Close Window

Removing Vegetation Canopy Bias from the Shuttle Radar Topography Mission Digital Elevation Model

Michael T. Coe, WHRC, mtcoe@whrc.org (Presenting)
Paul Lefebvre, WHRC, paul@whrc.org

We use the 90-meter horizontal resolution Shuttle Radar Topography Mission (SRTM) elevation data (Farr et al., 2007) to represent the surface morphology of the Amazon within our numerical modeling system. The absolute vertical accuracy of the SRTM data is estimated to be about +/- 6 m with relative errors believed to be less (Smith and Sandwell, 2003). The height reported in the data is a measure of the center of scattering within the vegetation canopy. It is not a direct measurement of the bald land surface, unless no vegetation is present. The height reported is therefore a complex function of canopy density and structure, and other land surface characteristics (Kellndorfer et al., 2004). Abrupt vegetation discontinuities, such as patchwork clear-cutting, produce errors that are clearly visible upon close regional inspection. The purpose of this study is to remove the absolute error imparted by the vegetation canopy to create a more accurate map of land surface elevation. Here we present the results of two attempts within the Xingu River basin at two resolutions. The first method is a direct correction of the 90-m resolution data on the 84000 hectare Fazenda Tanguru in the Upper Xingu River basin. We use 6 vegetation classes derived from year 2000 LandSat TM imagery, representing direct and indirect measurements of vegetation height to remove vegetation canopy bias for a sub-image containing the Fazenda. The second method uses 1-km resolution single-class forest data from Eva et al. (2002) and ground-based measurements of forest height throughout the Xingu Basin in Mato Grosso to remove the vegetation canopy bias. Both methods clearly improve the representation of land surface topography but errors are still visible, particularly along roads and where forest regrowth is active. Our next step is application of more explicit vegetation classification techniques to the entire Amazon basin.

Science Theme:  LC (Land Use and Land Cover Change)

Presentation Type:  Poster

Abstract ID: 107

Close Window