Remote Sensing Approach to Estimate Vegetation Aboveground Biomass Affected by Land Conversion in the Southern Area of Malang Regency

Accurate estimation of vegetation Aboveground Biomass (AGB) is crucial for understanding carbon dynamics and supporting climate change mitigation strategies. This study explores the integration of remote sensing datasets and machine learning models to estimate AGB in South Malang Regency, Indonesia, a region characterized by diverse landscapes and ongoing infrastructure development. By utilizing GEDI L4A AGBD data as the target variable, and predictors derived from Sentinel-2 and Copernicus GLO-30 DEM, the study leverages Random Forest regression to generate precise biomass predictions. The results demonstrate the model's high predictive accuracy, with an RMSE of 34.719 and an R2 of 0.709, indicating strong alignment between observed and predicted values. The total estimated AGB for the study area in 2024 is approximately 34,003.34 Mg, reflecting minimal changes compared to prior years and suggesting balanced land-use dynamics. These findings underscore the potential of remote sensing and machine learning to provide actionable insights for sustainable forest management and land-use planning, particularly in regions undergoing rapid development.