RegCM3模式在西北地区的应用研究Ⅱ:区域选择及参数化方案的敏感性

通过数值模拟方法,研究了降水对区域尺度、积云对流参数化等的敏感性.结果表明:由于较小区域尺度的模式内部场和大尺度分析场激发的外强迫之间有更强的束缚,这种束缚使得内外强迫更容易达到一致.小区域尺度模拟的降水型比大的区域尺度的模拟更好,但同时,大的区域尺度消弱了由于模式侧边界效应产生的虚假动力效应,模拟的降水在量值上更加接近于观测值.因此,进行区域气候模拟时,须根据需要对模式区域进行仔细的选择.结果同时表明,由于Grell方案倾向于模拟更多的对流降水,因此Kuo-Anthes方案对西北地区降水型和量值的模拟比Grell方案更接近于实际.由于地形对于降水的重要意义,在复杂地形下进行区域气候模拟时有必要在模式中仔细描述地形.

Application of Regional Climate Model(RegCM3) in Northwest China II: Sensitivity experiment for domain choice and cumulus convection parameterization

Sensitivities of precipitation to several key physical processes,such as domain scale and scheme of cumulus convection were studied using the method of numerical simulation.Results indicated that the pattern of precipitation with a small domain scale(such as 4000km in west-east direction) was better than the large domain scale because model with smaller domain size had stronger restrictions between external forcing exerted by large-scale reanalysis field and interior solution of model,which made them easier to be consistent with each other.However,large domain scale removed the ripple dynamical effect from lateral boundary and the simulated rainfall was closer to the observation.So the domain size must be carefully selected based on research purpose when we do some regional climate simulations.Results also indicated that cumulus parameterization scheme of Kuo-Anthes had the better performance in spatial pattern and rainfall than scheme of Grell because scheme of Grell had the tendency to simulate more convective rainfall.At last,we found it was necessary to describe the topography in detail when the complicated topography was simulated using a regional climate model since the deviation of topography can seriously affect precipitation simulation.