Empirical relationships between AIRSAR backscatter and LiDAR-derived forest biomass, queensland, Australia

To evaluate the use of multi-frequency, polarimetric Synthetic Aperture Radar (SAR) data for quantifying the above ground biomass (AGB) of open forests and woodlands, NASA JPL AIRSAR (POLSAR) data were acquired over a 37 x 60 kin area west of Injune, central Queensland, Australia. From field measurements recorded within 32 50 x 50 in plots, AGB was estimated by applying species-specific allometric equations to stand measurements. AGB was then scaled-up to the larger area using relationships established with Light Detection and Ranging (LiDAR) data acquired over 150 (10 columns, 15 rows) 500 x 150 in cells (or Primary Sampling Units, PSUs) spaced 4 x 4 kin apart in the north- and east-west directions. Large-scale (1 :4000) stereo aerial photographs were also acquired for each PSU to assess species composition. Based on the LiDAR extrapolations, the median AGB for the PSU grid was 82 Mg ha(-1) (maximum 164 Mg ha(-1)), with the higher levels associated with forests containing a high proportion of Angophora and Callitris species. Empirical relationships between AGB and SAR backscatter confirmed that C-, L- and P-band saturated at different levels and revealed a greater strength in the relationship at higher incidence angles and a larger dynamic range and consistency of relationships at HV polarizations. A higher level of saturation (above similar to 50 Mg ha(-1)) was observed at C-band HV compared to that reported for closed forests which was attributable to a link between foliage projected cover (FPC) and AGB. The study concludes that L-band HV backscatter data acquired at incidence angles approaching or exceeding 45 degrees are best suited for estimating the AGB up to the saturation level of similar to 80-85 Mg ha(-1). For regional mapping of biomass below the level of saturation, the use of the Japanese Space Exploration Agency (JAXA) Advanced Land Observing Satellite (ALOS) Phase Arrayed L-band SAR (PALSAR) is advocated. (C) 2005 Elsevier Inc. All rights reserved.