RGB-NDVI color composites for monitoring the change in mangrove area at the Maubesi Nature Reserve, Indonesia

The Maubesi Nature Reserve (MNR) is a protected lowland area in eastern Indonesia that mainly consists of mangrove forest. The objective of this paper was to demonstrate a simple technique to visualize and quantify the change in mangrove area using a 3-year dataset of Landsat TM images acquired in 1989, 2003 and 2009. The normalized difference vegetation index (NDVI) was calculated to determine high and low vegetation biomass in each image. Each NDVI extracted by Landsat image in 1989, 2003 and 2009 was assigned to red, green and blue (RGB) color, respectively, and then combined to make color composites. Additive color theory was applied to interpret mangrove changes within the MNR area on the RGB-NDVI color composite. Changed areas were quantified by performing an unsupervised classification on the RGB-NDVI image with 45 classes that were grouped into eight major mangrove change categories. An Indonesian land cover map was used to assess the accuracy of the classified image. The result showed that 77.13% of the MNR area was unchanged and 22.87% of the MNR area changed over 20 years (1989-2009).