TD886.5
.03
OOOR87104
Technical Report
Air Toxics Emissions From Motor Vehicles
By
Penny M. Carey
September, 1987
NOTICE
Technical Reports do not necessarily represent final EPA
decisions or positions. They are intended to present
technical analysis of issues using data which are currently
available. The purpose in the release of such reports is
to facilitate the exchange of technical information and to
inform the public of technical developments which may form
the basis for a final EPA decision, position or regulatory
action.
Technical Support Staff
Emission Control Technology Division
Office of Mobile Sources
Office of Air and Radiation
U. S. Environmental Protection Agency
-------
Preface
This report titled "Air Toxics Emissions from Motor
Vehicles" is designed to be a compilation of available
information on emission levels of potentially carcinogenic
substances from motor vehicles. While earlier EPA reports
(e.g., "The Air Toxics Problem in the United States; An
Analysis of Cancer Risks for Selected Pollutants") discussed
air toxics emissions in general, their main emphasis was not
mobile sources such as in this report.
EPA currently plans no further regulatory action as a
result of this report. EPA invites comments on this report and
plans to update this information as additional data become
available.
A draft version of this report was circulated prior to the
release of this final version. Users are cautioned that
Section 7.0 dealing with gasoline PIC/POM/organics has been
revised to correct an error due to a misinterpretation of some
of the references. The estimated risks from gasoline organics
are now much lower although still in the range of 100-200
cancers/year. The summary tables have been changed
accordingly. This is the only substantive change of note from
the draft version.
, II
-------
TABLE OF CONTENTS
PREFACE i
EXECUTIVE SUMMARY v
1.0 INTRODUCTION 1
2.0 DIESEL PARTICULATE 2
2.1 Formation and Composition 2
2.2 Control Technology 3
2.3 Emissions 4
2.3.1 Emission Standards 4
2.3.2 Emission Factors by Model Year 6
2.3.3 Emission Factors for Calendar Years
1986 and 1995 6
2.3.4 Nationwide Diesel Particulate Emissions. . 12
2.3.5 Other Sources of Particulate 17
2.4. Ambient Concentrations of Diesel Particulate . . 17
2.4.1 Estimation of Urban and Rural Exposure . . 17
2.4.2 Comparison to Other Exposure Estimates . . 24
2.4.2.1 1983 QMS Exposure Estimate .... 24
2.4.2.2 1981 Lovelace ITRI Exposure
Estimate 25
2.4.2.3 Exposure Estimate Based on a
Lead Surrogate Approach .... 26
2.4.3 Comparison to Particulate Monitoring Data. 27
2.5 Health Effects of Diesel Particulate and Unit
Risk Estimates 27
2.6 Current and Projected Health Risk 28
3.0 FORMALDEHYDE 30
3.1 Formation, Composition and Control Technology. . 30
3.2 Emissions 30
3.2.1 Emission Factors for Calendar Years
1986 and 1995 30
3.2.2 Nationwide Mobile Source Formaldehyde
Emissions 33
3.2.3 Other Sources of Formaldehyde 33
3.3 Ambient Concentrations of Formaldehyde
Emitted by Mobile Sources 34
3.3.1 Estimation of Urban and Rural Exposure . . 34
3.3.2 Contribution of Mobile Sources to Ambient
Formaldehyde Levels 36
3.4 Health Effects of Formaldehyde and Unit Risk
Estimates 38
3.5 Current and Projected Health Risk 39
3.6 Current Activities 41
4.0 BENZENE 43
4.1 Formation, Composition and Control Technology . 43
4.2 Emissions 43
4.2.1 Emission Factors for Calendar Years
1986 and 1995 43
ii
-------
TABLE OF CONTENTS (Continued)
4.2.2 Contribution of Mobile Sources to
Nationwide Benzene Emissions 45
4.3 Health Effects of Benzene and the Unit Risk
Estimate 45
4.4 Current and Projected Health Risk 46
4.5 CARB Analysis of Current and Projected
Health Risk 48
4.6 Comparison of EPA and CARB Health Risks .... 49
4.7 Current Activities 50
5.0 GASOLINE VAPORS 51
6.0 GAS PHASE ORGANICS 54
6.1 Formation and Control Technology 54
6.2 Composition 54
6.3 Mutagenicity of VOC 56
6.4 Risk Associated with Individual VOC 56
6.5 Reactivity of VOC 63
6.6 Current Activities 65
7.0 ORGANICS ASSOCIATED WITH NON-DIESEL PARTICULATE ... 66
7.1 Emission Rates and Composition 66
7.2 Risk from Particle-Associated Organics 66
8.0 DIOXINS 73
8.1 Composition 73
8.2 Emissions 73
8^3 Concentrations of Dioxins 73
8»4 Current Activities 73
9.0 ASBESTOS 74
9.1 Emissions and Ambient Concentrations 74
9.2 Cancer Risk 74
9.3 Current Activities 74
10.0 VEHICLE INTERIOR EMISSIONS 76
10.1 Composition and Concentration 76
10.2 Cancer Risk and Current Activities 76
11.0 METALS 77
11.1 Lead 77
11.1.1 Source and Emission Factors 77
11.1.2 Health Effects 78
11.2 Manganese 80
11.2.1 Source and Emission Factors 80
11.2.2 Health Effects 81
iii
-------
