An extension of the conditional nonlinear optimal parameter perturbation (CNOP-P) method is applied to the parameter optimization of the Common Land Model (CoLM) for the North China Plain with the differential evolution (DE) method. Using National Meteorological Center (NMC) Reanalysis 6-hourly surface flux data and National Center for Environmental Prediction/Department of Energy (NCEP/DOE) Atmospheric Model Intercomparison Project Ⅱ (AMIP-Ⅱ) 6-hourly Reanalysis Gaussian Grid data, two experiments (Ⅰ and Ⅱ) were designed to investigate the impact of the percentages of sand and clay in the shallow soil in CoLM on its ability to simulate shallow soil moisture. A third experiment (Ⅲ) was designed to study the shallow soil moisture and latent heat flux simultaneously. In all the three experiments, after the optimization stage, the percentages of sand and clay of the shallow soil were used to predict the shallow soil moisture in the following month. The results show that the optimal parameters can enable CoLM to better simulate shallow soil moisture, with the simulation results of CoLM after the double-parameter optimal experiment being better than the single-parameter optimal experiment in the optimization slot. Furthermore, the optimal parameters were able to significantly improve the prediction results of CoLM at the prediction stage. In addition, whether or not the atmospheric forcing and observational data are accurate can seriously affect the results of optimization, and the more accurate the data are, the more significant the results of optimization may be.
