A BAND-DIFFERENCED ANGULAR SIGNATURE TECHNIQUE FOR CIRRUS CLOUD DETECTION

A new approach to cloud detection is introduced that exploits the diff erence between two solar spectral reflectances as a function of view a ngle. The resulting band-differenced angular signature is sensitive to the contribution of Rayleigh scattering from above the tops of clouds and can be used to discriminate high clouds from lower level clouds a nd clear sky. We use model simulations to show that this technique cou ld be applied to measurements from the Multiangle Imaging SpectroRadio meter, scheduled for launch on the first platform of the Earth Observi ng System. Results show the technique to work best over ocean and snow surfaces. Over such surfaces, the minimum detectable high cloud optic al thickness (at 0.55 mum) would typically be 0.5 without the use of a ny a priori scene information, and lower if such information is availa ble.