Estimation of Forest Biomass in Temperate Forests of Northern Japan using Airborne Multifrequency Polarimetric Pi-SAR Data

Abstract

Dual frequency (Land X) Pi-SAR polarimetric SAR data was used to find out the sensitivity of biophysical parameters to radar backscattering in temperate forests of Northern Japan. The statistical relationships between the radar backscatter and different biophysical parameters are quite robust for L-band at all the polarizations. The X-band showed statistically poor sensitivity to most of the vegetation parameters including crown parameters. The experimental findings were compared with a radiative transfer based theoretical radar backscatter model to understand and explain these observed relationships. Further, a large sample of 186 stand age from coniferous species showed a robust relationship for all the three polarizations of the L-band data up to 40 years of age. L-band data provided very good retrieval accuracy for the dry biomass with the SEE = 22.52 tons ha-1. The addition of three data channels from the X-band enhanced the accuracy to 20.76 tons ha-1. On its own, X-band is unsuitable for the biomass retrieval as determined by its feeble relationships with biophysical parameters and thus imperfect retrieval accuracy.