不同边界层参数化方案在中国西南地区东部高分辨率气候模拟中的敏感性研究

为了进一步评估和提高区域模式对西南地区东部高分辨率气候的模拟能力，利用 WRF模式，采用 多种边界层参数化方案（下称“不同方案”）对西南地区东部 1998—2019年夏季降水和气温进行双重嵌套模拟 （外层为D01，内层为D02）。对比不同方案模拟结果表明：多年平均降水量在D01中基本为湿偏差；D02中在四 川盆地和重庆低海拔地区为干偏差，湿偏差主要位于贵州和重庆的城口、石柱和武隆一带的地形复杂区；总体 上D01中ACM2方案误差最小，D02中MYJ方案误差最小。对多年平均气温的模拟在D01中除了四川盆地一 带为暖偏差外其余大部地区基本为冷偏差，D02 中大部地区为暖偏差；总体上 D01 和 D02 中 MYJ方案误差最 小，YSU方案最大。对于降水量和平均气温年际变化的模拟技巧在D01和D02中相对较高的地区均集中在重 庆中西部和湖北大部地区；降水量总体为 YSU 方案最高，MYJ 方案最低；平均气温总体为 MYJ 方案最高， ACM2方案最低。因此，提升模式分辨率至对流尺度后对不同气象要素模拟技巧最优的方案存在差异，需根据 业务情况选择适合本地的参数化方案。

Sensitivity Analysis of Different Boundary Layer Parameterization Schemes in High-Resolution Climate Simulation over East of Southwestern China

To further assess and enhance the capabilities of regional models for high-resolution climate simulation in the east of Southwestern China (ESWC), the present study investigated two nested simulations (with D01 as the outer domain and D02 as the inner domain) of summer rainfall and temperature over the ESWC from 1998 to 2019. These simulations were conducted using the Weather Research and Forecasting (WRF) model coupled with various boundary layer parameterization schemes. The findings indicated that different parameterization schemes can lead to biases in the simulated mean annual rainfall. Specifically, the D01 domain exhibited a noticeable wet bias, while the D02 domain showed a dry bias in the Sichuan Basin and the lower altitudes of Chongqing. A wet bias was predominantly found in the complex terrain areas of Chengkou, Shizhu and Wulong in Chongqing , as well as in Guizhou. Overall, the YSU scheme in D01 and the MYJ scheme in D02 demonstrated the smallest bias in rainfall simulations. Regarding temperature simulations, a noticeable warm bias was present in D01 for the Sichuan Basin, with a cold bias in other areas. In D02, a warm bias was observed for most regions. The MYJ scheme exhibited the least bias, while the YSU scheme showed the greatest bias in both domains. The WRF model skillfully reproduced the interannual variability of rainfall and temperature in the midwest of Chongqing and most of Hubei. The YSU scheme outperformed others in rainfall simulation, while the MYJ scheme excelled in temperature simulation. These results suggest that the choice of an optimal scheme may vary depending on the meteorological parameter of interest, especially after model resolution is enhanced to the convective scale. Therefore, it is crucial to select the most suitable local scheme based on operational requirements.