2. AIR QUALITY DATA ANALYSIS

Ozone is a photochemical oxidant formed from reactions involving nitrogen oxides (NOx) and volatile organic compounds (VOC) in the presence of sunlight. In order to understand the nature of ozone levels and transboundary flows of ozone and precursor emissions in the border region, this section presents highlights of several existing analyses, assessments, and publications of air quality and meteorology. This work includes the existing and currently ongoing Canadian Multistakeholder NOx/VOC Science Assessment reports (Multistakeholder, 1997) and the OTAG final report (OTAG, 1997). In addition, new analyses that extend these data and analytical techniques into the Canada-U.S. transboundary region were developed for this report and are summarised below (Dann, 1999, Husar et al., 1999; Schichtel and Husar, 1999).

The sections presented here include: a snapshot of current ozone levels, regional maps showing episodic flows in the border region, emissions information, and an overview and analysis of meteorological factors affecting ozone concentrations and transport.

Air Quality Snapshot

The air quality snapshot depicts the regional extent of elevated ozone concentrations in the Canada-U.S. border area based on analysis and maps in an Environment Canada report (Dann, 1999). Data from 100 Canadian sites and 122 U.S. sites for the ozone season (May to September) for the period 1994 to 1996 were used to demonstrate regional patterns in ozone concentrations. Ozone concentrations were computed for running 8-hour periods and maximum 8-hour concentrations by day were determined. The maximum and the fourth highest daily maximum 8-hour ozone values were then computed by site by year.

Figure 1 shows the distribution of fourth highest daily maximum 8-hour ozone concentrations by monitoring site within each region using data for 1994 to 1996. The boxplot figure provides the median, the 95th, 75th, 25th and 5th percentile site ozone concentrations². Figure 2 maps the 4th highest daily maximum 8-hour ozone concentration for the north-eastern portion of North America averaged over the years 1994-1996. Figure 3 provides similar information but uses the average highest daily maximum 8-hour ozone value.

Figure 1. Distribution of 4th Highest Daily 8h Maximum Ozone (ppb) for Regional Sites (1994 to 1996) (Median, 5th, 25th, 75th and 95th Percentiles)³ The highest concentrations are recorded at the urban or industrialised sites in the United States (Michigan, New York, Ohio and Pennsylvania) and at the southwestern Ontario (SW) sites. The lowest ozone concentrations are at the Canadian Prairie (Pr) and Vancouver, British Columbia (VA) sites. The state of Maine also records high ozone values that are likely due to transport rather than local generation.

Figures 2 and 3 show large portions of the eastern United States exceeding the 8-hour 85-ppb level with some portions of Ontario also over this threshold. Many regions also exceed the range proposed for the Canada-wide Standard. A significant feature is that ozone concentrations in most of eastern North America, including further east along the coast, into Nova Scotia, and outside the major urban-industrialised areas are well above background concentrations. Maps of ozone episodes are shown in the following section.

Figure 2. Average of the 4th Highest Daily 8h Maximum Ozone Concentration (ppb) for 1994 to 1996⁴

Figure 3. Average of the Highest Daily 8h Maximum Ozone Concentration for 1994 to 1996⁴.

Ozone Episodes

Widespread regional episodes are a common feature of eastern North America and have the potential to contribute to exceedences of air quality objectives and standards. To illustrate how a regional episode develops and flows within the region of interest, measured ozone concentration data were compiled from monitoring sites located in eastern United States and eastern Canada for two regional ozone episodes and then mapped. The episode years, 1988 and 1995 were chosen because they show clearly ozone transport within the region of interest over the duration of the episodes. The 1988 episode illustrates ozone transport in both directions across the Canada-U.S. border whereas the 1995 episode illustrates a good example of transport from the United States to Canada. Both episodes were also used in the joint modelling scenarios presented later in this report.

During the summers of 1988 and 1995, ozone-rich plumes were transported across all of eastern North America. Many sites recorded multiple hours and days with ozone concentrations greater than the Canadian and U.S. air quality criteria. The following series of maps (Figures 4, 5 and 6) depict the levels of ozone concentrations in eastern United States and Canada at four progressively later hours in a day during an episode. The maps provide the magnitude and extent of high concentrations while demonstrating movement of ozone over time through the region.

