Evaluating CERES Terra and Aqua longwave and window top-of-atmosphere flux uncertainties using direct integration test

Understanding climate and its variability is largely dependent upon accurate and continuous monitoring of the Earth's radiation budget. Over many years, the Clouds and Earth's Radiant Energy System (CERES) instruments have provided satellite-measured radiances, from which top-of-atmosphere fluxes are inferred by applying angular distribution models (ADMs). This paper evaluates the uncertainties in regional mean TOA longwave and window (8-12 micron) fluxes from ADMs developed from Terra and Aqua CERES data. Terra and Aqua validation efforts have shown significant improvements in TOA flux accuracy relative to previous radiation budget products.

The uncertainties in regional mean TOA fluxes are evaluated by comparing ADM-derived fluxes with fluxes obtained by the direct integration (DI) of measured radiances on a region-by-region basis. Since DI technique is independent of scene identification and angular modeling, it is an effective tool in checking the accuracy of ADM-derived fluxes. Longwave bias errors are in the order of 0.2-0.4 W/m2 with regional RMS errors < 0.7 W/m2. Window bias errors are less than 0.25 W/m2 with regional RMS errors < 0.35 W/m2. Although uncertainties are generally similar for both Terra and Aqua, discrepancies were noted in the polar regions which is attributed to differences between CERES cloud masks applied to the imager (MODIS) in each platform.