Estimating Above-Ground Biomass Using Landsat 8 Imagery: A Case Study of Deciduous Broadleaf Forest in Dak Lak Province, Vietnam

Assessing the Above Ground Biomass (AGB) is vital for a better awareness of forest carbon sequestration potential. Traditional field-based methods for measuring the AGB are often long-lasting, expensive, and limited in their spatial scope. Recent studies have revealed that using a universal regression model to quantify AGB at a global or country extent is not feasible. Instead, it is necessary to create local regression models that can accurately estimate AGB on a provincial scale. However, the pattern of AGB in a deciduous broadleaf forest in Dak Lak province is currently unknown. Therefore, this study aims to develop regression analysis-based predictive models specifically for evaluating the spatial pattern of AGB in a deciduous broadleaf forest in Dak Lak province. The multiple bands and vegetation indices derived from Landsat 8 imagery were employed as independent variables, while AGB measurements were defined as the dependent variable. The statistical analysis results indicate a strong correlation between vegetation indices and measured AGB data, confirming the effectiveness of using Landsat 8 imagery for assessing AGB. The models show reasonably good performance, achieving R2 values varying from 0.61 to 0.62 and RMSE values varying from 30.68 to 31.54. The estimated AGB in the forests averages 100.80 Mg/ha with a standard deviation of 44.27. The regression model-derived spatial distribution of AGB in Dak Lak reveals the variation in AGB across the forest area, highlighting areas with high and low AGB in the province. This map can serve as baseline information for future AGB estimates in the area, contribute to carbon sequestration monitoring efforts, and support improved local forest management practices.