These figures illustrate that essentially all areas of eastern Canada and most areas of the eastern United States experience high concentrations of ozone. Although some areas experience very high 8-hour concentrations, widespread areas experience concentrations ranging from 60-80 ppb and from 80-100 ppb over 8 hours. The following section will discuss what factors contribute to high ozone concentrations locally and regionally.

Figure 4. Ozone Transport on July 13, 1995 10AM - 8PM. The first frame shows a regionally uniform pattern of ozone levels. Transport generally followed a north-easterly path, across the heavily industrialised and urbanised area of the U.S. Midwest, then across the Great Lakes (frame 2) into southern Ontario and out to the coast. After picking up local emissions, transport continues west along the St. Lawrence river basin (frame 3) and finally out to the North Atlantic.

Figure 5. Ozone Transport on July 7 1988, 10AM - 7PM. These frames show fewer urban-industrial peaks and an elevated non-urban pattern extending across most of eastern North America. Both characteristics are evidence of regional scale transport.

Figure 6. Ozone transport on July 10, 1988, 10AM - 11 PM. Figure 6 shows night time flows into eastern Canada when concentrations greater than 82 ppb occur late in the evening in the southern Atlantic region (frame 4) after high ozone concentrations have moved across the Northeast region (IND, WV, VA, OH, PA, ON) and along the Atlantic seaboard.

Factors that Influence Ozone Concentrations

The direction and spatial extent of transport and the relative contribution of transported ozone and precursors to individual ozone exceedences are highly variable. A number of factors influence site-to site differences in ozone concentrations, including sources of precursor emissions and large-scale and local meteorology. The following information comes from several sources, including the Canada-U.S. Air Quality Agreement 1998 Progress Report and analyses conducted for this report as cited above.

Sources of Ozone Precursors and their Influence on Ozone Formation

As discussed previously, ozone is not emitted directly, but formed in the atmosphere by reactions of "precursors" (NOx and VOCs). Both NOx and VOCs are emitted from a variety of sources. Anthropogenic sources of NOx emissions in the United States are 10 times larger and VOC emissions are 7 times larger in magnitude than in Canada, paralleling the relative population ratio between the two countries.

The relative distribution of major sources categories of NOx and VOC emissions in the U.S. and Canada are somewhat different, as shown in Figures 7 and 8. Although transportation is the largest source of both precursor emissions in both countries, there are some significant differences in source apportionment. For instance, electric utilities in the United States generate 27% of total NOx while contributing only 11% in Canada. Although not shown in these figures, natural emissions also play an important role with respect to summertime emissions of VOCs when emissions from vegetation, for example, have been found to equal or exceed manmade VOC emissions over large regions of eastern North America.

Figure 7. NOx Emissions in Canada and the United States (1995). Source: Canada-United States Air Quality Agreement 1998 Progress Report, p15.

Figure 8. VOC Emissions in Canada and the United States (1995). Source: Canada-United States Air Quality Agreement 1998 Progress Report, p15.

Although both VOC and NOx emissions contribute to ozone formation, the relative effectiveness of reductions of the two precursors can vary with location and atmospheric conditions. Many publications and reports have examined the impacts of reducing VOCs versus NOx in different urban and regional domains of the United States and Canada (NRC, 1991; OTAG, 1997; Multistakeholder, 1997; CEC, 1997). While the specifics of this situation can be quite complex, the following general conclusions can be reached based on these reports:

1. In urban conditions with relatively low VOC to NOx ratios, anthropogenic VOC reductions can be the most effective strategy for reducing ozone levels; in such conditions, NOx reductions can lead to localised increases in ozone, with decreases in ozone downwind.
2. In conditions with high VOCs relative to NOx, such as downwind of major urban centres, NOx controls may be the most effective strategy for reducing ozone. Due to the day-to-day variability in emission levels, background VOC and NOx concentrations and wind patterns, both NOx and VOC controls may be needed to reduce ozone in particular urban areas.
3. In rural areas of eastern North America, there is an abundance of naturally produced VOCs. In such areas, control of smaller anthropogenic VOC emissions may be ineffective, and NOx controls are an effective control approach.
4. Over a large multi-dimensional region - one that has both urban and rural areas with naturally produced VOCs and anthropogenic NOx emissions transported downwind of urban areas - regional reductions of NOx are the most effective control approach. As shown in this report, the eastern Canada-U.S. border region is such a multi-dimensional region. Therefore, NOx reductions should be the most effective control approach.

