A U.S. national scale modeling effort is underway to map height, biomass and carbon layers for the Year 2000 exploiting synergies among recently developed national-scale data sets derived from Shuttle Radar Topography Mission (SRTM) and Landsat ETM data. Due to the short radar wavelengths used for the SRTM mission (C-band and X-band), the SRTM InSAR signal represents a reflective surface rather than the ground elevation whenever vegetation or anthropogenic features are present. By differencing ground elevations from SRTM elevations over vegetated terrain a spatially continuous height signal, the mean height of the scattering phase center (MHSPC), can be extracted which is correlated with true vegetation canopy height. The MHSPC is dependent on the scattering characteristics of the observed canopy which are largely a function of vegetation cover type, density, and phenological state. Given a ground measured reference dataset of vegetation canopy height, the SRTM MHSPC signal can then be used with information on cover type and density to spatially estimate vegetation canopy height with empirically developed regression models. Subsequently, given spatial estimates of vegetation canopy height in conjunction with information on vegetation cover type and density, aboveground dry biomass can be modeled as well. Finally, spatial data layers of carbon distribution can be calculated with established aboveground dry biomass to carbon conversion factors. Within the conterminous United States, a timely confluence of spatial data sets provides the framework for the development of empirical regression models and their application at a national scale. In addition to the SRTM mission, national-scale data sets of ground surface elevation (i.e., National Elevation Dataset, NED), existing vegetation type (i.e., LANDFIRE) and canopy density (i.e., National Land Cover Database, NLCD 2001,) are currently being produced for the circa 2000 timeframe. These data sets are complemented by a database of ca. 150,000 forest plots provided by the USDA Forest Service Forest Inventory and Analysis (FIA) program. A novel approach based on object-oriented image analysis is employed to address SRTM inherent noise characteristics and to enhance the accuracy of the InSAR signal. As a statistical modeling framework a regression tree based approach is employed. A prototype study in central Utah covering 62,000 km(2) (i.e., NLCD 2001 mapping zone 16), which contains a diversity of vegetation types is now completed. Based on data from 508 FIA field plots, overall vegetation canopy height and aboveground dry biomass estimates at r(2) (and absolute error) values of 0.78 (2.1 m) and 0.56 (24 tons/ha) were obtained. The NBCD 2000 project is scheduled for completion in late 2008. Data will be accessible at 30 m postings via the U.S. Geological Survey seamless data server as mapping zones are completed.
