Air Toxic Emissions from Snowmobiles in Yellowstone National Park

Dec 2, 2009 - Climate Change Research Center, Institute for the Study of Earth, Oceans and Space, University of New Hampshire, Durham, New Hampshire ...
1 downloads 3 Views 1MB Size
Download PDF
Environ. Sci. Technol. 2010, 44, 222–228

Air Toxic Emissions from Snowmobiles in Yellowstone National Park Y O N G Z H O U , * ,† D A V I D S H I V E L Y , ‡ HUITING MAO,† RACHEL S. RUSSO,† BRUCE PAPE,§ RICHARD N. MOWER,§ ROBERT TALBOT,† AND BARKLEY C. SIVE† Climate Change Research Center, Institute for the Study of Earth, Oceans and Space, University of New Hampshire, Durham, New Hampshire 03824, Department of Geography, The University of Montana, Missoula, Montana 59812, and Department of Geography, Central Michigan University, Mount Pleasant, Michigan 48859

Received July 16, 2009. Revised manuscript received November 9, 2009. Accepted November 11, 2009.

A study on emissions associated with oversnow travel in Yellowstone National Park (YNP) was conducted for the time period of February 13-16, 2002 and February 12-16, 2003. Whole air and exhaust samples were characterized for 85 volatile organic compounds using gas chromatography. The toxics including benzene, toluene, ethylbenzene, xylenes (p-, m-, and o-xylene), and n-hexane, which are major components of twostroke engine exhaust, show large enhancements during sampling periods resulting from increased snowmobile traffic. Evaluation of the photochemical history of air masses sampled in YNP revealed that emissions of these air toxics were (i) recent, (ii) persistent throughout the region, and (iii) consistent with the two-stroke engine exhaust sample fingerprints. The annual fluxes were estimated to be 0.35, 1.12, 0.24, 1.45, and 0.36 Gg yr-1 for benzene, toluene, ethylbenzene, xylenes, and n-hexane, respectively, from snowmobile usage in YNP. These results are comparable to the flux estimates of 0.23, 0.77, 0.17, and 0.70 Gg yr-1 for benzene, toluene, ethylbenzene, and xylenes, respectively,thatwerederivedonthebasisof(i)actualsnowmobile counts in the Park and (ii) our ambient measurements conducted in 2003. Extrapolating these results, annual emissions from snowmobiles in the U.S. appear to be significantly higher than the values from the EPA National Emissions Inventory (1999). Snowmobile emissions represent a significant fraction (∼14-21%) of air toxics with respect to EPA estimates of emissions by nonroad vehicles. Further investigation is warranted to more rigorously quantify the difference between our estimates and emission inventories.

1. Introduction Yellowstone National Park (YNP) is designated a mandatory Class I Airshed under the federal Clean Air Act (CAA), which requires no air-pollution-induced impairment of visibility in the area. A steady increase in wintertime oversnow recreational vehicle use in the park since the late 1960s has * Corresponding author e-mail: [email protected]. † University of New Hampshire. ‡ The University of Montana. § Central Michigan University. 222

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 1, 2010

significantly increased mobile-source emissions (i.e., snowcoaches and snowmobiles). Exhaust from snowmobiles, especially those with two-stroke engines, contains numerous toxic compounds including benzene, toluene, ethylbenzene, xylenes (p-, m-, and o-xylene), and n-hexane. These toxics are classified by the Environmental Protection Agency (EPA) as toxics and/or hazardous air pollutants (HAPs), with serious environmental and health impacts (1). Furthermore, the accumulation of toxic organics in the snowpack may potentially contaminate the watershed when flushed into nearby streams and water bodies during the spring snowmelt (2, 3). The use of oversnow vehicles in YNP has been the subject of controversy for many years. Historically, an average of approximately 765 snowmobiles entered Yellowstone each day with most having two-stroke engines. In February 2003, the National Park Service (NPS) capped the number of snowmobiles to 920 per day. Additionally, for the 2003-04 winter use season the NPS required that 80% of the recreational snowmobiles operating in the Park be “best available technology” (BAT) equipped; by the 2004-05 winter season all snowmobiles were required to be BAT equipped. The NPS decision published on November 10, 2004 for the 2006-07 winter allowed 720 snowmobiles per day in YNP, all commercially guided and all required to meet the NPS’s BAT requirements. At the present time, BAT requirements include a 90% reduction in hydrocarbons and a 70% reduction in CO relative to the EPA’s baseline assumptions for uncontrolled snowmobiles that were published in the Federal Register on November 8, 2002 (4). These requirements translate into hydrocarbon emissions not exceeding 15 g/kW h and CO emissions not exceeding 120 g/kW h. The temporary winter use plan ensures that resources are protected and allows the NPS to complete the development of an Environmental Impact Statement (EIS) and a long-term plan for winter use that will guide the management of winter recreational use of the Park which began during the 2007-08 winter. During a 2-day study (February 17-18, 1999) conducted at the West Entrance station, a mean toluene mixing ratio of 1976 ppmv was measured in snowmobile exhaust (5). To put this value into perspective, the mean toluene level from four-stroke snowmobile exhaust in this study was 13 ppmv. The February 1999 study constructed a mobile source emissions inventory based on fuel use during winter 1998-99, which suggested that snowmobiles account for ∼27% and ∼77% of the annual emissions of CO and hydrocarbons (HC) in YNP, respectively (6). The NPS estimated 864 tons of CO and 608 tons of hydrocarbons per winter season in YNP for the baseline condition (920 two-stroke snowmobiles daily) according to the 2004-05 winter use plan (i.e., corresponding to 235 tons of CO/season and 23 tons of hydrocarbons/ season, based on 920 four-stroke snowmobiles daily) (7). The Southwest Research Institute (8) reported 95-98% less hydrocarbons and roughly 90% less toxic hydrocarbons in exhaust emissions from commercially available four-stroke snowmobiles than from two-stroke sleds. Bishop et al. reported the mean CO and HC emissions per mile per passenger from two-stroke snowmobiles in 1999 (CO ) 71 g and HC ) 92 g) and four-stroke snowmobiles in 2005 (CO ) 21 g and HC ) 2.5 g) and 2006 (CO ) 15 g and HC ) 0.8 g) (9, 10). Moreover, Kado et al. measured the highest levels of VOCs at the West Entrance (11). The majority of air-quality research conducted in YNP has focused on snowmobile exhaust emissions and human exposure to toxic air pollutants. Very little information exists 10.1021/es9018578