TABLE OF CONTENTS (Continued)
11.3 Platinum 82
11.3.1 Source and Emission Factors 82
11.3.2 Health Effects 82
11.4 Cadmium 83
11.4.1 Source and Emission Factors 83
11.4.2 Health Effects and Risk Estimate .... 83
12.0 SIX MONTH STUDY: SUMMARY AND COMPARISON OF RESULTS . 84
12.1 Purpose of Six Month Study and Summary of Results 84
12.2 Comparison of Six Month Study Results with Results
of This Study 86
12.2.1 Formaldehyde 87
12.2.2 Benzene 89
12.2.3 PIC, B(a)P, Diesel Particulate and
Gasoline PIC/POM 90
12.2.4 Gasoline Vapors 91
12.2.5 1,3-Butadiene 92
12.2.6 Ethylene 92
12.2.7 Asbestos 93
12.2.8 Ethylene Dibromide (EDB) and Cadmium ... 93
12.2.9 Total Aggregate Risk 94
13.0 SUMMARY AND LIMITATIONS 95
REFERENCES 97
GLOSSARY OF TERMS 105
IV
-------
EXECUTIVE SUMMARY
EPA completed a report in 1985, commonly referred to as
the Six Month Study, which contains estimated cancer risks for
a number of toxic air pollutants. The report indicated that
mobile sources may be responsible for a large portion of the
aggregate cancer incidence. The Six Month Study, however, was
broad in nature with the goal to obtain a quick assessment of
the air toxics problem in the United States and to guide
further studies.
The purpose of this report is to focus on air toxics
emissions from mobile sources. Specific pollutants and
pollutant categories which are discussed include diesel
particulate, formaldehyde, benzene, gasoline vapors, gas phase
organics, organics associated with non-diesel particulate,
dioxins, asbestos, vehicle interior emissions and metals. The
report considers all air carcinogens for which EPA has unit
risk estimates and are emitted from motor vehicles. Where
adequate information was available, quantitative estimates of
cancer incidence were made for calendar years 1986 and 1995.
The results were then compared to the results obtained in the
Six Month Study.
The unit risks used in this paper are defined as the
individual life time excess cancer risk from continuous
exposure to 1 ug carcinogen per m3 inhaled air. Assuming a
life time is 70 years, the excess lung cancer risk in 1 year is
derived by simply dividing the unit risk by 70. Using this
approach, latency is ignored. The unit risks used in this
paper are 95 percent upper confidence limits rather than best
estimates. This is consistent with current EPA practice. The
risk estimates presented should therefore be considered upper
bound estimates.
The risks obtained in this study are summarized in Table
S-l. The aggregate risk in 1986 for the total U.S. population
is estimated to range from 385 to 2286 cancer incidences and
drops roughly 40 percent by 1995. Reasons for the projected
decrease in risk in 1995 include 1) more stringent diesel
particulate standards for both light- and heavy-duty vehicles
and 2) the increasing use of 3-way catalyst-equipped vehicles
coupled with the phase out of non-catalyst-equipped vehicles.
As seen in Table S-l, there is a wide range of risk
estimates associated with each pollutant. For diesel
particulate, the range is due to the range of potency (or unit
risk) estimates which were used and, for 1995, a range of
assumptions regarding future diesel sales. The range for
formaldehyde can be attributed to the uncertainty regarding the
-------
Table S-l
Summary of Risk Estimates Contained in This Study3
U.S. Cancer Incidences/Year" Section of Report
Motor Vehicle Pollutant 1986
Diesel Particulate 178- 860
Formaldehyde 46- 131
Benzene 92- 223
Gasoline Vapors 65
Other Gas Phase Organics
1,3-Butadiene
Ethylene
Gasoline PIC/POM
Dioxins
Asbestos
Vehicle Interior Emissions
Cadmium
Ethylene Dibromide
Total: 385-2286
1995
92-576
29- 77
57-145
NDd
0-
0-
1.3 -
ND
0.41-
ND
0.18
1.8
656
60
176
113.4
0-460
0-31
0.78-136
ND
ND
ND
0
0.54
Discussing Pollutant
2
3
4
5
6
6
7
8
9
10
11
12
179-1426
The risk estimates are 95°o upper confidence limits.
The risk estimates for gasoline vapors, asbestos, cadmium and ethylene
dibromide are for urban exposure only. Risks for the other pollutants
include both urban and rural exposure.
The total risk in 1995 is slightly underestimated. Due to inadequate
information and the sensitivity of 1995 risk to control decisions which
have not yet been made, projected risk estimates were not made for some
of the pollutants.
ND=Not Determined.
-------
contribution of photochemically formed formaldehyde. The low
end of the range attempts to account only for formaldehyde
directly emitted from the exhaust of motor vehicles. The high
end of the range attempts to account for both formaldehyde
directly emitted and formaldehyde formed in the atmosphere from
other mobile source volatile organic compound (VOC) emissions.
