Quantifying Carbon Emissions from the Sundarbans: A Multi-Sensor Approach to Assessing Degradation, Land-Use Change, and Cyclone Impacts

The Sundarbans, the world’s largest contiguous mangrove forest, is an important blue-carbon ecosystem increasingly affected by chronic forest degradation, land-use conversion, and tropical cyclones. However, quantifying the carbon consequences of these disturbances remains challenging, particularly when distinguishing persistent land-cover change from transient cyclone-induced damage. This study developed a multi-sensor remote sensing framework to quantify carbon emissions associated with chronic degradation, permanent land-use conversion, and cyclone-induced disturbance across the Sundarbans during 2013–2025. Landsat 8/9 and Sentinel-2 imagery were used to characterize land-cover and vegetation changes, with the Normalized Difference Vegetation Index (NDVI) and Leaf Area Index (LAI) serving as proxies for vegetation condition and biomass dynamics. Sentinel-1 C-band Synthetic Aperture Radar (SAR) VV and VH backscatter changes from pre- and post-cyclone observations were used to characterize cyclone-induced canopy disturbance. A Markov–Cellular Automata (MCA) model was applied to characterize persistent land-cover transitions, while transient cyclone effects were identified from event-based SAR and vegetation-index changes followed by post-disturbance recovery. This combined approach enabled persistent land-use conversion to be differentiated from temporary cyclone-related vegetation disturbance. Remote sensing-derived disturbance areas were linked with published mangrove aboveground biomass estimates to approximate biomass losses. These losses were converted to carbon using a carbon fraction of 0.47 and subsequently expressed as CO₂ equivalents using the molecular-weight ratio of 44/12. The analysis indicates average gross emissions of approximately 9.6 million tonnes of CO₂ yr⁻¹ during 2013–2025, with chronic degradation accounting for 55% of the estimated emissions, permanent land-use conversion for 28%, and cyclone-induced damage for 17%. The principal sources of uncertainty include satellite-derived land-cover classification and vegetation-index variability, spatial variation in mangrove biomass, biomass-to-carbon conversion parameters, attribution of disturbance among overlapping drivers, and assumptions regarding post-cyclone vegetation recovery. Overall, the findings demonstrate the value of integrating optical and SAR observations with spatial land-change modelling to improve disturbance-specific carbon accounting and support targeted mangrove conservation and climate-mitigation strategies.