Testing and Improving the Performance of the Common Land Model: A Case Study for the Gobi Landscape

Land surface processes take place on the interface between the earth and atmosphere, exerting significant influences on the weather and climate. Correct modeling of these processes is important to numerical weather forecast and climate prediction. In order to obtain a more thorough understanding of the land surface processes over the Gobi landscape, we evaluated the performance of the Common Land Model (CoLM) at Dunhuang station in Gansu Province of China to determine whether the model formulation, driven by observational data, is capable of simulating surface fluxes over the underlying desert surface. In comparison with the enhanced observation data collected at Dunhuang station over the period 22–28 August 2008, the results showed that the surface albedo simulated by CoLM was larger than that in the observation, and the simulated surface temperature was lower than the observed. After the measured values were used to correct the surface albedo, the solar radiation absorbed by the ground surface was more consistent with the measurements. A new empirical relationship of the surface thermal exchange coefficient r_ah was used to modify the thermal aerodynamic impedance. The simulated soil surface temperature became significantly closer to the observed value, and the simulated surface sensible heat as well as net radiative fluxes were also improved.