| Like the ocean, land as the underlying surface of the Earth system can have a significant impact on the weather and the climate. Ongoing interactions in different scales exchange momentum energy and material between the atmosphere and land. Due to the complexity and diversity of the land surface processes and difficulties on getting the land surface observations data, the land surface model has become an important tool for researching land surface processes. The Community Land Model (CLM) as a mature land surface model, has been widely used by scientists. However, what kind of impact can be made by the differences between different versions of CLM models is still unclear. The simulation capability of the models in different regions has not been fully tested. Therefore, to compare different versions of the CLM model simulation results, it is necessary to test its performance and applicability in different regions.In this paper, three different versions of CLM models forced by the atmospheric forcing data from Princeton University are used for simulating global land surface processes during1948to2006. Simulation results from different models are compared in order to acquire a preliminary understanding of the impact from changes between the CLM models. To evaluate the three CLM models’capability on simulating soil temperature and moisture in China and provide some reference for improving the model simulation performance, the simulation data are also compared with Chinese national observation data from stations. The results show that:(1)By comparing different versions CLM model simulation results on the climate distribution of global soil temperature, soil moisture, sensible heat, latent heat and surface runoff, it is found that CLM4.0performs better than the other models especially in the Northern Hemisphere high latitudes areas and tropical regions. These are many differences among three CLM models on simulating seasonal variation in different regions. In each region, CLM3.0’ results on soil moisture and canopy evapotranspiration all have some shortcomings, while the CLM3.5and CLM4.0made a certain degree of correction.(2) The CLM series model can simulate the spatial distribution pattern of annual average soil temperature in China. Only in few parts of China, the simulated temperature is higher than the observation, and in other region, the simulated results is generally low. Three CLM models can reconstruction the interannual variability of soil temperature, the results of April are better than July, the results for surface are better than deep layers. In three models, CLM3.0’simulation results have the largest deviation from observation. On simulating soil temperature variation trend, three CLM models’performance show little difference. They can simulate the significantly decreasing in soil temperature to the south of the Yangtze River Basin, but all have some certain negative deviation.(3) The CLM series model can basically reflect the spatial distribution pattern-dry in North China, wet in northeast and southeast China-of annual average soil moisture in China. The results of CLM3.0are much higher than observation, CLM3.5and CLM4.0’s deviation is relatively small. Three models’simulated results for interannual variability of soil moisture are quiet like each other. Among the five districts, the correlation coefficients between observation and simulation are higher in the Huaihe River basin and the Yunnan-Guizhou region than the other district, while in the Northeast China, the simulation results are poor in April. The results are much closer to the observation in July. For soil moisture trend, three models’simulation is not good enough. They can basically reflect the wetting trend of soil moisture in Yunnan and Guizhou province. But for Northeast China, the simulation results are far away from observation. |
PDF Full Text Request