Estimating the Uncertainty of Land-use History Reconstructions in the Global Carbon Balance

Human land-use activities have had a marked affect on the land surface. Globally, cropland and pasture occupy more than 30 percent of the land surface, and much of the remaining forest area is in some stage of recovery from prior agriculture or logging. Landscapes are generally a heterogeneous mixture of patches with different land uses and land-use histories. Because of these changes and the potentially long-time scales of ecosystem responses to them, historical information on land-use activities is needed for ecosystem studies of the past, present, and future. Because of limited records globally, much of the information needed must be estimated in reconstructions. In this study, we evaluated the importance of several key aspects of uncertainty in land-use history reconstructions on estimates of the global terrestrial carbon balance. We first constructed an ensemble of global land-use history reconstructions that are consistent with major historical databases, but differ in important parameters unconstrained by these databases. We then propagated the ensemble of reconstructions through a set of ecosystem models, based on MIAMI-LU and ED family of models, to determine annual exchange of carbon between the terrestrial biosphere and the atmosphere. These models have been used previously in assessing the effects of land-use history on the U.S. carbon balance and are underdevelopment for global Earth System applications. Analyses identify the relative importance of key uncertainties in land-use history reconstructions for carbon flux estimates and suggest priorities for future research to reduce these uncertainties. In addition, a preferred land-use history product is presented for use in global earth system studies