National scale estimation of above-ground biomass density in Nepal's forests at 30 m resolution (circa 2015).

Accurate estimates of forest Above-Ground Biomass Density (AGBD) are pivotal for quantifying carbon stocks and guiding climate mitigation. However, existing global and national products often lack the accuracy and spatial detail required for mountainous, heterogeneous landscapes such as Nepal. This study provides Nepal’s national AGBD baseline for circa 2015 at 30 m resolution. We integrated 965 plot-based AGBD observations with 10 topographic and spectral predictors derived from a digital elevation model, Landsat-8, and ALOS-PALSAR. After removing outliers, we trained a decision tree algorithm, Classification and Regression Tree (CART), and assessed its performance using 5-fold cross-validation. The results were promising, with an R² of 0.88, RMSE of 40.93 Mg ha⁻¹, and MAE of 27.24 Mg ha⁻¹. The resulting map reveals marked spatial variability, with higher biomass in southern and central forests, a maximum of 524.67 Mg ha⁻¹, a mean of 173.13 Mg ha⁻¹, and a standard deviation of 112.48 Mg ha⁻¹. The total AGB across Nepal is 1.335 × 1010 Mg (13.35 Pg). An uncertainty surface, derived from cross-validation residuals, highlights elevated prediction variability in rugged mountain regions. This open, rigorously validated baseline supports transparent carbon accounting, REDD+ measurement, reporting, and verification, and forest management. The data products can be accessed at https://zenodo.org/records/17493215.