Forest Fragmentation Analysis in the Municipalities of Corumbiara and Buritis, Rondônia State, Brazil

ABSTRACT

The expansion of agriculture and pasture in Rondônia state was accelerated in the past decade due to fiscal incentives and higher external market demand. Consequently, serious forest degradation processes were initiated. The present study aims to analyze forest fragmentation in the municipalities of Corumbiara and Buritis. To this end, Landsat 5 TM images were used for forest and non-forest classification. The accuracy assessment for Maximum Likelihood supervised classification indicated an overall accuracy of 91% and Kappa of 78%. Results indicated an increase in the number of patches in the area of study in the past decade, with an increase in edge effects, significant reduction of forest class area, and a decrease in the connectivity of forest patches. Therefore, based on this study, a continuous increase in the fragmentation of native vegetation in these municipalities can be observed.

Keywords:
remote sensing; landscape analysis; land-use changes