 2010 American Chemical Society

Published on Web 12/02/2009

concerning the spatial variation of air toxics from snowmobile emissions in YNP. In this study (also discussed in ref 12), we use atmospheric measurements to characterize the spatial variation of air toxics associated with oversnow vehicular traffic. The emission rates of benzene, toluene, ethylbenzene, xylenes, and n-hexane from snowmobile use in the park are estimated on the basis of field measurements conducted in February 2002 and 2003. Our findings from this study provide baseline VOC data, which can ultimately be used by the NPS to assess the effectiveness of management decisions concerning YNP oversnow traffic control in the future.

2. Experimental Section Whole air samples were collected at 23 locations (20 primary and 3 alternate sites) to ensure complete spatial coverage throughout YNP (Figure S1a, Supporting Information). Samples were collected in the early morning and early afternoon on February 13 and 16, 2002 and February 12 and 15, 2003 to compare the relative impact of oversnow vehicle emissions on VOC distributions throughout the day and on low- and high-traffic days. February 13, 2002 and February 12, 2003 coincide with low-to-moderate traffic days (about 700-950 snowmobile entries), while February 16, 2002 and February 15, 2003 (about 1200-1500 snowmobile entries) were the President’s Day weekend, which has historically represented a high visitation period during the park’s winter season (J. Sacklin, personal communication, 2001). The morning sample collection periods began at 5 a.m. (Mountain Standard Time, MST) prior to the initiation of significant oversnow travel. Afternoon sampling began at approximately 12:00 pm (MST) and extended over a 4-5 h time period with significant oversnow travel in the park. In order to ensure acquisition of samples of well-mixed air, two separate samples were acquired at each sample site. Both were collected in an upwind direction of the nearest roadway with the first taken ∼500 m from the road (off-road) and the second ∼50 m from the road (near-road). Additionally, a full set of diurnal samples (one sample per hour) were acquired at the Lake Ranger Station, remotely located near the center of the Park (Figure S1a, Supporting Information). Samples were also collected in West Yellowstone, MT on February 16, 2003 to examine the impact of exhaust emissions from snowmobiles on the heavily traveled routes in town and at the West Entrance of YNP (Figure S1b, Supporting Information). Table S1 (Supporting Information) lists the representative snowmobiles, snowcoaches, and snowcats from which exhaust samples were collected to establish emission signatures of the various engine types. These samples were collected in canisters placed directly in the exhaust streams of oversnow vehicles and filled to ambient pressure. In most cases, the snowmobile had just completed a transit from West Yellowstone to Old Faithful. For a portion of the snowmobiles (Table S1, Supporting Information), the samples were collected at 3500 rpm, corresponding to the estimated average revolutions per minute (rpm) during transit. Although engine loading is not accurately simulated in this fashion, these samples are useful in determining relative emissions of speciated VOCs. Additionally, some samples were collected from snowmobiles that were idling with the engine still “hot”. These samples are significant and more accurately represent exhaust emissions when visitors in YNP pull over/stop to view the scenery and wildlife in the Park without turning off their oversnow vehicles. Upper air soundings to assess vertical stability of the atmosphere and potential for constraints on vertical mixing and air transport were acquired with the use of 2 m diameter meteorological balloons equipped with radiosondes during the sampling periods to determine boundary layer altitudes in the study area. Specific details regarding the meteorological conditions during the study periods are described in Shively

