• Air Toxics
• CMAQ
• Conformity
• It All Adds Up

• Laws and Regulations
• Policy and Guidance
• Research and Analysis
• Training

Recent Examinations of Mobile Source Air Toxics

A Methodology for Evaluating Mobile Source Air Toxic Emissions Among Transportation Project Alternatives

PRESENTATION OF RESULTS

Project-Level Emissions

Traffic data representative of a congestion-mitigation highway project was formulated; the most-recent official version of the MOBILE6.2 model (dated November 2003) was run; and the resulting emission inventory for the six priority mobile source air toxics was compiled as summarized in Table 2. The relative amount of total MSAT emissions attributable to the freeway and the crossing arterials and the emissions of each individual MSAT are also provided in Table 2.

Variability of MOBILE6.2 Emission Factors

MSAT emission factor predictions produced by the MOBILE6.2 model vary substantially over the typical evaluation period of a transportation project. The MSAT emission levels attributable to a group of transportation alternatives will be governed by project-specific and locale-specific circumstances such as external conditions, vehicle fleet characteristics, vehicle activity, and vehicle fuel specifications. Recognizing the importance of these factors, the EPA (2004) recommends which MOBILE6.2 input parameters should be based on locally-derived data in preparing emission inventories. Tianjia Tang (2003) of the FHWA Resource Center conducted an in-depth trend and sensitivity analysis of the air toxic function of the MOBILE6.2 model.

Table 2. Project-Level MSAT Emissions (tons per year).

Offered as a supplement is an evaluation of the range of MSAT emission factors obtained from the MOBILE6.2 model by varying the key parameters identified by Tang -- calendar year, temperature, fuel Reid vapor pressure, and vehicle speed. Calendar years ranging from 2005 to 2050 in 5 year increments were considered. No temperature variation over the day simulated was assumed; however, various temperatures were assessed where the minimum was set to the maximum to gauge the effects of discrete temperatures, ranging from 35 ºF to 95 ºF in 10 ºF increments. The effects of fuel RVP were calculated for four selected values: 7.8 psi (Class AA), 10.0 psi (Class B), 12.5 psi (Class C/D), and 15.0 psi (Class E). The effects of vehicle speeds for the freeway and arterial/collector roadway scenarios were evaluated in increments defined by the 14 speed bins established in MOBILE6.2 (i.e., 2.5 mph, 5 mph, and in 5 mph increments thereafter to 65 mph). This represents over 47,000 emission factor values calculated for the 6 priority MSAT compounds. The results are shown in Figure 2.

Figure 2. Variability of MOBILE6.2 Emission Factors for the Priority MSATs

DISCUSSION OF RESULTS

Project-Level Emissions

The analysis indicates that a significant decrease in MSAT emissions can be expected for a planned congestion-mitigation transportation project from current (e.g., 2005) levels through future (e.g., 2030 design year) levels. The emission trends obtained in this analysis are illustrated in Figure 3. Emissions of total MSATs are predicted to decrease by more than 56% in 2030 compared with 2005 levels. Differences in total MSAT emissions between the Build and No-Action alternates were found. The No-Action Alternate is expected to carry less traffic than the Build Alternates, but this is offset by an over-capacity traffic condition and breakdown of travel speeds during an extended peak period. As a result, less total MSAT emissions are associated with the Build Alternates compared to the No-Action Alternate (i.e., 2.2 to 6.2% less).

Figure 3. Projected Changes in MSAT Emissions

Most of the MSAT emissions are attributable to the freeway compared with the crossing arterials. But although the arterials account for only 11.1 to 11.4% of the total daily VMT; they contribute a disparate amount of the total MSAT emissions -- 12.3 to 14.4%. While this may not seem significant, arterial travel accounts for 7.9 to 29.6% more MSAT emissions than indicated by its portion of the total VMT. This is attributable to the congestion reflected for the arterial facilities represented in the analysis. Since emissions are directly proportional to the assumed daily VMT, the results by facility type presented in Table may be adjusted to reflect different assumptions with respect to the length of the freeway corridor or crossing arterials and to a lesser extent, the AADT on the two facility types. Changing traffic volume assumptions would affect traffic speeds and emission factors that would not be reflected in a simple proportional adjustment of facility VMT.