Both ends use a single unit risk estimate which is based on
formation of malignant tumors.
For benzene, the lower limit is based on ambient
concentrations predicted by a model, whereas the upper limit is
based on actual monitoring data, with a mobile source fraction
assigned based on the mobile source emissions contribution.
The ranges for 1,3-butadiene and asbestos account for
uncertainties in emission factors; for ethylene, the
uncertainty is based on the unit risk estimate. The range for
gasoline products of incomplete combustion (PIC) is due to a
number of different assumptions regarding both emission factors
and unit risk estimates.
Mobile source emissions are extremely complex. Hundreds
of compounds, both in the gas phase and associated with
particles are present. The lack of emissions data and/or
health data and/or exposure data prevented quantitative risk
estimates for any additional pollutants. Of particular concern
are pollutants which are formed photochemically from mobile
source emissions. This category of pollutants could have
considerable impact but not enough is known to make a
quantitative estimate.
A comparison of the results obtained in this study with
those obtained in the Six Month Study is given in Table S-2.
In the Six Month Study, the mobile source contribution to the
air toxics risk was determined based on dispersion modeling of
emissions in 35 highly populated counties (total population of
45 million). They represent a variety of industrial and
population distributions, but are not considered a
statistically representative sample of the country. When
comparing the results of the Six Month Study with the results
of this study, cancer incidences are expressed per million
urban people exposed.
The aggregate risk from mobile sources in the Six Month
Study is 2.65 per million. In this study, the aggregate risk
ranges from 1.80 to 10.58 per urban million. As seen in Table
S-2, a few of the unit risk estimates have increased
considerably since the release of the Six Month Study, most
notably formaldehyde and 1,3-butadiene. If the formaldehyde
risk in the Six Month Study is increased to reflect the updated
unit risk, the aggregate risk in the Six Month Study would
vii
-------
I
00
CO
a;
o
01
a
§
u
0) J5
U *» >i
G C T)
03 O 3
O -H 4J
rH C/J
(!) r-H
U -rl 01
1-1 6 -rt
3 JS
O C H
O
r-
o
i
(N
.
0
I/I 03
CU U
rH 3
O 0) X
S 044J
c
rH O
03 X S
3 w
C -* \o
[-^
O
o
c a
^
3
4-1
I/I
01
H
X
i
0
r t
^
v£)
in
^
-o
0) 3
rH 4->
rt M
E
*^ 01
Cn -rt
** jC
CO EH
^*
o
o
1
01
^i
o
0
4-1 U
rt 0
0} 4-1
RO
03
O »C
CJ C -M
o a
rt O
01 S
01
rt v£)
CO
co
0
0
1
CO
TJ
4->
C
^J
3
r i
r 1
O
Ol
f^
0)
-a
03
Ui
o
fo
O
rH
1
^
Q
0 2
i i r~
co o
. .
O OJ
<£
f
O
r-H
^
O
Q
oo 2
to
1 l
O O
rH rH
<£ OJ
v£» «#
0
O
rH -0
Q
0 2
Ul
)
O
i-l
rH X
ro VJD
i-l vD
*
O rH
CO
C
CO
N u
G O
CO M
CQ P[
0)
OJ
o
o
n
1
o
r-H
X
oo
.
CO
0) 0)
G C
rt 'rt
r-H r-H
0 0
01 01
OJ 03
O O
vO
r-H
CO
1
O
O
IT
I
O
r-l
X
00
.
01
r-
1
O
r-H
X
VO
.
«*
\o
01
o
o
1
01
o
0
m
1
0
rH
X
0
.
r-l
CO
C
CO
rt
T3
0)
3
n
i
CO
V
r-l
01
o
1
o
a
2
f
i
o
rH
X
r-
01
P
2
i-H
VO
01
0
1
co
01
o
Q
J2»
0)
C
CO
r 1
£
^j
w
co
VO
o
1
01
O
o
o
Q
^
«
o
r-l
X
o
{Q
01
1
*JD
vO
Q
2
to
1
o
i-H
X
00
01
1
^J*
Q
£5
Ol
o
01
CO
01
i-H
o
o
o
1-H
0
o
M
1
o
r-l
X
00
.
i-H
1
o
i-H
X
CO
01
to
1
o
rH
X
rj\
.
rH
to
i
0
rH
X
in
00
rt
1
TJ
CJ
rH
O
O
01
0
o
T
1
O
i-H
X
rH
.
in
r
O
rH
X
rH
.
in
in
i
o
rH
X
i-H
.
f.^
T
1
0
rH
X
CO
.
rH
CO
T3
rt
E
0
rt
Q
CO
C
CO
1 1
J51
4-)
OO
in
o
i-H
1
O
oo
.
rH
in
vO
.
01
. .
1 1
03
4-1
O
H
limits.
CO
o
C
co
-d
H
a
o
o
ui
CO
cu
3
oN°
m
0)
t-i
m
01
CO
ft)
£~
H
j i
01
0)
*^
01
H
Ul
(1)
.c
H
ro
TJ
0)
01
3
01
03
C
O
r-H
l~^
rt
£
O
OO
rH
|[ |
O
C
o
rl
4J
03
rH
3
CU
a
G
(0
J_|
3
£
J3
organics from gasoline-fueled vehicles were considered to
PIC risk was not determined in this study.
