Abstract
The ability of a satellite‐based model (Gautier et al., 1980) to resolve the mesoscale variability of solar irradiance over the lower Fraser Valley is assessed. The estimated and observed irradiances are found to have a similar spatial coherence. The estimates do not reproduce the anisotropy associated with mountain‐lowland variations in cloudiness. This discrepancy occurs as a result of spatial averaging inherent in the satellite‐based methodology. Large errors of estimate of the hourly solar irradiance induce large errors of optimal interpolation. However, over the long‐term, errors are small and results suggest that differences in interpolated irradiances determined from observed and estimated data may be minor.