Of the six priority MSAT compounds, benzene contributes the most to the emissions total. The amount of diesel particulate matter emitted in 2005 is comparable to the amount of benzene emitted -- 21.0 versus 25.6 tpy for the 6-lane No-Action Alternate. But for future years, a substantial decline in benzene is anticipated (more than a 41% decrease from 2005 to 2030); and an even larger reduction in diesel particulate matter emissions is predicted (about an 88% decrease from 2005 to 2030). In 2005, the largest portion of the total MSAT emissions is due to benzene (40%) followed by diesel particulate matter (33%), formaldehyde (15%), acetaldehyde (5%), butadiene (5%), and acrolein (1%). By 2030, the rank order and proportions change as follows: benzene (54%), formaldehyde (21%), diesel particulate matter (9%), acetaldehyde (8%), butadiene (7%), and acrolein (1%). This is displayed graphically in Figure 4.

Figure 4. Changes in the Proportional Mix of MSAT Emissions

Associated with each MSAT compound is its inherent toxicity. While huge uncertainties and inconsistencies exist in assigning absolute toxicity values, considering their relative toxicities may be an effective technique to gauge the relative importance of the magnitudes or increases and decreases in emissions among individual MSAT species, transportation alternatives, and/or mitigation measures. A sampling of acute and chronic toxicities associated with the six priority MSATs are provided in Table 3. Included in this table are short-term worker exposures provided by the National Institute for Occupational Safety and Health (NIOSH) and the U.S. Department of Labor, Occupational Safety and Health Administration (OSHA); and short-term and long-term air quality exposures provided by the U.S. EPA and the states of New York and California. The range of variability in toxicities from the short-term to the long-term is six to eight orders of magnitude. No National Ambient Air Quality Standards (NAAQS) have been set for any of the priority MSATs. California is the only entity included in this sampling to adopt a level for diesel particulate matter.

Notes: All Concentrations in µg/m³.

NIOSH IDLH -- National Institute for Occupational Safety and Health, Immediately Dangerous to Life or Health, 15-minute average.

OSHA PEL -- U.S. Deparment of Labor, Occupational Safety and Health Administration Permissible Exposure Limits, 8-hour Time Weighted Average.

NAAQS -- National Ambient Air Quality Standard.

RAC -- U.S. EPA Reference Air Concentration (annual average), 40 CFR 266, Appendix IV.

RsD -- Risk Specific Dose of a 10^-5 increased cancer risk due to a lifetime exposure (70-year average) via the inhalation pathway, 40 CFR 266, App. IV.

New York State Air Guide values: SGC -- Short-term Guideline Concentration (1-hour average); AGC - Annual Guideline Concentration.

California: Acute Inhalation -- 1-hour average Reference Exposure Limits (except for arsenic, 4-hour average and benzene, 6-hour average);

Chronic Inhalation -- Annual average Refererence Exposure Limits; Inhalation Risk Specific Dose = 1 / Inhalation Unit Risk * 10^-5.

^A As PM-2.5.

While the apparent decrease in MSAT emissions projected in 2030 compared to 2005 levels is more than 56%, the effective decrease in MSAT emissions is greater on a toxicity-weighted basis. Employing the appropriate EPA air exposure criteria (e.g., RAC or RsD) and California's RsD for diesel particulate matter, emissions can be expressed on a common basis -- as diesel particulate matter or benzene or any of the other priority MSATs. On a toxicity-weighted basis, the effective decrease in MSAT emissions is 81% from current to design year levels. Figure 5 provides a comparison of these emissions decreases on an un-weighted versus toxicity-weighted basis.

Figure 5. Predicted Decreases in MSAT Emissions

Several of the findings have alluded to the significant effect that traffic congestion has on predicted MSAT emission levels. Consequently, as a practical consideration, comes a question of the required level of detail of vehicle activity data to accurately characterize the amount of congestion on an affected transportation network. One aspect of this question was examined by determining if an average hourly congested speed for the day may be used as an adequate surrogate for the hour-by-hour variation in congested speeds. For facilities operating close to and above capacity, an average hourly congested speed is a marginal to poor indicator of congestion as it relates to the prediction of emissions for all priority MSAT compounds except DPM, which is insensitive to changes in speed. Use of an average hourly congested speed results in an underestimation of MSAT emissions minus DPM for the No-Action Alternate by 6 to 7% for 2005 and 2010 to as much as 16% for 2030. In contrast, for the 2010 Build Alternates where the freeway V-to-C ratios are less than 0.82, a closer match (within 2%) is obtained. The 2030 Build Alternates operate with peak V-to-C ratios ranging from 0.93 to 1.16. Under-estimations of MSAT emissions minus DPM in the range of 3 to 7% are obtained using an average hourly congested speed as a surrogate for the hourly variation.