The spatial patterns of NOx emission sources in Canada and the United States are shown in Figures 9 and 10. The figures show high NOx emission densities in urban-industrialised areas, although the order of magnitude is different from Canada to the United States.

Figure 9. 1995 NOx Emission Densities (in kg/km²) for Eastern Canada.
Source: Environment Canada Pollution Data Branch.

Figure 10. NOx Emission Densities (in tons/hr/grid) in the Eastern United States and Canada. The emissions included on this map for the U.S. reflect the 1995/96 base year emissions that were used in the NOx SIP Call. The emissions for Canada are the 1990 emissions that were used by OTAG.

In the transborder region, the areas of highest emission densities are along the Canada-U.S. border and along the Atlantic coast in the United States. Taken together, these twin corridors of dense population and precursor emissions run from the Southwest to the Northeast, in parallel to weather patterns that frequently occur in the summer. The metropolitan areas along the Canada-U.S. border also have high emission densities of ozone precursors. The following sections present data on transport meteorology, and show the relationship between transport and ozone concentrations within the region.

Ozone as a Function of Wind Speed and Direction

Analysis of ozone a function of wind speed and direction can help provide insight into the relative importance of local and distant sources under varying meteorological conditions. Previous analyses, including those done for OTAG (OTAG, 1997) and for the Canadian Multistakeholder NOx/VOC Science Assessment (Multistakeholder, 1997) indicate that these factors are important influences on the transport of ozone. The series of analyses for the OTAG region has been extended for this report to include Canada to provide more specific insights into transboundary issues (Husar et al., 1999).

In order to analyse the effects of wind speed and direction on ozone concentrations, analysts sorted 11 years of measured ozone concentrations (1989-1996) and averaged for specific wind direction and speed ranges. The average ozone concentration was computed for each wind direction range in 90o increments, starting with 0-90*, i.e. when the wind blew from the North or Northeast. This resulted in four wind directional concentration bins. The average concentrations for each directional bin were further classified by wind speed, ranging between 0-2, 2-4, 4-6, 6-8 metre/second (m/s) increments. Thus, there were four directional and four wind speed bins, yielding a total of 16 concentration bins.

The results of this analysis are presented in maps of average ozone concentration for the four wind directions and three wind speeds - low, medium and high (Figures 11, 12, 13). The detailed results are discussed in a supporting technical paper (Husar et al., 1999).

In summary, average ozone concentration maps at low wind speeds (<3 m/s, Figure 11) show elevated levels of ozone throughout the eastern North American domain. Ozone concentration hot-spots appear over the major metropolitan areas in the United States and the Ohio River Valley but the concentrations are virtually the same regardless of the wind direction. Ozone concentrations in metropolitan areas of Canada are similar to surrounding sites. At intermediate wind speeds (3-6 m/s, Figure 12) the overall concentrations are lower, and the higher ozone concentrations appear to be displaced up to 500 km downwind of the major source areas. At high wind speeds (>6 m/s, Figure 13) most metropolitan source areas do not cause elevated ozone in their own vicinity. Rather, higher concentrations appear in the downwind corners of the eastern North American domain, up to 1000 km from the domain centre.

The ozone concentration pattern at different wind directions and speeds are consistent with an atmospheric ozone lifetime of about one day and a corresponding transport distance of 200, 500 and 800 km at 2, 5, and 8 m/s respectively. Therefore, at low wind speeds, ozone accumulates near precursor emission source areas. Higher wind speeds cause increased dilution of local concentrations and increased transport from one source region to another.

