Evaluation of the effects of a multiphysics ensemble on the simulation of an extremely hot summer in 2003 over the CORDEX-EA-II region
Evaluation of the effects of a multiphysics ensemble on the simulation of an extremely hot summer in 2003 over the CORDEX-EA-II region
复制标题
评估多物理场系综对 2003 年 CORDEX-EA-II 地区极热夏季模拟的影响
DOI：
10.1002/joc.6028
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
Congbing Fu
中科院分区：
文献类型：
--
作者：
Linyun Yang;Shuyu Wang;Jianping Tang;Xiaorui Niu;Congbing Fu
在本文中，我们使用天气研究与预报（WRF）模型评估了2003年CORDEX-EA-II域JJA极端降水和温度的48个成员多物理系综。在大尺度环流控制的分区中，模拟的降水和温度是合理的，但在中尺度过程影响重要的分区中，降水和温度的偏差是明显的。不同WRF物理方案组合对江淮降水的模拟效果因地区而异。同时，积云方案和微物理方案对降水的模拟有重要影响，陆面模式和积云方案对地表温度的模拟起着决定性的作用。分析表明，采用Noah方案模拟陆面过程，Lin方案模拟微物理过程，G3 D方案模拟积云参数化，CAM方案模拟辐射方案模拟中国地区极端降水和极端温度，模拟的极端气候往往伴随着较大的集合扩散，这意味着模式对某些次区域的物理过程较为敏感。模拟的中低层大气风场显示了对陆面模式和积云方案选择的响应。陆-气相互作用对极端天气模拟的重要影响很大程度上归因于活跃的对流过程。WRF难以再现观测到的降雨过程的时间演变，其中包括在模拟期间的观测数据集中连续的大降雨事件。
In this paper, we evaluate a 48‐member multiphysics ensemble using the Weather Research and Forecasting (WRF) model for the JJA extreme precipitation and temperature in 2003 over the CORDEX‐EA‐II domain. The simulated precipitation and temperature are reasonable in the subregions controlled by a large‐scale circulation, yet the biases for both precipitation and temperature are evident over the subregions where the effects of mesoscale processes are important. The performance of various combinations of WRF physical schemes for simulating the JJA precipitation is dependent on the region. Meanwhile, the cumulus and microphysical schemes have substantial influences on the simulation of precipitation, and the land surface models and cumulus schemes play crucial roles in the surface temperature. Our analysis shows that the combination of Noah for the land surface process, Lin for the microphysics, G3D for the cumulus parameterization, and CAM for the radiation scheme can provide the most reliable reproduction of both precipitation and temperature extremes over China.Ensemble analysis shows that the simulated climate extremes are usually accompanied by a large ensemble spread, implying sensitivities to the model physical processes in some subregions. The simulated wind fields at low‐to‐middle atmospheric levels display responses to the options of the land surface models and cumulus schemes. The essential impact of the land–atmospheric interaction on simulating the extremes can be largely attributed to the active convective processes. WRF has difficulties in reproducing the observed temporal evolution of the rainfall process, which consists of continuous large rainfall episodes in the observation data set during the simulation period.