The level of detail of the vehicle activity data employed in a project-level MSAT emission analysis extends to other factors that may either mitigate or adversely affect congestion. Factors such as variations in vehicle activity by weekday/weekend, month or season, and directional split should be considered.

Emission Factor Variability

The variability of MSAT emission factors projected for the variety of conditions that may be representative of transportation projects and locales throughout the U.S. is huge, ranging by an order of magnitude for each individual compound:

• Acetaldehyde -- 0.00133 to 0.0238 g/VMT;
• Acrolein -- 0.000165 to 0.00365 g/VMT;
• Benzene -- 0.0102 to 0.422 g/VMT;
• Butadiene -- 0.00131 to 0.0249 g/VMT;
• Diesel Particulate Matter -- 0.00207 to 0.0285 g/VMT; and
• Formaldehyde -- 0.00371 to 0.0665 g/VMT.

The highest emission factors are associated with the current year (2005), the minimum speed (2.5 mph), the maximum temperature (95 ºF), and the top end fuel RVP (12.5 or 15.0 psi). The lowest emission factors are associated with years extending into the future (2035 to 2050), higher speeds (55 to 65 mph), moderate temperature (75 ºF), and minimum fuel RVP (7.8 psi).

Some conditions are unlikely to occur concurrently -- such as high temperatures and high volatility gasoline, low temperatures and low volatility gasoline, high vehicle speeds on arterial facilities, or minimal vehicle speeds on freeways and arterials for extended periods of time. A series of graphs were prepared to identify practical ranges of MSAT emission factors by calendar year considering the common operating speeds of vehicles on freeways and arterials and a combination of temperatures with relatively high volatility gasoline.

The vehicle operating speeds selected were 15 to 65 mph on freeways and 10 to 40 mph on arterials. These speeds were chosen considering the congested speeds calculated in the project-level emissions analysis and the average speeds of the test cycles used in developing the speed correction factors in MOBILE6.2. Temperature/fuel RVP combinations were selected considering that automotive gasoline is adjusted seasonally by manufacturers to meet EPA's volatility regulation and ASTM fuel volatility specification D-4814. Suppliers generally publish specification schedules for their gasoline shipments. Typically, high volatility gasoline is supplied during the winter months when high temperatures are unlikely to occur and low volatility gasoline is supplied during the summer months when low temperatures are unlikely to occur. The spring and fall months may be more representative of the minimum/maximum temperature range for periods when relatively high volatility gasoline (Class C or D) is in use. A temperature range of 35 to 95 ºF in combination with a fuel RVP of 12.5 psi was evaluated, as well as a fuel RVP range of 7.8 to 15.0 psi in combination with a temperature of 55 ºF.

Figures 6 though 9 show the variability of emission factors predicted by the MOBILE6.2 model as a function of calendar year for acetaldehyde, acrolein, benzene, butadiene, diesel particulate matter, and formaldehyde. What's represented by the floating bar graphs in each figure is the full extent of projected emission factors by calendar year considering vehicle speeds from 2.5 to 65 mph, temperatures from 35 to 95 ºF, and fuel RVP from 7.8 to 15.0 psi. Superimposed are line graphs illustrating the range of results expected considering practical assumptions of vehicle speed, temperature, and fuel RVP. Each figure consists of six panels, one for each MSAT, delineated by line graphs representing:

• The freeway operating speed range of 15 to 65 mph for a temperature of 55 ºF and fuel RVP of 12.5 psi in Figure 5;
• The arterial operating speed range of 10 to 40 mph for a temperature of 55 ºF and fuel RVP of 12.5 psi in Figure 6;
• The temperature range of 35 to 95 ºF for a fuel RVP of 12.5 psi and freeway operating speed of 55 mph in Figure 7; and
• The fuel RVP range of 7.8 to 15.0 psi for a temperature of 55 ºF and freeway speed operating of 55 mph in Figure 8.