Integrating Satellite and UAV Imagery for Mangrove Aboveground Biomass and Carbon Stock Modeling

This study aimed to: (1) quantify aboveground biomass (AGB) and carbon (AGC) stocks in the Banlaem mangrove forest, Nakhon Si Thammarat, Thailand; and (2) construct an AGB estimation model using vegetation indices (VIs) derived from Sentinel-2, Landsat-8, and unmanned aerial vehicle (UAV) imagery. On-the-ground measurements were carried out to evaluate the AGB and AGC stocks of the mangrove forest. VIs were then calculated using passive remote sensing data, including satellite and UAV imagery. These indices were compared through multiple regression analysis with the ground-truthed AGB for evaluation. Three mangrove species were found: Rhizophora mucronata, R. apiculata, and Avicennia Marina. Overall, the AGB and AGC stocks ranged from 0 to 179.78 tons•ha¹ (56.30 ± 51.81 tons•ha¹) and 0 to 89.89 tons•ha¹ (28.15 ± 25.90 tons•ha¹), respectively. The best AGB model exhibited an R² of 0.73 and an RMSE of 22.0 tons•ha¹. This study presents a novel approach for estimating AGB and AGC stocks in the Thai mangrove ecosystem by integrating a UAV with two open-access satellite imagery sources. Combining multiple VIs (NDVI, SAVI, and GNDVI) with CHM provides better accuracy for the mangrove AGB estimation model than using a single variable.