Page 254 - 《高原气象》2022年第5期

5 期 谭 洁等：BCC-CSM2-MR模式对中国陆面过程模拟能力评估 1347
拟性能评估以及误差分析［J］. 地球物理学报，63（6）：2184- 12006/j. issn. 1673-1719. 2019. 039.
2198. DOI：10. 6038/cjg2020N0129. 续昱，高艳红，2020. 基于 GLDAS与再分析资料的青藏高原内循环
辛晓歌，吴统文，张洁，等，2019. BCC 模式及其开展的 CMIP6 试 降水率分析［J］. 高原气象，39（3）：499-510. DOI：10. 7522/j.
验介绍［J］. 气候变化研究进展，15（5）：533-539. DOI：10. issn. 1000-0534. 2019. 00013.

Evaluating the Performance of BCC-CSM2-MR Model in
Simulating the Land Surface Processes in China

2
4
3
TAN Jie ，HUANG Anning ，SHI Xueli ，ZHANG Yu ，ZHANG Yanwu ，CAO Lu ，WU Yang 1
1
2
1
（1. School of Atmospheric Sciences，Nanjing University，Nanjing 210023，Jiangsu，China；
2. National Climate Center，Beijing 100081，China；
3. School of Atmospheric Sciences，Chengdu University of Information Technology，Chengdu 610225，Sichuan，China；
4. Meteorological Bureau of Jiangsu Province，Nanjing 210008，Jiangsu，China）
Abstract：This study compares the CMIP6（Coupled Model Intercomparison Project Phase 6）historical experi‐
ment results of BCC-CSM2-MR（Beijing Climate Center-Climate System Model-Medium Reslution）with GL‐
DAS（Global Land Data Assimilation System）dataset and site-observed data to systematically evaluate the per‐
formance of BCC-CSM2-MR in simulating the land surface variables，such as surface soil temperature，upper
soil moisture and surface energy balance components in China. And the causes of the model biases are also deep‐
ly discussed. The spatial correlation coefficient，temporal correlation coefficient，Taylor score and root mean
square error between the observed data and GLDAS data and the model data are calculated to quantitatively ana‐
lyze the simulation ability of BCC-CSM2-MR to land surface variables. The results show that the model can well
simulate the spatial distribution and variability of land surface variables，but the model biases in quantity are still
obvious. The spatial correlation coefficient and Taylor score between the simulated land surface soil temperature，
upward net long wave radiation flux and surface upward latent heat flux by BCC-CSM2-MR and GLDAS are
above 0. 8 in each season，and it shows that the simulation performance of above variables is relatively great.
Compared with GLDAS data，the model overestimates the summer surface soil temperature over southeastern
China，but it tends to underestimate the surface soil temperature over most China in all seasons with much larger
underestimation over the Qinghai-Xizang plateau in winter and spring. From the model error analysis，it can be
concluded that the precipitation underestimated by the model leads to the overestimated surface downward short‐
wave radiation over southeastern China in summer，which leads to the overestimated surface soil temperature
there. The surface albedo overestimated by the model results in the underestimated surface net downward short‐
wave radiation over the Qinghai-Xizang plateau，which further leads to the underestimated surface soil tempera‐
ture，obviously in winter and spring. In addition，the model seriously underestimates（overestimates）the upper
soil moisture over southeastern China in all seasons（Qinghai-Xizang plateau in winter and spring），and the sim‐
ulation effect of the characteristics of time evolution of deep soil time is better than that of upper soil，this is
mainly resulted from the biases in the modeled precipitation. Meanwhile，the overestimation of upper soil mois‐
ture and 10m wind speed leads to the overestimated surface upward latent heat flux over Qinghai-Xizang Plateau
in the winter and spring. After the model evaluation，it can be found that the simulation performance of BCC-
CSM2-MR to land surface variables still needs to be improved and the causes of the model biases are complicated.
Key words：Qinghai-Xizang Plateau；BCC-CSM2-MR；model evaluation；land surface process；China