As stated in Section 1.1, the goal of this work effort was to address several issues:
What is the contribution of on-road mobile sources to regional air toxics problems?
What is the relative importance of on-road mobile sources to local, or hot spot, air toxic pollution problems?
What key on-road mobile source air toxics findings are available from the MATES-II study and how transferable are those findings from Los Angeles to Las Vegas?
The study also sought to identify, for the most important issues raised by the Sierra Club regarding US 95, key technical considerations that would assist FHWA in better understanding the issues raised. This section summarizes our findings on regional air toxics issues, microscale air toxics issues, PM2.5, and the transferability of MATES-II findings to Las Vegas. It closes with a summary of the overall concerns raised by the Sierra Club and technical considerations appropriate for each of the issues raised.
Emission inventories prepared by SCAQMD, CARB, and the EPA all indicate that mobile sources typically contribute 50% or more of the emissions of five air toxics: benzene, 1,3-butadiene, DPM, acetaldehyde, and formaldehyde. The EPA also identified mobile sources as an important contributor (39%) to total emissions of acrolein (which is also an aldehyde), although this pollutant was not addressed by SCAQMD or CARB. For each of these six pollutants, which we refer to as priority MSAT, both on-road and non-road mobile sources are important contributors to overall emission inventories. SCAQMD, CARB, and the EPA all project mobile source emissions to decrease in the future as a result of new vehicle technologies and cleaner fuels (in spite of increases in the number of vehicles and miles traveled), and ambient monitoring data demonstrate that concentrations of these pollutants have been decreasing over the last 10 years.
In addition to preparing an emission inventory, the SCAQMD MATES-II investigators collected monitoring data and conducted regional air quality modeling. The study noted that modeling results agreed with monitoring data within ±50% to 60%. Ambient air quality data showed relatively little spatial variability among different communities in the South Coast Air Basin, which is consistent with data analyses conducted in the San Diego area.
The MATES-II study noted that aldehydes are formed by chemical reactions in the air (secondary formation) in addition to being emitted directly (primary emissions), and attributed only 50% of the monitored concentrations to mobile sources, despite the fact that mobile sources accounted for over 90% of the direct emissions of these pollutants.
Although risk assessments depend upon the URFs assumed by each of the agencies, risk assessments conducted by SCAQMD, CARB, and the EPA all agree that the excess cancer risk due to mobile sources is a large fraction of the overall cancer risk due to toxic air pollution. It is important to note, however, that the conclusion of the SCAQMD MATES-II study that DPM concentrations contribute to approximately 1,000 excess cancers per million people (i.e., 70% of the total calculated cancer risk due to air pollution) is dependent upon two important assumptions:
The diesel PM URF: MATES-II includes use of the DPM URF developed by the California OEHHA. However, the scientific community has yet to develop a consensus view of an appropriate DPM URF. The EPA's May 2002 diesel exhaust health assessment document (U.S. Environmental Protection Agency, 2002a) does not identify a DPM URF owing to the uncertainties in the scientific literature, although OEHHA's URF falls roughly in the middle of EPA's approximate range of potential risks.
The association between measured EC and DPM: Because DPM cannot be measured directly in ambient air, MATES-II measured EC as a surrogate for DPM. Interpretation of the MATES-II DPM analysis results is complicated, however, by the fact that different EC assessment methodologies do not yield consistent results, and the proportion of EC in DPM appears to vary by engine technology and driving patterns (see Appendix A).
Models predict and field studies confirm that short-term concentrations of air toxics can be elevated for receptors downwind of and very near roadways, particularly within approximately 100 m. However, relatively little can generically be concluded about the differences in long-term concentrations as a function of distances from roadways, which depend largely upon wind speeds and wind directions. For example, the MATES-II study located two monitors within 100 m of interstates and a third monitor approximately 200 m from an interstate, but prevailing winds were not from the direction of the closest point on the roadways. MATES-II did not find a statistically significant difference between concentrations measured at these monitors and concentrations measured at other monitors further from the interstates (with the exception of acetaldehyde, which SCAQMD noted may have resulted from differences in measurement methodologies). For more persistent winds and/or monitor locations significantly closer than 100 m, higher pollutant concentrations would be expected.
