The Study Of The Uncertainty Of Numerical Forecasts And Ensemble Forecasts Method Of Heavy Rainfall In North And South China

To meet the demand of the development of mesoscale ensemble prediction of heavy rainfall in China. By use of the numerical simulation, statistics method and diagnostic technique, The uncertainty of heavy rainfall forecast of non-hydrodynamic model (GRAPES) ,especially the difference impact of convective parameterization schemes on forecasts of heavy rainfall events occurred respectively in north and south China are investigated. The method of the generation of initial perturbation based on TIGGE data was established for the study of the initial perturbation effects on heavy rain from different TIGGE center. The goal of this research is to obtain the deeper understanding of the uncertainty of GRAPES model on heavy rainfall events and provide the evidence for GRAPS ensemble predictions. The major results of this study are as follows:(1)The impact of convective parameterization schemes on the local precipitation of heavy rainfall is obvious in the south China with weaker large scale forcing and in the north China with stronger large scale forcing. The impact of convective parameterization schemes on forecast uncertainty of heavy rainfall shows distinct diurnal variation for the south case, with strong uncertainty during the day and relatively weak in the evening. It has not apparent diurnal for the north case.(2)The uncertainties of model variables of south heavy rainfall were different from north heavy rainfall. The convective parameterization schemes have less effect on the large scale variables, but greater influences on mesoscale variables, such as vorticity, divergence. Furthermore, the influences is distinct on the upper and middle (the middle and low) level variables during the early (later) integral period for the south (north) heavy rainfall case.the convective parameterization schemes have more influences on the high (low)-level variables than the low (high)-level variables for the temperature filed for the south (north) heavy rainfall case.(3)On the whole, the convective parameterization schemes have more influences on the south heavy rainfall than the north heavy rainfall. The characteristics of the forecast uncertainty for the south (north) heavy rainfall case is (not) consensus to the variance of the vorticity and vertical velocity in 200hPa level.(4)The initial perturbation generated from TIGGE is much different, which means that the ensemble members can reflect the uncertainty of initial fields effectively. This is positive to the increase of the ensemble spread. The results show that the ensemble members improve the ability that GRAPES-MESO ensemble prediction system can capture the extreme precipitation events.(5)From the view of the skill of precipitation forecast, the performance of the GRAPES-MESO ensemble prediction system constructed by muti-initial value and muti-physics is excel to the control forecast. The probability forecast gives the high probability area, where the heavy rainfall occurred in reality. On the whole,the ensemble spread increases during the integral period. The ensemble verification results indicate that the relation between the spread and rmse demonstrates the rationality of the GRAPES-MESO ensemble prediction system based on TIGGE data during the early integral period.