Relating Solar Resource Variability to Satellite-retrieved Cloud Properties

Power production from renewable energy resources is rapidly increasing. Studies of the impact of renewables on the transmission grid require estimates of high temporal and spatial resolution power output under various scenarios. Satellite-based solar resource estimates are the best source of long-term irradiance data but are generally available at lower temporal and spatial resolution than needed and thus require downscaling. Likewise, weather forecast models cannot provide high spatial or temporal irradiance predictions. Downscaling requires information about solar irradiance variability in both space and time, which is primarily a function of cloud properties.

In this study, we analyze the relationships between solar resource variability and satellite-based cloud properties. One-minute resolution surface solar irradiance data were obtained from the National Oceanic and Atmospheric Administration's Surface Radiation (SURFRAD) network. These sites are spread across the United States and thus cover a range of meteorological conditions. Cloud information at a nominal 4 km resolution and half hour intervals was retrieved from NOAA's Geostationary Operation Environmental Satellites (GOES). The retrieved cloud properties were then used to select and composite irradiance data from the measurement sites with the goal of identifying which properties exert the strongest control over short-term irradiance variability. The irradiance variability was characterized using both statistics of the irradiances themselves and their variability in time, as represented by differences computed for short time scales (minutes). The statistical relationships derived using this method will be presented, comparing and contrasting the statistics computed for the different cloud properties. The implications for downscaling irradiances from satellites or forecast models will also be discussed.