Estimation and Monitoring of Tropical Forest Biomass Using Polarimetric Interferometric SAR Data

The purpose of the proposed work is to examine the feasibility of using Polarimetric Interferometric SAR (PolInSAR) techniques on ALOS PALSAR data to extract forest canopy heights with the ultimate objective of deriving biomass estimates. Previous work [1] has shown that in homogeneous European forest stands, tree height is a reasonably robust estimator of biomass through a simple allometric relationship. Moreover PolInSAR has proven itself as a valuable technology for tree height estimation at L-Band frequencies. The RVoG model (Random Volume over Ground) proposed in [2] and elaborated in [3], permits a separation of ground and canopy scattering components of the interferometric phase. Through model inversion, the canopy height can be derived [4]. Additionally, the bare earth elevation beneath canopy can also be recovered at least in airborne repeat-pass cases where temporal de-correlation is not prohibitive. While the PolInSAR results to date from airborne repeat-pass L-Band campaigns have been impressive, it is not clear to what extent it will be possible to derive similar results from ALOS data, owing to the much longer temporal baseline between acquisitions. In this work we primarily focus our research on the extraction of DEM underneath canopy and the estimation of tree height and other forest parameters used in biomass calculation. The principal research site is chosen to be Kalimantan, Indonesia in order to overlap the area covered in the INDREX-II airborne campaign for which the PolInSAR technologies have been developed. We use the ground measurements taken during the Indrex-II campaign [5] for validation and verification of our results. Unfortunately at the time this report is being prepared, we do not yet have repeat pass, fully polarimetric ALOS data for the desired area. Therefore we will limit this report to a brief description of the background to the PolInSAR technology and some results extracted from the INDREXII data set by the authors using this technology.