Epidemiological studies of communities in the vicinity of roadways, such as the Denver study cited by the Sierra Club, have been based upon assumptions regarding exposures (rather than actual exposure measurements) and have reported mixed results. Many have also been based upon a relatively small number of subjects, limiting their statistical ability to discern cancer risks in the 10-3 to 10-6 range.
Several qualitative conclusions can be drawn from MATES-II which are likely to be applicable to Las Vegas:
Mobile sources are likely to be the major contributors to emissions and ambient concentrations of the priority MSAT in Las Vegas. The EPA's emission inventory estimates mobile sources to be the primary sources of priority MSAT. In MATES-II, measured concentrations in the ambient air were within ±50-60% of concentrations that were modeled (based upon the emission inventory).
It is reasonable to assume that the risk associated with ambient concentrations of the priority MSAT represent a large fraction of the total risk associated with regional air pollution in Las Vegas. Although the Las Vegas and Los Angeles areas are likely to have different emission inventories (Las Vegas has not prepared an MSAT emission inventory), the risks calculated by MATES-II for pollutants other than diesel particulate matter are consistent with those calculated by EPA nationally (the EPA, lacking a URF for DPM, does not estimate DPM-related risks).
Between different communities in the urbanized Las Vegas area, it is likely that there are not large spatial variations in long-term concentrations of air toxics. MATES-II found relatively uniform concentrations at the regional scale, a finding supported by other southern California research in the Barrio Logan area of San Diego (Main and Stiefer, 2001).
Within the scope of this study, it was not possible to quantitatively transfer the results from MATES-II to Las Vegas. Las Vegas has no ambient air toxics monitoring network and no air toxics emission inventory, but limited monitoring data for EC indicate that DPM concentrations may be lower in Las Vegas than in southern California. PM2.5 mass concentrations in Las Vegas are lower than in southern California, an indication that PM2.5 emission inventories and/or meteorological variables are different for the two areas.
The Sierra Club stated that short-term (24-hour average) PM2.5 concentrations adversely affect human health at levels above 15 mg/m3, although the EPA has set the 24-hour PM2.5 NAAQS at 65 mg/m3 (the EPA set the annual PM2.5 NAAQS at 15 mg/m3). A review of the large quantity of studies upon which the NAAQS were based is not within the scope of this white paper, but it is useful to note that: (1) section 109(b)(1) of the Clean Air Act requires EPA to set the NAAQS to be protective of human health with "an adequate margin of safety," and (2) to date, the standard has survived legal challenges. Research into health effects from particulate matter is continuing. Some studies suggest that ultrafine particles (PM0.1) or particle number densities (i.e., number of particles per unit volume of air, rather than mass of particles per unit volume of air) may be more relevant to human health effects than mass concentrations of PM2.5 (e.g., Zhu et al., 2002a,b); a more comprehensive review of the literature in this area would assist in further detailing emerging insights.
It is estimated that a large fraction of mobile source PM2.5 emissions are originating from heavy-duty diesel vehicles, even though these vehicles are less numerous than light-duty passenger vehicles. The EPA's emissions estimation model for on-road mobile sources, MOBILE, has been revised to incorporate methodologies for estimating PM2.5 emissions, although these algorithms are relatively new and do not yet incorporate factors that are known to affect heavy-duty diesel PM2.5 emissions. For example, exhaust PM2.5 emissions factors estimated by MOBILE are independent of speed, whereas speed/driving patterns are known to be one of the most important factors influencing heavy-duty diesel exhaust PM emissions (Clark et al., 2002).
Several ambient monitoring sites for PM2.5 are present in both Los Angeles and Las Vegas; as shown in Figure 5-2, PM2.5 data values are generally lower in Las Vegas and comply with the 24-hour PM2.5 NAAQS, although annual concentrations appear to be close to the annual PM2.5 NAAQS. Limited roadway studies have indicated that mobile sources increase PM2.5 concentrations very near roadways. However, given the known deficiencies in the ability to estimate PM2.5 emissions from heavy-duty diesels, particularly with respect to factors that influence emissions, utilization of existing modeling tools to analyze PM2.5 impacts from various transportation alternatives may not yield meaningful results.
