Preview of 1994 Ozone Precursor Concentrations in the Norteastern U.S.
An important use of PAMS data is to identify the influence from a wide variety of local and distant mobile, area and point source emissions. With the exception of isoprene, all the most abundant hydrocarbon species (Figure 5.2) are major components of motor vehicle exhaust and/or related evaporative emissions. For example, it has been shown that >60% of gasoline vapor (partial evaporation, as in refueling/headspace losses) is composed of the compounds n-butane, isopentane, n-pentane and propane 9, 10. The 1994 PAMS data shows that these four compounds accounted for 28%, 35%, 31%, 36% and 38% of the total VOCs at Chicopee, E. Hartford, Lynn, Stafford and Cape Elizabeth respectively.
Although the air at all five sites has a similar flavor, it is not entirely due to gasoline vapor. Toluene, ethane and the m&p-xylenes comprise a large portion of the total VOCs measured, but make up only one percent of gasoline vapor. These three compounds however account for up to 20% of TNMHC tailpipe emissions from catalyst equipped automobiles 9, indicating that auto exhaust may also a leading source of the VOCs measured in the Northeast during the period under discussion. In the same site order as above, toluene, ethane and m&p-xylenes accounted for 28%, 25%, 23%, 24% and 26% (only 2 of 3 available) of the total targeted VOCs.
Toluene and m&p xylenes are also major components (up to 20% of TNMHC emissions) of whole gasoline (released to the atmosphere from complete gasoline evaporation - such as from spills, leaks and vehicle hot-soak emissions). Given the extreme high temperatures during the selected July, 1994 episode periods, it is likely that whole gasoline evaporation was also a significant contributor to ambient VOC concentrations. In addition to the above-mentioned motor vehicle-related sources, many of these same compounds are also emitted by a variety of other stationary and area source activities, and/or are formed in the atmosphere by photochemical processes. Over time, fresh emissions from any of these source types will age, depleting more reactive compounds, while compounds subject to secondary formation processes increase. Also, the composition of gasoline and other fuels varies from year to year, and on a seasonal and spatial basis. Taken together, these factors pose a substantial challenge to the use of PAMS data for source identification.
Despite these difficulties, a number of recent investigations have successfully employed PAMS data to distinguish among different source types. Henry et al. (1994)11 employed a combined graphical and multi variate statistical analysis (GRACE/SAFER method) to successfully reproduce a series of vehicle-related source profiles from ambient hydrocarbon data collected at a downtown Atlanta site during the summer of 1990. While this sophisticated analysis technique is well beyond the scope of the current report, Henry's source profiles may provide a useful starting point for source-related investigation of the 1994 NESCAUM episode PAMS results.
The authors first developed measurement-based emissions profiles for 37 compounds from 3 motor vehicle-related source types: roadways (tailpipe plus running losses from highway tunnel measurements); gasoline headspace vapor (partial evaporation such as from storage tanks or vehicle diurnal evaporation); and whole gasoline (complete evaporation such as from spills, leaks and vehicle hot-soak emissions). The authors then reproduced these emissions profiles from ambient measurement data collected at the Georgia Tech site in downtown Atlanta.
Selected results from the emissions-based source profiles are reproduced here in Table 6.1. This table also includes Carter's MIR5 values (see Section 5.4) - to provide an approximate indication of the relative reactivities of the individual species. The relative abundances are normalized to n-Butane, which composed 6.22%, 4.83%, and 24.2% of the (37 species) total TNMHC from roadway, gasoline, and headspace emissions respectively. The estimated contributions from the 3 source types to the (37 species) ambient TNMHC levels were: roadway (62%); gasoline (15%) and headspace (4%).11
Table 6.1 Selected Reactivities and Source Composition Profiles from Atlanta
Species MIR5 Roadway11 Gasoline11 Headspace11 ethene 7.4 1.06 0.00 0.00 acetylene 0.5 0.92 0.00 0.00 n-Butane 1.02 1.00 1.00 1.00 n-Pentane 1.04 0.65 1.04 0.44 isopentane 1.38 2.09 2.52 1.40 benzene 0.42 0.67 0.55 0.05 toluene 2.70 1.75 2.83 0.06 m/p-Xylene 7.4 1.06 1.83 0.02
5. Maximum Incremental Reactivity estimates (g. Ozone/g. HC) from Carter (1994)
11. Estimated Source Composition (normalized to n-Butane) from Henry et al. (1994)
Benzene, toluene and m/p-xylenes are evident in all three motor vehicle-related source profiles in Table 6.1, but are present in different proportions in each source-type. These compounds also vary considerably in their relative reactivities (MIR) and have estimated atmospheric lifetimes ranging from more than a week (benzene) to several days (toluene) to less than 24 hours (m/p-xylene). Consequently, these compounds may prove useful in distinguishing among source types and in identifying influences from fresh vs. aged emissions. To initiate an analysis of the July, 1994 NESCAUM episode data, we might begin with a set of assumptions based on the Atlanta source profiles, predict certain relationships based on these assumptions, and then test these predictions by comparison with the ambient NESCAUM PAMS measurements.
