What Does the 1996 National-scale Air Toxics Assessment Mean to Vermont?

To better assist the public and its understanding of the 1996 National-scale Air Toxics Assessment, the Vermont Air Toxics Program summarized the results and general implications of the study.  To view a general summary of the results, please click here. The estimated risks posed to Vermont's population are summarized as well.

To view a detailed description of the uses, sources, chemical properties, and risk characterization of each of the 33 Hazardous Air Pollutants considered in the NATA, please click here.

To view the U.S. EPA's 1996 U.S. EPA Summary of the NATA Modeled Risk, please click here. For a more detailed summary of the U.S. EPA's NATA results, please click on the following web link: http://www.epa.gov/ttn/atw/nata/risksum.html

Vermont General Summary of the 1996
National-scale Air Toxics Assessment

The 1996 National-scale Air Toxics Assessment (NATA) represents a great step forward in the evaluation of key air toxic pollutants and their health effects. The US EPA put forth a tremendous effort into this immense project. Yet, there are some important issues that must be addressed before the results may be perceived as entirely credible.

It is difficult to perform a thorough and reasonable comparison of the 1996 NATA results for Vermont. Due to the nature of Vermont's industry (i.e., the smaller size), many of Vermont's stationary sources are included in the federally-classified "area source" category along with some more typical area sources like wood stoves, landfills, auto body shops, etc. Therefore it is difficult for the State to assess the accuracy of EPA's estimates of Vermont's various area sources. In addition, the Vermont Air Toxics Program feels that the quality and reliability of some of the emissions data and emission factors used in the 1996 National Toxics Inventory are questionable.

Vermont has collected ambient air monitoring data for 21 of the 34 NATA HAPs (see Appendices A & D). Sample results for a several of the pollutants consistently fall below the method detection limit. When possible, a monitor-to-model comparison at the census-tract level was completed. Both a table and graphical comparisons were developed to demonstrate the differences. In most cases, EPA's modeled estimate of the mean annual concentration at a census-tract level was an order to two orders of magnitude less than the mean annual monitored concentration. EPA recognizes that "pollutants examined are typically lower than the measured ambient annual average concentrations when evaluated at the exact location of the monitors." (Monitors are often located in area where the highest levels of pollutants are anticipated.) EPA states that modeled estimates are much closer to monitored concentrations "when the maximum modeled estimate for distances up to 10-20 km from the monitoring location are compared to the measured concentrations."

For certain chemicals with "higher" EPA confidence (e.g., benzene & perchloroethylene), mean annual modeled concentrations were close to the mean annual monitored concentrations (Appendices A & B). Additionally, concentrations of several of the NATA pollutants in Vermont are dominated by regional or global background concentrations. Examples of these are: benzene, carbon tetrachloride, chloroform, ethylene dibromide, ethylene dichloride, hexachlorobenzene, methylene chloride, polychlorinated biphenyls (PCBs), perchloroethylene (perc), and trichloroethylene. This does not imply, however, that these pollutants are not emitted in significant quantities in Vermont. In some cases, they are.

Several NATA results evoke further questioning.

• Unusually high concentrations of acrolein, acetaldehyde, benzene and 1,3-butadiene (primarily mobile source related pollutants) were modeled for Orange County, a very rural Vermont county. There are approximately 55 miles of interstate highway in the Orange County but Windsor County has more miles of highway, greater vehicle miles traveled (VMT), but lower modeled concentrations of these pollutants. Orange County is not densely populated and does not have many lakes. Emissions from snowmobiles in Orange County may be somewhat similar to emissions in northeastern Vermont counties.
• Background dominates concentrations of many pollutants in Vermont.
• EPA's carbon tetrachloride background value equals 0.88 µg/m³. This is 0.34 µg/m³ greater than Vermont's mean concentration of 0.54 µg/m³. As carbon tetrachloride is one of the primary cancer risk drivers in Vermont (and the US), this would imply that EPA overestimates cancer risk relative to the Vermont HAAS. A concentration of 0.88 µg/m³ yields a risk of 13.1 in a million. 0.54 µg/m³ yields a risk of 8.1 in a million creating an overestimation of 5 in a million deaths due to carbon tetrachloride in Vermont. The apparent overestimation would likely extend to most of the U.S. and may give the public the wrong impression of the true risk. (see page 10 for details)
• For some pollutants dominated by background (carbon tetrachloride, chloroform, hexachlorobenzene, PCBs), the distribution percentile scales are often misleading and seem to be very artificial. Some concentration and emission distribution scales seem to be very artificial as well and are artifacts of the modeling e.g., hydrazine's 25^th to 50^th percentile ranges from 0.000000000063 µg/m³ to 0.0000000013 µg/m³. These levels are below the femtogram level. This may be statistically correct, but what literal value is there when percentiles are separated by such small fractions? Where's the reality in this?
• 1,1,2,2-tetrachloroethane has a half-life of > 2 years and is practically inert in the troposphere, why is the background concentration not elevated above zero? Is it all assumed to go to the stratosphere?

