Modeling of Flow and Photochemistry in Complex Urban Environments

Frequent exceedances of air quality standards for many primary and secondary pollutants continues to be a problem in complex urban environments. High vehicle-generated emissions can lead to high primary species concentrations within the urban canopy (e.g., street canyons and intersections), where the atmospheric flow and turbulence is disturbed by buildings or other obstacles. While high vehicle NOx emissions from urban centers serves to lower O3 levels locally, accompanying NO2 levels may be extremely high in such cases, and questions arise about the role of urban street canyons as "pre-reaction chambers" that may further enhance NO2 formation. To understand the ramifications of proposed emission reduction strategies on the street canyon level, photochemical models for high resolution simulation of urban core air quality are needed. Previous modeling on the urban street canyon scale has almost always considered only a single street canyon and has either focussed on non-reactive pollutants or has invoked simple approximations for dealing with simplified NO-NO2-O3 chemistry (Yamartino and Wiegand, 1986; Berkowicz et al., 1997). While such simplifications may be appropriate within a single canyon for receptors nearby the emission sources, they are not reasonable for addressing questions associated with the series of canyons or neighborhoods as are found in the urban core.

To address these issues, the German Federal Environmental Agency (Umweltbundesamt) supported the linkage of an existing Navier-Stokes (N-S) flow and turbulence solver to a photochemical model capable of operating a resolutions of a few meters in an obstacle-filled environment. As such a photochemical model did not exist, modifications were made to the urban-scale photochemical model, CALGRID (Yamartino et al., 1992), including:

· Modification of CALGRID's framework to allow for simulation of 2-d domains (i.e., downwind and vertical planes assuming infinite crosswind domain extent), as well as the more traditional 3-d domains, thereby enabling model-ing of large spatial domains with spatial resolutions down to a few meters;

· Conversion of the vertical grid system from terrain following coordinates to absolute vertical coordinates to accommodate the vertically walled buildings of urban cores, and allow for part of the modeling domain to contain impenetrable surfaces;

· Addition of a detailled TKE sub-model for vehicle-generated turbulence; and

· Addition of a simple NO-NO2-O3-chemistry scheme, as an alternative to the full SAPRC or CB-IV chemistries, to examine sensitivity to the short traveling times of pollutants within modeling domains extending over only one or a just a few canyons.

This new model, Micro-CALGRID, was then linked to the flow and turbulence fields produced by the N-S flow model MISCAM, which can treat geometries of arbitrary complexity, and further linked to CALGRID to facilitate integrated (one-way nested) assessments of the urban- and micro-scales.

This paper summarizes Micro-CALGRID development and features, and reports the findings of initial 2-d and 3-d applications of this coupled flow-photochemistry modeling system as a precursor to its more comprehensive evaluation with full-scale, urban field programs.