Integrating multi-index remote sensing and machine learning for mangrove dynamics assessment using sentinel-2 imagery

Mangrove ecosystems in the Indus Delta provide critical coastal protection, carbon storage, and biodiversity support but remain vulnerable to environmental change. This study integrated sentinel-2 imagery, nine spectral indices, and random forest machine learning to evaluate mangrove dynamics between 2018 and 2024. Multi-temporal classification achieved high accuracy and enabled detailed assessment of land-cover transitions across the delta. Mangrove extent increased from 70.69 km² in 2018 to 117.05 km² in 2024, representing a 65.6% expansion. Change detection revealed that regeneration (50.54 km²) substantially exceeded degradation (3.43 km²), indicating strong ecosystem recovery. Most expansion occurred through colonization of previously bare tidal flats and succession from non-mangrove vegetation. These findings demonstrate the effectiveness of integrating multi-index remote sensing with machine learning for coastal ecosystem monitoring and provide valuable information for mangrove conservation, restoration planning, and sustainable management of vulnerable deltaic environments.