Use of remotely sensed radar data to assess tropical deforestation in Guyana.

The purpose of this study was to assess tropical deforestation in Guyana using ERS-1 SAR data. The ERS-1 data were analyzed using a general sequence of digital image processing and analysis procedures. The image was acquired on June 3, 1992. Although a series of preprocessing and image enhancement operations were used, it was only through visual interpretation that areas of deforestation were able to be identified. Supervised classification was not successful at delineating these areas based on their spectral properties and produced low accuracies under the conditions presented in this study. While system and surface factors were acknowledged to limit the utility of ERS-1 data for forestry studies, many other variables influenced the results of the study. Logging methods in Guyana may be a important reason why deforested areas are not easily detected on satellite imagery. Also, when considering that forestry production increased almost 200 percent from 1993 to 1994, it is apparent that the 1992 image could not reveal the accelerated amount of timber harvesting that has been occurring in recent years. Even though extensive deforestation was not visible in the study area, indicators suggest that there is increasing pressure being placed on the forests of Guyana. Based on the enhanced utility of SAR data when multiple images are employed, further research may yield significant results. The potential of SAR data for tropical deforestation studies is very promising. (Abstract shortened by UMI.)Dept. of Geography. Paper copy at Leddy Library: Theses & Major Papers - Basement, West Bldg. / Call Number: Thesis1996 .K553. Source: Masters Abstracts International, Volume: 37-01, page: 0220. Adviser: V. C. Lakhan. Thesis (M.A.)--University of Windsor (Canada), 1996