The ultimate goal of this research is to identify and assess the contribution of atmospheric general circulation model (AGCM) differences to the uncertainties in regional climate change to forcing by tropical sea surface temperatures (SSTs) in different perspectives. The proposed research will utilize multiple climate models to examine the regional climate sensitivity to tropical forcing through a series of ensemble experiments. The focus of this work is on the relationship – or in mathematical terms, the operator -- between forcing and response. Major objectives:

1. Compute the global teleconnection operator (GTO) that relates a large class of Tropical SST variations to the regional climate change response. This is done through the use of idealized forcing scenarios such as SST and/or heat flux anomaly forcing scenarios
2. Identify and quantify uncertainty in regional climate change related to model physical parameterizations. This is accomplished by analyzing the differences among model specific GTOs, including models from a perturbed-physics experiment
3. Develop regional detection and attribution methodology that incorporates the results from objectives 1. and 2. including a comparison of these uncertainties to other uncertainty components.

An additional aim of this project is to help bring these methods into common use in the climate modeling community for both model diagnosis and regional assessments. Too often, the uncertainties in regional prediction problem are deemed too difficult to quantify and the climate model results are taken at face value. To move forward on understanding uncertainties in regional predictions, many approaches are necessary and here we initially focus on the influence of tropical SSTs and the dynamical response of atmospheric GCMs. Gaining perspective in this area, we will then consider how the coupled ocean-atmosphere system could differ from an AGCM alone. This project is directly addressing the issue of climate change risk at local and regional scales by analyzing the components of regional prediction related to the uncertainty in large-scale dynamical forcing of the regional environment. Policy and decision makers require robust and practical assessments of local climate change impacts and hence, we must provide such assessments of climate change, both mean state and variability, for a large range of variables. To date, the standard method for assessing uncertainty in this situation is to consider the ad hoc ensembles of global AOGCMs to provide a statistical distribution of possible outcomes. This ad hoc approach does not provide an assessment of the sources of the uncertainties in these predictions based on current understanding of the climate system and its dynamical behavior. This project will provide the climate modeling community a method to address this shortcoming for one component of regional predictions.