QUANTIFYING DEFORESTATION IN A PERMANENT FOREST RESERVE USING VECTORISED LANDSAT TM

Abstract

Vectorising Landsat TM image (127/56 path/row) was carried out in this study to quantify deforestation using integration of remote sensing and geographic information system (GIS) in Permanent Forest Reserve (PFR) of northern Kelantan, Malaysia (latitudes 04° 30' N to 06° 15' N and longitudes 101° 20' E to 102° 40' E). The PFR boundary map in 1989 was digitised to analyse physical factors that affect deforestation. By using minimum distance supervised
classification, Landsat TM is an effective tool for distinguishing forested area from non-forested area such as rubber, oil palm and paddy with overall accuracy of about 84.8% and mean accuracy of between 58.3 and 97.7%. The rate of deforestation was about 2.3% per year with 73 236 ha of the PFR in the northern part of Kelantan being converted into non-forested purposes within eight years (1989-1997). The change from forest into mixed crop or rubber within eight years was estimated to be 19 252.1 ha. However, 53 984.1 ha of forest were converted into shrubs, while approximately 1178.5 ha of forest were converted into grassland/open areas. This study implies that RS/GIS technique is a useful tool in quantifying deforestation in PFR.