Aboveground Biomass in Tajikistan

We develop a dynamic sampling scale-up method for AGB estimation by integrating multi-source data from Sentinel-2 MSI, Unmanned Aerial Vehicle (UAV), and ground observations. Initially, UAV and Sentinel-2 multispectral images are separately combined with ground-measured data for feature interaction analysis and factors selection. Subsequently, eight machine-learning models are constructed and optimized for AGB estimation at different scales. The dynamic sampling scale-up algorithm based on UAV and ground data enhances AGB estimation accuracy based on the optimal models.