'O flj
(U
4-1 -
(0 CO
H S-l
0 O
O MH
01 0)
01 U
(0 (U
1 JC
CO H
u
r-l
Ui 'O
Ol 4J
01
a
c 01
0) -H
JC
(1) 4J
4J
03 C
r-H -r-|
3
U O
H M
4J Cli
fO 4-)
Oj G
0)
r-l 01
u
CU Qj
H CO
a u
u
T3
0)
C
'e
0)
4J
0)
Q
4J
o
2
II
Q
o
C*
O
H
4_)
8
03
t-H
Qj
X
CO
U
0
4J
X
CO
4->
(U
0)
00
01
-------
increase to 2.73 per million. If the 1,3-butadiene risk in the
Six Month Study is increased to reflect the updated unit risk,
the risk from 1,3-butadiene would no longer be negligible.
The Six Month Study did not explicitly include diesel or
gasoline-fueled particulate or gas phase exhaust organics (with
the exception of formaldehyde and benzene). It did include a
broad category of pollutants referred to as products of
incomplete combustion (PIC). PIC are postulated to be
primarily polynuclear aromatic hydrocarbons. PIC therefore
includes most of the organics on motor vehicle particulate and
some gas phase exhaust organics. The unit risk for PIC was
derived from epidemiological studies of the general population
and studies of occupational exposure to PIC (e.g., coke oven
emissions). The PIC unit risk is expressed per unit of
exposure of benzo(a)pyrene (B(a)P).
In the Six Month Study, the B(a)P emission factor for
mobile sources was used to calculate an annual average
exposure. The B(a)P exposure was then multiplied by the PIC
unit risk to estimate annual cancer incidence due to PIC from
mobile sources. The same process was performed for all other
sources emitting B(a)P to determine cumulative annual cancer
incidence from PIC. Since roughly 75 percent of the total
B(a)P emissions were said to be attributable to mobile sources,
75 percent of the total PIC cancer incidence is due to mobile
sources. The PIC risk from mobile sources in the Six Month
Study is estimated to be 2.07 per urban million.
In this study, risks from diesel particulate and gasoline
particle-associated organics (referred to as gasoline PIC/POM
where POM is polycyclic organic matter) were analyzed
separately. The sum of the diesel particulate and gasoline
PIC/POM risks will be considered to represent PIC. Using this
approach, the total PIC risk ranges from 0.77-4.43 per urban
million. The PIC risk obtained in the Six Month Study lies
within this range.
When reviewing the results of this study, the following
are a few of the important limitations that should be
considered. The report only accounts for a small number of the
mobile source pollutants known to be emitted, and does not
consider reactions of mobile source pollutants in the
atmosphere. Resulting secondary pollutants may be more or less
carcinogenic than what was originally emitted. The report also
does not fully and accurately take into account seasonal
variations in emissions. This may result in underestimating
the risk since it is likely some air toxics emissions will
increase with decreasing temperature. The emission factors
were developed using 75°F as the ambient temperature. The
risks presented are assumed to be additive. The risk
projections for 1995 are based on the emission standards
ix
-------
currently in place. Changes in fuel composition are not
considered. The risk projections do not consider use of
alternative fuels, such as methanol. Neither do the current or
projected risk estimates consider use of alcohol/gasoline
blends.
The following is a brief summary of each pollutant/
pollutant category covered in this report. It includes current
research and regulatory activity. It should be read when
reviewing the tables, since it describes important
uncertainties. Literature references are not given in this
summary, but are in the body of the report.
Diesel Particulate - Diesel particulate exhaust is
composed of an elemental carbon core with hundreds of adsorbed
organic compounds ranging from C14 to about C40. Over 90
percent of diesel particulate is less than 1 micron in size.
The light-duty vehicle and truck emissions standard is
currently 0.6 gram/mile. New standards, effective in 1987, are
0.20 gram/mile for light-duty vehicles and 0.26 gram/mile for
light-duty trucks. For heavy-duty diesel engines, there is
currently no diesel particulate emission standard. A standard
of 0.6 gram/brake horsepower-hour (g/bhp-hr) begins in 1988,
with increasingly more stringent standards effective in 1991
and 1994. These increasingly stringent standards are accounted
for in the 1995 projections.
Diesel particulate was found to be mutagenic in the late
1970's. Subsequent studies revealed that nitropolynuclear
aromatic hydrocarbons (nitro-PAH), specifically, nitropyrenes,
dinitropyrenes and nitrohydroxypyrenes together account for
much of the mutagenicity observed. The organics extracted from
diesel particulate and other known carcinogens such as coke
oven emissions were tested in a battery of bioassays. These
included bacteria and mammalian cell bioassays, and one skin
painting study with SENCAR mice. Animal inhalation studies
were conducted at that time but gave negative or inconclusive
results. The unit risk for diesel particulate was determined
by comparing the potency of diesel particulate with the
potencies of the other carcinogens determined in these tests.
