EPA's ambient air quality monitoring program provides the data needed to track air quality throughout the United States. The data gathered by the existing monitoring networks provide a major source of information for the designation of future nonattainment areas, tracking compliance, and developing resources such as emissions modeling tools, emissions inventories, and control programs. There are many questions of interest in this area, including whether the ability to monitor and measure volatile organic compounds in the ambient air exists, and if so, what are we measuring and how?
It should be noted that EPA and FHWA are far from the only organizations interested in the information that can be gained by analyzing air toxics monitoring data. In addition to local and state stakeholders, at the national level, organizations such as the Office of Homeland Security are interested in the impacts of various types of air contaminants. Though not strictly relevant to the workplan, workshop participants did encourage FHWA to consider efforts and methods of pooling technological and information resources with such organizations to provide a more complete picture of the air toxics situation. Along with shared resources, FHWA should be aware of concerns that are common to all involved in air quality monitoring, for example, the need for standard methods of measurement and reporting of air toxics.
When considering research recommendations, it is important to note that there is distinct overlap between the subdivisions presented in this report. Therefore, when a priority recommendation seems equally valid under two categories, it is presented under the category associated with the sessions under which it was developed.
One of the primary considerations regarding ambient monitoring is the availability of data. For air toxics, there is both archived historical data and several active networks providing relevant data. The archived monitoring data has information on hundreds of hazardous air pollutants from approximately 2500 sites and across 37 years. Unfortunately, the archive only maintains limited information about the monitoring methods, as well as, duplicate records and unclear minimum detection limit information.
There are three major monitoring networks of note actively providing data. The air toxics monitoring pilot city project study focused on a year long study of 18 core pollutants, including acetaldehyde, formaldehyde, 1,3-butadiene, and benzene. The Photochemical Assessment Monitoring Stations (PAMS) network focuses on areas with persistently high ozone levels. The network collects information on VOCs, several carbonyls, nitrogen oxides, ozone levels, and meteorology. Finally, the National Air Toxics Trends Stations (NATTS) network is composed of 22 sites, sited in both urban and rural locations. The NATTS network samples once every six days and measures, among other compounds, 1,3-butadiene, benzene, acrolein, arsenic, hexavalent chromium, and formaldehyde. Note that EPA is also investigating the potential of conducting community scale monitoring in conjunction with the trends sites.
In addition to monitoring networks, information can be gleaned regarding highway air toxics from at least two other sources. Tunnel studies can provide controlled information about vehicle emissions. Since the volume of traffic through and dimensions of a tunnel can be known precisely, a tunnel can represent an almost ideal laboratory to the air toxics researcher. Tunnel studies can thus be used to validate existing model forecasts, as well as to evaluate the effects of new regulations over time. On the other hand, biomonitoring, that is, the use of biological organisms to monitor air toxics, may also have some use. Namely, biological monitors may provide lower cost surrogate measures of certain highway air toxics.
It should be noted that the following discussion summary does not necessarily explicitly include all conversations had during the workshop. An attempt has been made to limit the comments to only those associated most directly with transportation-related air toxics. For instance, while discussions of air quality often lead to discussion of particulate matter, we have removed certain PM-specific comments from this summary.
Participants were interested in there being more consideration of the various uses of air quality monitoring when studies are planned and executed. In the opinion of the participants, too often the ambient monitoring studies do not include sufficient emissions-specific information. This is true of bio-monitoring studies, as well as studies to measure ambient air concentrations of toxics.
For measuring outdoor air toxic levels, it was suggested that a modest air monitoring network be designed to measure roadway-related contributions. This might include measuring air toxic concentrations versus downwind distance from the roadway. Other vehicle-related air toxic concentration measurement sites might include truck rest stops, warehouses with large truck populations, border crossings, and highway construction projects. Some monitors need to be sited so that their data can be used to assess air toxic concentration trends, while others need to have a project orientation. This would likely be a collaborative transportation-air pollution agency research effort.
Ambient monitoring needs for toxics include shorter sampling times, such as one hour averages, that allow us to distinguish long range transport/secondary formation from local, primary emissions. Pollutants for which we need shorter sampling times in urban areas include formaldehyde and acetaldehyde. We also need the ability to be able to distinguish weekday from weekend emission and concentration patterns. The ambient data collection strategy needs to give analysts the ability to understand emissions differences and the atmospheric response during these periods. In addition, we need more information from ambient monitoring studies to assist in improving our understanding of the sources of black carbon/elemental carbon. While some research indicates that diesel engines are the dominant source, other studies have indicated a significant contribution from SI engines, wood smoke, wildfires, agricultural burning, and other combustion sources.
