Mangroves, wetland plant communities thriving in intertidal zones and along tidal rivers, play a crucial role in carbon sequestration, coastline protection, and biodiversity conservation. However, they are currently facing numerous threats. The advancements in remote sensing big data and cloud computing platforms have provided new opportunities for spatial mapping of mangroves. In ths study, we analyze global mangrove changes using a deep learning classification method based on spectral and morphological features. Our findings indicated that from 1985 to 2020, a significant loss of mangroves occurred across most regions of the world, particularly in Southeast Asia. Sea-level rise and human activities have negative impacts on this decline to some extent. 

multi-source data,mangrove changes,remote sensing