Improving ozone simulations in the Great Lakes Region: The role of emissions, chemistry, and dry deposition

Abstract

The Great Lakes Region of the US continues experiencing exceedances of the ozone (O3) standards, despite years of emissions controls. In part, this is due to interactions between emissions from surrounding large cities (e.g., Chicago) and meteorology, which is heavily influenced by the presence of the Great Lakes. These complex meteorology-emissions interactions pose a challenge to fully capture O3 dynamics, particularly near shores of the lakes, where high O3 levels are often experienced. In a simulation with the Community Multiscale Air Quality (CMAQ) model, using inputs as typically constructed, the model tends to be biased high. A literature review indicated that NOx emissions from mobile sources, possibly overestimated in the 2011 National Emission Inventory (NEI), or the version of the Carbon Bond chemical mechanism used in CMAQ could be responsible for high biases of O3. As such, a series of sensitivity tests was conducted to identify potential causes for this bias, including emissions biases (e.g., biogenic VOCs, anthropogenic NOx), chemical mechanism choice, and O3 dry deposition to fresh water, for high O3 periods in July 2011. The base model/emissions configuration used the following: Carbon Bond mechanism CB05, biogenic emissions using Biogenic Emission Inventory System (BEIS), and anthropogenic emissions from the 2011 National Emissions Inventory (NEI). Meteorological inputs were developed using the Weather Research and Forecasting (WRF) model (version 3.8.1). Simulated daily maximum 8-h average O3 without and with a cutoff of 60 ppb (referred to as MDA8 O3 and elevated MDA8 O3 hereafter, respectively) were evaluated against measurements. The evaluation showed a high bias in MDA8 O3 across the domain, particularly at coastal sites (by ∼6 ppb), while elevated MDA8 O3 (i.e., greater than 60 ppb) was biased low, with exceptions centered along the shore of Lake Michigan. Using the CB6 chemical mechanism or 50% reduction of NOx emissions from on-road mobile sources led to substantial domain-wide decreases in O3 from the base case, and the model performance improved, particularly along the Lake Michigan shoreline and for the western domain. However, elevated MDA8 O3 was more biased against measurements, compared to the model performance in the base case, except at a few sites along the shoreline. Using the Model of Emissions of Gases and Aerosols from Nature (MEGAN) instead of BEIS to estimate biogenic emissions, or increasing dry deposition of O3 to fresh water by a factor of ten (which is unrealistic), had minor impacts on simulated O3 over land. But, combining MEGAN with CB6 resulted in improved elevated MDA8 O3 simulation along the western coast of Lake Michigan. Finally, using CB6 combined with a 30% reduction of on-road mobile NOx emissions and MEGAN led to the best performance. Two companion papers investigate how meteorological modeling can be improved. Together, the recommended modeling system could serve as a starting point for future O3 modeling in the region.