Figure 11. Maps of average ozone concentration at low (< 3 m/s) wind speed. a) 270-360 degrees, b) 0-90 degrees, c) 90-180 degrees, d) 180-270 degrees. At low wind speeds, ozone concentrations tend to be somewhat higher just downwind of urban areas. Concentrations tend to be fairly similar, regardless of wind direction. In such cases, local sources likely dominate ozone formation.

Figure 12. Maps of average ozone concentration at intermediate (3-6 m/s) wind speed. a) 270-360 degrees, b) 0-90 degrees, c) 90-180 degrees, d) 180-270 degrees. At intermediate wind speeds, there are substantial differences between the maps depending on the wind direction. Northerly flows (frames a and b) show low concentrations throughout Canada and the northern United States. Southerly flows result in higher concentrations in the north, especially in the Michigan/Ontario/New York region.

Figure 13. Maps of average ozone concentration at high (>6 m/s) wind speed. a) 270-360 degrees, b) 0-90 degrees, c) 90-180 degrees, d) 180-270 degrees. At high wind speeds, the eastern North American domain appears as a regional domain, although there are still some near-urban areas of elevated ozone.

Graphs of individual urban areas (Detroit, Chicago and Toronto) in the north-central domain show the average ozone concentrations at four characteristic speeds (1,3,5,7 m/s) from wind speed ranges of 0-2, 2-4, 4-6, 6-8 m/s respectively. The data are further stratified by four wind directional quadrants from 0-90 degrees through 270-360 degrees. A fifth line represents wind speed dependence of ozone, regardless of direction.

Figure 14. Dependence of ozone concentration on wind speed and direction at different metropolitan areas. The metropolitan areas of Chicago, Detroit and Toronto show a remarkable directionality of the ozone-wind speed dependence. Northerly winds (270-90 degrees) are associated with declining ozone with wind speed. This is interpreted as evidence of substantial local contributions. On the other hand, the wind directions from the centre of the eastern United States are associated with speed-independent ozone levels, implying transported ozone. Taking into account all wind directions, transport is clearly an important component of the ozone problem in each of these areas.

The results for the three urban areas (Figure 14) show remarkable directionality of the ozone-wind speed dependence. For these areas, when the wind blows from the North (270-90 degrees), ozone concentrations decrease with increasing wind speed. This pattern is consistent with locally generated ozone. When the wind blows from the South (90-270 degrees), the graphs show speed-independent ozone concentrations, indicative of levels dominated by transport.

The following section examines the transport of ozone on days of regionally high and low ozone concentrations.

Transport During High and Low Ozone Days

Ozone transport on days of regionally high and low ozone concentrations was also investigated in the analyses conducted for this report. Ozone transport, precursor emissions, and the influence of wind were examined in two ways. First, from the transport climatology, regions were identified where transport conditions are conducive to the accumulation of ozone from local or sub-regional sources, as well as where influence by regional scale transport can be seen. Second, by contrasting the transport conditions during periods of high and low ozone concentrations, unique transport pathways for a given region, as well as common pathways for multiple regions, were identified. The results of this analysis are summarised below and presented fully in a background paper (Schichtel and Husar, 1999).

Transport conditions were established for regionally high (90^th percentile) and low (10^th percentile) daily maximum 1-hour ozone concentrations. Figure 15 shows source regions of influence (SRI) overlaid with transport wind vectors. The wind vectors convey the direction and magnitude of the air mass transport while the SRIs represent the area encompassing the source impact and resultant air mass transport direction and speed. The SRIs are for the nearest modelled regions to Atlanta, Houston, Chicago, the Ohio River Valley, and New York City. At each source region, the highest and lowest daily maximum ozone values usually occurred on different days. Therefore, the transport conditions at each source region represent transport over different time periods.

