Sentinel-2 Madagascar mangrove cover density

Data images are acquired at 1C level and pre-processed at the SEAS-OI station, Réunion island. We used an object-based image analysis method based on NDVI value to classify three level of mangrove density (very dense, dense and less dense).

The attribute table is composed of the object identifier, the vegetation density and the administrative region to which the object belongs. The analysis method at national scale was based on field data collection in three reference sites: in north-western of Madagascar (Mahajamba Bay representing 10% of the malagasy mangrove and Mahavavy Delta ) and in south-western of Madagascar (Tsingilofilo Bay). The overall accuracy of the analysis exceeds 90% for all three sites making it an operational product.