et al. (12). However, it is worth noting that the prevailing conditions for sampling in 2003 resulted in approximately 50% of the samples being collected under calm conditions (wind speed ∼ 0 m s-1), while ∼40% of samples were obtained when wind speeds ranged from 1 to 2 m s-1. The remaining samples (∼10%) were collected when wind speeds ranged from 2 to 5 m s-1; only two samples were collected when wind speeds were above 5 m s-1. A total of 96 and 218 whole air samples were collected in YNP in our 2002 and 2003 campaigns, respectively. The 1-L Silonite canisters (Entech Instruments Inc., Simi Valley, CA) and 2-L electropolished stainless steel canisters (University of California, Irvine, CA) were filled to ambient pressure and returned to the University of New Hampshire to be analyzed for a full suite of nonmethane hydrocarbons (NMHCs), halocarbons, and alkyl nitrates. Details regarding the system used for sample analysis have been described previously (13–15). The trace gas analytical system utilizes three Shimadzu GC17-A gas chromatographs (GCs). The samples were analyzed by cryotrapping 1000 cm3 (STP) of air from the 2-L canisters (650 cm3 (STP) of air for the 1-L canisters) on a 1/4-in. o.d. Silicosteel (Restek Corp., Bellfonte, PA) concentration loop filled with 1 mm diameter glass beads, immersed in liquid nitrogen. The total volume sampled was measured by pressure difference using a capacitance manometer. After the sample was trapped, the concentration loop was isolated and warmed to 80 °C. Helium carrier gas then flushed the contents of the loop and the stream was then quantitatively split into four, with each substream feeding a separate GC column. Specific details regarding the separation column, the corresponding detector, and the compounds measured for each column-detector pair are given in Zhou (14) and Shively et al. (12). The measurement precision for the C2-C6 NMHCs ranged from 0.4 to 3%, while the precision for the C7-C10 NMHCs varied from 2 to 10%. For the halocarbons and alkyl nitrates, the measurement precision ranged from 1 to 10% (14). For exhaust samples, a 0.5 cm3 (STP) aliquot of each sample was analyzed by direct injection using the analytical system described previously and mixing ratios for 60 different NMHCs were determined for each engine type. For the 2003 samples, additional analyses were conducted for CO and CH4 by D. Blake at the University of California, Irvine (UCI). Further details regarding the CO and CH4 measurements from the UCI group can be found in Blake and Rowland (16) and Blake (17).

3. Results and Discussion 3.1. Source Signatures from Exhaust Samples. Results from the 2002 and 2003 exhaust sample analyses revealed that all engines sampled had very similar emission ratios for both ethene and ethane. Thus, for this comparison we have normalized all of the emission ratios to ethene in order to meaningfully compare the emission ratios between the different engine types and also between years. However, it is worth noting that the results do not differ if the emission ratios are normalized to ethane. The 2003 exhaust samples revealed markedly different exhaust composition for each engine type as indicated by the emission ratios for a subset of NMHCs (ppmv) relative to ethene (ppmv) for the twostroke, four-stroke, and diesel engines (Figure 1a). The 2002 exhaust measurements exhibited a strikingly similar fingerprint to the 2003 exhaust measurements (Figure 1b). Minor differences were found in the exhaust composition between samples collected from idling and revved engines (to the appropriate RPM). Henceforth, average emission ratios for each engine type were used. Additionally, the composition of the exhaust from the Bombadier snowcoach was similar to the four-stroke snowmobiles and is included in this category. Figure 1a shows that the average VOC emission VOL. 44, NO. 1, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

223

FIGURE 1. Average NMHC exhaust emission ratios relative to ethene (ppmv/ppmv) for (a) two-stroke, four-stroke, and diesel engines in 2003 and (b) two-stroke and four-stroke engines in 2002. ratio relative to ethene is significantly larger for the twostroke engine exhaust samples than the four-stroke engines, with the largest difference in i-pentane and toluene emissions (approximately 16 and 7 times larger for the two-stroke than the four-stroke, respectively). Relatively large amounts of ethyne, n-butane and n-pentane were also present in exhaust emissions. For all three engine types, only a small fraction of the exhaust consisted of ethane, propane, propene and i-butane. The two-stroke engines emitted significantly larger quantities of air toxics (i.e., benzene, toluene, ethylbenzene, xylenes, and hexane) than either the four-stroke or diesel engines, as shown clearly in Figure 1. 3.2. Distributions of Air Toxics in YNP. The distribution of VOCs observed in YNP during February 2002 and 2003 provides striking evidence of anthropogenic influence on the air mass composition throughout the Park region. Benzene, toluene, ethylbenzene, xylenes, and n-hexane are the major components of two-stroke snowmobile exhaust and show large enhancements between the morning and afternoon sampling periods, with the largest increases in toluene, ethyne, and n-pentane (Figures 2 and S3, Supporting Information). Because the road from Mammoth Hot Springs to the Northeast Entrance is limited to automobiles, the mixing ratios of these gases remain essentially unchanged in this region during all sampling periods (Figures 2 and S3, Supporting Information). The impact of increased snowmobile traffic in the Park is illustrated in Table S2 (Supporting Information), which lists the mean and median mixing ratios in pptv of all toxics and percent difference in median mixing ratios between each sampling period (morning and afternoon) and location (nearroad versus off-road; median values are used for comparison rather than mean values so that the results are not skewed by a small number of samples with very high NMHC concentrations). In order to rule out the influence of urban emissions on the samples collected in the Park, the distribution of tetrachloroethene (C2Cl4), an industrial solvent, was first examined. For all samples, the mean C2Cl4 mixing ratio was 8.0 pptv ((0.8 pptv, 1σ) (Table S2, Supporting Information), indicating little or no influence on the Park air from 224

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 1, 2010

FIGURE 2. Spatial distributions of the off-road samples for toluene in YNP. Results are shown for the morning of February 12, 2003 (top left), the afternoon of February 12, 2003 (top right), the morning of February 15, 2003 (bottom left), and the afternoon of February 15, 2003 (bottom right), with February 12, 2003 and February 15, 2003 corresponding to low- and high-traffic days, respectively. urban areas directly upwind and that the VOC enhancements observed at the Park were caused by local emissions.

