A probabilistic scenario approach for developing improved Reduced Emissions from Deforestation and Degradation (REDD plus ) baselines

Performance-based payments are widely seen as a promising tool for Reduced Emissions from Deforestation and forest Degradation (REDD+) in tropical forests. Despite great advances in international REDD+ negotiations, there is a lack of consensus around the development of business-as-usual (BAU) reference scenarios or baselines to derive and quantify net carbon emission reductions. In this paper, we explore a novel approach for developing baselines (point forecasts) using exponential smoothing. Further, we introduce the concept of probabilistic BAU scenario ranges developed using this approach. We compare predictive performance with the linear trend and historical averages approaches conventionally used in policy proposals and REDD+ pilots. Weempirically test the relative performance of all three approaches by forecasting BAU baselines and scenario ranges in 36 sites (consisting of 20 countries and 8 Amazonian states with and 8 countries without REDD+ schemes). Based on two predictive performance measures (the root mean squared error and mean absolute percentage error), we find that exponential smoothing outperforms the linear trend and historical average models at predicting forest cover changes. In addition, we show how prediction intervals based on a desired confidence level generated through exponential smoothing can be used in novel ways to determine likely baseline scenario ranges. In this way it is possible to quantify the degree of variability and uncertainty in datasets. Importantly, this also provides a statistical measure of confidence to determine if REDD+ interventions have been effective. By generating robust probabilistic baseline scenarios, exponential smoothing models can facilitate the effectiveness of REDD+ payments, support a more efficient allocation of scarce conservation resources, and improve our understanding of effective forest conservation investments, also beyond REDD+. (C) 2015 The Authors. Published by Elsevier B.V.