Forest Structure and Biomass Estimation from Polarimetric SAR Interferometry

Maxim Neumann, JPL, maxim.neumann@jpl.nasa.gov (Presenter)
Sassan Saatchi, CALTECH/JPL, sassan.saatchi@jpl.nasa.gov

Synthetic aperture radar (SAR) provides the means for forest monitoring at high

resolution (meter scale) and global coverage, independent of the cloud cover and

the time of the day. Polarimetric SAR interferometry (PolInSAR) provides the

means for forest structure and biomass retrieval, as it is sensitive to the

vertical structure and physical characteristics of the forest layers. In the

past, we have developed a model and inversion approach which allows to retrieve

forest structure characteristics from repeat-pass PolInSAR, related to the

vertical forest structure, the density, and the morphology of the trees. This

approach does not require a-priori information, and is able to compensate

for temporal decorrelation if multiple baselines are acquired. We have

demonstrated the performance over temperate, boreal and tropical forests using

air-borne PolInSAR data at L- and P-band frequencies.

In this study, within the frame of forest carbon stocks and change monitoring in

NASA's planned DESDynI mission (Deformation, Ecosystem Structure and Dynamics of

Ice), biomass estimation performance from model-based PolInSAR data using

parametric and non-parametric regression methods is evaluated. PolInSAR data is

decomposed into ground and volume contributions, estimating forest structure,

and using a set of obtained parameters for biomass regression. The considered

estimation techniques include multiple linear regression, support vector

machines and random forests. The biomass estimation performance is evaluated on

airborne SAR data (DLR's E-SAR sensor) at L- and P-bands over Krycklan

Catchment, a boreal forest test site in Northern Sweden. The combination of

polarimetric indicators and estimated structure information has improved the

root mean square error (RMSE) of biomass estimation up to 28% at L-band and up

to 46% at P-band in comparison to using only SAR backscatter values. The

cross-validated biomass RMSE was reduced to 20 tons/ha.

Presentation Type:  Poster

Session:  Other   (Tue 11:30 AM)

Associated Project(s): 

• Saatchi, Sassan: Vegetation Structure Studies ...details

Poster Location ID: 262