Research

Evaluation of the public health impacts of traffic congestion: a health risk assessment

Jonathan I Levy12*, Jonathan J Buonocore2 and Katherine von Stackelberg3

Author Affiliations

1 Department of Environmental Health, Boston University School of Public Health, Boston, MA, USA

2 Department of Environmental Health, Harvard School of Public Health, Boston, MA, USA

3 Harvard Center for Risk Analysis, Boston, MA, USA

For all author emails, please log on.

Environmental Health 2010, 9:65  doi:10.1186/1476-069X-9-65

 Received: 3 June 2010 Accepted: 27 October 2010 Published: 27 October 2010

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Background

Traffic congestion is a significant issue in urban areas in the United States and around the world. Previous analyses have estimated the economic costs of congestion, related to fuel and time wasted, but few have quantified the public health impacts or determined how these impacts compare in magnitude to the economic costs. Moreover, the relative magnitudes of economic and public health impacts of congestion would be expected to vary significantly across urban areas, as a function of road infrastructure, population density, and atmospheric conditions influencing pollutant formation, but this variability has not been explored.

Methods

In this study, we evaluate the public health impacts of ambient exposures to fine particulate matter (PM2.5) concentrations associated with a business-as-usual scenario of predicted traffic congestion. We evaluate 83 individual urban areas using traffic demand models to estimate the degree of congestion in each area from 2000 to 2030. We link traffic volume and speed data with the MOBILE6 model to characterize emissions of PM2.5 and particle precursors attributable to congestion, and we use a source-receptor matrix to evaluate the impact of these emissions on ambient PM2.5 concentrations. Marginal concentration changes are related to a concentration-response function for mortality, with a value of statistical life approach used to monetize the impacts.

Comparisons between Public Health and Economic Impacts

One of our primary objectives is to compare the monetized public health damages with the economic damages associated with congestion, over time and across urban areas. Although economic damages from fuel and time wasted have been derived previously, we re-estimated these values to correspond with our VMT estimates and to provide consistency with our public health estimates. We applied algorithms from TTI [1,14] to calculate time and fuel spent in traffic at the modeled speeds and at free-flow. We calculated the difference between the time and fuel consumed at modeled speeds and at free-flow, which gave us the time and fuel wasted as a result of time spent in congestion, allowing us to compare economic and public health costs from each of these sources.

While the structure of our analyses did not allow for formal uncertainty propagation, given the numerous parameters with no plausible uncertainty quantification, we recognize that the uncertainties in our monetized public health damage estimates are significant. In our core results, we present public health damages based on central estimates for each parameter to provide a first approximation of impacts, and within limited sensitivity analyses, focus on the influence of selected parameter values on qualitative conclusions regarding the relative magnitudes of public health and economic damages.

Results

Vehicle-Miles Traveled and Degree of Congestion

In total, across the 83 urban areas modeled, VMT is projected to increase 33% from 2000 to 2030 (an increase from 2.97 billion daily VMT to 3.94 billion daily VMT), closely paralleling projected population growth in the urban areas of 32% (an increase from 133 million people to 176 million). There are some clear geographic patterns in the areas with high and low percentage changes in VMT. Multiple urban areas in the South and West are projected to have VMT growth in excess of 40%, including Raleigh NC, Oxnard CA, Richmond VA, and San Antonio TX. Those urban areas with limited VMT growth or declines are generally found in the industrial Midwest, including Cleveland OH, Dayton OH, Detroit MI, and Toledo OH-MI, but also in New Orleans LA and Salem OR. Intermediate values are generally seen in the largest urban areas, with 7-10% projected growth in New York-Newark, Los Angeles, and Chicago. A complete table of VMT estimates by urban area is provided in Additional file 1, Table S4.

Trajectories of the degree of congestion by urban area closely pattern trajectories of VMT, although not in a linear fashion given the maximum value of 50% of time in congestion in the traffic model and other non-linearities in the system. Whereas 18 of the 83 urban areas were estimated to have 50% of time in congestion in 2000, 40 urban areas reached this threshold by 2030, given our baseline scenario of population growth with no corresponding changes in infrastructure or policy.

