Model assessment and verification - Air Pollution Modeling and Its Application - page 297

3. Results and Discussion
The hourly average modeled NO x concentrations in the grid cell covering
Washington, DC and at an urban monitoring site in Fig. 2 depict the typical
temporal variation found at numerous locations with the highest concentrations
occurring from 6:00 to 9:00 AM. Peak concentrations are clearly associated with
the highest mobile emissions displayed in Fig. 1. The dramatic decrease in
modeled and observed NO x concentrations during the late morning period is
primarily attributed to a combination of meteorological processes, (i.e., greater
vertical mixing and increasing horizontal transport) associated with the growth of
the convective mixing layer and increasing photochemical activity which converts
NO x to other nitrogen products. The temporal pattern of hourly modeled concent-
rations ( Fig. 2a ) closely resembles the observed results in Fig. 2b indicating
CMAQ captures the dynamic and chemical processes governing concentrations
during this time period. Modeled and observed results indicate lower NO x
concentrations in summer 2005.
Fig. 1. Modeled mobile NO x emissions in the grid cell of Washington, DC during summers 2002
and 2005
Weekday morning periods from 6:00 to 9:00 are of particular interest since the
high NO x concentrations are strongly governed by mobile emissions. Observed
and modeled 3-h average NO x concentrations were computed from hourly
concentrations for each weekday. Cumulative frequency distributions (CFDs) of
modeled 3-h NO x concentrations from an urban grid cell displayed in Fig. 3 depict
the range in concentrations and the differences between these two summer periods.
A notable decrease in NO x is apparent across a broad range of concentrations
which was also evident at other urban locations (not shown). A weekday mean
concentration was computed from values over the 50th to 95th percentiles in each
CFD to focus on change in the upper portion of the modeled and observed
distributions for each urban site. A comparison of modeled results from summer
2002 versus summer 2005 in Fig. 4a reveals that, with few exceptions, NO x in 2005
was lower than in 2002 at sites exhibiting a broad range of NO x concentration
levels. Figure 4b displays similar results from the urban measurement sites. These