The range of upper confidence limit unit risks used in this
study, 0.2-1.OxlO~4, is based on various analyses of the
comparative potency data. Several animal inhalation
experiments have been recently completed which, in contrast to
the earlier studies, show that diesel exhaust causes lung
tumors in rats. After being analyzed, these experiments may
help narrow the range of unit risks.
Total diesel particulate emissions in 1986 were estimated
to be 274,000 metric tons, or roughly 3.9 percent of the total
suspended particulate (TSP) emissions. Diesel particulate
emissions are projected to drop to 125,000-154,000 metric
-------
tons/year in 1995. This is due to the more stringent diesel
particulate standards. The range of emissions estimates in
1995 is due to the range of diesel sales assumed.
The annual mean exposure level, estimated using a modified
version of the NAAQS exposure model (NEM) for CO, was 2.6
ug/m3 in 1986. This exposure level accounts for both indoor
and outdoor exposure to diesel particulate. This represents
roughly 5 percent of the 1984 annual geometric mean TSP
concentration. This level drops to 1.2-1.6 ug/m3 in 1995.
The resulting annual lung cancer risk from diesel particulate
exposure for the U.S. population is 178-860 in 1986 and 92-443
in 1995.
Formaldehyde - Formaldehyde is emitted in the exhaust of
both gasoline- and diesel-fueled vehicles. It has the chemical
formula CH2O. Formaldehyde is of interest due to its
photochemical reactivity in ozone formation and suspected
carcinogenicity. An upper confidence limit unit risk of
1.3x10 was used. It is based on a single study in which
rats exposed to formaldehyde developed malignant and benign
tumors in the nasal cavities; the unit risk is based on
malignant tumor formation only. The current consensus in EPA
favors use of malignant tumors only. EPA's Office of Toxic
Substances is currently using the unit risk based on malignant
tumors only in assessing the risk to garment workers and home
residents. The consideration of benign tumors would increase
the formaldehyde risk presented in this report by a factor of
15.
Formaldehyde exhaust emissions from motor vehicles
correlate well with exhaust hydrocarbon (HC) emissions. For
this analysis, formaldehyde emissions were expressed as a
weight percentage of exhaust HC. These percentages were then
applied to the exhaust HC output from the MOBILES emissions
model for 1986 and 1995 to obtain the formaldehyde emission
factors. In this way, deterioration and other effects are
included. The percentages generally vary from 1 to 4 percent,
depending on the vehicle class. Mobile source formaldehyde
emissions in 1986 were estimated to be roughly 71,000 metric
tons, or roughly 28 percent of the total formaldehyde emissions
in the U.S. Mobile source formaldehyde emissions are expected
to drop to 41,000 metric tons in 1995. This is due to the
increasing use of 3-way and 3-way plus oxidation
catalyst-equipped gasoline-fueled vehicles together with the
phase out of non-catalyst-equipped vehicles. The result is a
marked decrease in projected HC and, by association,
formaldehyde emissions.
Nationwide exposure levels, using the modified CO NEM, are
roughly 1.04-1.13 ug/m3 and 0.59-0.65 ug/m3 for 1986 and
1995, respectively. The resulting risk is 46-50 in 1986 and
29-31 in 1995. The range accounts for both the presence and
XI
-------
absence of an Inspection/Maintenance program. This attempts to
account for direct emissions only and does not account for
either the destruction or photochemical formation of
formaldehyde in the atmosphere.
Another approach intended to include photochemistry was
also developed. With this approach, a mobile source fraction
was applied to an annual average formaldehyde concentration
developed by OAQPS using available ambient monitoring data.
Mobile sources account for 28 percent of the total VOC
emissions and 30 percent of the formaldehyde emitted directly.
Assuming that the VOC from all sources have the equivalent
potential to form formaldehyde, a mobile source fraction of
0.30 was selected. This fraction was applied to an urban
population weighted average of 12.71 ug/mj (based on data
obtained in 4 cities) and a rural concentration of 1.50
ug/m1. Since the summer concentrations used to calculate the
urban concentration probably represent maximum rather than
average values, the risk estimates can be used to represent a
plausible upper limit. Using this approach, the annual cancer
risk from mobile source formaldehyde is 131 in 1986 and 77 in
1995. Combining both approaches, the cancer incidences due to
mobile sources range from 46-131 in 1986 and 29-77 in 1995.
Research activity is planned or underway in three areas:
1) emissions characterization, 2) photochemistry, specifically,
factors affecting formaldehyde formation in the atmosphere, and
3) ambient monitoring.
Benzene - Benzene is an aromatic hydrocarbon with the
formula C6H6. It is present in both exhaust and
evaporative emissions. Several epidemiologic studies have
associated benzene with an increased incidence of leukemia.
The upper confidence limit unit risk estimate determined from
these studies is 8.0 x 10"6.
Mobile sources (including refueling emissions) dominate
the nationwide benzene emission inventory. In 1982, mobile
source benzene emissions were roughly 250,000 metric tons, or
85 percent of the total benzene emissions. Of the mobile
source contribution, 70 percent comes from exhaust, 14 percent
from evaporative emissions and 1 percent from motor vehicle
refueling.
