Ma Xiaogang, Jin Jiming, Zhu Lingjing, Liu Jian
Key Laboratory of Agricultural Soil and Water Engineering in Arid and Semiarid Areas, Ministry of Education, Northwest A&F Universtiy, Yangling, Shaanxi Province, China.
College of Water Resources and Architectural Engineering, Northwest A&F University, Yangling, Shaanxi Province, China.
PeerJ. 2021 Mar 16;9:e11040. doi: 10.7717/peerj.11040. eCollection 2021.
This study evaluated and improved the ability of the Community Land Model version 5.0 (CLM5.0) in simulating the diurnal land surface temperature (LST) cycle for the whole Tibetan Plateau (TP) by comparing it with Moderate Resolution Imaging Spectroradiometer satellite observations. During daytime, the model underestimated the LST on sparsely vegetated areas in summer, whereas cold biases occurred over the whole TP in winter. The lower simulated daytime LST resulted from weaker heat transfer resistances and greater soil thermal conductivity in the model, which generated a stronger heat flux transferred to the deep soil. During nighttime, CLM5.0 overestimated LST for the whole TP in both two seasons. These warm biases were mainly due to the greater soil thermal inertia, which is also related to greater soil thermal conductivity and wetter surface soil layer in the model. We employed the sensible heat roughness length scheme from Zeng, Wang & Wang (2012), the recommended soil thermal conductivity scheme from Dai et al. (2019), and the modified soil evaporation resistance parameterization, which was appropriate for the TP soil texture, to improve simulated daytime and nighttime LST, evapotranspiration, and surface (0-10 cm) soil moisture. In addition, the model produced lower daytime LST in winter because of overestimation of the snow cover fraction and an inaccurate atmospheric forcing dataset in the northwestern TP. In summary, this study reveals the reasons for biases when simulating LST variation, improves the simulations of turbulent fluxes and LST, and further shows that satellite-based observations can help enhance the land surface model parameterization and unobservable land surface processes on the TP.
本研究通过将社区土地模型第5.0版（CLM5.0）与中分辨率成像光谱仪卫星观测数据进行比较，评估并改进了该模型对整个青藏高原（TP）陆地表面温度（LST）昼夜循环的模拟能力。在白天，该模型低估了夏季植被稀疏地区的LST，而在冬季，整个青藏高原都出现了冷偏差。模拟的白天较低LST是由于模型中传热阻力较弱和土壤热导率较大，这导致了更强的热通量传递到深层土壤。在夜间，CLM5.0在两个季节中都高估了整个青藏高原的LST。这些暖偏差主要是由于土壤热惯性较大，这也与模型中较大的土壤热导率和较湿的表层土壤有关。我们采用了曾、王和王（2012年）提出的感热粗糙度长度方案、戴等人（2019年）推荐的土壤热导率方案以及适合青藏高原土壤质地的改良土壤蒸发阻力参数化方法，以改进对白天和夜间LST、蒸散和地表（0-10厘米）土壤湿度的模拟。此外，由于高估了积雪覆盖率以及青藏高原西北部大气强迫数据集不准确，该模型在冬季产生了较低的白天LST。总之，本研究揭示了模拟LST变化时偏差产生的原因，改进了湍流通量和LST的模拟，并进一步表明基于卫星的观测有助于增强青藏高原陆地表面模型参数化和不可观测的陆地表面过程。
