<![CDATA[Mangroves play a crucial role in maintaining the stability of coastal ecosystems by providing habitats for diverse species, protecting shorelines from erosion, and acting as a carbon sink. The importance of conserving and developing mangrove areas can be effectively monitored using remote sensing data and classification methods, such as Random Forest (RF), ensuring an accurate assessment and management of these vital ecosystems. This research aims to develop and evaluate an RF classification model to produce accurate spatial information on mangrove canopy cover. The research area, Benoa Bay in Bali, Indonesia, is known for its dynamic and ecologically complex mangrove habitats. The inputs for RF classification are bands on Sentinel-2A satellite imagery, Mangrove Vegetation Index (MVI), Normalized Difference Vegetation Index (NDVI), Enhanced Mangrove Index (EMI), Modified Normalized Difference Water Index (MNDWI), Normalized Difference Moisture Index (NDMI), and the Normalized Difference Salinity Index (NDSalI), along with topographic variables such as elevation and slope. Model validation was conducted using high-resolution imagery from Google Earth Pro and cross-referenced with the 2024 National Mangrove Map. The classification of coastal land cover is divided into water bodies, mangroves, open land, built-up land, and non-mangrove vegetation, with an overall accuracy of 0.98 and a kappa statistic of 0.98. In contrast, the accuracy of the classification of mangrove canopy cover concerning the national mangrove map produces an overall accuracy of 0.97 and a kappa value of 0.86. These findings demonstrate the robustness of the RF model and its potential for supporting data-driven coastal management practices.]]>
Copyrights © 2024
