Modeling Uncertainty in Climate using Ensembles of Regional and Global Climate Models and Multiple Observation-based Data Sets

Because a vast array of data sets related to climate change are being produced by a multitude of research groups, statistical methodology is needed to combine the information across these data sets so as to accurately quantify a consensus uncertainty about climate projections. To this end, a hierarchical model is proposed here that combines climate change information from observation-based data sets, general circulation models, and regional climate models. In order to capture multiple aspects of climate change, the model combines information from these data sources to obtain an estimate and measure of uncertainty for the average, temporal trend, and inter-annual variability of seasonal average temperatures for predefined climate regions. Results are presented for three distinct climate regions. For each region, the combined information projected a warming in average temperatures. However, changes in rate of temperature increase (linear trend) and changes in the inter-annual variability of seasonal average temperatures varied by region and season.