Comparing Measured and Satellite-Derived Surface Irradiance

The purpose of this study is two-fold: 1) To examine the performance of the Global Solar Insolation Project (GSIP) physics-based model in characterizing global horizontal solar radiation across the United States by comparing to the ground measured data, and 2) to examine improvements of the GSIP data to address temporal and spatial variations. The study enumerates and examines the spatial and temporal limitations of the GSIP model. Most comparisons demonstrate relatively good statistical agreement. However, the methodology used in the satellite model to distinguish microclimate conditions presents significant challenges, and the model requires refinement in addressing aerosol estimates, water vapor estimates, and clear sky optical properties. Satellite derived datasets are only available at half-hour intervals. Surface measurement can easily be made at temporal resolution in the order of seconds. Therefore intra-hour variability, an important quantity for understanding how power production in power plants will vary, cannot be directly derived from satellites. This paper illustrates how intra-hour variability in ground measurements cannot be captured by the satellite based datasets. We also discuss the potential for improved next-generation geostationary satellite data to improve the accuracy of surface radiation estimates.