Ozone, oxides of nitrogen, and carbon monoxide during pollution events over the eastern United States: An evaluation of emissions and vertical mixing

Chemical transport models such as the Community Multiscale Air Quality (CMAQ) model provide useful guidance on air pollution control strategies. We evaluate the performance of a 12 km resolution CMAQ simulation with surface and aircraft observations of CO, O-3, and NOx during the summer of 2002. When all data are considered, on average, modeled and observed CO total column contents (surface to 3,000 m) agreed to within 14% in the morning and 22% in the afternoon. Reducing the deposition velocity for CO improves model-measurement agreement but did not eliminate the model bias. The majority of observed vertical profiles have a maximum near the surface. Although many observed spirals had a secondary maximum at the top of the boundary layer, indicating subgrid-scale shallow convection. The model was not able to replicate these vertical structures. Water vapor profiles likewise showed greater vertical variability in the observations than in the model. General conclusions from these model-measurement comparisons: total CO emissions estimates are either adequate or underestimated, but there is no evidence of gross error; NOx emissions from mobile sources may be overestimated while the lifetime of NOx may be underestimated in CMAQ 4.5.1 with CBIV, and vertical mixing in the model boundary layer may be too fast, but venting out of the boundary layer into the lower free troposphere may be too slow.