The Effects of Vortex Initialization to TC Track and Intensity Prediction of Regional Model

The Global models have been undergoing dramatic improvements with higher resolution close to 10km and advanced data assimilation techniques (such as NCEP-GFS and ECMWF). But the TC vortex initialization including position relocation and intensity correction are routinely applied in regional TC models for decades even in a cold start model which directly uses the global analyses as initial condition. Is it still that necessary now and how the vortex initialization effects TC track and intensity prediction of regional model with cold start?

This paper will compare the differences of TC track and intensity prediction of regional model GRAPES-TYM between two different initial conditions from the two major global models in the world: NCEP-GFS and ECMWF. The main TCs In 2016 are used and the statistical analyses show that the mean track errors with NCEP-GFS as initial condition are slightly larger than those with ECMWF as the initial condition, but the intensity errors are much smaller with NCEP-GFS as the initial condition especially for the TCs with faster intensifying rate..

The effects to track and intensity prediction of TC vortex initialization including relocation and intensity correction are studied using NCEP-GFS as the initial condition. The results show that vortex relocation in regional model has no notable effects to TC track and intensity from the average perspective. The intensity correction is essential to the deepening rate when TCs are in their week stage including TD and TS and mainly impact the absolute intensity when TCs are stronger than TY.