Forest Biomass Modeling Based on Landsat-8 Spectral Indices Using Google Earth Engine

Abstrak

Estimating forest biomass is essential for sustainable forest management and carbon monitoring. This study aimed to develop an aboveground biomass (AGB) estimation model by integrating multispectral Landsat-8 OLI imagery and field measurements in a 95.76 ha rehabilitation area near Rindam II Sriwijaya, Muara Enim, South Sumatra. Field data were collected using the National Forest Inventory Protocol, recording tree diameter and height to calculate AGB through species-specific allometric equations. Several vegetation indices (NDVI, EVI, SAVI, MSAVI, RVI, TVI, NDWI) were derived and analyzed on the Google Earth Engine (GEE) platform to identify the most responsive spectral indicator for biomass estimation.The analysis showed that AGB and carbon stocks varied across the rehabilitation site, reflecting differences in stand structure and vegetation moisture. Among all tested indices, NDWI demonstrated the highest correlation with AGB, indicating its effectiveness in capturing canopy water content and biomass variation under humid, mixed-vegetation conditions. These results emphasize the potential of GEE-based vegetation indices as a cost-efficient and replicable approach for monitoring biomass in tropical rehabilitation forests. NDWI proved to be the most suitable index for modeling forest biomass, offering a practical reference for applying similar remote sensing methods in other tropical regions to support large-scale forest carbon assessments