Biomass estimation of Thetford forest from L-band SAR data: potential and limitations

This paper summarises the results of two pieces of work (i) a comparison of the relative ability of the AirSAR and E-SAR sensors, to estimate stand volume, age and height, (ii) an analysis of the sensitivity of observed backscatter change to changes in stand height and volume between 1991 (AirSAR data) and 2000 (E-SAR data). In both cases the method employs neural network and linear multivariate models to generate relationships between backscatter and stand volume/age/height. Experiment 1 shows the estimates of tree height, age and timber volume, from the AirSAR and E-SAR sensors have similar levels of accuracy, with the use of the E-SAR coherence data appearing to compensate for the lack of P-band data. The results of experiment 2 show change in height and percentage volume are estimated with a percentage RMSE of less than 10%.