A dual-frequency SAR mosaic of the Amazon

The development and availability of accurate landcover information is critical to the ecological science questions [Cerri, et al., 1995; Hall, et al., 1995], such as defining the rate and extent of deforestation within the Amazon. This need can only be met by techniques that are sensitive to phenomena at the scale of the landscape patch and can be efficiently applied in a consistent fashion over very large regions. For example, estimation of Amazonia´s carbon budget will require a large-scale mosaic of the location and spatial extent of landcover types, forest regrowth biomass, and selective logging. Vegetation type and surface condition exert control over carbon and trace gas fluxes. Estimates of the biotic carbon pool are also important since carbon accumulation rates vary not only with vegetation types but with successional stage and management practices as well. Both the extent and carbon accumulation rates of regenerating tropical forest are poorly known. If the ages of forest disturbance can be established (i.e., from the Landsat time series), then the SAR-derived estimates of aboveground biomass can be used to estimate annual sequestering of carbon. Though these data were acquired at higher resolution, it is possible to use this data to map landcover, biomass and selective logging at a spatial scale of 100m for the Amazon basin.