Sean McGinnis, 2011, Geosciences

Climate models from around the world are used by the Intergovermental Panel for Climate Change in order to assess effect of human activity on the global climate. A relatively new direction is to include chemistry in these models, allowing for projections of climate change on air quality (ozone and aerosols in surface air). To date, the ability of global chemistry-climate models to reproduce observed correlations between temperature (or any meteorological index) and air quality has not been evaluated. Furthermore, many of these models (including the GFDL chemistry-climate model) exhibit a large positive bias in surface ozone over the eastern United States, casting doubts on their reliability for this application. It is well known that surface ozone correlates strongly with temperature over the eastern United States in summer.  I analyzed observations of surface ozone and temperature over the United States to explore whether there are broad regional patterns in this relationship that we can use to test the models. This information could also tell us if some of the model bias is associated with inaccuracies in the surface temperature simulation. I also looked to both interannual and seasonal variability across the US.   Some of the findings include the different sensitivities to temperature in ozone production in different regions of the US, which might tell us something about the processes going on in those regions.   To further evaluate the current model, I developed diagnostic code in order to quickly assess the bias of the current model run.