There was general interest in using monitors to better understand air toxics concentrations near roadways, and how these air toxic concentrations and characteristics change with distance from the roadway. This would provide a base of knowledge for evaluating project-level impacts. Two questions that were of specific interest and that might be addressed in this manner were: What are the differences in ambient air toxics characteristics with distance from roadways? How can such information be used to evaluate public safety?
It was noted that most air toxic monitoring networks are designed to not be overly influenced by roadway contributions. Most monitor siting is designed for health effects studies. At the same time, there is little measurement in passenger compartments of cars and transit buses. Exposures there may be to higher air toxic concentrations than in outdoor air. For example, an Environment Canada study showed that the highest wintertime passenger compartment exposures were to windshield washer fluid solvents. As a result, more careful monitoring of roadway emissions and vehicle specific air toxics levels may be in order.
Related to ambient monitoring is local activity data collection. Traffic counts are needed at sites near monitor locations. Such information is relevant to associating air toxics levels to sources or classes of sources. For example, stations might be established in locations that facilitate learning more about the difference between light-duty and heavy-duty dominated situations.
Another approach might be to measure concentrations inside vehicles during actual operation, as well as measuring tailpipe emissions, near highway concentrations, and concentrations at various downwind distances. The results of such measurements could be used to estimate exposures to motor vehicle-emitted air toxics in different locations, and could potentially be used to validate air dispersion models. In addition, research is needed to define the conditions and driver actions that exacerbate or mitigate exposure to toxics while driving/riding.
A potentially underutilized air toxic concentration data source is the PAMS sites, which provide speciated HC concentration estimates in urban areas that are serious, severe or extreme ozone nonattainment areas for the one hour ozone NAAQS. It was suggested that these data might be better utilized for assessing vehicle contributions in the urban areas with PAMS sites. (Note that some of these data have been used for related criteria pollutant - VOC and NOx emissions - assessments.) The PAMS data have 1 to 3-hour average resolution, so they are more time resolved than many air toxic concentration estimates.
In establishing monitoring networks and instrumentation, it was noted that no perfect system is or will be available for air toxics. More work is needed to improve our ability to measure acrolein, and specific constituents of PM, for example. We need to be able to determine the level of data quality and quantity that is needed to assess air toxic concentrations, exposures and control strategies. The participants believe there is a need to identify and evaluate current technologies in order to make the desired innovations.
Participants also expressed interest in knowing whether the existing/planned set of air toxic monitors will be able to measure the changes associated with new technologies and fuels. They expressed a desire that any monitoring system provide data to help assess the impact of such changes.
Plant-based bio-monitoring of highway air toxics was discussed. Bio-monitoring deals with using biological organisms, living or dead, to assess vehicle emission levels. Bio-monitors may be able to provide information about historical changes in air toxic levels; however, it was noted that they may be limited in their ability to separate out the effects of temperature and rainfall.
A logical way to proceed might be to have an initial effort that resulted in setting monitoring objectives that serve the transportation community (and others). Investigation should be done into the specific pollutants and monitoring objectives of interest for highway air toxics research. It should be determined if monitoring data can be used to identify the spatial impact of air toxic exposure (regional versus hot spot) and if currently collected data is sufficient to validate air quality dispersion models. Guidance should be developed for the design and implementation of studies to address highway air toxics specific questions. Existing technologies should be reviewed and improved with respect to monitoring air toxics of concern to the highway community. This should include the identification of compounds not currently measurable and the investigation of monitors to measure them, as well as, the development of continuous, low maintenance monitors routine deployment by agencies.
An additional suggestion was made to begin by identifying one or more surrogates for the air toxics of interest. Then simple monitoring methods could be used to set priorities. Then, more sophisticated equipment could be used to provide a more complex analysis.
Is there a potential application for remote sensing in measuring near roadway air toxic concentrations, or a surrogate for these MSATs? This could be considered a supplement to standard ambient monitoring networks.
The participants expressed a need for some established data quality objectives for highway needs. EPA has some guidelines that could serve as a starting point, but these should not be handcuffs.
There was a general interest in the current monitor siting strategies, and if they were the same for criteria pollutants and air toxics. The general answer to this question is that the initial air toxics monitoring strategy was to identify hot spot situations (peak concentrations). Now the emphasis is on choosing representative sites.