Large enhancements (∼100 pptv in total) in the mixing ratios of benzene, toluene, ethylbenzene, and the xylenes resulted from increased snowmobile usage between the morning and afternoon sampling periods. On the contrary, mixing ratios of most NMHCs, such as ethane, propane, and isobutane, decreased from the morning to afternoon sample collection (not shown) and is speculated to be associated with an increase in the boundary layer height. Radiosonde data suggested a stable planetary boundary layer at the sampling times on both days, as evidenced by a strong increase of potential temperature with height. Vertical mixing in the boundary layer probably did not begin until about noon because of the extremely strong inversion at surface. On the basis of the vertical profiles of potential temperature (not shown), the mixed layer thickness was 100 m (06:00) and 250 m (13:00) on February 12, 2003 and 88 m (6:00), 720 m (13:00), and 995 m (16:00) on Feburary 15, 2003 (12). This variability suggests that mixing and dilution took place with relatively clean free tropospheric air from aloft mixing down into the boundary layer. The median mixing ratios of air toxics observed in the town of West Yellowstone, MT (Figure S1b, Supporting Information) on February 16, 2003 were much higher on average than the levels observed throughout the rest of the study region on February 12 and 15, 2003 as a direct result of the widespread snowmobile usage throughout the town. Additionally, two ambient samples were collected at the Park’s West Gate Entrance Station on February 16, 2003, adjacent to the main office (under the roof enclosure); extremely high levels (∼0.46-8.71 ppmv) were measured for the air toxics associated with snowmobile exhaust (Table S3, Supporting Information). Although the time required to fill each of the 2-L sample canisters was on the order of 1 min, benzene levels may have approached the NIOSH STEL regulatory standard (1 ppmv) in both the morning and afternoon samples. It is likely that the area under the roof enclosure at the West Gate Entrance Station experienced elevated mixing ratios because of sustained traffic at the Entrance Gate throughout the day. It should be noted that Park employees working in the vicinity of this site were likely exposed to significantly higher levels of toxic compounds, which eventually can have profound health impacts. From the samples collected throughout the Park, the maximum values for the air toxics were found on the route between the West Gate Entrance and Old Faithful (Figure S1a, Supporting Information). Mixing ratios ranged from 2-5 ppmv for CO and ∼5-30 ppbv for benzene, toluene, xylenes, and ethylbenzene. While these mixing ratios are considerably lower than the regulatory standards set by OSHA or NIOSH, they are significantly higher than typical background levels on the West Coast during the same time of year and at similar latitudes as the Park (17). 3.3. Correlations of Air Toxics. The correlations of benzene, toluene, ethylbenzene, m-xylene, p-xylene, and o-xylene in YNP during the afternoon sampling periods are shown in Figure S4 (Supporting Information). Samples collected in the northern part of the Park (the road between Mammoth Hot Springs and the Northeast Entrance, Figure S1a, Supporting Information) are not included, as snowmobiles are prohibited from traveling on this road during the winter. The aromatic hydrocarbons exhibited strong correlations, with r2 ) 0.99 or 1.00, indicating that these gases came from the same or similar local sources, presumably oversnow vehicle exhaust. Aromatic hydrocarbon ratios are useful for identifying the sources of anthropogenic emissions because various sources have well-known hydrocarbon emission ratios (18). The slopes of toluene vs benzene and benzene vs ethyl-

benzene from our measurements were 2.78 and 2.15, respectively (Figure S4a,b, Supporting Information), similar to the slopes from traffic samples (2.30 for toluene vs benzene and 2.16 for benzene vs ethylbenzene) (18). Our slope value of toluene vs benzene in the afternoon samples in YNP compared to the slope for the traffic samples in Monod et al. (18) is likely the result of the high toluene levels emitted from the two-stroke snowmobile engines. Moreover, the toluene/benzene and benzene/ethylbenzene ratios were consistent with the ratios observed in two-stroke snowmobile exhaust, having values of 2.82 and 2.00, respectively. The slope values of m-xylene vs o-xylene and m-xylene vs p-xylene were 1.85 and 1.98, respectively (Figure S4c,d, Supporting Information), and also similar to those obtained in urban and traffic samples (1.81-1.91 for m-xylene vs o-xylene and 2.23-2.42 for m-xylene vs p-xylene) (18). The m-xylene/o-xylene and m-xylene/p-xylene ratios were also consistent with the ratios observed in two-stroke snowmobile exhaust, 1.95 for m-xylene/o-xylene and 1.87 for m-xylene/p-xylene. The strong correlations of the aromatic hydrocarbons effectively suggested that the air mass sampled in YNP appeared to be dominated by fresh emissions from two-stroke engine snowmobile traffic. 3.4. Photochemical Age of Air Masses Sampled in YNP. Recent and localized influence on air quality in YNP from snowmobile emissions can be demonstrated by the photochemical age of air masses. Ratios of alkyl nitrates to their parent alkanes and trace gases to benzene (particularly toluene/benzene ratios) were used to estimate the photochemical ages of air masses sampled during this study. Alkyl nitrates (RONO2) are photochemically produced from the oxidation of alkanes by OH in the presence of nitrogen oxides. As photochemical processing of an air mass occurs, it becomes more “aged” with increases in alkyl nitrate mixing ratios (19, 20). Thus, alkyl nitrates can provide key insight on the photochemical history of air masses as well as NMHC-NOx-O3 photochemistry (19, 21–23). The photochemical age of an air mass is commonly estimated from the deviation of the RONO2/RH ratio relative to a predicted photochemical production line. Additional information regarding the air mass age calculations is provided in the Supporting Information. As illustrated in Shively et al. (12), the slopes of the 2-PenONO2/n-pentane and 3-PenONO2/n-pentane versus 2-BuONO2/n-butane data are consistent with the slope of the predicted line, indicating a primarily photochemical source of the pentyl nitrates. The results from this method indicate that emissions of these gases were relatively recent with a photochemical air mass age of 2 days to ∼5 days old (12, 17). In addition to alkyl nitrate/parent alkane ratios, ratios of anthropogenic NMHCs, commonly paired in contrasting lifetimes, are also effective indicators to gauge the photochemical age of an air mass and source distributions, assuming minimal mixing during transport (25, 26). Ratios of toluene, ethylbenzene, m-xylene, p-xylene, o-xylene, and n-hexane to benzene were examined to characterize air VOL. 44, NO. 1, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

