Forest Height and Aboveground Biomass Mapping by synergistic use of GEDI and Sentinel Data using Random Forest Algorithm in the Indian Himalayan Region

Forests play a vital role in the global climate and carbon budget regulation. The present study aims to map the forest canopy height and aboveground biomass (AGB) by integrating Global Ecosystem Dynamics Investigation (GEDI) and Sentinel data using Random Forest (RF), a machine learning algorithm in a part of the Pauri Garhwal district of Uttarakhand, India. The vegetation types/land uses of the study area were mapped using Sentinel-2 data with an overall accuracy of 91% using the RF algorithm. GEDI was integrated with 16 texture and 2 backscatter variables derived from Sentinel-1 SAR data to map the height of the forest canopy. A combination of LiDAR and SAR variables was found to be efficient in predicting forest canopy height. The forest canopy height map was validated using field-measured canopy height values from outside the LiDAR footprints with an R2 of 0.81, an RMSE of 1.61 m, and a %RMSE of 8.20. The modelled canopy height was further used with Sentinel-2 data to estimate forest AGB in the study area. The forest AGB was mapped by integrating field-measured AGB with Sentinel-2 data-derived spectral and texture variables and modelled forest canopy height derived from GEDI and Sentinel-1 data, using the RF model. The combination of spectral, texture, and height variables was able to predict the spatial distribution of AGB with an R2 = 0.88, RMSE = 22.05 Mg ha-1, and %RMSE = 15.11%. The study highlighted that a synergistic approach involving multi-sensor data can effectively predict the forest canopy height and AGB. It also highlighted the utility of machine learning algorithm in mapping forest biophysical parameters. The study presented an effective approach for forest canopy height and AGB mapping using multi-sensor earth observation data.