What does the available information about health effects and exposure periods under which those effects are found tell us about what time periods are important to capture in our air monitoring efforts? Are we concerned about acute exposures to elevated levels or chronic exposures to long-term average concentrations? Monitoring strategies designed to capture peaks might be considerably different from those designed to estimate long-term average concentrations in residential neighborhoods. In addition, we need to know whether there are seasonal issues that might influence network design, and needs for short-term versus continuous monitoring. There might also be continued work to evaluate how well motor vehicle-emitted pollutants like carbon monoxide serve as a marker for specific MSATs, such as benzene, or 1,3-butadiene. Given the cost and sparseness of the existing/planned air toxics monitoring network, perhaps a more commonly monitored pollutant like CO, could serve to fill data gaps where air toxic concentrations are not monitored nearby.
Another transportation-related issue is how a transportation agency might use monitoring data. We need to evaluate what the potential uses are for project-level analyses. These might include identifying affected populations in the vicinity of roadway projects, with this affected population differing according to whether peak or long-term average exposures are of interest. For criteria pollutants, the interest is in keeping ambient concentrations below the NAAQS. For toxics, although there are no regulatory requirements at this point, the long-term goal is to reduce the level of risk to public health. Another important reason for monitoring is validation of dispersion models for air toxics.
There need to be guidelines for performing project-level air toxics studies. These guidelines should address the highway construction phase of the project as well, with recommendations for investigating and mitigating emissions from construction equipment. Thresholds need to be established so that transportation agencies know where there is likely to be a measurable difference in ambient air toxics concentrations near a project. Guidelines are needed to help identify when a project environmental impact study is warranted for toxic air pollutants.
The set of current air toxic monitoring sites were established by State and local agencies. More recently, EPA has interfaced their national program with the programs previously established by States/locals. If other agency initiatives, including those at FHWA, are going to produce more monitoring, then these new activities need to be integrated into existing objectives, and be consistent with EPA goals.
Three major ambient air toxics monitoring research areas seemed to be consistently important to experts and stakeholders in the area of transportation air toxics. Experts are interested in the development of possibly modest, possibly tiered transportation air toxics monitoring networks that address temporal and spatial distributions. Regional end-users are interested in a suite of practical tools to help define and implement policies and procedures. Among these tools, there is interest in enhanced technologies, and practical guidance and protocols. Finally, stakeholders expressed an interest in surrogates, linkages, indicators, and synergies. While accurate, specific, direct measurements are preferable, there is a desire to further define relationships and technologies to link air toxics to various continuous measurements with validation.
Proposed Programmatic InitiativesProgrammatic Initiative P4: Fund research to identify what additional monitoring information needs to be collected to enhance existing monitoring networks to meet transportation sector needs.
EPA and other national, State, regional, and local organizations sponsor monitoring networks to measure various air toxics nationwide. It is desirable for FHWA to incorporate the data from such networks into their ambient monitoring research; however, transportation needs are sufficiently unique that additional information may be desirable. Key questions might include:
Investigation of tiered network design might facilitate the collection of data that meets unmet transportation sector needs. By tiered network, we refer to a network designed to measure both spatially and temporally at various scales. One scale of the network might be distributed broadly and collect integrated measurements. A finer scale might be distributed at a narrow spatial scale and collect continuous measurements, but only be located in a small overall range of locations. Such a network might also include portable monitors to be used in specific locations for a limited time before being moved to a new location and re-used.
Estimated Cost: $250,000
Duration: 1 year
Programmatic Initiative P5: Fund research to understand existing, and develop new, practical tools for local and regional organizations to assess MSAT impacts and evaluate results.
It is important that such practical tools do not limit the ability of transportation personnel to collect data, but instead, provide them with assistance to facilitate meaningful collection.
Estimated Cost: $250,000
Duration: 1 year
Programmatic Initiative P6: Determine whether the existing knowledge about transportation air toxics can be enhanced by further examining current data.
A significant amount of data already exists from current ambient monitoring networks. However, often this data exists separately from information that helps put the data into context. It is important that more complex relationships be identified, validated, and be applied to the development of future analyses and policies. Thus, it is recommended that the relationships of individual air toxics be more carefully identified. Further, investigation into the use of measures of CO, traffic flow and other transportation characterizations, and even meteorological or other non-toxics measurements as surrogates for specific transportation related air toxics is desirable.
While accurate, specific, direct measurements are always preferred, surrogate relationships may be useful to facilitate future monitoring in place of specific technology development.
Estimated Cost: $200,000
Duration: 18 months