Evaluation of forest loss in Balikpapan Bay in the end of 2015 based on Sentinel-1A polarimetric analysis

Abstract

Satellite remote sensing has proved to be efficient for forest change monitoring. In tropical areas, polarimetric satellite images have a great potential given their ability to see through clouds, smoke and atmospheric haze. For Balikpapan Bay (Borneo, Indonesia), Sentinel-1A acquired images every 24 days during 2015 in both vertically co-polarized and cross-polarized modes. Using series of polarimetric radar images taken before and after an observed event (in this case a fire), information about changes in native forest can be delivered. In this work we detect and delineate areas burnt or damaged by catastrophic fires in autumn 2015. This work demonstrates a potential of satellite radar imagery using a relatively simple method for identification of forest changes. The whole processing chain as presented has been prepared for using open-source software (mainly ESA SNAP). Presented results were compared to both global services (GLAD and FIRMS databases) and local observation (UAV image over burnt area at Bugis canal).