Assumptions for Interpreting NESCAUM PAMS data (in relation to Atlanta source profiles):
- benzene, toluene and m/p-Xylene are emitted predominantly by motor vehicle-related sources,
- motor vehicle-related source emissions profiles in the Northeast are similar to those in Atlanta,
- the relative mix of motor vehicle-related sources in the Northeast is similar to the mix in Atlanta.
Predictions for NESCAUM Urban Sites (influenced by fresh motor vehicle-related emissions):
- benzene, toluene and m/p-xylene concentrations will be strongly inter-correlated,
- ratio of benzene to toluene (B/T) will approximate 0.4,
- ratio of m/p-Xylene to toluene (X/T) will approximate 0.6.
Predictions for NESCAUM Rural Sites (influenced by aged motor vehicle-related emissions)
- benzene, toluene and m/p-xylene will be less well correlated than in urban centers,
- B/T ratio will be greater than 0.4 (as toluene is differentially removed during transport),
- X/T ratio will be less than 0.6 (as xylenes are differentially removed during transport),
- The effects of aging will be more evident during daylight, and less evident at night.
Figure 6.1 displays the estimated B/T and X/T ratios based on 1990 Atlanta source profiles and source mix. Figure 6.2 shows the measured hourly ratios of benzene and m/p-xylene to toluene from the urban (type 2) PAMS site in E. Hartford, CT during the July, 1994 episode periods.
Figure 6.1 Estimated Urban B/T and X/T Ratios from Atlanta, Source profiles (from Henry et al., 1994)
Figure 6.2 Measured Urban B/T and X/T Ratios from E. Hartford, CT PAMS Site during July '94 Episodes
The scatter plot in Figure 6.2 shows that toluene levels at E. Hartford were highly correlated with both benzene and m/p-xylenes - in a manner consistent with a "common source" hypothesis. The B/T and X/T ratios at E. Hartford are also consistent with the predicted ratios displayed in Figure 6.1 derived from the Atlanta source profile data - suggesting that the E. Hartford source profiles and source mix are consistent with those in Atlanta. Figure 6.3 shows the hypothetical effect of aging on Atlanta B/T and X/T ratios, while Figure 6.4 shows the measured B/T and X/T ratios for the rural type 3 PAMS site in Stafford, CT site (about 20 miles downwind of E. Hartford) during the July, 1994 episodes.
Figure 6.3 Predicted Changes in B/T and X/T Ratios at Rural Sites (Resulting from Airmass Aging)
Figure 6.4 Measured Rural B/T and X/T Ratios from Stafford, CT PAMS Site during July '94 Episodes
The scatter plot in Figure 6.4 shows that the B/T ratio at the downwind, rural Stafford site has increased and the X/T ratio has decreased in comparison to the urban E. artford site (Figure 6.2). This is consistent with the predicted effect of airmass aging, as the more reactive species are differentially removed during transport.
The points plotted in Figure 6.4 also exhibit greater scatter (B/T and X/T correlations are poorer) than Figure 6.2. While common (motor vehicle-related) sources are still anticipated to be a predominant cause of benzene, toluene and m/p-xylenes at the Stafford site, the species intercorrelations are diminished during transport, as the degree of aging depends on variable factors such as wind speed direction, solar radiation, OH, NOx, etc.
