Performance of model for predictions of inversion layer in the snow & storm event in 2008

There are four major processes involved in the snow storm event over southern China in 2008. Our previous study show the variability of every process has the significant relationship with strength of the inversion layer. Basing on this, we discuss the validation of five models data which has 1-15 days forecast from China, America, England, ECMWF and Canada. In qualitative aspects the results show the UK model is the best one at forecasting the inversion layer existing, and its maximum predictable time scale reaches 10 days. And for the single warm (cold) layer, the T213 (UK) model has the 15 days forecasting ability. That is to say, it is possible to forecast the existence of reversion layer in extended range which is helpful to forecast ice rain beyond 10 days. Moreover, there are three models except of T213 have good ability and get the 9 days prediction time in quantitative aspects. This work reveals it is necessary to analysis multi-models result in operational forecast.