Understanding how climate models reproduce phenomena known to influence weather and climate is a crucial aspect to model evaluation. Utilizing a type of functional analysis of variance, a statistical approach for connecting process-related phenomena to outcomes of interest (e.g., temperature, precipitation, etc.) in climate model ensembles will be presented. The methodology allows for not only novel ways of evaluating models based on well they reproduce processes but also incorporate uncertainty to the evaluations. These ideas will be demonstrated on initial analyses of the model output from the North American Regional Climate Change Assessment Program (NARCCAP).

Keywords: Climate model ensemble; Markov random field; Statistical computing

Biography: Steve is the head of the Geophysical Statistics Project and a Scientist in the Institute for Mathematics Applied to Geosciences at the National Center for Atmospheric Research in Boulder, Colorado, USA. His research interests include uncertainty quantification in the analysis of climate model ensembles, the assessment of the impacts of climate change, the design and analysis of computer experiments.