Figures 6.5 and 6.6 show the relationships between toluene and m/p-xylenes at E. Hartford and Stafford, with different symbols to distinguish between daytime and nighttime samples. At the urban E. Hartford site, there is a relatively small difference in the X/T ratios between nighttime samples (when reactivity is minimal) and daytime samples (when reactivity is maximal). This is consistent with a predominant, continuous influence of fresh, local, motor vehicle-related emissions at this site.
Figure 6.5 m/p-Xylene vs. Toluene at E. Hartford, CT During July, 1994 Episodes
Figure 6.6 m/p- Xylene vs. Toluene at Stafford, CT During July, 1994 Episodes
At the rural Stafford site (Figure 6.6), there's a more distinct difference between the daytime and nighttime X/T ratios. Nighttime ratios show a stronger correlation, and a slope similar to (East Hartford's and) the predicted value of 0.6 for fresh emissions. For daytime samples at Stafford, there's a clear downward shift in the X/T slope (and a much poorer X/T correlation). This is consistent with a predominant influence of transported, motor vehicle-related emissions, which are photochemically aged (in a highly variable way) during the day, but which remain relatively unaged in the absence of sunlight.
Figure 6.7 Average Diurnal B/T Ratios at CT PAMS Sites during July, 1994 Episodes
A similar pattern is evident in the Figure 6.7 plot of average diurnal benzene to toluene ratios at the E. Hartford and Stafford sites. The E. Hartford B/T ratio is relatively constant over the course of a day, while the Stafford ratio is slightly higher than E. Hartford at night and substantially higher during the daylight hours of maximum photochemical reactivity.
In the examples above, we began with a number of assumptions relating to source profiles and source mixtures - based on 1990 Atlanta data. Our predicted benzene to toluene ratio of 0.4 (0.37) for fresh motor vehicle-related emissions was based on estimated Atlanta B/T ratios of 0.38, 0.19, and 0.83 for "roadway"' "gasoline" and "headspace" emissions, and an estimated source mixture of 62% roadway, 14% gasoline and 4% headspace. While the 1994 NESCAUM episode data from East Hartford and Stafford, CT are consistent with these assumptions, they do not necessarily demonstrate that the assumptions are correct. Figure 6.8 shows the average benzene to toluene ratios for all NESCAUM sites averaged over the two July, 1994 episode periods.
Figure 6.8 Average Benzene to Toluene Ratios For NESCAUM PAMS Sites During July, 1994 Episodes
The average B/T ratios at the urban E. Hartford, E. Providence and Lynn sites are close to the predicted value of 0.37 from Atlanta. The rural Stafford and Cape Elizabeth sites have substantially higher B/T ratios - consistent with the predicted effect of aging. The Rider University, NJ and Chicopee, MA sites exhibit substantially lower B/T ratios - suggesting that local emissions at these sites are relatively enriched in toluene in comparison to the Atlanta estimate (and other NESCAUM PAMS sites). The New Jersey data set is too limited (only 2 days of 3-hour samples) to draw meaningful conclusions. However, the Chicopee data includes hourly coverage for each of the 2 3-day episodes - and so the Chicopee average B/T ratio in Figure 6.8 is based on a sample size of > 100.
Figure 6.9 plots the individual hourly B/T ratios for all NESCAUM sites during the July, 1994 episodes. The data points from the Chicopee, MA site are indicated with different symbols.
Figure 6.9 B/T Ratios for NE PAMS Sites During 7/94 Episodes, with Estimated Atlanta Source ratios for Headspace (H), Roadway (R) and Gasoline (G)
The estimated B/T ratios for Headspace (H), Roadway (R) and Gasoline (G) emissions from Atlanta (Table 6.1) are also displayed. The effect of aging would be to increase these source B/T ratios (shift points toward upper left). With the exception of the Chicopee site, nearly all the plotted points fall within the range indicated by the 3 Atlanta motor vehicle-related source profiles. Most of the Chicopee B/T points fall outside (below) all of the Atlanta profiles - suggesting influence from a local, non-auto-related source enriched in toluene.
Figure 6.10 n-Pentane vs. n-Butane at NE Sites During 7/94 Episodes (with Atlanta Source Ratios)
Figure 6.11 Isopentane vs. n-Butane at NE Sites During 7/94 Episodes (with Atlanta Source Ratios)
Similar scatter plots in Figures 6.10 and 6.11 show ratios of n-Pentane and isopentane to n-butane, with Chicopee distinguished separately, and Atlanta MV-related source profiles for comparison.