For 2005, nationwide estimates of traffic emissions attributable to time spent in congestion include approximately 1.2 million tons of NOx, 34,000 tons of SO2, and 23,000 tons of PM2.5. These emissions are associated with approximately 3,000 premature deaths in 2005 (Figure 1), with an economic valuation of $24 billion (in 2007 dollars). Overall, approximately 48% of the impact over the 83 urban areas is attributable to NOx emissions, with 42% attributable to primary PM2.5 emissions and 11% attributable to SO2 emissions. However, the relative proportion of the impact attributable to different pollutants varies significantly across urban areas. For example, the proportion due to NOx ranges from 6% in multiple Northeast cities (Hartford, CT; Boston, MA; New Haven, CT; Springfield, MA) to over 70% in less densely populated areas of Texas (Brownsville, Austin) and Washington State (Spokane). Similarly, the proportion due to primary PM2.5 emissions is highest in densely-populated urban areas of the Northeast (approximately 80%) and below 20% in Brownsville. The proportion attributable to SO2 emissions is highest in California, with four urban areas in California constituting the only places with more than 20% of the mortality risk from SO2 emissions. These relative proportions are attributable in part to high ambient sulfate in the eastern US, which tends to reduce particulate nitrate formation, and to atmospheric conditions in California favoring the secondary formation of particulate sulfates. Figure 1. PM2.5-related premature mortality attributable to traffic congestion nationwide. Pollutants in legend indicate emitted pollutants, and the figure indicates the effects of these emissions on PM2.5 concentrations. Examining time trends in mortality risk attributable to congestion (Figure 1), there is a steady decline from 2000 until 2020, with an estimated 4,000 premature deaths in 2000 and an estimated 1,600 premature deaths in 2020. This decline is driven by significant reductions in projected per-vehicle emissions that offset the growth in population and general increase in per-capita mortality rates (based on the aging of the population). However, after this point, there is a steady increase in mortality risk (to 1,900 premature deaths in 2030), as population growth continues but projected pre-vehicle emissions reductions level off. Using economic values assigned to premature mortality in each year, with all values in 2007 dollars, the impact decreases to approximately$13 billion in 2020, after which point it increases up to $17 billion by 2030. Changes occur in the relative contribution of various particle constituents, given that NOx emissions per vehicle-mile are projected to decrease by 83% on arterial streets and 86% on freeways between 2005 and 2030, as opposed to a 64% decrease in primary PM2.5 emissions on both road types and a 24% decrease in SO2 emissions on both road types. By 2030, the relative contributions of the different pollutants to mortality risk attributable to congestion have changed to 24% attributable to NOx emissions, 27% attributable to SO2 emissions, and 49% attributable to primary PM2.5 emissions. Figure 2 presents the monetized health impacts over time for selected urban areas, illustrating that the trajectories differ as a function of differential population growth, congestion, population density and atmospheric chemistry. For example, monetized health impacts increase steadily over time in cities such as Raleigh NC and San Diego CA, in which VMT and population growth are significant and primary PM2.5 makes a substantial contribution to health risk. In contrast, Chicago and other cities in the Midwest are projected to have small VMT growth and have more substantial contributions to public health damages from NOx emissions, and therefore show a steady decline in health risks over time given the larger decline in NOx emissions per vehicle-mile. Figure 2. Monetized estimates of the PM2.5-related mortality risks from traffic congestion in selected urban areas. Pollutants in legend indicate emitted pollutants, and the figure indicates the effects of these emissions on PM2.5 concentrations. Comparison of Economic Impacts and Public Health Impacts Figure 3 presents the economic costs from time and fuel wasted and monetized estimates of premature mortality attributable to traffic congestion across the 83 urban areas. Overall, time wasted accounts for the bulk of the economic cost associated with traffic congestion, and the cost of time wasted increases from$56 billion in 2000 to $96 billion in 2030, as this is directly proportional to the extent of congestion. In contrast, reductions in per-vehicle emissions contribute to declines in economic costs associated with premature mortality from$31 billion in 2000 to $13 billion in 2020, with modest increases after that point to$17 billion in 2030.

