Impact of geolocations of validation data on the evaluation of surface incident shortwave radiation from Earth System Models

Ground‐based observations of surface incident solar radiation ( R s ) have been used to evaluate simulations of global climate models. Existing studies have shown that biases in simulated clouds have a significant spatial pattern, which may be transferred to the simulated R s . Therefore, the evaluation results of R s simulations may depend on the locations of the ground‐based observations. In this study, R s simulations of 48 models participating in the fifth phase of the Coupled Model Intercomparison Project (CMIP5) were first evaluated with ground‐based observations from different networks (446 stations in total) from 2000 to 2005. The global mean biases of the CMIP5 R s simulations were found to vary from 4.8 to 11.9 W m −2 when R s observations from different networks were used as reference data. To reduce the location impact on the evaluation results, CMIP5 simulated R s was then evaluated with the latest satellite R s retrieval at 1° × 1° spatial resolution by the Clouds and the Earth's Radiant Energy System, Energy Balanced and Filled (CERES EBAF). It was found that the CMIP5 simulated multimodel mean R s has a small bias of 2.6 W m −2 compared with CERES EBAF over the globe, 4.7 W m −2 and 1.7 W m −2 over land and oceans, respectively. CERES EBAF R s was found to have a positive bias of 1.3 W m −2 compared with ground‐based observations. After removing this bias of CERES EBAF R s , global mean R s was estimated to be 185 W m −2 .