ESTIMATION OF SURFACE ALBEDO FROM SPACE - A PARAMETERIZATION FOR GLOBAL APPLICATION

This study proposes a simple parameterization to derive surface broadband albedo from satellite observations, using the results of detailed radiative transfer computations for a variety of atmospheric and surface conditions. The end result is a single equation that directly yields surface albedo from observed albedo at the top of the atmosphere, solar zenith angle, and total precipitable water. It was demonstrated that the parameterization is valid for retrieval of both instantaneous and daily mean surface albedo. Sensitivity tests were conducted for precipitable water, aerosol, CO2, 03, and temperature profile. Preliminary validation using collocated satellite and tower measurements indicates that the absolute accuracy requirement of 5% for climate studies is well satisfied. A global monthly climatology (excluding polar areas) of surface albedo is then developed from 5 years of Earth Radiation Budget Experiment clear-sky satellite data and European Centre for Medium-Range Weather Forecasts humidity analysis data. Examination of month-to-month differences for specific 2.5-degrees x 2.5-degrees areas indicate that the absolute random error on monthly estimates is less than 1%. Seasonal variation of surface albedo exceeding 1% can thus be detected. Comparisons with other satellite estimates show much closer agreement than with the values used in some general circulation models and numerical weather prediction models especially over snow/ice and deserts.