Figure 3. Comparison of the economic costs of congestion with the monetized estimates of PM2.5-related mortality risks (in billions of 2007 dollars).

As a result, whereas the public health impacts contributed approximately 34% of the total cost of congestion in 2000, this decreases to 14% by 2030. However, the proportion of the cost of congestion attributable to premature mortality varies substantially across urban areas. For example, in 2000, 17 urban areas had health impacts contributing less than 20% of the total cost of congestion, whereas 19 urban areas had contributions in excess of 50%. Those urban areas with relatively small contributions from public health had very high levels of congestion (near or at the 50% threshold) but did not have correspondingly high population density, including Laredo TX, Eugene OR, and Las Vegas NV. In contrast, those urban areas where public health impacts dominated had smaller percentage of time spent in congestion but greater public health benefits per ton of emissions.

Sensitivity Analyses

Our findings should be considered quite uncertain and are clearly sensitive to a number of the model assumptions, many of which could not be formally quantified or propagated. However, we note that the relative economic values assigned to the cost of fuel, the value of time per hour, and the value of statistical life, as well as future fuel efficiency, are key assumptions that could greatly influence our conclusions. In addition, we note that the magnitude of the public health damages is sensitive to the concentration-response function applied. Because our model results scale linearly with these values, we can provide some first-order approximations of the relative values that would be needed to alter our conclusions about the significance of public health relative to economic impacts, providing a sense of the likelihood of significantly different conclusions by year.