225

masses in YNP, where benzene has the longest lifetime (7.2 days) compared to the 1.3, 1.2, and ∼0.4 day lifetime of toluene, ethylbenzene, and the xylenes, respectively (http:// www.iupac-kinetic.ch.cam.ac.uk/ (27)) (Figure S5, Supporting Information). For samples collected in the Park and West Yellowstone, the higher toluene/benzene ratios coincide with larger toluene mixing ratios (Figure S5a, Supporting Information) and indicate relatively fresh emissions. The toluene/benzene ratios at West Yellowstone were consistently the highest of all locations sampled because of the large amount of continuous snowmobile traffic generating fresh exhaust emissions. Also, a significant number of the lowest toluene/ benzene ratios were found to be 0.01) (12). The ratios of alkyl nitrates/parent alkanes and aromatic hydrocarbons provided comparable results in determining the photochemical age of an air mass. The ethylbenzene/ benzene, m-xylene/benzene, p-xylene/benzene, o-xylene/ benzene, and n-hexane/benzene ratios all increased with the toluene/benzene ratio (Figure S5b-f, Supporting Information). These results, although expected, arise from the shorter atmospheric lifetimes of toluene, ethylbenzene, and the xylenes compared to that of benzene. For ethylbenzene and n-hexane, the slopes of the correlations are in agreement with those predicted by their reactions with OH. However, the slopes for the xylenes are smaller than those predicted by reactions with OH, suggesting that the emissions of xylenes are larger relative to the other aromatic hydrocarbons along the transport path of the air mass. Overall, anthropogenic NMHC ratios were generally larger at West Yellowstone than within the Park, reflecting again that West Yellowstone was subject to fresh emissions from large amounts of snowmobile traffic. The photochemical age of an air mass was estimated using eq 1

[|

[toluene] 1 ln [OH](ktoluene - kbenzene) [benzene]

t)

|

t)0

- ln

[toluene] )] ( [benzene]

(1)

where t is the photochemical age of an air mass, [OH] is the average concentration of hydroxyl radical, and ktoluene and kbenzene are the temperature-dependent rate coefficients for the reaction of OH with toluene [1.8 × 10-12 exp(340/T) cm3 molecule-1 s-1, where T is temperature in K] and benzene [2.3 × 10-12 exp(-190/T) cm3 molecule-1 s-1] (http:// www.iupac-kinetic.ch.cam.ac.uk/). The emission ratio of toluene to benzene (2.8) was derived from the exhaust sample data from the 2003 campaign. Hydroxyl radical concentration is dependent on factors such as the dew point, water vapor, and surface albedo. The daytime average OH concentration (assuming 8 h of daylight) is estimated to be 1.4 × 106 molecules cm-3 for this study based on the increase in surface albedo from the snow-covered surface and the associated increase in J(O1D) production (28, 29). Our estimates suggest a “younger” air mass age for the Park during the afternoon (19 ( 22 h) compared to the morning (34 ( 20 h) on February 12 and 15, 2003, indicating influences from fresh snowmobile emissions in the afternoon. Between the 2 days, there was more traffic on February 15, 2003 and subsequently the air mass photochemical age was 24 ( 15 and 13 ( 22 h in the morning and afternoon, respectively, compared to 43 ( 25 h in the morning and 24 ( 22 h in the afternoon on February 12, 2003. At the diurnal sampling site, the air mass was slightly more aged (54 ( 15 h in the morning and 49 ( 22 h in the afternoon of February 226

