Research Progress on China typhoon numerical prediction models and associated major techniques

China is one of the countries affected by the most serious typhoon disasters.The promotion of the capability of typhoon disaster prevention mainly depends on the accuracy of typhoon prediction, which relies on the performance of numerical prediction models.In this paper,history of the development of typhoon numerical models and associated techniques in China is briefly reviewed.Subsequently,major features of numerical prediction techniques on typhoon initialization and physics parameterization are addressed,in which the techniques developed by the China Typhoon 973 project during 2009-2013 are emphasized.Vortex Initialization scheme based on Retrieved Variables(VIRV)was developed to enhance the analysis of typhoon intensity.The VIRV scheme was verified effective in enhancing the dynamical balance of typhoon,which is the basis of realistic typhoon numerical prediction.To better describe the air-sea interaction process,a regional air-sea coupled model is constructed by using Global and Regional Assimilation and Prediction System-Tropical Cyclone Model(GRAPES-TCM)and Estuarine,Coastal and Ocean Model(ECOM).Verifications show that TC intensity produced by the coupled model is more accurate than that by single atmospheric model which ignores ocean effect.Convection and radiation schemes are modified to reproduce the genesis,growth of convection and its subsequent interaction with atmospheric radiation. Moreover, the transfer of momentum between the atmosphere and the ocean is described in terms of a drag coefficient levels off as the wind speeds increase above typhoon force.Finally,major problems for the current typhoon numerical models are presented.It is stressed that,although bogus vortex has been found to be effective in prediction of typical typhoon tracks,more satellite data should be employed in typhoon initialization to enhance the prediction of unusual phenomenon of typhoons.The paper also suggests that more effors should be put into physical parameterization for high-resolution model to better describe mesoscale which leads to the difficulty of prediction on small scale processes and typhoon rainfall and intensity change.