An efficient and effective method to simulate the earth spectral reflectance over large temporal and spatial scales

[1] Atmospheric and surface properties have been measured from space with various spatial resolutions for decades. It is very challenging to derive the mean solar spectral radiance or reflectance over large temporal and spatial scales by explicit radiative transfer computations from the large volume of instantaneous data, especially at high spectral resolution. We propose a procedurally simple but effective method to compute the solar spectral reflectance in large climate domains, in which the probability distribution function (PDF) of cloud optical depth is used to account for the wide variation of cloud properties in different sensor footprints, and to avoid the repeated computations for footprints with similar conditions. This approach is tested with MODIS/CERES data and evaluated with SCIAMACHY measured spectral reflectance. The mean difference between model and observation is about 3% for the monthly global mean reflectance. This PDF-based approach provides a simple, fast, and effective way to simulate the mean spectral reflectance over large time and space scales with a large volume of high-resolution satellite data.