Comparative Analysis of Major Global Forestry Datasets for Country Level Estimates of C Stock Change in Living Biomass

Carbon (C) stock change and related Carbon Dioxide (CO2) emissions/removals from living biomass for Forest Land remaining Forest Land (IPCC sector 5-FL-1) are one of the most difficult to estimate because of the large natural fluxes of CO2 and the high accuracy requirements for deriving net changes of the large C stock. Following the Good Practice Guide (GPG) methodology of IPCC (2003) with IPCC default values and data from FAO (FRA and ForesSTAT 2010) for harvested woodfuel and industrial roundwood and with fires emissions from the Global Fires Emission Database GFED v.3, a Tier-1 level estimate of country-specific C stock for sector 5-FL-1 was carried out for the years 1990, 2000 and 2005, which mounted to about 6 Pg C or about half of the mean total CO2 emissions in that period. The resulting bottom up globally harmonized CO2 inventory for 5-FL-1 from 1990 to 2010 completes further the Land Use, Land-Use Change and Forestry sector in the Emission Database for Global Atmospheric Research (EDGAR). Even though the GPG methodology of IPCC (2003) is assumed adequate and detailed enough to account well for C stock changes, the estimates of sector 5-FL-1 in the national GHG inventories remain highly uncertain. In all three years ~35% of the national inventories for 5-FL-1 show similarities but the values differ in most cases by an average factor of -5 to +5. In all other cases, Tier 1 values are the opposite of what countries report regarding sink and source. Vis-à-vis forested area the average difference for Annex I countries between Tier 1 calculations and FRA2010 and UNFCCC mount to 9% and 16% respectively. It is concluded that differences are mainly due to the activity data explicitly forested area, the definition of forests used by each country and losses which are not always accounted by countries or perhaps double counted. The feasibility for establishing national inventories in a consistent way for all world countries was demonstrated but only with a limited accuracy of Tier 1. Better forestry data, such as annual activity data from satellite images are needed to derive estimates at higher Tier, but are too scarce at the moment for obtaining more accurate estimates of global level C stock budgets.