Data and Methods to Estimate National Historical Deforestation Baselines in Support of UNFCCC REDD

Author(s): Lydia P. Olander, Holly Gibbs, Marc Steininger, Jennifer Swenson and Brian C. Murray

Published: July 2007

download: working paper (.pdf) >

Global climate policy initiatives are now being proposed to compensate tropical forest nations for reducing carbon emissions from deforestation and forest degradation (REDD). These proposals have the potential to include developing countries more actively in international greenhouse gas mitigation and to address a substantial share of the world’s emissions which come from tropical deforestation. For such a policy to be viable it must have a credible benchmark against which emissions reduction can be calculated. This benchmark, sometimes termed a baseline or reference emissions scenario, can be based directly on historical emissions or can use historical emissions as input for business as usual projections. Here, we review existing data and methods that could be used to measure historical deforestation and degradation baselines including FAO national statistics and various remote-sensing sources. The freely available and corrected global Landsat imagery for 1990, 2000 and soon to come for 2005, may be the best primary data source for most developing countries with MODIS or other coarser high frequency data as a valuable complement for addressing problems with cloud cover and for distinguishing larger scale degradation. While sampling of imagery has been effectively useful for pan-tropical and continental estimates of deforestation, wall-to wall (or full coverage) assessments may be best for measuring national-level reference emissions. It is possible to measure historical deforestation and forest degradation with sufficient certainty for determining reference emissions, but there must be continued calls at the international level for making high-resolution imagery freely available, and for financial and technical assistance to help countries determine credible baselines.