Using Cluster-Derived Diurnal Ozone Groups to Validate Model Surface Ozone Predictions

Air-quality forecasts for pollutants, such as tropospheric ozone, continue to be a necessity for the air quality control of major metropolitan regions. The ability to forecast high ozone events still proves to be difficult due in part to the deficiency of representing chemistry and meteorological processes on the local scale. For example, air quality models are known to struggle with predicting ozone concentrations in cases where local forcing dominate (e.g.: sea/bay breeze) due to complex boundary layer processes. The effects of these processes are noticeable in the day-to-day variability of the diurnal surface ozone cycle, which suggests that the diurnal ozone cycle can be used to validate air quality models. Using clustering methods we will group surface ozone concentration based on their diurnal variability, focusing on regions affected by sea/bay breezes such as Houston, Texas and Baltimore, Maryland (Chesapeake Bay area). The cluster-derived groups are then linked with hourly ozone predictions from the Community Multiscale Air Quality (CMAQ) model to better understand which conditions the model performs best. The study will analyze meteorological processes such as sea/bay breeze circulations, which favor the buildup of high ozone concentrations. The clustering analysis will be used to identify unique diurnal ozone patterns and evaluate their associated meteorological patterns as well as their impacts on air quality model predictions. Meteorological models such as NOAA’s operational North American Mesoscale (NAM) 12 km weather predictions, which are used to drive the CMAQ model, will be evaluated to understand air quality model performance under complex meteorological processes.