Angular Distribution Models for Top-of-Atmosphere Radiative Flux Estimation from the Clouds and the Earth's Radiant Energy System Instrument on the Tropical Rainfall Measuring Mission Satellite. Part II: Validation

Abstract

Top-of-atmosphere (TOA) radiative fluxes from the Clouds and the Earth's Radiant Energy System (CERES) are estimated from empirical angular distribution models (ADMs) that convert instantaneous radiance measurements to TOA fluxes. This paper evaluates the accuracy of CERES TOA fluxes obtained from a new set of ADMs developed for the CERES instrument onboard the Tropical Rainfall Measuring Mission (TRMM). The uncertainty in regional monthly mean reflected shortwave (SW) and emitted longwave (LW) TOA fluxes is less than 0.5 W m?2, based on comparisons with TOA fluxes evaluated by direct integration of the measured radiances. When stratified by viewing geometry, TOA fluxes from different angles are consistent to within 2% in the SW and 0.7% (or 2 W m?2) in the LW. In contrast, TOA fluxes based on ADMs from the Earth Radiation Budget Experiment (ERBE) applied to the same CERES radiance measurements show a 10% relative increase with viewing zenith angle in the SW and a 3.5% (9 W m?2) decrease with viewing zenith angle in the LW. Based on multiangle CERES radiance measurements, 1° regional instantaneous TOA flux errors from the new CERES ADMs are estimated to be <10 W m?2 in the SW and <3.5 W m?2 in the LW. The errors show little or no dependence on cloud phase, cloud optical depth, and cloud infrared emissivity. An analysis of cloud radiative forcing (CRF) sensitivity to differences between ERBE and CERES TRMM ADMs, scene identification, and directional models of albedo as a function of solar zenith angle shows that ADM and clear-sky scene identification differences can lead to an 8 W m?2 root-mean-square (rms) difference in 1° daily mean SW CRF and a 4 W m?2 rms difference in LW CRF. In contrast, monthly mean SW and LW CRF differences reach 3 W m?2. CRF is found to be relatively insensitive to differences between the ERBE and CERES TRMM directional models.