Sensitivity of Backscatter Intensity of ALOS/PALSAR to Above-ground Biomass and Other Biophysical Parameters of Boreal Forests in Alaska and Japan
Abstract
For the better understanding of the carbon cycle in the global environment, investigations on the spatio-temporal variation of the carbon stock which is stored as vegetation biomass is important. The backscatter intensity of "Phased Array type L-band Synthetic Aperture Radar (PALSAR)" onboard the satellite "Advanced Land Observing Satellite (ALOS)" provides us the information which is applicable to estimate the forest above-ground biomass (AGB). This study examines the sensitivity of the backscatter intensity of ALOS/PALSAR to the forest AGB and other biophysical parameters (tree height, tree diameter at breast height (DBH), and tree stand density) for boreal forests in two geographical regions of Alaska and Kushiro, northern Japan, and compares the sensitivities in two regions. In Alaska, a forest survey was executed in the south-north transect (about 300 km long) along a trans-Alaska pipeline which profiles the ecotone from the boreal forest to tundra in 2007. Forest AGBs and other biophysical parameters at 29 forests along the transect were measured by Bitterlich method. In Kushiro, a forest survey was carried out at 42 forests in 2011 and those parameters were similarly obtained by Bitterlich method. 20 and 2 scenes of ALOS/PALSAR FBD Level 1.5 data that cover the regions in Alaska and Kushiro, respectively, were collected and mosaicked. Backscatter intensities of ALOS/PALSAR in HH (horizontally polarized transmitted and horizontally polarized received) and HV (horizontally polarized transmitted and vertically polarized received) modes were compared with the forest AGB and other biophysical parameters. The intensity generally increased with the increase of those biophysical parameters in both HV and HH modes, but the intensity in HV mode generally had a stronger correlation to those parameters than in HH mode in both Alaska and Kushiro. The HV intensity had strong correlation to the forest AGB and DBH, while weak correlation to the tree stand density in Alaska. Also the analyses in Alaska and Kushiro revealed that the HV intensity had the sensitivity to the forest AGB up to 150 Mg/ha, but almost saturated above the value of AGB. According to the logarithmic regression function of the HV intensity against the forest AGB, the forest AGBs were estimated and mapped in Alaska and Kushiro, and their distributions were delineated in detail. However the erroneous estimation due to so-called "slope effect", which is inherently involved SAR data, was apparent, and we will challenge to reduce the effect.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2012
- Bibcode:
- 2012AGUFM.B41E0367S
- Keywords:
-
- 0428 BIOGEOSCIENCES / Carbon cycling;
- 0480 BIOGEOSCIENCES / Remote sensing;
- 0758 CRYOSPHERE / Remote sensing