Top-of-Atmosphere Radiative Fluxes: Validation of ERBE Scanner Inversion Algorithm Using Nimbus-7 ERB Data

John T. Suttles Atmospheric Sciences Division, NASA Langley Research Center, Hampton, Virginia

Search for other papers by John T. Suttles in
Current site
Google Scholar
PubMed
Close
,
Bruce A. Wielicki Atmospheric Sciences Division, NASA Langley Research Center, Hampton, Virginia

Search for other papers by Bruce A. Wielicki in
Current site
Google Scholar
PubMed
Close
, and
Sastri Vemury Scientific Management and Applied Research Technologies, Inc, Silver Spring, Maryland

Search for other papers by Sastri Vemury in
Current site
Google Scholar
PubMed
Close
Restricted access

Abstract

To derive top-of-atmosphere radiative fluxes from measurements of the Earth Radiation Budget Experiment (ERBE) scanning radiometers, it is necessary to convert radiance observations to fluxes. Two key steps of the ERBE radiance-to-flux conversion are the identification of the scene being viewed and the application of an angular model for the particular scene. In this study, the ERBE algorithm is applied to the Nimbus-7 earth radiation budget (ERB) scanner data for June 1979 to analyze the performance of this inversion method in deriving top-of-atmosphere albedos and longwave radiative fluxes. The performance is assessed by comparing ERBE algorithm results with appropriate results derived using the sorting-by-angular-bins (SAB) method, the ERB MATRIX algorithm, and the “new-cloud ERB” (NCLE) algorithm. Comparisons are made for top-of-atmosphere albedos, longwave fluxes, viewing zenith-angle dependence of derived albedos and longwave fluxes, and cloud fractional coverage. Using the SAB method as a reference, the rms accuracy of monthly average ERBE-derived results are estimated to be 0.0165 (5.6 W m−2) for albedos (shorlwave fluxes) and 3.0 W m−2 for longwave fluxes. The ERBE-derived results were found to depend systematically on the viewing zenith angle, varying from new nadir to near the limb by about 10% for albedos and by 6%–7% for longwave fluxes. Analyses indicated that the ERBE angular models are the most likely source of the systematic angular dependence. Comparison of the ERBE-derived cloud fractions, based on a maximum-likelihood estimation method, with results from the NCLE showed agreement within about 10%. Little correlation was found between regional albedo and longwave-flux differences and corresponding cloud-fraction differences.

Abstract

To derive top-of-atmosphere radiative fluxes from measurements of the Earth Radiation Budget Experiment (ERBE) scanning radiometers, it is necessary to convert radiance observations to fluxes. Two key steps of the ERBE radiance-to-flux conversion are the identification of the scene being viewed and the application of an angular model for the particular scene. In this study, the ERBE algorithm is applied to the Nimbus-7 earth radiation budget (ERB) scanner data for June 1979 to analyze the performance of this inversion method in deriving top-of-atmosphere albedos and longwave radiative fluxes. The performance is assessed by comparing ERBE algorithm results with appropriate results derived using the sorting-by-angular-bins (SAB) method, the ERB MATRIX algorithm, and the “new-cloud ERB” (NCLE) algorithm. Comparisons are made for top-of-atmosphere albedos, longwave fluxes, viewing zenith-angle dependence of derived albedos and longwave fluxes, and cloud fractional coverage. Using the SAB method as a reference, the rms accuracy of monthly average ERBE-derived results are estimated to be 0.0165 (5.6 W m−2) for albedos (shorlwave fluxes) and 3.0 W m−2 for longwave fluxes. The ERBE-derived results were found to depend systematically on the viewing zenith angle, varying from new nadir to near the limb by about 10% for albedos and by 6%–7% for longwave fluxes. Analyses indicated that the ERBE angular models are the most likely source of the systematic angular dependence. Comparison of the ERBE-derived cloud fractions, based on a maximum-likelihood estimation method, with results from the NCLE showed agreement within about 10%. Little correlation was found between regional albedo and longwave-flux differences and corresponding cloud-fraction differences.

Save