Global high-resolution land surface albedo monthly maps from NOAA/AVHRR for use in climate models

A set of algorithms was combined for a simple derivation of land surface albedo from measurements of the Advanced Very High Resolution Radiometer (AVHRR) on board NOAA satellites. The system consists of a narrow-to-broadband conversion and bi-directional correction at the top of the atmosphere, and an atmospheric correction. We demonstrate the results with one month worth of data from the NOAA/NESDIS global vegetation index (GVI) weekly dataset and the NOAA/NASA pathfinder atmosphere (PATMOS) project daily data with the original global area coverage (GAC) resolution. Error analysis of the methodology indicates that the surface albedo can be retrieved with 15-20% relative accuracy. Monthly albedos derived from September 1989 GVI and PATMOS data agree well except for small discrepancies attributed mainly to different cloud screening techniques. 5-year mean September albedos derived from the GVI multi-annual time series are consistent with those derived from low-resolution Earth Radiation Budget Experiment data as well as with a September map compiled from ground observations and used in many numerical weather and climate models. Instantaneous GVI-derived albedos were validated against surface measurements. The discrepancies were found to be within the expected 15% relative error, and a 1-2% absolute bias can be attributed to differences in areal coverage and representativeness of the satellite and ground data. The presented pilot study is a prototype for a real-time production of global surface albedo maps from the original GAC resolution data within the NOAA/NESDIS modernized AVHRR processing system (MAPS).