An empirical solution for the AIRS scene mirror contamination from adjacent footprints

Abstract

AIRS has provided highly stable and accurate radiances since 2002, which has exceeded the expected instrument noise level. Despite the highly accurate data our analysis has shown warming as high as 5 Kelvin in the shorter wavelength portions of the AIRS spectrum for extremely cold scenes (less then 230 Kelvin), such as when AIRS views deep convective clouds. The warming only exists when AIRS is viewing a cold scene with neighboring warmer scenes. This work will demonstrate the spatial characteristics of these trends in relation to where the warm area is to the scene AIRS is viewing. Lastly, we will show that a linear regression correction approach that will account for the trend in the shorter wave portions of the AIRS spectrum.