Vermont Risk Summary:

According to EPA's 1996 Median Cancer Risk and Median Noncancer Hazard Quotient Assessment, of the 33 HAPs assessed, the following pollutants are driving the risk in Vermont. A full summary of median modeled 1996 cancer risk and noncarcinogen hazard quotients in Vermont may be viewed at the following link.

• Benzene, carbon tetrachloride and formaldehyde are driving the 1996 estimated median cancer lifetime risk in Vermont, according to EPA estimates. Benzene poses a 10 to 30 in-a-million median cancer risk in Chittenden County and 3 to 10 in-a-million risk in all other counties. Carbon tetrachloride poses a 3 to 10 in-a-million risk in the entire state and across most of the US. (However, as noted above, EPA's background concentration may lead to an overestimation of risk of 5 in a million.) Modeled risk for formaldehyde is estimated to be between 3 to 10 in-a-million in the entire state as well.
• Acrolein is the major noncarcinogenic risk driver in Vermont and the entire U.S. In Chittenden County, the acrolein hazard quotient is 3 to 10, and in all other counties, it is 1 to 3.
• EPA Estimated Median Cancer Risk for all carcinogens:

o    Essex & Orleans - 18 to 24 in-a-million

o    Grand Isle, Lamoille, Caledonia, Addison & Windham Counties - 24 to 27 in-a-million

o    Franklin, Washington, Orange, Rutland, Windham & Bennington - 27 to 34 in-a-million

•     Chittenden County - 34 to 45 in-a-million
• EPA Estimated Median Noncancer Hazard for all noncarcinogens:

o    Grand Isle & Orleans Counties - 0.81 to 1.6 Hazard Quotient

o    All other Counties except Washington & Chittenden - 1.6 to 2.6 Hazard      Quotient

o    Washington County - 2.6 to 3.9 Hazard Quotient

o    Chittenden County - 3.9 to 5.1 Hazard Quotient

1996 U.S. EPA Summary of the NATA Modeled Risk

1996 Estimated County Median Cancer Risk for All Carcinogens - VERMONT Counties

• 1996 Risk Characterization - Distribution of lifetime cancer risk for the US population. Based on 1996 exposure to 29 carcinogenic air pollutants from various source sectors.
• From Major US Sources                        Median Risk (50^th Percentile) ~ 0.8 in 10⁶
• From Area & Other Sources                  Median Risk (50^th Percentile) ~ 8.0 in 10⁶
• From Onroad Mobile Sources               Median Risk (50^th Percentile) ~ 10.0 in 10⁶
• From Onroad Mobile Sources               Median Risk (50^th Percentile) ~ 6.0 in 10⁶
• From Background Concentrations        Median Risk (50^th Percentile) ~ 30.0 in 10⁶

• Total Distribution Ranges from 18 to 240 in a million
• Total US Median Risk = 27 in a million

1996 Estimated County Median Noncancer Hazard for All Noncarcinogens - VERMONT Counties

Note: A hazard quotient less than one suggests that exposures are likely to be without an appreciable risk of non-cancer effects during a lifetime. A hazard quotient greater than one can be best described as only indicating that a potential may exist for adverse health effects.

• National Noncarcinogenic Hazard Quotient Risk Drivers for Respiratory Irritation
• Acrolein, formaldehyde, acetaldehyde and 1,3-butadiene (in order of greater to lesser importance) are the four primary chemicals driving the hazard quotient risk for respiratory irritation.
• Manganese compounds, nickel compounds and benzene are also of importance for respiratory irritation according to the 1996 NATA assessment. 

Back to Top

Return to Air Toxics Program Page

Last Updated: 1/22/03