The effect of subpixel clouds on remote sensing

Abstract

The effect of subpixel clouds on remote sensing of the surface reflectivity and of the vegetation index was numerically simulated for a thin layer of rectangular clouds. The simulation is for a 4 × 4km² field of view (e.g. the low resolution NOAA-AVHRR images). A fixed cloud reflectivity was used as well as an empirical cloud model for eight cloud types. In the empirical models the cloud reflectivity varied with cloud fraction. A cloud fraction of 20 percent, in a pixel with surface reflectance of 005, can increase the apparent surface reflectance by 002 for a model of strato-cumulus and by 0.08 for a fixed cloud reflectance of R_cequals;0.5. By using measured cloud fraction probability distributions in several climatic regions (not including a tropical climate) it was found that the cloud effect can be eliminated if the best remote sensing case out of about four independent cases is chosen. This corresponds to 8-16 observation days. In order to estimate the cloud effect on remote sensing of a particular area it is necessary to measure the probability distribution of the cloud fraction and the dependence of the cloud reflectance on the cloud size. Subpixel clouds were shown to affect the variation of upward radiance across an image. Therefore the standard deviation of the radiances over a uniform area can be used to sense the presence of the residual cloud effect.