For example, in 2000, the price of gasoline is known, so the estimate of the cost of fuel wasted is relatively less uncertain. The cost of delay depends on the value of time, which we derived from TTI models using a non-commercial value of time (approximately $16/hr in year 2000 dollars), which is likely a lower bound on the true value of time. The VSL and the concentration-response function represent the most significant quantifiable uncertainties for public health damages, so we can explore the combinations of values needed for public health damages to exceed economic damages. Using the central estimate for the concentration-response function, the VSL would need to be approximately$14 million in 2007 dollars for the public health damages to exceed the value of time wasted in 2000, or roughly the 95th percentile of the uncertainty distribution for VSL reported by US EPA [32]. Using an upper-bound concentration-response function of 2% would lead the public health damages with a central estimate VSL to be comparable to the economic damages, whereas a lower-bound concentration-response function of 0.3% would require an extremely large VSL for the magnitudes to be comparable. By 2005, the VSL would need to have been $18.5 million in 2007 dollars for public health damages to exceed the value of time using the central estimate concentration-response function, which is outside of the reported confidence intervals. When considering future years, uncertainties exist regarding both the value of time and the cost of fuel, with potentially large uncertainties in the latter case. However, across all model years, the cost of fuel contributes an order of magnitude less than the value of time, indicating that the cost of fuel would need to increase appreciably relative to income in order to strongly influence our conclusions (and such a change in the cost of fuel would appreciably influence driving patterns and resulting congestion). As the contribution from public health damages decreases relative to the value of time in future years, the necessary VSL and concentration-response function to equalize these damage estimates becomes quite high, although significant uncertainties regarding the time trajectory of emissions and effect of changing background concentrations on pollutant formation add to uncertainties in this comparison, and public health damages are the dominant contributors in selected urban areas even in 2030. Discussion Our modeling illustrates that the public health impacts of traffic during periods of congestion, associated with premature mortality from primary and secondary PM2.5 concentrations, are appreciable, with thousands of deaths per year and a monetized value of tens of billions of dollars per year. While the monetized public health damages are smaller than the economic value of time wasted, with the differential anticipated to grow over time, there are some geographic areas where public health damages represent a significant proportion of the total damages, even in future years when per-vehicle emissions are expected to be substantially less. Prior analyses of population exposure per unit emissions from motor vehicles [13] demonstrated that these values were highest in dense urban areas for primary PM2.5 and secondary sulfate, especially in California, the mid-Atlantic states, and the industrial Midwest, and were highest in the Southeast and Midwest for secondary nitrate. The urban areas with the greatest proportion of damages from public health were often found in parts of California and the Midwest, where the damages per ton of emissions were greater and the projected future population growth was lower. These findings provide an indication that considering only the direct economic costs of congestion will underestimate societal benefits of mitigating congestion, significantly so in certain urban areas. There are some clear limitations in our models and their interpretation, which provide some potential guidance for future studies. First, our results assume that the infrastructure is fixed at 2005 levels but that population continues to increase out to 2030. This is relatively unlikely to be the case, but given our modeling framework and available information, it is difficult to evaluate future changes in infrastructure in a manner that is not misleading. Increasing the available infrastructure (whether through roadway capacity or public transportation) increases average speed. Because of limitations in the MOBILE6 model, which assumes an average speed and does not incorporate stop-and-go traffic or fast acceleration and deceleration, increased speed leads to higher MOBILE6 emissions estimates, particularly for NOx. In contrast, Frey et al. [33] illustrated that when comparing traveling conditions of congestion and free-flowing traffic in which the estimated average speed is similar, emissions during congested driving conditions are 50% higher. Other studies have shown that NOx emissions from heavy-duty diesel vehicles are three times higher at low speeds than when at highway speed [34], and PM, elemental carbon (EC) and organic carbon (OC), and air toxics emissions are heightened during start-and-stop traffic [35,36], but these insights are not captured in our emissions model. We therefore could only characterize the public health impacts of congestion as the total impacts during periods of congestion, rather than the marginal difference between these impacts and the impacts if the infrastructure were expanded, a potentially significant limitation. Future studies should explore other emissions modeling frameworks that better capture the effects of congestion on emissions, linked to more detailed traffic flow models. For example, micro-simulation techniques have been developed to evaluate the impact of improved traffic flow on vehicle emissions [37], capturing both the near-term decrease in emissions and the long-term effect of induced vehicle traffic. However, this model does not capture all pollutants or provide insight about the effects of changes in technology over time. The MOVES2010 emissions model [38] became available subsequent to our analysis, which provides some insight regarding start-and-stop traffic, and this modeling framework could be explored in future studies. There are further limitations with MOBILE6 that should be acknowledged, including difficulties in addressing malfunctioning vehicles and in projecting fleet composition over an extended period of time. For comparability with other work and lacking other quantitative values, we used default MOBILE6 values where possible, including for fleet mixtures (by age and class), mileage accumulation, inspection and maintenance programs, vehicle fuel economy and emissions performance, starting emissions, reformulated gasoline and oxygenated fuels programs, and low-sulfur diesel regulations, among others. To the extent that there are significant regulatory or technology changes over time, which would influence per-vehicle emissions as well as fleet composition, our analyses would become progressively more uncertain over time. In an extreme case, given significant penetration of electric vehicles (which were not formally considered in this analysis), the impact of congestion would need to consider emissions from electricity production, which was well beyond the scope of this analysis. Regardless, this emphasizes that there are some large uncertainties in emissions characterization, which could not be quantified in our sensitivity analyses but which should be considered in formulating policy strategies. A related limitation has to do with a difference in how congestion is treated within the economic and public health impact models. Our estimates of time and fuel wasted are derived from the difference between modeled speed and free-flow speed on arterial roads and freeways (assumed to be the posted speed limit), while the health impact estimates are based on the roadway congestion index, a piecewise linear function providing an estimate of the proportion of the daily traffic that occurs in congestion, which we then use to attribute emissions to congestion. Because changes in speed are only predicted once a threshold relative congestion index is reached, time and fuel wasted therefore have a congestion threshold below which there are no predicted impacts, although the roadway congestion index predicts emissions attributable to congestion at all levels. As a result, some areas will show public health impacts but little to no economic damages from wasted time and fuel. These situations should be interpreted as largely artifacts of the modeling structure, and a more refined characterization of traffic flows and infrastructure-based congestion would likely eliminate these predictions in future assessments. Additionally, the congestion model treats the entire urban area as a whole, so the model does not explicitly distinguish between arterial and freeway traffic conditions the way that the speed model does. Both this issue and the previous issue regarding infrastructure emphasize the potential need for more refined assessments in targeted geographic areas, in which modeling can go beyond what is feasible when looking at 83 urban areas concurrently. That said, modeling 83 urban areas concurrently provides first-order insight about geographic areas necessitating further study, allowing for targeted future studies. An additional limitation arises in our focus solely on the mortality risks of PM2.5. There are clearly numerous other health endpoints or pollutants that may contribute to the public health burden of congestion, including morbidity endpoints associated with PM2.5, mortality and morbidity from ozone, and effects of multiple air toxics. As described earlier, previous studies have demonstrated that PM2.5-related mortality dominates monetized public health damages, but if lower concentration-response functions are applied for mortality or lower economic values are applied to premature mortality (e.g., using a life-year approach), the significance of PM2.5-related morbidity would be enhanced. In addition, in years and urban areas where the contributions from NOx emissions are large relative to the contributions from SO2 or primary PM2.5 emissions, inclusion of ozone concentrations may be warranted. We also omitted potential impacts of ultrafine particulate matter, which would be challenging to quantify given limitations with emissions inventories, dispersion models, and available concentration-response functions, but which could exhibit significant health effects in the near field and may merit future consideration. One additional pollution-related endpoint of interest would be carbon dioxide emissions and corresponding effects of global climate change. While estimating this pathway is highly uncertain and was beyond the scope of our analysis, we can make a first-order approximation to determine the likelihood that this would contribute appreciably to damages. In 2005, we estimated approximately 3.12 billion daily VMT across the 83 urban areas during periods of congestion. According to a recent report [39], gasoline contributes climate-related damages on the order of 0.06 to 6 cents/VMT, which would correspond with between$680 million and $68 billion in damages, bounding the PM-related public health damages of$24 billion in 2005. While both categories are quite uncertain, this indicates that modeling global climate impacts clearly merits further consideration.

