WRF单柱模式中单参数方法对热带对流模拟能力的影响

Effects of One-Moment Parameterization on the Capacity to Simulate Tropical Convection Using a Single-Column Model

利用耦合Milbrandt 2-mon(MY)双参数微物理方案的WRF中的单柱模式,对TWP-ICE试验(Tropical Warm Pool International Cloud Experiment)期间的个例进行数值模拟和敏感性试验。通过与观测资料和云分辨率模式的模拟结果进行对比发现:MY方案默认的双参数版本和单参数版本均能够再现TWP-ICE期间的热带云系的总体宏观和微观特征。MY方案的双参数版本模拟的降水率的演变特征同观测十分吻合,冰相粒子的微观特征同观测事实较为一致。单参数默认版本的降水率、液态云的构成及冰相粒子微观特征方面同观测事实存在明显差距。然而实际业务应用中单参数方案由于计算量较小应用更为广泛,但模拟效果有待改善。为了使方案保持计算量较为合理的同时具有较好的模拟效果,参考双参数控制试验中的冰相物质的微观特征,尝试对单参数方案中冰相粒子的单参数方法进行改进。冰晶单参数改进试验中虽然对于冰晶数浓度采用两种不同的处理方法,但模拟效果均未明显改善。其中冰云总含量更加接近观测,且冰云构成发生显著变化,主要归因于冰晶有效半径的减小间接削弱了雪和霰的发展。云滴含量的异常增强导致液态水含量比观测偏高约一个量级,暖云异常增厚则与上升运动的增强直接相关。雨水含量明显增强及雨滴有效半径减小综合导致了降水率仅有微弱改善。雪的单参数改进试验中,雪的截距值增加及环境场过饱和条件改善促进了冰云的发展。通过适当调整雪的截距的经验诊断公式,雪的截距、液态水含量以及降水率均得到较好的改善;而指定雪截距为常数的处理方式使液态云更为偏厚,降水率演变细节同观测仍然差异显著。改进试验结果表明,单参数方案中采用适当的经验公式诊断雪的截距的处理方法对改善单参数方案的模拟能力具有一定的可行性。

A double-moment bulk microphysics scheme, Milbrandt 2-mon（MY）, was evaluated using the WRF singlecolumn model during the period of the Tropical Warm Pool International Cloud Experiment（TWP-ICE）. Results from separate simulations using the double-moment and sample one-moment（1M） versions of the microphysics scheme with the default settings in WRF were able to reproduce the general characteristics of moisture variables from the macro to the micro scale, as compared with observations and cloud resolving model（CRM） results. The rain rate and microscale features of the ice phase from the double-moment control simulation（CTL） were very close to observations, but showed significant deviations from the sample 1M version simulation. One-moment bulk microphysics schemes have been widely applied in real business situations with little computation, but their simulating abilities need to be improved. In order to maintain a reasonable computational expense, but with better simulating ability, several measures were applied to improve the sample 1M versions of the MY scheme on the basis of the microscale features of the ice phase from CTL. The results from the runs with two ice 1M modified versions were similar. Both showed non-significant improvements, despite their different treatments in terms of the number concentration of ice crystals. The frozen water path and observations showed more agreement, but significant changes occurred in the composition of ice clouds, mainly attributed to the weakened growth of snow and graupel indirectly affected by smaller ice crystals. The liquid water path showed an order of magnitude larger than observations as a result of abnormal enhancement in cloud droplet content, related to enhancements in upward movement. The liquid water content and raindrop effective radius changing in an opposite direction led to a tiny improvement in the rain rate. Sensitivity tests using two snow 1M modified versions of the MY scheme showed that ice-phase cloud was enhanced after adding the snow intercept and improving the environmental conditions of superstation. The characteristics of the snow intercept, liquid water content, and rain rate became significantly improved after adjusting the empirical formula of the snow intercept properly. Meanwhile, the method of adopting the reference values of the snow intercept in CTL did not show the same performance in ice-phase clouds, resulting in thicker warm clouds and a significant discrepancy between the simulated and observed rain rate. Optimized simulations with 1M modified versions suggested that the treatment of adjusting the empirical formula of the snow intercept properly in the 1M version of the microphysics scheme has a certain feasibility.