Nationwide exposure levels from exhaust and evaporative
emissions were estimated using the modified CO NEM. Benzene
emissions were expressed as percentages of exhaust and
evaporative HC. Percentages of exhaust HC vary from 1.1-5.12
percent for the vehicle classes; for evaporative emissions, the
percentage varies from 0.35-1.53 percent. Nationwide exposure
levels from both exhaust and evaporative emissions are roughly
3.1-3.2 ug/m3 and 1.7-1.8 ug/mj for 1985 and 1995,
XII
-------
respectively. (A previous analysis was relied on which used
1985 as the base year. It was assumed in this report that the
calendar year emission factors for 1985 and 1986 would not
differ significantly.) This assumed the presence of a
standard, minimum I/M program. The range is due to
consideration of both a low and high range evaporative
emissions estimate for light-duty gasoline-fueled vehicles.
Annual cancer incidences from exhaust and evaporative emissions
are estimated to be 84-89 in 1985 and 50-52 in 1995. The
reason for this marked decrease is the decrease in projected HC
in 1995 and, thus, benzene emissions.
Exposure to benzene during refueling includes self-service
refueling, occupational exposure (service station attendants)
and community exposure in an urban area. Exposure to
self-service refueling and occupational exposure was determined
by measuring benzene levels in the region of the face of a
person refueling a vehicle tank. The exposure in a typical
urban area was estimated using a dispersion model. Annual
cancer incidences from benzene refueling are estimated to be 8
in 1985 and 7 in 1995. Total cancer incidences from benzene
exhaust, evaporative and refueling emissions are 92-97 in 1985
and 57-59 in 1995.
An alternative approach, similar to that used for
formaldehyde, was also developed. With this approach, a mobile
source fraction was applied to estimated urban and rural
concentrations developed using available ambient monitoring
data. Mobile sources account for 85 percent of the total
benzene emissions. Therefore, a fraction of 0.85 was applied
to an urban population weighted average of 10.24 ug/m1 and a
rural concentration of 7.52 ug/m3. Using this approach, the
annual lung cancer risk from mobile source benzene is 223 in
1985. Based on the NEM modeling, emissions of benzene from
mobile sources are projected to decrease roughly 40 percent
from 1985 to 1995. Accounting for this decrease and the
projected population increase, the annual lung cancer risk is
145 in 1995. Combining both approaches, the lung cancer
incidences due to mobile sources range from 92-223 in 1985 and
57-145 in 1995.
The California Air Resources Board (CARB) has also
attempted to determine the risk posed by benzene emissions from
various sources in 1984 and 2000. The vehicular contribution
is estimated to be 21-166 cancer cases in California in 1984
and 20-154 cancer cases in 2000. Expressed as individual risk,
the vehicular contribution is 8.1-64.3x10"7 cancers/person in
1984 and 6.4-49.0x10"7 cancers/person in 2000. In this case,
the range is due to a range of unit risks used by CARB. The
range of risk estimates given in this report, when expressed as
individual risk, are 3.8-9.3x10"7 cancers/year in 1986 and
2.2-5.6xlO"7 cancers/year in 1995. The upper bound of these
Xlll
-------
risk estimates are roughly equivalent to the lower bound of the
GARB estimates because, for these estimates, the ambient
concentrations and unit risks are roughly equivalent. The
upper bound CARB estimates use an upper bound unit risk that is
roughly 7 times higher than the EPA-CAG unit risk used in this
report.
The California Air Resources Board (CARB) is considering
implementing regulations requiring control of motor vehicle
benzene emissions. OAQPS has designated benzene as a hazardous
air pollutant under Section 112 of the Clean Air Act and is
implementing necessary controls for stationary sources. QMS is
including benzene impacts in its assessment of VOC control
options.
Gasoline Vapors - Totally vaporized gasoline has been
found to cause a statistically significant increase in kidney
tumors in male rats and liver tumors in female mice. An upper
confidence limit based on the rat data is 1.18 x 10"s.
Exposure to gasoline vapors during refueling was estimated
based on an American Petroleum Institute (API) study that
involved measuring gasoline and vapor levels in the region of
the face of a person refueling a vehicle tank. The exposure in
a typical urban area for these refueling emissions was also
estimated using the Industrial Source Complex dispersion model
to calculate annual concentrations. Based on these exposures,
the risk from gasoline vapors (excluding benzene) was estimated
as 65 lung cancer incidences per year.
EPA has not made a decision on what controls should be
proposed for gasoline vapors.
Gas Phase Organics - Gas phase organics, or volatile
organic compounds (VOC), are present in both exhaust and
evaporative emissions. Over 300 VOC have been identified. The
majority of VOC consist of unsaturated and saturated
hydrocarbons along with benzene, alkyl benzenes, aliphatic
aldehydes and a variety of polycyclic aromatic hydrocarbons.
Most of the known mutagenicity of motor vehicle emissions is
associated with the particulate phase, however.