In addition, while we have included the effects of emissions on ambient concentrations, we have not captured the effects on in-vehicle concentrations and personal exposures. As drivers and passengers spend more time in traffic, their in-vehicle exposures to traffic-related pollutants will increase due to time-activity changes as well as increases in roadway concentrations, which can sharply increase personal exposures and doses. For example, Alm et al. [40] found that added exposures (due to commuting in addition to background levels) to carbon monoxide and PM2.5 were considerably higher during morning commutes and higher under slower average speed conditions than in faster driving conditions. Riediker et al. [41] measured PM and volatile organic compounds in vehicles at various times and found that elevated contaminant levels were related to locations with high traffic volumes. These and other studies suggest that in-vehicle exposures may significantly contribute to personal exposures, particularly during times of congestion, and these would add to the estimates of health impacts developed here. We could not quantify this pathway, given challenges in linking epidemiological evidence from central site monitors with personal exposures during commuting, as well as because of complications related to differences between ambient and in-vehicle concentrations, but we anticipate that this pathway would be appreciable enough to merit consideration in future analyses.

Although these and other factors are significant limitations, it is important to note that more advanced modeling approaches would be unable to characterize 83 urban areas individually and simultaneously, and would therefore lack insight on relative differences among urban areas. Moreover, our VMT estimates and resulting economic costs of congestion are quite similar to those previously determined [1], and our estimated public health damages per ton of emissions compare quite favorably to those estimated previously in spite of the simplified atmospheric model applied. For example, a recent study [42] used a response surface model derived from CMAQ runs to determine health impact estimates for nine urban areas in 2015. The estimated damage values for mobile sources were $550,000 per ton of primary PM2.5 and$9,700 per ton of NOx, with $73,000 per ton of SO2 for area sources (no value was reported for mobile sources). The corresponding values from our study (for all 83 urban areas) for 2015 are approximately$530,000 per ton of primary PM2.5, $11,600 per ton of NOx, and$100,000 per ton of SO2.

