不同初始场条件对GRAPES模式数值预报的影响

利用中国新一代数值预报模式GRAPES(Global/Regional Assimilation and Prediction Enhanced System),分别使用T213 L31及NCEP FNL 6 h间隔的资料作为初始和边界条件,对2005年7月9—10日淮河流域一次致洪暴雨过程进行了个例试验,初步探讨了GRAPES模式的数值预报产品对不同初始场的敏感程度,以及三维变分同化对数值预报结果的可能影响。结果表明,T213和NCEP初始场中存在着差异较大的次天气尺度特征,并由此造成了此后GRAPES模式预报场的差异,且此差异不会随时间消失;同化对GRAPES模式积分结果的影响主要表现在最初的24 h内;模式对此次致洪暴雨过程的暴雨区分布、强度均有一定的模拟能力,但模拟的强暴雨区与实况仍存在着较大差异。由此可知,GRAPES模式的数值预报能力对不同的初始场和侧边界条件存在不同程度的依赖性,初始场的差异决定了模拟结果的差异。

EFFECTS OF DIFFERENT INITIAL FIELDS ON GRAPES NUMERICAL PREDICTION

The case of flush-flood-producing rainstorm in Huaihe River Basin during 9-10 July 2005 was selected,and by using the new generation numerical weather prediction model system GRAPES,four groups of contrast experiments were conducted with the initial fields and lateral boundaries provided by the T213 L31 and NCEP FNL data,respectively,to investigate the sensitivity of GRAPES numerical prediction products to different initial fields and the effect of 3-D variational assimilation on the results of the GARPES forecasts.After analyzing the differences between the T213 and NCEP initial fields and corresponding experiment results,the mnemonic ability of the model to initial fields and the influence of different initial fields on the precipitation forecast were concentratedly analyzed.The results show that there existed obvious differences in sub-synoptic scale characteristics between the T213 and NCEP initial fields,thus leading to different simulation results,and moreover the differences did not disappear when the integration went on.It has also been shown that the 3-D variational assimilation of the initial fields only markedly influenced the GRAPES model results in the first 24 hours,and afterwards the effect became diminished.In addition,both the location and intensity of the rainstorm forecasted by the GRAPES numeric model are very close to the observed,but there are some differences in the forecast of exceptional heavy rain.Further more,the 9-12 h,12-24 h and 0-24 h precipitation forecasts of the model are better in the experiments with the 3-D assimilation of the initial fields than those without assimilation.All these suggest that the prediction ability of GRAPES depends to some extent on different initial fields and lateral boundary conditions,and the differences of initial fields will determine the differences of GRAPES simulated results.