Of all the VOC emitted from motor vehicles, only benzene,
formaldehyde, benzo(a)pyrene (B(a)P), ethylene, and
1,3-butadiene have unit risks. Gas phase B(a)P was considered
with particle-associated B(a)P since the majority of B(a)P is
in the particulate phase. Ethylene emissions are present in
vehicle exhaust and constitute from 6 to 13 percent of the
exhaust HC emissions. Exposures were estimated using the
modified NEM model. Based on the upper confidence limit unit
risk provided in the Six Month Study (2.7 x 10"6), risk
estimates of lung cancer incidence for 1986 and 1995 are 55-60
and 29-31, respectively. The range accounts for the presence
and absence of an I/M program. The unit risk is extremely
xiv
-------
tentative, however, since there is no available direct evidence
that ethylene is carcinogenic. The unit risk for ethylene was
estimated based on assumptions regarding its potency relative
to ethylene oxide, a metabolite of ethylene and an animal
carcinogen. For this reason, a lower risk estimate of zero is
used.
1,3-Butadiene is a photochemically reactive compound
present in vehicle exhaust. The Six Month Study found no risk
associated with emissions of 1,3-butadiene from mobile sources;
however, since the release of the Six Month Study the unit risk
has increased by a factor of 1000. Updated emission factors
for 1,3-butadiene also appear higher but determination of an
accurate emission factor is difficult because 1,3-butadiene and
n-butane coelute and thus have the same retention point on the
gas chromatograph. Emission characterization studies to date
have not attempted to determine the percentage of the peak due
to 1,3-butadiene. Therefore, assumptions must be made about
the percentage each compound contributes to this peak. It was
assumed in this report that 15 percent of the peak was due to
1,3-butadiene, based on data collected in New York's Lincoln
Tunnel, although ambient data indicate that the actual
percentage could well be much lower. Fifteen percent was
chosen as an upper limit. Based on data from in-use
gasoline-fueled vehicles, 1,3-butadiene is roughly 0.94 percent
of the total exhaust HC as measured by the flame ionization
detector (FID). Due to the lack of data for the other vehicle
classes, this percentage was simply applied to the MOBILES
composite exhaust HC emission factor. It was further assumed
that the percentage would remain the same from 1986 to 1995.
The modified NEM was used to estimate exposures.
Nationwide urban exposure in 1986 is estimated to be 0.60-0.66
ug/ms. The range accounts for the presence and absence of an
I/M program. These exposure estimates are for direct emissions
of 1,3-butadiene and do not account for reactions of
1,3-butadiene in the atmosphere. Available ambient monitoring
data were reviewed and compared to the exposure estimates.
Average mean values in urban settings range from 0.24-24.23
ug/ms, although the accuracy of the analytical methods used
is uncertain. The NEM urban exposure estimate lies within this
range.
Using the exposure estimates in conjunction with the upper
confidence limit unit risk estimate (2.8xlO~4), estimates of
lung cancer incidence for 1986 and 1995 are 593-656 and
391-460, respectively. Preliminary emission characterization
results indicate the presence of 1,3-butadiene, but the amount
has not yet been quantified. Therefore, a lower risk estimate
of zero will also be used. The resulting ranges of cancer
incidences for 1986 and 1995 given in Table S-l are 0-656 and
0-460, respectively. OAQPS is currently working on a source
assessment document for 1,3-butadiene which should be completed
sometime in late 1987.
xv
-------
The atmospheric photochemical reaction products of mobile
source VOC are largely unknown. Smog chamber experiments have
produced unidentified gas phase mutagens starting with simple
VOC compounds that are present in vehicle emissions. It
appears that, as non-catalyst-equipped vehicles are phased out
of the fleet, the reactivity of vehicle exhaust will decrease.
This is evidenced by the decreasing percentages of reactive
olefins and aromatics, coupled with the increasing percentage
of less reactive paraffins (particularly methane) in the
exhaust. The Integrated Air Cancer Project (IACP) is a
long-term effort by EPA ORD to identify the principal airborne
carcinogens and their sources, and may clarify the risk posed
by reaction products.
Organics Associated with Non-Diesel Particulate
Gasoline-fueled vehicles emit far less particulate than their
diesel counterparts. It is thought that a number of
nitro-polycyclic aromatic hydrocarbon (PAH) compounds account
for much of the mutagenicity of diesel particulate emissions.
Particulate emissions from gasoline-fueled vehicles contain
significantly less of these nitro-PAH's; however, the
mutagenicity of the gasoline soluble organic fraction (SOF),
expressed as revertants/ug SOF, is greater than diesel SOF.
Also, unlike diesel SOF, the mutagenic activity of gasoline SOF
increases with the addition of S9 activation, indicating
indirect-acting activity. This suggests that the classical
PAH's may be responsible for the mutagenicity of gasoline SOF,
rather than the nitro-PAH's.