More generally, though capturing the congestion-related contribution from traffic emissions was challenging given available models, this framework was necessary to provide direct comparisons with the economic impacts of congestion. This would also allow for future evaluations of policy measures addressing congestion, including both infrastructure enhancements and other measures to reduce congestion (e.g., congestion pricing and traffic management). Thus, despite the limitations and somewhat simplified modeling approaches in our study, our findings can be used to inform both public policy and future investigations of the economic consequences of traffic congestion.

Conclusions

Our model results show that health impacts (based on estimates of premature mortality following exposure to primary and secondary PM2.5) are likely significant enough to necessitate inclusion in a comprehensive evaluation of the benefits of measures to reduce congestion. While health impacts are projected to decrease relative to economic impacts, given reductions in per-vehicle emission rates, there remain multiple urban areas in which the public health impacts are still appreciable in 2030. We recommend that future investigations develop refined estimates of emissions in congestion and linkages with personal exposures, which could make a significant contribution to the estimated public health burdens of congestion in multiple urban areas.

List of abbreviations

CMAQ: Community Multiscale Air Quality model; CUTR: Center for Urban Transportation Research; EC: elemental carbon; EPA: Environmental Protection Agency; ESPN: EPA Speciation Network; FRM: Federal Reference Method; GDP: gross domestic product; NEI: National Emissions Inventory; NOx: nitrogen oxides; OC: organic carbon; PM2.5: fine particulate matter; S-R: source-receptor; SO2: sulfur dioxide; TTI: Texas Transportation Institute; US: United States; VMT: vehicle miles traveled; VSL: value of statistical life

Competing interests

The authors declare that they have no competing interests.

Authors' contributions

JIL participated in study design, contributed to the concentration modeling and health impact estimates, and helped to draft the manuscript. JJB carried out the modeling of traffic, congestion, and emissions and led all analytical efforts. KVS conceived of the study, participated in its design and coordination, and helped to draft the manuscript. All authors read and approved the final manuscript.

Acknowledgements

Funding was provided by the American Road and Transportation Builders Association (ARTBA). ARTBA was involved in suggesting the topic for research but played no role in data collection, data analysis, or manuscript writing, and was not involved in the manuscript submission. We would like to thank Steve Polzin from the Center for Urban Transportation Research at the University of Southern Florida for sending us a copy of the spreadsheet model for the prediction of traffic demand, and Dr. Tom Smith of the Harvard School of Public Health for many helpful planning discussions and comments on earlier drafts.

References

1. Schrank D, Lomax T: The 2007 Urban Mobility Report. [http://mobility.tamu.edu]

Texas Transportation Institute; 2007.

2. Victoria Transport Policy Institute: Transportation Cost and Benefit Analysis: Techniques, Estimates and Implications. Victoria, BC, Canada; 2009.

3. Rizzo MJ, Scheff PA: Fine particulate source apportionment using data from the US EPA speciation trends network in Chicago, Illinois: Comparison of two source apportionment models.

Atmos Environ 2007, 41:6276-6288. Publisher Full Text

4. Hammond DM, Dvonch JT, Keeler GJ, Parker EA, Kamal AS, Barres JA, Yip FY, Brakefield-Caldwell W: Sources of ambient fine particulate matter at two community sites in Detroit, Michigan.

Atmos Environ 2008, 42:720-732. Publisher Full Text

5. Lee S, Liu W, Wang YH, Russell AG, Edgerton ES: Source apportionment of PM2.5: Comparing PMF and CMB results for four ambient monitoring sites in the southeastern United States.

Atmos Environ 2008, 42:4126-4137. Publisher Full Text

6. Pope CA, Burnett RT, Thun MJ, Calle EE, Krewski D, Ito K, Thurston GD: Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution.

