Air quality models, such as the Community Multiscale Air Quality (CMAQ) model require gridded meteorological fields, such as those available from the 5th Generation Mesoscale Model (MM5). The air quality models are highly dependent on the predictions from the meteorological model (e.g. temperature, wind speed and direction) for predicting many things, such as the formation of clouds, chemical and photochemical reaction rates, aerosol equilibrium, advection of chemical species throughout the modeling domain and the removal of chemical species via wet and dry deposition. Typically, the air quality predictions are evaluated independent of the performance of the meteorological model, which makes it difficult to assess whether biases in the air quality predictions are caused by errors in the air quality model or due to some systematic bias/error present in the meteorological predictions. Using annual MM5 and CMAQ simulations, we examine the impact that the meteorological performance, particularly any systematic biases/errors that are present, has on the performance of the air quality predictions. Using an integrated database of containing both meteorological and air quality observations and predictions, the performance of CMAQ predictions of ozone and fine particles will be evaluated based on both the performance of the meteorological predictions and on the specific meteorological conditions (e.g. clouds versus no clouds). The goal of this work is to explore any associations that exist between meteorological prediction biases/errors and the air quality model performance. Ultimately, the hope is that this work will lead to a more rigorous evaluation of the CMAQ model that helps identify biases in the CMAQ model that are specific to certain meteorological biases/errors and conditions, which will in-turn help prioritize meteorological model issues that impact air quality.