Mapping Mexico's Forest Lands with Advanced Very High Resolution Radiometer

Data from the Advanced Very High Resolution Radiometer (AVHRR) were used in a program sponsored by the U.S. Department of Agriculture, Forest Service, and the United Nations Food and Agriculture Organization to help scientists from Mexico generate forest-cover maps of that country. Two near-cloud-free composite images were generated for December and March 1990 from AVHRR Local Area Coverage data sets. A supervised training and classification technique was used to process data in nine physiographic regions of the country. Regional classifications were combined to produce the final map. Classification accuracy for 60 test polygons was 78 percent correct for the polygons and 84 percent correct as evaluated by pixels in selected polygons. More detailed physiographic or ecologic partitioning of the country was cited as a way to improve results of the classification technique used in this study.