Long-term monitoring and predictive modeling of mangrove ecosystem health using Sentinel-2 data in coastal regions of Pakistan (2015–2024)

Mangrove ecosystems along Pakistan’s coastline provide critical ecological services, including coastal protection, carbon sequestration, and biodiversity support, but face increasing threats from sea-level rise, salinity intrusion, and reduced freshwater flows. This study integrates multi-index remote sensing (RS), machine learning (ML), and climate projection modeling to assess mangrove health, spatial dynamics, and future vulnerability from 2015 to 2024. Sentinel-2 imagery was analyzed using the Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), Modified Soil-Adjusted Vegetation Index (MSAVI), Normalized Difference Water Index (NDWI), Modified Normalized Difference Water Index (MNDWI), Normalized Difference Moisture Index (NDMI), Moisture Vegetation Index (MVI), Salinity Index (SI), and Bare Soil Index (BSI). A Random Forest (RF) classifier implemented in Google Earth Engine (GEE) mapped mangrove extent and quantified temporal change. Results show a net loss of ~10 km² over the decade, despite 104 km² of gains and 114 km² of losses. Approximately 458 km² remained stable, while 166 km² exhibited transitional dynamics, indicating localized resilience. Climate projections from the Coupled Model Intercomparison Project Phase 6 (CMIP6) under Shared Socioeconomic Pathways (SSP2-4.5 and SSP5-8.5) suggest relative stability under moderate emissions but heightened vulnerability under high-emission scenarios. This integrated framework supports climate-adaptive mangrove conservation and long-term coastal resilience planning.