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 1, 2010

FIGURE 3. Mixing ratios of (a) benzene, toluene, and CO and (b) ethylbenzene, xylenes, and n-hexane at the Lake Ranger Station diurnal sampling site on February 15, 2003. The shaded region corresponds to the high-traffic period. 12; 27 ( 9 h in the morning and 22 ( 22 h in the afternoon of February 15) than those at other sites because of its relatively remote location and less direct influence from snowmobile traffic. In contrast to the samples collected in the Park, the median air mass photochemical age for the samples collected in West Yellowstone was 0.6 ( 14 h, demonstrating again the pronounced influence of the snowmobile traffic in town. 3.5. Estimates of Air Toxic Emission Fluxes from Snowmobiles in YNP. To quantify the regional impact of the YNP snowmobile emissions on air quality, we calculated fluxes of air toxics on the basis of diurnal samples from Lake Ranger Station, our most remote sampling site. Air toxics at the Lake Ranger Station site showed increasing mixing ratios from local noon to 4 pm on February 15, 2003 (Figure 3). During this time period, heavy snowmobile traffic occurred throughout the Park, leading to the observed mixing ratio enhancements of the air toxics. Toxics mixing ratios correlated positively with CO, a marker of fuel combustion, indicating that these gases were emitted from the same and/ or co-located sources. Using eq 2, daily and annual fluxes of the air toxics were estimated from snowmobile use in YNP. The flux of compound x (Fx, Gg yr-1) is estimated by using (i) the last (16:00 LT) (C1, ppbv or pptv) and initial sampling period mixing ratios (noon) (C0, ppbv or pptv), (ii) an averaged boundary layer height of 1000 m (H) (30), (iii) the air pressure (p, atm), (iv) the ambient air temperature (T, K), (v) the molecular weight (mwx, g/mol), and (vi) sampling time (t, h) as follows:

Fx )

(C1 - C0) × H ×

( RTp ) × mw

x

×T

t

(2)

The emission flux estimates for benzene, toluene, ethylbenzene, xylenes, and n-hexane were 5.82, 18.6, 4.1, 24.1, and 6.0 tons day-1, respectively. The emission flux estimate is assumed to represent the emission rates during high-traffic days (approximately 1200 snowmobiles). YNP reported that approximately 600 snowmobiles were run in the Park per day during the 120 day snowmobile winter season. By scaling the emission fluxes from the high-traffic day to average conditions, the annual emission fluxes for benzene, toluene, ethylbenzene, xylenes, and n-hexane are estimated to be 0.35, 1.12, 0.24, 1.45, and 0.36 Gg yr-1. The toxic emission fluxes were also obtained on the basis of the actual use of snowmobiles and the emission measurements (Table 1) made during the 2003 campaign using eq 3: F ) Etwo-strokeCtwo-strokeNtwo-stroke + Efour-strokeCfour-strokeNfour-stroke

(3)

TABLE 1. Emission Fluxes of Toxics from Snowmobiles via a Box Model and Snowmobile Usage in Yellowstone (Gg yr-1) benzene

toluene

ethylbenzene

xylenes

n-hexane

0.35

1.12

0.24

1.45

0.36

267

753

140

591

240

8

17

3

14

3

0.23

0.77

0.17

0.7

0.23

emission flux by using box model average mixing ratios in exhaust from two-stroke snowmobiles (ppm) average mixing ratios in exhaust from four-stroke snowmobiles (ppm) emission flux from snowmobile usage in Yellowstone

TABLE 2. Emission Estimates of Air Toxics from Snowmobile Usage in the U.S. (2002) and the EPA Estimates of Annual Emissions from Nonroad Vehicles for 1999 (Tg yr-1)

a

NEI (2002) snowmobile US 2002b EPA (1999)(30)

benzene

toluene

ethylbenzene

xylenes

n-hexane

0.0018 0.007-0.04 0.32

0.040 0.025-0.15 0.9

0.001 0.005-0.03 0.14

0.004 0.02-0.13 0.65

0.007-0.04 0.22

a Emission estimates of annual emissions in 2002 by snowmobiles in the NEI were based on 29 northern, western, and southwestern U.S. states (not including AK and CA). b The low limits and high limits of emission fluxes are estimated on the basis of exhaust flow rates at 1000 and 6000 rpm, respectively.

where F is the emission rate, E the exhaust flow rate (http:// www.nett.ca/tools/size/index.html), C the concentration of a gas, and N the number of snowmobiles used in the Park. As stated previously, YNP reported approximately 600 machines per day running in the Park during the 120 day snowmobile season prior to 2004; 80% of the fleet was comprised of two-stroke engines and 20% with four-stroke engines. If we assume that each snowmobile was operated approximately 8 h per day at ∼30 miles per hour, these conditions yield fluxes of 0.23, 0.77, 0.17, 0.70, and 0.23 Gg yr-1 for benzene, toluene, ethylbenzene, xylenes, and nhexane, which are comparable to the fluxes estimated using eq 2. To quantify the relative importance of snowmobile emissions, we extrapolated our YNP flux numbers to total snowmobile usage in the U.S. Snowmobile usage has been dominated by four-stroke snowmobiles in YNP since the winters of 2004 and 2005 (ref 9 and references therein), whereas most of the snowmobiles used in the rest of this country are two-stroke snowmobiles. According to the Snowmobile Organization of America (http://www. snowmobile.org) 1 652 754 snowmobiles were registered in the United States in 2002, which was used in the following calculation. However, it should be noted that this number was very likely an underestimate of the total number of snowmobiles owned and operated in reality. For example, in Montana, the total number of snowmobiles in the state was ultimately found to be underestimated by ∼65% because a large number of the snowmobiles were unregistered (31). Typically, an average snowmobile runs >990 miles at a conservative speed of 30 miles per hour (31, 32). The total snowmobile emissions were calculated on the basis of the exhaust rates of two-stroke and four-stroke engines used in snowmobiles (http://www.nett.ca/tools/size/index.html) and our emission measurements from the 2003 YNP field campaign. Using these parameters and eq 3, the snowmobile emissions in the U.S. were estimated to be 0.064-0.39 Tg yr-1 (Table 2). Table 2 also provides the annual emissions for air toxics from snowmobiles based on the EPA 2002 National Emission Inventory (NEI) (http://www.epa.gov/ttn/chief/ net/2002inventory.html) and the EPA 1999 emissions including 29 states in the north, west, and southwest (AK and CA excluded). The results from this study were significantly higher than the values from the NEI inventory. Snowmobile emissions might represent a significant fraction (∼14-21%) of air toxics with respect to the EPA estimates of annual

