3-Dimensional Tropical-Forest Structure from Interferometric SAR
Robert
Treuhaft, Jet Propulsion Laboratory, California Institute of Technology, robert.treuhaft@jpl.nasa.gov
(Presenting)
Bruce
Chapman, Jet Propulsion Laboratory, California Institute of Technology, bruce.chapman@jpl.nasa.gov
Luciano
Dutra, Insituto Nacional de Pesquisas Espaciais, dutra@dpi.inpe.br
João
Roberto
dos Santos, Insituto Nacional de Pesquisas Espaciais, jroberto@ltid.inpe.br
Jose
Claudio
Mura, Insituto Nacional de Pesquisas Espaciais, mura@dpi.inpe.br
Fábio
Gonçalves, Insituto Nacional de Pesquisas Espaciais, fabiogg@dsr.inpe.br
Paulo
Alencastro
Graça, National Institute for Research in the Amazon - INPA, pmlag@inpa.gov.br
Corina
da Costa
Freitas, Insituto Nacional de Pesquisas Espaciais, corina@dpi.inpe.br
Jason
Drake, USDA Forest Service, jasondrake@fs.fed.us
Global monitoring of the 3-dimensional (3-D) structure of vegetation bears on both the carbon cycle and the functioning and productivity of ecosystems. Beyond its relevance to fire susceptibility and habitat characterization, detailed 3-D information on the vertical distribution of vegetation seems to correlate strongly with forest biomass, perhaps more so than traditionally remotely sensed radar power, optical radiance, or tree height alone. Recent evidence suggests that the vertical profiles of vegetation needed to supply the third remotely sensed dimension can be estimated from multiple interferometric synthetic aperture radar (InSAR) observations—with multiple baselines, frequencies, or polarizations. Multialitude InSAR data taken with AirSAR over La Selva Biological Station in Costa Rica effectively provide a multibaseline InSAR data set from which to estimate vegetation density profiles. First InSAR profiles from La Selva primary and secondary forests as well as abandoned pastures demonstrate the current profile sensitivity and level of error. Comparisons to extensive field-measured profiles and lidar measurements show the potential of InSAR profiling and suggest spatial scales over which tropical-forest profiles change. Comparing InSAR, field, and lidar, each with their own errors, also helps to underscore the complexity of the characterization and use of the third dimension in forest remote sensing. The exact specifications of which quantities should or can be measured remotely, and which lateral and vertical resolutions would be most advantageous, are active areas of inquiry supported by this research.
Remote sensing measurements of forest properties are rarely direct. Biophysical estimates result from observations of electromagnetic power and phase. The means of connecting the electromagnetic signals to the underlying biophysics vary from statistical regression to physical models. The profiles in this study are derived from physical models, the approximations of which are quantifiable and beacons to the next observational and analytic steps. In addition to the 3-D profiles they produce, they provide a solid, quantitative paradigm for future NASA remote sensing.