物理过程参数化方案对中尺度暴雨数值模拟影响的研究

利用中尺度非静力MM 5模式和中国 2 0 0 1年 8月的 4个暴雨个例 ,研究了非绝热物理过程对中国暴雨动力和热力场预报的影响 ,深入分析了对流参数化方案在中尺度暴雨预报中的作用 ,讨论了利用模式扰动方法开展中国暴雨集合预报的可行性。结果表明 ,在短期数值预报中 ,非绝热物理过程对高度场预报影响较小 ,但边界层方案和对流参数化方案对产生暴雨的 3个基本条件即水汽通量散度、垂直速度、不稳定层结的影响很明显。不同对流参数化方案所预报的中尺度热力、动力场离差的结构特征与所预报降水的离差特征相似 ,且主要是在模式积分初期迅速增加 ,其后即趋于稳定。对中国热力场较均匀的暴雨过程 ,可以通过扰动模式的边界层和对流参数化方案 ,构造集合预报模式

THE IMPACT OF PHYSICS PARAMETERIZATION SCHEMES ON MESOSCALE HEAVY RAINFALL SIMULATION

Based on the non-hydrostatic version of Mesoscale Model version 5(MM5) and the data sets of four heavy rainfall scenarios occurring in Aug. 2001 in China, the impacts of three non-adiabatic physic processes including cumulus parameterization scheme, planetary boundary layer schemes and radiation parameterization schemes on dynamic and thermodynamic element predictions of heavy rain in China are investigated,with deeply analyzing the effect of cumulus parameterization schemes on mesoscale heavy rain simulations and discussing the feasibility of using model physics perturbations in ensemble simulation of heavy rain. The results show that non-adiabatic physic processes have less effects on the short-range forecast of geopotential height. Planetary boundary layer schemes and cumulus convection schemes have more impacts on heavy rain than radiation schemes, the most influences of two schemes of planetary boundary layer and cumulus convection are on the divergence of moisture flux, vertical velocity,unstable stratification ,which are the three basic conditions of torrential rain. The great influence of planetary boundary layer is on the thermal fields in the low levels. The forecast variance between convection schemes grows fast in the beginning to 12-h time periods of simulation , with the variance structure being a good correlation to that of sub-grid rain, in the later periods of simulation with less correlation and decrease during 12-18h time periods, then oscillates in the subsequent integral period . The structures of dynamic and thermal fields and the initial field using for model integral are all different since 12-h. Planetary boundary layer and cumulus convection schemes affect basic conditions of triggering convection.The accuracy of heavy rain forecasts is very sensitive to these parameterization schemes. The key of success of heavy rain ensemble prediction is to find out these sensitive factors and perturb them. For the storm rain events with a more homogeneous thermal environmental condition in China,the numerical model ensembles could be created by perturbing the planetary parameterization scheme and convection parameterization. Ensemble Prediction System of heavy rainfall can be constructed by two methods, one is by using different PBL and convection process parameterization schemes , the other is by using different parameter in Grell scheme. Different PBL and convection process parameterization schemes affect not only the mesoscale weather system, but also the "initial condition " of the model, reflecting the uncertain impact of non-adiabatic physics in model on heavy rain simulations.