Top-of-atmosphere (TOA) radiative fluxes are estimated from satellite broadband radiance measurements by applying correction factors that account for the angular dependence of the observed scenes. This study uses 8 months of Clouds and the Earth's Radiant Energy System (CERES) measurements aboard the Tropical Rainfall Measuring Mission (TRMM) spacecraft to determine uncertainties in TOA LW flux estimates from new CERES LW Angular Distribution Models (ADMs) for radiance-to-flux conversion. The new CERES ADMs are based on coincident CERES broadband radiances and high-resolution cloud retrievals from the Visible Infrared Scanner (VIRS). Several hundred ADM scene types are defined for different combinations of imager-based retrievals and atmospheric state parameters provided by ECMWF data assimilation analysis. Fluxes from the new CERES ADMs are compared with those based on ADMs from the Earth Radiation Budget Experiment (ERBE) and a theoretical model [Stubenrauch et al. 1993].

Longwave all-sky regional mean flux errors from the new CERES ADMs remain smaller than 0.5 W m-2, compared to 1.5 W m-2 for the ERBE ADMs. This reduction in error is shown to be largely due to increased stratification of the new ADMs by scene type. The theoretical model provides accurate clear-sky average LW fluxes, but overestimates the flux for overcast scenes by approximately 3 W m-2. To demonstrate the influence of a variable CERES field-of-view (FOV) size with viewing geometry on TOA flux estimates, full-resolution TOA fluxes are compared with those from 60-km reduced-resolution CERES FOVs.