Stochastic cloud clearing of hyperspectral radiances observed by the Atmospheric InfraRed Sounder (AIRS) on the Aqua satellite

This paper presents a stochastic approach to estimating cloud perturbations of hyperspectral radiances observed by the Atmospheric InfraRed Sounder (AIRS) on the NASA Aqua satellite, launched in May 2002. The initial cloud-clearing algorithm uses the AIRS/AMSU/HSB infrared and microwave observed radiances as input, estimates the effects of clouds based on 316 infrared radiances and 5 microwave channels in various combinations, and then compensates these radiances for the cloud effect. Each cloud-cleared radiance spectrum corresponds to a ~50-km cell for which the infrared spectra of 9 independent subcells with ~15-km resolution are examined. The nine subcells are arranged in a 3X3 array. Each 50-km estimate is based on spectra from two or more 15-km subcells. Metrics for the efficacy of these cloud-clearing algorithms are presented and then used to compare the proposed stochastic cloud-clearing algorithms and alternatives. AIRS observes 2378 spectral channels in the range 3.7 to 15 microns wavelength. AMSU/HSB observes 19 microwave channels at frequencies from ~23.8 to 190 GHz with comparable spatial resolution.