Discriminating Cloud Over ice Using MISR Data

The uncertainties in the radiative feedback to climate by clouds pose the most formidable obstacle to climate prediction by General Circulation Models (GCMs). Particularly, in polar regions where the ice and snow cover the ground, the net radiative impact of clouds is uncertain (Charlock and Ramanathan 1985; Li and Leighton 1991). One of the reasons for this uncertainty is that scene identification and cloud detection remain difficult over snow- and ice-covered surface. MISR (Multi-angle Imaging SpectroRadiometer) is a sensor in EOS, and contains nine cameras looking at the earth from different angles simultaneously. Its multi-angle information allows a relatively new approach for discriminating cloud from snow/ice. In this talk, we propose a novel method to discriminate cloud over ice based on MISR observations. Demonstrations and validations of the proposed method will also be given based on MISR data from the polar regions.