BSS) relative to climatology and in comparison with other statistical models, for both a training period (2000–2018) and an independent forecast period (2019–2020).