Mangrove Forests Monitoring Using Google Earth Engine Platform in the Extreme North-East of Madagascar
摘要
2% of the global world’s mangrove forests are located in Madagascar. This ecosystem provides essential goods and services to millions of people residing in coastal communities. However, mangrove forest decreases rapidly due to anthropogenic pressures and climate change. Remote sensing observation is very useful and provides precious information on the land cover changes. In this study, we present the Google Earth Engine (GEE) based approach for mangrove forests mapping in the extreme North-East of Madagascar using Sentinel 2 satellite imagery. We used the RandomForest classifier and a post classification change detection approach. The overall classification accuracies for the years 2016 and 2020 of land cover maps were 91% and 92%, respectively. The change detection results showed a significant loss of 144 ha (4%) of mangroves during the study period. This encouraging result contributes to the understanding of the dynamic of mangrove change and presents a potential tool for monitoring and managing wetland ecosystem.