DEVELOPING REGIONALLY DOWNSCALED PROBABILISTIC CLIMATE CHANGE PROJECTIONS FOR THE SOUTHEAST REGIONAL ASSESSMENT PROJECT

The Southeast US contains the highest levels of biodiversity in North America outside of the tropics (Jose et al. 2006). This is partly due to the climate over the last few millennia, characterized by abundant precipitation, mild temperatures, and low climatic variability. Recently, the Intergovernmental Panel on Climate Change Fourth Assessment Report (IPCC AR4) concluded that it is very likely that humans are largely responsible for increasing the global average surface temperature by one degree celsius in the 20th century through the release of greenhouse gasses (GHG) such as CO2 into the atmosphere. This warming is expected to continue well into the future and is projected to cause sizeable impacts on managed and unmanaged ecosystems. Thus, mitigation of, and adaptation to the impacts of climate change on ecosystems in the Southeast will likely be the key challenge confronting natural resource managers in the coming decades. Central to this is how to best implement an adaptive management strategy given the large uncertainty associated with climate change projections. This requires a careful treatment of this uncertainty as well as methods to downscale climate projections to the scale of ecosystem processes because of the coarse resolution of the models. To date, most studies use the range of GCM output to represent the full predictive uncertainty; thus underestimating the actual structural and parametric uncertainty associated with these projections. This underestimation will then propagate through all levels of analysis requiring climate change projections, leading to overconfident predictions. As a result, decision-makers may insufficiently hedge against the risks associated with extreme climatic events that have a low probability of occurrence, but are high impact events. We address this by developing a suite of regional probabilistic climate change projections for the Southeast Regional Assessment Project (SERAP). Two core climatic datasets are used for base projections: (1) GCM simulations from the IPCC AR4 for fully coupled global-scale climate simulations; and (2) an Earth Model of Intermediate Complexity (EMIC) to sample the parametric uncertainty of key climate system variables such as ocean diffusivity. These datasets are further post-processed through: (1) Bayesian ensemble dressing methods to estimate structural uncertainty and the accuracy of the GCMs; and (2) statistically downscaled simulations forced by boundary conditions from the GCM and EMIC runs. The probabilistic projections generated through these methods enable other SERAP researchers to propagate uncertainty to other models, thus forming the basis for projecting ecosystem changes in the Southeast over the next century. References: Jose, S., E.J. Jokela, and D.L. Miller (eds.), 2006: The Longleaf Pine Ecosystem: Ecology, Silviculture, and Restoration. New York, NY: Springer Science.