本文利用复杂地形条件下嵌套细网格预报模式^[1],使用1979年6月2日20时-6日20时的FGGEⅢ_b级资料序列,进行了大尺度稳定性降水预报的对比试验.文中采用修改了的饱和凝结法计算大尺度稳定性降水。各组对比试验结果指出:(1)以不同初始场作预报时,24小时预报比48小时预报更好,特别是在中低纬度;(2)每步调用大尺度降水过程,从物理上讲较合理,其预报结果也较每小时调用大尺度降水过程的预报更好;(3)在高温高湿的南方或季节,降水预报对Fr特别敏感;(4)模式考虑地形以后,对500hPa及以下层次的形势和流场改进明显;(5)湿模式的预报较干模式有所改进,大尺度降水潜热加热使系统的移速加快,使对流层低层降压,也影响了低层的流场,但这种加热主要集中在880-740hPa。

By using a nested fine mesh model over complex terrain and the FGGE level Ⅲ_b dataset from 2 to 6 June 1979, a series of contrast experiments for large scale precipitation (LSP) have done in this paper. A scheme of the LSP which uses modified saturation removal method was adopted. Results of the experiments show; 1) In the runs with the different initial conditions, the 24-hour forecast is better than the 48-hour one, especially in mid-and low-lati- tudes; 2) The LSP can be set on either at each time step or once per model hour, but the former run is more reasonable in physics and gives a better forecast in rainfall than the latter run; 3) In the southreu part or in the season with high temperature and humidity, the rainfr.ll forecast is sensitive to the changes in saturation condensation ratio, Fr; 4) There are significant improYement s in the forecast of situations and flow fields in the model run with topography) 5) The somewhat improvements of the moist model run over dry one are that the LSP latent heating makes weather system moving faster, which cause the pressure fall and affect the flow fields in the lower level. The heating mainly occur in the levels from 740 to 800hPa.