Cloud screening for determination of land surface characteristics in a reduced resolution satellite data set

Publisher Website
Google Scholar
Abstract
A spatial coherence method is developed for cloud screening of a reduced resolution data set over the U.S. Great Plains, The method is based upon a comparison of the spatial standard deviation of scene radiance for a particular observation time with that for clear sky conditions (the background variability). The procedure relies on the assumption that the presence of partial cloudiness increases spatial variability. The standard deviation thresholds are tabulated functions of space and time and are presented as the background variability maps. Complete cloud covers are screened by testing the computed skin temperatures and albedos against constant thresholds, whereas night-time cirrus and low levels of cloudiness are delected using the 3.7-11 μm brightness temperature differences. The cloud-screening scheme is tested on daytime and night-time data of July 1986.