Forest Biomass Estimation Using S-Band SAR and Lidar Data

A random forest regression strategy was used to identify an appropriate combination of variables for aboveground biomass prediction by integrating SAR backscatter data from S-band NovaSAR and forest stand height from LiDAR. The results show that combining backscattering of HH and HV polarizations from S-band with LiDAR-interpolated stand height produced favorable results with a relative root mean square error (rmse) of 25.17%. This study demonstrates the effectiveness of combining different characteristics from different sensors to improve the precision of aboveground biomass estimation.