A global land cover database was developed for use in a wide variety of applications, including modeling, sustainable development, and operational tasks. The database was based on 1992-93 1-km resolution Advanced Very High Resolution Radiometer (AVHRR) satellite data supplemented with other earth science information. The classification methodology was based on an unsupervised classification with extensive interactive post-classification refinement. The resulting land cover data have been translated into several conventional land cover schemes that are commonly used in many applications. The land cover data have been validated using a stratified random sample of higher resolution imagery. Results indicate accuracy in the order of 60-75% overall with higher accuracy reported in tropical and subtropical regions and less accuracy in temperate and boreal zones.