Country-level evaluation of solar radiation data sets using ground measurements in China

Solar radiation is a crucial parameter that affects the thermal environment in buildings. The spatial and temporal distributions of solar radiation data are important for energy-efficient building design. Models that correlate solar radiation with other parameters can address the lack of solar radiation data. Satellite-derived products and reanalysis data sets have been produced using solar radiation models. The accuracy of these products directly affects building thermal environment design. To choose the most appropriate data set, it is necessary to evaluate the deviation in different data sets based on ground measurements. We used data acquired between 2001 and 2016 from 98 solar radiation measurement stations in China to verify two satellite-derived products (SARAH-E and CERES-SYN1deg) and two reanalysis data sets (ERA5 and MERRA-2). The CERES-SYN1deg and SARAH-E products performed better than the ERA5 and MERRA-2 data sets at estimating the daily global solar radiation. The daily global radiation products were more accurate than direct, diffuse, and hourly global solar radiation products. The models merged ground measurements show good performance. Further improvement in solar radiation estimation especially direct and diffuse in areas where there are no ground measurements and taking into account the effect of inadequate weather conditions on the hourly solar radiation is required. These findings may provide the basis for solar radiation models and products, especially applications in the building industry. • Global solar radiation in four products validated at 98 stations. • Satellite-derived products more effective than reanalysis data sets. • Accuracy of global solar radiation data sets better than direct solar radiation data sets.