emissions by nonroad vehicles, although the differences between our estimates and emission inventory need to be further studied. Reducing the overall number of two-stroke snowmobiles, using clean fuel alternatives, and implementing state-of-the-art emission control strategies will likely reduce wintertime emissions of air toxics in the U.S.

Acknowledgments This research received funding from the Planning Office of the National Park Service.

Supporting Information Available Sample site locations; YNP vehicle entries; spatial distributions of off-road samples; correlations of air toxics, oversnow vehicles, and conditions for exhaust samples; mean and median mixing ratios in YNP; air toxic mixing ratios at the West Gate Entrance Station and alkyl nitrate air mass age calculation details. This information is available free of charge via the Internet at http://pubs.acs.org/.

Literature Cited (1) U.S. EPA. Air toxics from motor vehicles; EPA 400-F-92-004; U.S. EPA, Office of Mobile Sources: Ann Arbor, MI, 1994. (2) Ingersoll, G. Effects of snowmobile use on snowpack chemistry in Yellowstone National Park, 1998. Water-Resources Investigation Report 99-4148; USGS, Department of Interior: Denver, 1999. (3) Hagemann, M.; VanMouwerik, M. Potential water quality concerns related to snowmobile usage. Internal memo; National Park Service, Water Resources Division, 1999. (4) U.S. EPA. Control of emissions from nonroad large spark-ignition engines, and recreational engines (marine and land-based); final rule. Federal Regist. 2002, 67 (217), 68241-68290. (5) Morris, J. A.; Bishop, G. A.; Stedman, D. H. Real-time remote sensing of snowmobile emissions at Yellowstone National Park. An oxygenated fuel study; Western Regional Biomass Energy Program, University of Denver: Denver, 1999. (6) Bishop, G. A.; Morris, J. A.; Stedman, D. H. Snowmobile contributions to mobile source emissions in Yellowstone National Park. Environ. Sci. Technol. 2001, 35, 2874–2881. (7) National Park Service. Yellowstone National Park 2004-05 winter use plan air quality analysis of snowmobile and snowcoach emissions; Air Resource Specialists: Fort Collins, CO, 2004. (8) Southwest Research Institute. Laboratory testing of snowmobile emissions; Final Report prepared for Yellowtone National Park and Montana Department of Environmental Quality, 2002. (9) Bishop, G. A.; Burgard, D. A.; Dalton, T. R.; Stedman, D. H.; Ray, J. D. Winter motor-vehicle emissions in Yellowstone National Park. Environ. Sci. Technol. 2006, 40, 2505–2510. VOL. 44, NO. 1, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

227

(10) Bishop, G. A.; Stadtmuller, R.; Stedman, D. H. Portable emission measurements of Yellowstone Park snowcoaches and snowmobiles. Air Waste Manage. Assoc. 2009, 59, 936–942. (11) Kado, N. Y.; Kuzmicky, P. A.; Okamoto, R. A. Environmental and occupational exposure to toxic air pollutants from winter snowmobile use in Yellowstone National Park; Final Report prepared for the Yellowstone Park Foundation, Inc., the Pew Charitable Trusts, National Park Service, Montana Department of Environmental Quality, California Air Resource Board by the Department of Environmental Toxicology, University of California, Davis, 2001. (12) Shively, D. D.; Pape, B. M.; Mower, C. R.; Zhou, Y.; Russo, R.; Sive, B. C. Blowing smoke in Yellowstone: Air quality impacts of oversnow motorized recreation in Yellowstone National Park. Environ. Manage. 2008, 41 (2), 183–199. (13) Zhou, Y.; Varner, R. K.; Russo, R. S.; Wingenter, O. W.; Haase, K. B.; Talbot, R.; Sive, B. C. Coastal water source of short-lived halocarbons in New England. J. Geophys. Res. 2005, 110, D21302 (doi:10.1029/2004JD005603). (14) Zhou, Y. Atmospheric volatile organic compound measurements: Distributions and effects on air quality in coastal marine, rural and remote continental environments. Ph.D. thesis, University of New Hampshire, Durham, NH, 2006. (15) Zhou, Y.; Mao, H.; Russo, R. S.; Blake, D. R.; Wingenter, O. W.; Haase, K. B.; Ambrose, J.; Varner, R. K.; Talbot, R.; Sive, B. C. Bromoform and dibromomethane measurements in the seacoast region of New Hampshire, 2002-2004. J. Geophys. Res. 2008, 113, D08305 (doi:10.1029/2007JD009103). . (16) Blake, D. R.; Rowland, F. S. Continuing world-wide increase in tropospheric methane, 1978-1987. Science 1988, 239, 1129– 1131. (17) Blake, D. Methane, nonmethane hydrocarbons, alkyl nitrates, and chlorinated carbon compounds including 3 chlorofluorocarbons (CFC-11, CFC-12, and CFC-113) in whole-air samples. In Trends: A Compendium of Data on Global Change, Carbon Dioxide; Information Analysis Center, Oak Ridge National Laboratory: Oak Ridge, TN, 2005; http://cdiac.ornl.gov/trends/ otheratg/blake/blake.html. (18) Monod, A.; Sive, B. C.; Avino, P.; Chen, T.; Blake, D. R.; Rowland, F. S. Monoaromatic compounds in ambient air of various cities: A focus on correlations between the xylenes and ethylbenzene. Atmos. Environ. 2001, 35, 135–149. (19) Bertman, S. B.; Roberts, J. M.; Parrish, D. D.; Buhr, M. P.; Goldan, P. D.; Kuster, K. C.; Fehsenfeld, F. C.; Montzka, S. A.; Westberg, H. Evolution of alkyl nitrates with air mass age. J. Geophys. Res. 1995, 100, 22805–22813. (20) Simpson, I. J.; Blake, N. J.; Blake, D. R.; Atlas, E.; Flocke, F.; Crawford, J. H.; Fuelberg, H. E.; Kiley, C. M.; Meinardi, S.; Rowland, F. S. Photochemical production and evolution of selected C2-C5 alkyl nitrates in tropospheric air influenced by Asian outflow. J. Geophys. Res. 2003, 108, 8808 (doi:10.1029/ 2002JD002830).

