Forest Mapping of Central America and Mexico with AVHRR Data

Abstract

Concerns over changes in global forest resource distributions have prompted a number of studies to examine and map forest areas at continental scales with various types of satellite data. The work described here details the use of Advanced Very High Resolution Radiometer (AVHRR) data in concert with Landsat Thematic Mapper (TM) and Systeme Probatoire d'Observation de la Terre (SPOT) imagery to identify and map forest land in Mexico and Central America. A two stage approach was used to accomplish the study objectives. First, a modeling procedure was used to estimate percent forest cover in AVHRR pixels based on enumeration of forest area with Landsat TM and SPOT data of the same region covered by the AVHRR data. This product was used to subset the original AVHRR data into areas of probable forest lands. The AVHRR spectral data of the subset area was then classified by forest type and the results were compiled to produce the final product. The implications of this work are discussed in the context of global forest monitoring and landscape management.