Automated estimation of forest height and underlying topography over a Brazilian tropical forest with single-baseline single-polarization TanDEM-X SAR interferometry

Forest height is an important variable for modeling terrestrial carbon storage and global carbon cycle dynamics. Spaceborne SAR Interferometry (InSAR) has the sensitivity to measure canopy height and the underlying topography. In this paper, we refine and automate an interferometric ground finding approach that exploits few look (2to 4-look) averaged interferograms and incorporates the use of a coherent electromagnetic simulator and field inventory data. Using the coherent electromagnetic simulator, an InSAR simulation based on field data is performed to study the true ground position as a function of the statistics of few-look InSAR phase heights from a model perspective. With this statistical model, both the underlying topography and the canopy height (mean and top canopy height) can be estimated. Using German Aerospace Center's (DLR) single-baseline single-polarization TanDEM-X InSAR data, we validate the approach over a Brazilian tropical forest (Tapajos National Forest) with both field inventory and lidar data. As validated against lidar data, the underlying topography is estimated to an accuracy of 3 m. At one hectare aggregated pixel size, InSAR phase-center height is best compared with the field/ lidar mean canopy height with an accuracy of 2-3 m, while InSAR-inverted total height best characterizes the lidar top canopy height with an accuracy of 4-5 m. Given the global data availability of TanDEM-X and the future TanDEM-L, this approach has the potential for wall-to-wall mapping of forest height as well as underlying topography and also serves as a complementary tool to other existing InSAR, Polarimetric InSAR (PolInSAR) and SAR tomography (TomoSAR) methods when only single polarization and/or baseline data are available.