Forest mapping in tropical region using multitemporal and interferometric ERS-1/2 data

Abstract

The paper presents a study to assess the use of ERS SAR data to map and monitor deforestation in Sumatra, a region characterized by rapid depletion of the rain forest, mainly due to logging and permanent or semi-permanent settling of transmigrated population. Two approaches were used and compared (with reference to optical data) for forest / non forest mapping : 1- Temporal Change approach : in the frame of the TREES/ERS-1 SAR'94 project, previous studies have shown that up to a certain limit forested areas could be distinguished from non forested areas using the particularly stable temporal behavior of the backscatter values of the forest. 2- Interferometric approach : coherence images are realized from the interferometric pairs. Forest/non forest maps are realized based on the contrast between coherence of forest compared to that of the other land covers. Both methodology have been applied over different test sites in Central Sumatra and South Sumatra. This paper presents the results obtained over one specific site . Kayuagung in Southern Sumatra which currently undergoes active deforestation.