Within the context of air toxics and PM2.5 problems, the Sierra Club and its consultants raised a variety of concerns related to US 95. Many of these have been addressed in the preceding sections. This section categorizes the issues raised by the Sierra Club under eight broad topics (see Table 6-1) and provides brief observations concerning each of the comments raised.
"Due to the evidence demonstrating the relationship between road proximity and risk, the FHWA should evaluate the potential risks associated with air toxics and fine particulate matter from completion of the US 95 highway project. The 21 mobile source air toxics identified by the EPA should be addressed in the evaluation." The Sierra Club articulated the need for the US 95 project EIS to evaluate air toxics and PM2.5 impacts from the proposed improvements. Inherent in the Sierra Club request is a need to have appropriate analytical tools to conduct air toxics and PM2.5 evaluations related to highway projects. This white paper identifies a few of the technical issues related to using existing modeling tools to address air toxics and PM2.5. The extent to which evaluations can quantitatively be conducted may be compromised by known inadequacies in the available emissions estimation and dispersion modeling tools. However, it was beyond the scope of this project to prepare a complete analysis of the existing analytical tools available. A more complete evaluation of the appropriateness of existing tools is a research need. Studies completed by the EPA, CARB, and other agencies indicate that air toxics analyses for on-road mobile sources should focus on six MSAT; these include acetaldehyde, acrolein, benzene, 1,3-butadiene, diesel particulate matter and organic gases (DPM and DEOG), and formaldehyde (see U.S. Environmental Protection Agency,2001).
"US 95 will result in higher traffic densities, which in turn will produce increased emissions and pollutant concentrations." Emissions are the product of traffic activity (e.g., miles traveled per day or per year) and emission factors (e.g., grams per mile). Although traffic activity is projected to increase in the future, emissions factors are projected to decrease. In general, most metropolitan areas project declining on-road emissions over the next 10 to 20 years, even after accounting for increased VMT (examples of this are available in most conformity analyses prepared for transportation improvement programs and regional transportation plans; an important exception includes NOx emissions). The most relevant comparison is of the future conditions with the US 95 improvements and the future conditions without them; this type of analysis is usually referred to as "build vs. no-build." The US 95 EIS listed both "future build" and "future no-build" conditions, so that alternatives could be weighed. For example, for CO, although the EIS shows that "build" scenario concentrations are often higher than "nobuild" concentrations, the highest concentration predicted corresponds to an intersection evaluated for the "no-build" scenario.
"Existing NAAQS and NAAQS-associated air quality control programs are not sufficient to provide public health protection against exposure to air toxics." NAAQS have not been established for air toxics. It is important to note that existing NAAQS control programs are resulting in air toxics emissions reductions. The EPA found in its final MSAT rulemaking (U.S. Environmental Protection Agency, 2001) that control programs to reduce emissions contributing to criteria pollutant problems have also reduced MSAT emissions. As of 2001, the EPA stated that the MSAT emission reductions benefits accruing from its other control programs were sufficient to avoid establishing additional control requirements; the EPA also committed to continued MSAT evaluations, and to periodic assessments of the need for additional controls (U.S. Environmental Protection Agency, 2001).
"There is evidence to indicate that health risk increases with proximity to roads and traffic, and especially heavy-duty diesel vehicle traffic." Real-world monitoring studies indicate that roads contribute to elevated pollutant concentrations in close proximity to the road site (especially within 100 m). Epidemiological studies have produced mixed findings; some studies find evidence of increased health risk for individuals (including children) who live near roads, other studies find no statistically significant association. The epidemiological work is often handicapped by the inability to obtain large enough sample sizes to produce statistically significant findings. Recent research suggests that the particulate fraction of diesel exhaust is the dominant contributor to urban air toxic-related health risks (e.g., SCAQMD, 2000). Additional studies (e.g., Zhu et al., 2002a, 2002b) observe that ultrafine PM may be even more toxic than larger size PM, when comparing PM of identical mass and composition. Diesel-powered vehicles produce far more PM, including PM2.5, than gasoline-powered vehicles (on a g/mi basis). The degree to which these factors produce microscale problems is a function of the number of heavy-duty vehicles operating in the affected area, the distance to receptors, and the unit risks associated with exposure to diesel exhaust. The unit risks used by the SCAQMD to estimate excess cancer risk during the MATES-II research have not been agreed to by the EPA; the uncertainty around risk from diesel exhaust is such that the MATES-II risk estimates may be an order of magnitude under or overestimated.
