Using Dynamic Evaluation to Assess Changes in Modeled and Observed Maximum Ozone Concentrations in Response to NOx Emissions Reductions

Dynamic evaluation investigates a photochemical model's ability to replicate changes in ozone air quality that can primarily be attributed to discernable changes in emissions. Thus, it is particularly relevant to air quality management applications where various emission control strategies are employed in regional modeling simulations which must demonstrate realistic responses in ozone levels to comply with a national air quality standard. A dynamic model evaluation study involving the current version of the Community Multiscale Air Quality Model (CMAQv5.0) using the Carbon Bond 2005 (CB-05) chemical mechanism was undertaken to assess the magnitude of modeled and observed maximum 8-h ozone concentration changes in light of notable real-world emission reductions of nitrogen oxides (NOx) at major elevated point sources and to concurrent gradual decreases in mobile source NOx emissions. To accomplish this task, model simulations were performed for two historical 3-month periods during the summers of 2002 and 2006, which preceded and followed, respectively, the implementation of major point source NOx emission reductions in conjunction with the US EPA's NOx SIP Call program. The CMAQ simulations were conducted for a continental domain that also included large parts of Canada and Mexico with a 12-km grid cell spacing and 35 vertical layers. Meteorological fields for both periods were supplied by the Weather Research and Forecasting (WRFv3.3) model that was applied with an updated four-dimensional data assimilation (FDDA) approach by incorporating all available upper air measurements to generate more accurate temporally-varying 3-D fields of winds and other meteorological parameters. Gridded emission inputs were derived from annual National Emissions Inventories (NEI) for ground-level area source sectors, MOBILE6 processing for the on-road source sector, Continuous Emissions Monitoring System (CEMS) hourly data for major point sources, and the Biogenic Emissions Inventory System (BEISv3.14) modeling of natural sources.

Hourly modeled concentrations were paired in time and space with the Clean Air Status and Trends Network (CASTNET) and Air Quality System (AQS) network site observations which permits an examination of differences in O3 concentrations over the diurnal cycle between these summer periods. In addition, spatial analysis results of the change in modeled and observed maximum daily 8-h O3 concentrations demonstrate the spatial extent and variability of improvements in ozone levels across the eastern US due to the emissions reductions between these two summer periods. Based on cumulative concentration distributions of maximum 8-h O3 values, observed and modeled changes between these summer periods were computed at the median and at the 95th percentile levels. Statistical metrics, which include the absolute (ppb) and relative (%) changes, in modeled and observed maximum 8-h O3 levels are also presented. The modeled response at the median was comparable to the observed change, however, at high ozone levels (ex., 95th percentile) the model tended to underestimate the ozone change between these periods. Key factors impacting model response are also discussed.