Info

Reactive organics (% control)

FIGURE 16.16 Peak ozone isopleths calculated for downtown Los Angeles (DTLA) and Rubidoux, approximately 100 km east and downwind of DTLA under typical meteorological conditions. Spatially uniform reductions of VOC and NO^ were employed in an airshed model by Milford et al. (1989). The top shows isopleths in two dimensions as presented by Milford et al. (1989), and the bottom shows these data extrapolated to three dimensions (from Finlayson-Pitts and Pitts, 1993).

point A does not produce a significant change in 03. It is important, however, to note that DTLA is not where the 03 peaks for the airshed as a whole are experienced.

In contrast, at downwind locations where the airshed ozone peaks are more typically encountered, reduction in VOC starting at point B without concurrent NOx control does not lead to the rapid decrease in 03 observed for DTLA; indeed, even an 80% reduction in VOC alone is not predicted to reach the U.S. air quality standard of 0.f2 ppm 03. In this case, control of NOx reduces 03 more rapidly than comparable control of VOC.

While the results in Fig. 16.16 were developed for the Los Angeles area, the same general features have been found in a number of other studies as well (e.g., see National Research Council, 1991; and Tesche and McNally, 1991). The approach and issues are thus qualitatively applicable to urban/suburban areas worldwide where pollutants are transported downwind to less densely populated areas and the VOC/NOx ratio changes as the air mass is transported.

There are, however, some important changes in our understanding that have come about since the predictions in Fig. 16.16. First, as discussed in more detail shortly, the emissions inventory for VOC was underestimated at the time of the model calculations by a factor of two or more, at least for mobile source emissions. This tends to introduce a bias in favor of VOC control. Furthermore, this study only addressed one set of meteorological conditions characteristic of the particular episode modeled. For example, in a more recent study of a different air pollution episode in the same air basin where the meteorology was not as conducive to formation of high ozone levels, ozone was shown to be sensitive to increasing VOC from vehicle hot exhaust as far east as San Bernardino (Harley et al., 1993a). This sensitivity to the particular meteorological conditions illustrates that the effectiveness of a particular control strategy may depend on whether one chooses to focus on the most severe episodes or, alternatively, on more typical conditions.

While recognizing these limitations, such isopleths are useful, however, in examining the differing responses of various locations within an air basin to control of VOC and NOx. Since the air mass that starts upwind in the morning traverses the air basin during the day, a multidimensional approach to ozone control is clearly needed. Isopleths characteristic of different locations in an air basin such as those in Fig. 16.16 demonstrate that a combination of VOC and NOx control is essential if air quality throughout a major air basin is to improve consistently (although not necessarily by equal amounts) in all locations.

3. Models

Ultimately, it is not possible to mimic experimentally all conditions of potential interest for control strategies. In addition, even if one assumed that one could carry out environmental chamber studies covering all of these conditions, as discussed earlier, there still remain a number of uncertainties in how to extrapolate these to the "real world" with its different surfaces and potential for heterogeneous reactions, complex meteorology, deposition processes, and varying new emissions. As a result, the development and application of mathematical models describing the chemistry, meteorology, and deposition are critical elements in the development and assessment of effective control strategies.

Of course, in developing these models, it is necessary to test them extensively against experimental data to ensure that their predictions are the result of appropriately simulating the chemical and physical processes involved, and not to a fortuitous cancellation of errors. For example, the chemical submodels are tested against environmental chamber data, and the final model results against ambient air measurements.

There are a variety of mathematical models used to describe the relationship between the precursors and the secondary pollutants they form upon reaction. In addition, there are some simple, often empirical, models that have been developed for application in particular areas. An example of these is also discussed in the following section.

a. Simple Models

(1) Linear rollback As the name implies, linear rollback is based on the assumption that pollutant concentrations will decrease proportionally to a decrease in the precursor emissions. For a pollutant such as CO, for example, the percentage reduction in emissions required to meet air quality standard for CO (A co) in a region that currently has observed concentrations as high as Cco is given by linear rollback as:

Percentage reduction in emissions

Cco ^co

CcO flee

0 0

Post a comment