"The MATES-II study in southern California offers an opportunity to approximate air toxics-related health risks from roads." MATES-II was not designed to provide accurate approximations of risk as a function of proximity to roads. Monitoring data near freeways were limited to three sites, and modeling results were not finely resolved to provide concentration gradients near roads. The MATES-II monitoring results are consistent with other research indicating that pollutant concentrations are often close to or approximately the same as background conditions beyond 100 m from a road.
"MATES-II modeling analyses underestimate the pollutant concentrations likely to occur in close proximity to roads." The MATES-II modeling results likely underestimated short-term toxics concentrations within 100 m of the freeways studied. As discussed in Section 4.3, the MATES-II modeling results cannot be used to interpret long-term pollutant concentrations in close proximity to roads. Short-term data have shown that there are strong concentration gradients within 100 m downwind of roads; however, the extent to which long-term data (which reflect variations in traffic and meteorology) would show the same gradients has not been determined.
"The risks estimated by the MATES-II study exceed federal health risk guidelines for excess cancer risk." The EPA generally acts to reduce cancer risks greater than 10-4 and considers risks less than 10-6 to be acceptable. Even in remote areas of the United States, the EPA has estimated that risks from background levels of air pollution are 10-5 (Guinnup, 2003). We agree that the risks estimated by MATES-II are greater than 10-4, regardless of the value assumed for the DPM URF.
"Future conditions once the US 95 highway project is complete remain uncertain. Although congestion may decrease and speeds increase, it is unclear whether that will prove beneficial with respect to lowering air toxics emissions. In addition, there will likely be increased traffic density over time, resulting in increased emissions. The implementation of various mobile source controls may lower emissions, but those benefits may be many years distant (e.g., heavy-duty diesel vehicle standards are not scheduled to take effect until 2007)." In general, as congestion decreases and traffic flow improves, emissions decrease. Both MOBILE6, the latest version of the EPA motor vehicle emissions model, and EMFAC2002, the California motor vehicle emission factor model, include speed correction factors that generally estimate g/mi emission reductions as speeds accelerate from the 10 to 30 mph range, towards the 40 to 60 mph range. To the extent that MSAT are HC, and HC emissions are reduced with improved traffic flow, we would expect transportation projects that improve traffic flow to reduce per-vehicle MSAT emissions on a g/mi basis. Given the importance of diesel PM emissions to estimated risk, it is noteworthy that research indicates that diesel vehicle emissions are substantially affected by the variability of driving behavior, and that stop-and-go activity produces higher emissions compared to free-flow travel (Clark et al., 2002). As discussed earlier, an important consideration as traffic volumes increase over time is the build vs. no-build emissions comparison. If failure to implement the US 95 improvements results in more surface street and freeway congestion, then (assuming similar total traffic volumes in both scenarios) no-build emissions would likely be higher compared to a build scenario. The US 95 EIS provides evidence that at the worst impacted microscale sites, the build scenario lowers CO emissions. More research might assist in documenting the correlation between CO and air toxics emissions. Although future controls (e.g., heavy-duty vehicle fuel and emission standards that phase in beginning 2007) will take years to become fully implemented, existing control programs are already reducing on-road emissions. Light-duty vehicle emissions, for example, have dropped and will continue to drop over time (e.g., Eisinger et al., 2002). Already enacted heavy-duty vehicle requirements are yielding annual emission reductions as well (one study of 21 in-use heavy-duty vehicles found, for example, that PM emissions dropped from 8% to 16% per model year from 1988 to 1995; Yanowitz et al., 1999). The EPA and CARB both forecast substantial reductions in diesel PM over time (U.S. Environmental Protection Agency, 2001; California Air Resources Board, 2000).