228

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 1, 2010

(21) Flocke, F.; Volz-Thomas, A.; Buers, H.-J.; Patz, W.; Garthe, H.-J.; Kley, D. Long-term measurements of alkyl nitrates in southern Germany 1. General behavior and seasonal and diurnal variation. J. Geophys. Res. 1998, 103, 5729–5746. (22) Roberts, J. M.; Bertman, S. B.; Parrish, D. D.; Fehsenfeld, F. C.; Jobson, B. T.; Niki, H. Measurement of alkyl nitrates at Chebogue Point, Nova Scotia during the 1993 North Atlantic Regional Experiment (NARE) intensive. J. Geophys. Res. 1998, 103, 13569– 13580. (23) Stroud, C. A.; Roberts, J. M.; Williams, J.; Goldan, P. D.; Kuster, W. C.; Ryerson, T. B.; Sueper, D.; Parrish, D. D.; Trainer, M.; Fehsenfeld, F. C.; Flocke, F.; Schauffler, S. M.; Stroud, V. R. F.; Atlas, E. Alkyl nitrate measurements during STERAO 1996 and NARE 1997: Intercomparison and survey of results. J. Geophys. Res. 2001, 106, 23043–23053. (24) Russo, R. S.; Zhou, Y.; Haase, K. B.; Wingenter, O. W.; Frinak, E. K.; Mao, H.; Talbot, R. W.; Sive, B. C. Temporal variability, sources, and sinks of C1-C5 alkyl nitrates in Coastal New England. Atmos. Chem. Phys. Discuss. 2009, 9, 23371–23418. (25) Roberts, J. M.; Fehsenfeld, F. C.; Liu, S. C.; Bollinger, M. J.; Hahn, C.; Albritton, D. L.; Sievers, R. E. Measurements of aromatic hydrocarbon ratios and NOx concentrations in the rural troposphere: Estimates of air mass photochemical age and NOx removal rate. Atmos. Environ. 1984, 18, 2421–2432. (26) Roberts, J. M.; Hutte, R. S.; Fehsenfeld, F. C.; Albritton, D. L.; Sievers, R. E. Measurements of anthropogenic hydrocarbon concentration ratios in the rural troposphere: Discrimination between background and urban sources. Atmos. Environ. 1985, 19, 1945–1950. (27) Calvert, J. G.; Derwent, R. G.; Orlando, J. J.; Tyndall, G. S.; Wallington, T. J. The mechanisms of atmospheric oxidation of aromatic hydrocarbons; Oxford University Press: New York, 2002. (28) Blumthaler, M.; Ambach, W. Solar UVB-albedo of various surfaces. Photochem. Photobiol. 1988, 48 (1), 85–88. (29) Mao, H.; Wang, W.-C.; Liang, X.-Z.; Talbot, R. W. Global and seasonal variations of O3 and NO2 photodissociation rate coefficients. J. Geophys. Res. 2003, 108 (D7), 4216, (doi:10.1029/ 2002JD002760). (30) Blake, N. J.; Blake, D. R.; Sive, B. C.; Chen, T. Y.; Collins, J. E.; Sachse, G. W.; Anderson, B. E.; Rowland, F. S. Biomass burning emissions and vertical distribution of atmospheric methyl halides and other reduced carbon gases in the South Atlantic region. J. Geophys. Res. 1996, 101 (24), 151–24, 164. (31) Sylvester, J. T. Snowmobiling in Montana: A 1998 Update, October 1998; Bureau of Business and Economic Research at the University of Montana, 2001. (32) Eriksson, K.; Tja¨rner, D.; Marqvardsen, I.; Ja¨rvholm, B. Exposure to benzene, toluene, xylenes and total hydrocarbons among snowmobile drivers in Sweden. Chemosphere 2003, 50 (10), 1265–1397.

ES9018578