JAMA 2002, 287:1132-1141. PubMed Abstract | Publisher Full Text

7. Laden F, Schwartz J, Speizer FE, Dockery DW: Reduction in fine particulate air pollution and mortality: Extended follow-up of the Harvard Six Cities study.

Am J Respir Crit Care Med 2006, 173:667-672. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

8. Schwartz J, Coull B, Laden F, Ryan L: The effect of dose and timing of dose on the association between airborne particles and survival.

Environ Health Perspect 2008, 116:64-69. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

9. Muller NZ, Mendelsohn R: Measuring the damages of air pollution in the United States.

J Environ Econ Manage 2007, 54:1-14. Publisher Full Text

10. HEI Panel on the Health Effects of Traffic-Related Air Pollution: Traffic-Related Air Pollution: A Critical Review of the Literature on Emissions, Exposure, and Health Effects. Boston, MA: Health Effects Institute; 2009.

11. Homburger W, Kell J, Perkins D: Fundamentals of Traffic Engineering. 13th edition. University of California, Berkeley; 1992.

12. Levy JI, Wolff SK, Evans JS: A regression-based approach for estimating primary and secondary particulate matter intake fractions.

Risk Anal 2002, 22:895-904. PubMed Abstract | Publisher Full Text

13. Greco SL, Wilson AM, Spengler JD, Levy JI: Spatial patterns of mobile source particulate matter emissions-to-exposure relationships across the United States.

Atmos Environ 2007, 41:1011-1025. Publisher Full Text

14. Polzin S, Chu X, McGuckin N: Long-range personal VMT forecasting scenario tool. Center for Urban Transportation Research (CUTR); 2007.

15. Woods and Poole: Complete Economic and Demographic Data Source, 2008. [http://www.woodsandpoole.com]

16. Noland RB, Lem LL: A review of the evidence for induced travel and changes in transportation and environmental policy in the US and the UK.

Transport Res Part D 2002, 7:1-26. Publisher Full Text

17. Noland RB, Cowart WA: Analysis of metropolitan highway capacity and the growth in vehicle miles of travel.

Transport 2000, 27:363-390. Publisher Full Text

18. Noland RB: Relationships between highway capacity and induced vehicle travel.

Transport Res Part A 2001, 35:47-72.

19. Graham DJ, Glaister S: Road traffic demand elasticity estimates: A review.

Transport Rev 2004, 24:261-274. Publisher Full Text

20. Florida Department of Transportation: Florida Traffic Monitoring Handbook. [http://www.dot.state.fl.us/planning/statistics/tmh/ch4.pdf]

2007.

21. US Environmental Protection Agency: MOBILE6 Vehicle Emissions Modeling Software. [http://www.epa.gov/oms/m6.htm]

22. US Environmental Protection Agency: Regulatory Impact Analysis - Control of Air Pollution from New Motor Vehicles: Tier 2 Motor Vehicle Emissions Standards and Gasoline Sulfur Control Requirements. Washington, DC: Office of Air and Radiation; 1999.

23. Abt Associates: User's Manual for the National Co-Benefits Risk Assessment Model, Beta Version 2.0. Bethesda, MD: US EPA State and Local Capacity Building Branch; 2004.

24. Levy JI, Wilson AM, Zwack LM: Quantifying the efficiency and equity implications of power plant air pollution control strategies in the United States.

Environ Health Perspect 2007, 115:743-750. PubMed Abstract | Publisher Full Text | PubMed Central Full Text

25. Levy JI, Baxter LK, Schwartz J: Uncertainty and variability in health-related damages from coal-fired power plants in the United States.

Risk Anal 2009, 29:1000-1014. PubMed Abstract | Publisher Full Text

26. Levy JI, Wilson AM, Evans JS, Spengler JD: Estimation of primary and secondary particulate matter intake fractions for power plants in Georgia.

Environ Sci Technol 2003, 37:5528-5536. PubMed Abstract | Publisher Full Text

27. US Environmental Protection Agency: Regulatory Impact Analysis for the Final Clean Air Interstate Rule. Washington, DC: Office of Air and Radiation; 2005.