N-pentane and n-butane are relatively non-reactive, and have similar, low MIR values - so we would anticipate minimal effects of aging on P/B ratios. Isopentane is slightly more reactive - so we might anticipate an effect of aging to decrease the I/B ratios in Figure 6.11 (shift points to the lower right). As in Figure 6.9, the majority of data points in Figures 6.10 and 6.11 fall within the range suggested by the Atlanta motor-vehicle-related source profiles. The Chicopee, MA site also exhibits a number of data points with P/B and I/B ratios similar to the other regional sites. However, Chicopee also exhibits a number of extreme outliers, where n-pentane and isopentane are highly enriched (relative to n-butane) - indicating occasional (but not continuous) influence from a local, non-motor-vehicle-related source or sources. It is also noteworthy that the "Chicopee outliers" in Figures 6.10 and 6.11 occur during the same sample hours. That is, isopentane and n-pentane are enriched concurrently - suggesting a common, local source for these compounds. This is further illustrated by the average diurnal cycle plots in Figures 6.12 and 6.13.
Figure 6.12 Average Diurnal n-Pentane at NE PAMS Sites During 7/94 Episodes
Figure 6.13 Average Diurnal Isopentane at NE PAMS Sites During 7/94 Episodes
N-pentane and isopentane are relatively non-reactive - and so we expect, and find, relatively little diurnal variation at all sites except Chicopee, where the unique local source of these compounds is active only during the day between 10 AM and 5 PM. As shown in Figure 6.14, cyclopentane and 2,2-dimethylbutane also exhibit a similar mid-day increase at the Chicopee site during the same hours.
Figure 6.14 Average Diurnal Patterns of Selected VOCs at Chicopee, MA During 7/94 Episodes
Figure 6.15 Average Diurnal Toluene Patterns at NE PAMS Sites During 7/94 Episodes
We had also noted that toluene was relatively enriched at the Chicopee site (Figure 6.9), in comparison to other NESCAUM sites and to expected Atlanta MV-related source profiles. The diurnal cycle plots in Figure 6.15 illustrate this unique toluene influence at Chicopee. However, in contrast to Figures 6.12, 6.13 and 6.14, the Chicopee toluene enrichment is obvious only at night. It would appear that there are at least 2 unique, local, non-MV-related sources near Chicopee. One operates only during the day, emitting n-pentane, isopentane, cyclopentane and 2,2-dimethyl butane. In comparison to their "baseline" levels (6 am), these relative concentrations from this daytime source are approximately: (1) 2,2-dimethyl butane to (6) cyclopentane to (7) iso-pentane to (14) n-pentane. The second source operates (or its influence is most notable) only at night, and emits toluene. This concept of (at least) 2 unique, local sources at Chicopee is further illustrated by the toluene vs. isopentane plot in Figure 6.16.
Figure 6.16 Isopentane vs. Toluene at NE PAMS Sites
In this case, the effect of aging (depletion of the more reactive toluene) would tend to shift points toward the upper left. This plot may be thought of as showing (at least) three different distributions:
- one for the majority of points from all sites except Chicopee (consistent with Atlanta MV-related source profiles);
- one with much higher I/T ratios at Chicopee ("Source 1"); and
- one with much Lower I/T source ratios at Chicopee ("Source 2").
The nature of these hypothetical local sources at Chicopee can be further investigated by examining some of the on-site meteorological data associated with the extreme Chicopee outliers. "Source 1" (Figure 6.17) influence occurs only during the day, at a wide range of wind speeds (3.2 to 10 mph), but at a very consistent wind direction (about 200 degrees). "Source 2" influence occurs predominantly at night, at a variety of wind directions, but only at very low wind speeds (< 3.2 mph).
Figure 6.17 Surface Winds for Chicopee and Springfield, MA for Hours with High I/T Ratios (Source 1)
Additional investigations of these predicted local source influences are currently underway.
Figure 6.18 Surface Winds for Chicopee and Springfield, MA for Hours with Low I/T Ratios (Source 2)
