Remote Sensing Monitoring of Seagrass Beds as Marine Carbon Sinks and Assessment of Ecological Restoration Benefits,—A Case Study of Longdao Island, Caofeidian

To accurately assess the carbon sequestration status and ecological restoration potential of seagrass beds at Caofeidian Longdao Island, and to support coastal wetland blue carbon management and the ”dual carbon” goals, this study uses Sentinel-2 imagery as the core data source. Based on the Google Earth Engine (GEE) platform, we propose an integrated seagrass extraction method combining multi-temporal bright green period compositing, multispectral index enhancement, and random forest classification. Coupling the InVEST model with field survey data, we establish a full-process digital technical system for temperate seas: remote sensing monitoring–carbon stock estimation–restoration simulation–benefit assessment.We preprocessed long-term Sentinel-2 images (2019–2024) on GEE; the extraction model achieved an overall accuracy above 90%, enabling high-precision automated mapping of seagrass distribution. By integrating EVI with the InVEST carbon module, we refined vegetation and soil carbon stock estimation, improving the spatial representativeness of traditional methods.Results show that from 2019 to 2024, seagrass area expanded from 2,922.44 ha to 4,426.36 ha, and total carbon stocks rose from 34,300 t to 51,900 t, reflecting a virtuous cycle of area expansion, biomass growth, and enhanced carbon sequestration. This study is the first long-term dynamic monitoring and restoration benefit assessment for temperate seagrass beds in Caofeidian Longdao. The ”GEE–Sentinel-2–InVEST” framework fills a technical gap in northern temperate seagrass carbon monitoring, supporting seagrass conservation, blue carbon development, and ecological management in northern coastal zones.