UC San Diego

A method to estimate global horizontal irradiance (GHI) at the surface using a geostationary satellite is presented. The spatial variation of ground and cloud albedo of California is characterized in the 0.55-0.75 [mu]m wavelength spectral region by analyzing a series of images from the visible channel on the GOES West satellite. Using these two characteristic albedo maps a cloud index is generated for each pixel in subsequent images as an estimate of the cloud fraction in that pixel. The cloud index is converted into a clear sky index which is then coupled with the modeled GHI under a clear sky to generate estimates of the GHI under the current cloud conditions. The model was applied to 914 images over 65 days and validated with ground truth measurements in California. The mean bias difference from 119 stations in the California Irrigation Management Information System was 6.7 W/m² (3.2%) and the root mean square difference was 78.6 W/m² (21.9%) which is consistent with previously reported results for satellite algorithms. A Matlab-Java code to execute the model is documented