Title: Temporal and spatial downscaling for solar radiation

Abstract: Global and regional climate model projections are useful for gauging future patterns of climate variables, including solar radiation, but data from these models is often too coarse to assess local impacts. Within the context of solar radiation, the changing climate may have an effect on photovoltaic (PV) production, especially as the PV industry moves to extend plant lifetimes to 50 years. Predicting PV production while taking into account a changing climate requires data at a resolution that is useful for building PV plants. Although temporal and spatial downscaling of solar radiation data is widely studied, we present a novel method to downscale solar radiation data from daily averages to hourly profiles, while maintaining spatial correlation of parameters characterizing the diurnal profile of solar radiation. The method focuses on the use of a diurnal template which can be shifted and scaled according to the time or year and location and the use of thin plate splines for spatial downscaling. This analysis is applied to data from the National Solar Radiation Database housed at the National Renewable Energy Lab and a case study of the mentioned methods over several sub-regions of continental United States is presented.