Spatial Modeling of Forest Carbon Stocks: Integrating Engineering Framework, Field Data, and Sentinel-2 Imagery
نویسندگان
1 Space Technology and Geo-Informatics Research Unit, Faculty of Science, Mahasarakham University, Maha Sarakham, Thailand
2 Department of Biology, Faculty of Science, Mahasarakham University, Maha Sarakham, Thailand
3 Department of Physics, Faculty of Science, Mahasarakham University, Maha Sarakham, Thailand
4 Greenhouse Gas Research Center and Operations, Faculty of Science, Mahasarakham University, Maha Sarakham, Thailand
5 Department of Physics, Faculty of Science, Mahasarakham University, Maha Sarakham, Thailand
6 Greenhouse Gas Research Center and Operations, Faculty of Science, Mahasarakham University, Maha Sarakham, Thailand
doi
10.5829/ije.2026.39.11b.12چکیده
Reliable estimation of forest carbon stocks is essential for sustainable land management and climate-change mitigation. This study presents an engineering-based spatial framework that integrates field-measured aboveground biomass with Sentinel-2 satellite imagery to estimate and validate the spatial distribution of aboveground carbon in Tat Ton National Park, northeastern Thailand. Three vegetation indices—NDVI, MSAVI2, and GRVI—were derived from Sentinel-2 data and converted to fractional vegetation cover (FC) using a linear scaling approach based on vegetation and bare-soil reference conditions. The FC metrics were then used as predictors in linear regression-based carbon models. Field plots were systematically distributed across representative forest types and environmental gradients within the park to capture heterogeneous forest structures and support robust model validation. The results indicate moderate to strong correlations between satellite-derived and field-measured carbon stocks, with coefficients of determination (R²) ranging from 0.6964 to 0.7885. Among the tested indices, NDVI_FC showed the closest agreement with field measurements, while GRVI_FC exhibited the highest statistical correlation. MSAVI2_FC tended to overestimate carbon stocks under dense canopy conditions.
