陆面模型Noah-MP的不同参数化方案在沙漠区域的适用性研究

Applicability of different parameterization schemes of land surface model Noah-MP in desert areas

摘要:

为研究陆面模型Noah-MP在沙漠下垫面的最优参数化方案组合，本文利用塔克拉玛干沙漠大气环境监测站观测数据，根据沙漠环境特征进行不同参数化方案组合的三组模拟实验，并基于观测数据对10 cm土壤温湿度、感热、潜热通量模拟值对比分析，得出最优组合。研究表明：第三组对10 cm的土壤温度模拟效果最好，主要原因是Chen97感热交换系数和全网格二流近似（gap=0）辐射传输方案比较符合塔中站的环境特征。三组试验对土壤湿度模拟效果差，其主要原因是塔中站的地表环境和土壤信息未能在体现在模式中，第二组选择CLM方案对土壤类型影响蒸发方面有一定考虑，其模拟结果相对较好。对于感热通量，第一、二组模拟值在波峰存在高估，尤其是第二组模拟值在降水后出现了明显低估情况，第三组模拟效果最好，主要得益于选择了感热交换系数Chen97方案，能够较为真实的刻画Ch变化特征。潜热通量在四个特征量中模拟效果最差，主要原因是沙漠土壤水分极低，观测降水和实际进入土壤的水量有差异，另外没有植被和植物根系，模式无法准确计算土壤蒸发和植被蒸散，因此对沙漠区域的潜热通量计算不够理想。综合统计分析和泰勒图可知，第三组模拟性能最好，能够更好地还原沙漠区域的陆面过程。

Abstract:

In order to study the optimal parameterization scheme of the land surface model Noah-MP in the desert underlying surface, this paper set up three sets of simulation experiments combining different parameterization schemes according to the characteristics of the desert environment, which based on the observation data over the Taklimakan Desert Atmospheric Environment Monitoring Station (TDAEMS), and then get the best combination. The results show that the 3rd group was the best simulation effect on soil temperature of 10 cm. The main reason is that the Chen97 sensible heat exchange coefficient and the entire grid cell two-stream approximation (gap=0) radiation transfer scheme are in line with the environmental characteristics of the TDAEMS. The three groups of experiments have poor performance soil moisture simulation . The main reason is that the surface environment and soil information of TDAEMS are not reflected in the model. The second group chooses CLM scheme to have some considerations on the influence of soil type evaporation, and has a better correlation. As for the sensible heat flux, the 1st and 2nd sets of simulation values are overestimated at the peak, especially the second set of simulated values are significantly underestimated after precipitation, and the 3rd set of simulations is the best, mainly due to the choice of sense. The heat exchange coefficient Chen97 scheme could describe the Ch variation characteristics more realistically. The latent heat flux has the worst simulation effect among the four characteristic variables. The main reason is that the desert soil moisture is extremely low. There is a bias between the observed precipitation and the actual amount of water entering the soil. In addition, there is no vegetation and plant roots. The model cannot accurately calculate soil evaporation and vegetation evapotranspiration. Therefore, the calculation of latent heat flux in the desert area is not ideal. Based on the comprehensive statistical analysis and the Taylor diagram, the third set of simulations has the best performance and can better restore the land surface process in the desert area.