28. Levy JI, Greco SL, Melly SJ, Mukhi N: Evaluating efficiency-equality tradeoffs for mobile source control strategies in an urban area.

Risk Anal 2009, 29:34-47. PubMed Abstract | Publisher Full Text

29. Krewski D, Jerrett M, Burnett RT, Ma R, Hughes E, Shi Y, Turner MC, Pope CA III, Thurston G, Calle EE, Thun MJ, Beckerman B, DeLuca P, Finkelstein N, Ito K, Moore DK, Newbold KB, Ramsay T, Ross Z, Shin H, Tempalski B: Extended follow-up and spatial analysis of the American Cancer Society study linking particulate air pollution and mortality.

Res Rep Health Eff Inst 2009, 140:5-114. PubMed Abstract

30. Jerrett M, Burnett RT, Ma R, Pope CA III, Krewski D, Newbold KB, Thurston G, Shi Y, Finkelstein N, Calle EE, Thun MJ: Spatial analysis of air pollution and mortality in Los Angeles.

Epidemiol 2005, 16:727-736. Publisher Full Text

31. Industrial Economics: Expanded expert judgment assessment of the concentration-response relationship between PM2.5 exposure and mortality. Cambridge, MA: Prepared for Office of Air Quality Planning and Standards, US Environmental Protection Agency; 2006.

32. US Environmental Protection Agency: 2006 National Ambient Air Quality Standards for Particle Pollution. Washington, DC: US Environmental Protection Agency; 2006.

33. Frey C, Rouphail N, Unal A, Colyar J: Emissions reduction through better traffic management: An empirical evaluation based upon on-road measurements. US Department of Transportation and the North Carolina Department of Transportation; 2001.

34. Shah SD, Johnson KC, Miller JW, Cocker DRI: Emission rates of regulated pollutants from on-road heavy-duty diesel vehicles.

Atmos Environ 2006, 40:147-153. Publisher Full Text

35. Shah SD, Cocker DRI, Miller JW, Norbeck JM: Emission rates of particulate matter and elemental and organic carbon from in-use diesel engines.

Environ Sci Technol 2004, 38:2544-2550. PubMed Abstract | Publisher Full Text

36. Cocker DRI, Shah SD, Johnson KC, Zhu X, Miller JW, Norbeck JM: Development and application of a mobile laboratory for measuring emissions from diesel engines. 2. Sampling for toxics and particulate matter.

Environ Sci Technol 2004, 38:6809-6816. PubMed Abstract | Publisher Full Text

37. Noland RB, Quddus MA: Flow improvements and vehicle emissions: Effects of trip generation and emission control technology.

Transport Res Part D 2006, 11:1-14. Publisher Full Text

38. US Environmental Protection Agency: MOVES: Motor Vehicle Emission Simulator. [http://www.epa.gov/otaq/models/moves/index.htm]

39. Committee on Health Environmental and Other External Costs and Benefits of Energy Production and Consumption: Hidden Costs of Energy: Unpriced Consequences of Energy Production and Use. Washington, DC: National Academies Press; 2010.

40. Alm S, Jantunen MJ, Vartiainen M: Urban commuter exposure to particle matter and carbon monoxide inside an automobile.

J Expo Anal Environ Epidemiol 1999, 9:237-244. PubMed Abstract | Publisher Full Text

41. Riediker M, Williams R, Devlin R, Griggs T, Bromberg P: Exposure to particulate matter, volatile organic compounds, and other air pollutants inside patrol cars.

Environ Sci Technol 2003, 37:2084-2093. PubMed Abstract | Publisher Full Text

42. Fann N, Fulcher CM, Hubbell BJ: The influence of location, source, and emission type in estimates of the human health benefits of reducing a ton of air pollution.

Air Qual Atmos Health 2009, 2:169-176. PubMed Abstract | Publisher Full Text | PubMed Central Full Text