The performance of the new tropical cyclone track prediction system of the China National Meteorological Center

Summary

The global spectral model (GSM) at T213 horizontal resolution with 31 vertical levels has been used operationally in the China National Meteorological Center (CNMC) since 2002. The results show that the GSM is skillful in predicting large-scale circulations because of both the high spatial resolution and the improved model physics. However, because of the low skill in predicting the tropical cyclone (TC) genesis, the model has low skill in predicting TC tracks. In order to improve the prediction skill for TC tracks, an asymmetric bogus vortex scheme was implemented in the CNMCs operational GSM, and has being run in an experimental mode in 2003 and in an operational mode since 2004.

In this paper, the performance of the new system in predicting TC tracks over the western North Pacific in the last 3 years are evaluated together with a description of the construction of the asymmetric vortex bogus scheme and its implementation into the CNMCs GSM. It is shown that the asymmetric bogus vortex scheme can improve the track forecast skill not only in short term but also beyond 72 h compared with the corresponding axisymmetric vortex bogus scheme. The results also show that the performance of the GSM with the asymmetric vortex bogus scheme is superior to that of the current operational limited area TC model used at CNMC. However, systematic errors are found in the new TC prediction system: the model predicted tracks tend to deflect westward when TCs travel northeastward while northward when TCs travel westward or northwestward. The plan for future improvement and development of the TC prediction system at CNMC is also briefly discussed.