Regionally high ozone days (Figure 15A) were associated with slow meandering or recirculating transport over Kentucky, Tennessee, and West Virginia, with strong clockwise transport around this region. It is clear that transport from sources in the southern Great Lakes border region moves from the United States into Canada, over the most dense emission regions of Southern Ontario, and back into New York and the New England States. This flow pattern is consistent with that of a large high-pressure system over eastern North America. The regionally low ozone days (Figure 15B) had northerly flow into Canada from Wisconsin and Michigan that converged over Kentucky and Tennessee with swift westerly-southwesterly flow in the Southeast. In New England, substantial transport occurred in all directions with resultant mass transport to the East.

Figure 15. Transport wind vectors and source regions of influence for the highest (A) and lowest (B) 10% of regional ozone days during June - August, 1991 - 1995. Transport vectors in Figure A, regionally high ozone days, show wind speed and directions consistent with regional-scale episode, i.e., strong clockwise transport. There is substantial transport into southern Ontario, particularly from the Chicago source region. Figure B, low regional ozone days, show transport vectors from the Great Plains into the Prairie provinces, east towards southern Ontario and south into the New England states. The New York source region shows substantial transport in all directions, while transport is primarily south in the Chicago source region.

Air Quality Analysis Conclusions

Ozone is not emitted directly, but produced in photochemical reactions involving NOx and VOC emissions. High ozone concentrations occur in and around many of the urban-industrialised areas in the transboundary region in both countries, resulting in frequent exceedences of current and proposed air quality objectives and standards. Elevated concentrations also occur over areas several hundred kilometres downwind of urban areas, causing exceedences of the objectives and standards in relatively less populated non-industrial areas. Some of these exceedences can occur at night as a result of ozone transport.

Transport of ozone and precursors has no boundaries. Polluted air masses can travel across states and provinces and between the United States and Canada. High ozone concentrations are typically located downwind of areas with the highest emissions. In the Canada-U.S. transboundary region, a more uniform pattern of ozone concentrations occurs across the region. The result is a regional "sea" of elevated ozone, extending east from the Mississippi River to the Atlantic coast and north-northeast into the Windsor-Québec City Corridor, New Brunswick, and Nova Scotia, punctuated by "hot-spots" associated with dense emissions areas. Emissions of NOx in the area were shown to form a boundary around the populated transboundary region, with dense emissions in the Windsor-Quebec corridor, and along the Atlantic coast from New Jersey to Massachusetts. The Ohio River Valley area, dense in precursor emissions, sits at the "entrance" of these twin transboundary emission corridors.

Ozone transboundary flux data, presented in the section on wind speed and direction, illustrate flows in both directions, but are consistent with greater transport of ozone from the United States to Canada than from Canada to the United States. Concentrations along the Detroit-Windsor-Québec City corridor, and eastward, are increased when the wind blows from the "entrance" of the corridor, both towards the northeast, and north-northeast around the Great Lakes and out to the Atlantic.

Increasing wind speeds generally bring about reductions in locally produced ozone concentrations in urban areas. Ozone concentrations in the northeast urban corridor, along the Atlantic coast, are reduced when the ozone and precursor emissions are blown out to sea. For border areas where local concentrations remain constant with wind speed, it appears that the regional transport dominates ozone concentrations. This is observed in several of the urban areas along the border (e.g., Detroit, Windsor, and Toronto) where ozone concentrations were not reduced with increasing wind speeds.

Designing ozone control strategies is complicated because the effectiveness of strategies depends on factors such as meteorological conditions, the absolute and relative amounts of VOCs and NOx, the spatial and temporal distribution of anthropogenic and natural emissions, and background concentrations. The air quality data analysis section discussed the interrelationships between these factors and how each influences ozone concentrations. Some general conclusions can be made, therefore, regarding the effectiveness of strategies in the Canada-U.S. transborder region, based upon the conclusions from earlier reports and the information presented in this report. In the urban areas, a combination of VOC and NOx emission reductions are expected to lead to reductions in high ozone levels locally and downwind. In the areas affected by transport of ozone and precursor emissions, the downwind urban, suburban and rural areas, i.e. the Canada-U.S. transborder region, NOx reductions are expected to be more effective.

The next section presents the results of air quality modelling using Canadian and U.S. data and forecasts of planned reduction program, focusing on NOx emission reductions to show the likely impact of emission control scenarios in the transboundary region.