The National Solar Radiation Database (NSRDB), which is a spatially-dense modeled dataset intended to accurately represent long-term statistics, was used to construct a 15-year climatology of variations in the direct-beam and diffuse radiation fields across the United States (US). Continental scale variations in the radiation field vary predominantly by latitude during the winter and longitude during the summer. Eight sub-regions, defined by similar climate forcing and geography, demonstrate that the largest seasonal and year-to-year variations in the diffuse-to-direct ratio are in the Northwest and Southwest, while the smallest are in the Northeast and Southeast. The year-to-year variability across the western US is attributed to large scale climate signals, such as the El Nino Southern Oscillation (ENSO). A preponderance of diffuse radiation during the summer in the Midwest, South-Central, Southeast and Northeast is demonstrated. It is suggested that harvesting procedures be tuned to accommodate this preponderance. Orienting a fraction of fixed panels to optimize the diffuse beam during the summer in these regions is suggested to increase energy yield. Adopting a seasonal and year-to-year approach in the harvesting procedure in the Northwest and Southwest that is tuned to large scale climate signals is likely to optimize harvest yield in those regions. (C) 2018 The Authors. Published by Elsevier Ltd.