Pantropical Forest Carbon Mapped with Satellite and Field Observations

The ability to estimate the distribution and total amount of carbon stored in woody biomass across the tropics is important for compensation mechanisms for Reducing Emissions from Deforestation and Degradation (REDD) under the United Nations Framework Convention on Climate Change (UNFCCC). Errors in biomass estimation of tropical forests are, however, large due to of a paucity of systematic surveys, together with large spatial variability and diversity of tree species (Clark et al. 2001). Because systematic surveys are sparse, plot-level measurements of biomass density (thus carbon) are extended to large areas using a variety of methods including (i) classifcation of land cover types, each assigned an estimated average value of biomass density based on estimates from the literature or forestry surveys, (ii) calculation of biomass density from regressions based on gridded environmental variables such as mean annual temperature and precipitation, (iii) determination of relationships between in situ biomass density and remote sensing characteristics that can be consistently mapped over large regions. Maps derived using the latter approach are a substantial departure from the other more traditional methods, and have the advantage of providing robust, spatially consistent and continuous values of the magnitude (amount) of carbon stock at any given location (Baccini et al. 2008, Houghton et al. 2009). They thereby also provide a basis for monitoring stock changes through time (Goetz et al. 2009). Here we report on the "rst pantropical map of forest carbon derived using this approach, informed with extensive satellite canopy structure sampling calibrated with co-located "field measurements.