The organics associated with gasoline particulate were
considered to represent gasoline products of incomplete
combustion (PIC). Three different approaches were taken to
estimate the risk from gasoline PIC. The first approach
assumes the risk from B(a)P emissions adequately represents the
risk of all gasoline PIC. B(a)P is emitted primarily from
gasoline-fueled vehicles. The annual cancer risk of B(a)P from
gasoline-fueled vehicles was determined by adjusting the upper
confidence limit B(a)P risk obtained in the Six Month Study to
account for the difference in emission factors. The resulting
annual cancer risk in 1986 is 0.007 per urban million, or 1.3
cancer incidences, assuming an urban population of 180
million. B(a)P emissions should decrease with the phase-out of
leaded fuel. In 1995, the cancer risk is projected to decrease
to 0.004 per urban million, or 0.78 cancer incidences, assuming
an urban population of 195 million.
The second approach uses B(a)P emission factors from
gasoline-fueled vehicles together with the PIC upper confidence
limit unit risk used in the Six Month Study (which is expressed
per unit of exposure of B(a)P) to estimate the annual cancer
risk of PIC from gasoline-fueled vehicles. This approach
xvi
-------
assumes that B(a)P is an adequate surrogate for all PIC
compounds. It should be noted that this approach is rather
uncertain as the mix of PIC compounds differs among source
types. The annual cancer risk was determined by adjusting the
PIC risk obtained in the Six Month Study to account for the
difference in B(a)P emission factors. The resulting annual
cancer risk is 0.68 per urban million, or 122 cancer incidences
in 1986. In 1995, the cancer risk is projected to decrease to
0.37 per urban million, or 72 cancer incidences.
The third approach uses estimated emission rates of
gasoline particle-associated organics (as an unspeciated
mixture) together with an upper confidence limit unit risk for
these organics. Exposures were estimated using the modified
NEM model.
Estimated composite emission factors for gasoline
particle-associated organics, with and without an I/M program,
are 0.0075-0.0082 g/mile in 1986 and 0.0048-0.0058 g/mile in
1995. Total organic emissions are not projected to decrease as
much as B(a)P emissions. This is because emission data
indicate that use of a catalyst reduces B(a)P emissions to a
greater extent than total organic emissions.
A unit risk estimate for gasoline particle-associated
organics has been estimated, based on data from the only
catalyst-equipped vehicle tested for particle organic
mutagenicity. The vehicle had exceptionally high exhaust
emissions, comparable to those from a non-catalyst equipped
vehicle. It was originally chosen for testing in 1979 on this
basis since it was easier to collect enough extractable
organics for analysis. The mutagenic activity of the
particle-associated organics from this vehicle, as indicated by
the Ames Salmonella strain TA-98 bioassay, is on the low end of
the range, when compared with other catalyst-equipped
vehicles. As a result, the vehicle should be considered to be
of uncertain representativeness. An upper confidence limit
unit risk estimate based on this vehicle is 2.5xlO~4.
The total risk in 1986, accounting for both urban and
rural exposure, is 163-176 cancer incidences and drops to
115-136 cancer incidences in 1995.
For this report, a range of risk estimates for gasoline
PIC will be reported, which encompasses the results of all
three approaches. The resulting range of cancer incidences is
1.3-176 in 1986 and 0.78-136 in 1995.
Dioxins - The major dioxin compound of interest and the
one considered in this report is 2,3,7,8-tetrachlorodibenzo-
p-dioxin. This dioxin compound exists in the particulate state
xvn
-------
or is adsorbed onto particulates. Some qualitative analytical
measurements have found dioxin to be present in the muffler
scrapings of vehicles using either leaded or unleaded
gasoline. It appears to be emitted in only trace quantities
(e.g., 10~9 g/mile) in vehicle exhaust. OAQPS plans to make
a decision within the coming year on whether to list dioxin as
a hazardous air pollutant.
Asbestos - Asbestos is used in brake linings, clutch
facings and automatic transmissions. About 22 percent of the
total asbestos used in the U.S. in 1984 was used in motor
vehicles. Asbestos emissions from vehicles with front disc
brakes and rear drum brakes ranged from 4-28 ug/mile. Based on
these emission rates, maximum annual average asbestos levels in
urban areas due to motor vehicles are estimated to range from
0.25 to 1.75 nanograms per cubic meter (ng/m3)- Asbestos
from mobile sources appears to account for roughly 1-10 percent
of urban asbestos concentrations, although mobile sources could
be responsible for as much as 70 percent, under worst case
emissions conditions.
EPA has not attempted to quantify the excess mortality
from asbestos exposure. The National Academy of Sciences (NAS)
has, however, estimated life time risks for persons in urban
areas. Based on the data in the NAS report, the individual
annual cancer risk from urban levels of asbestos is estimated
to range from 9 x 10 "9 - 3.6 x 10"7 per ng/m3 exposure.
Assuming an urban population of 180 million, mobile source
asbestos emissions could be responsible for as many as 113
cancer incidences per year.
EPA's Office of Pesticides and Toxic Substances (OPTS) has
proposed regulations under Section 6 of TSCA to ban certain
uses of asbestos and to allocate permits to mine and import
asbestos which would restrict its remaining uses. EPA is also
considering a ban on asbestos friction products about 5 years
after the final rules are promulgated. The fact that there are
not yet available good substitutes to replace asbestos in
certain automobile and truck brakes, however, may push back
EPA's goal for banning and phasing out asbestos for such uses.
Vehicle Interior Emissions - A total of 147 compounds have
been identified in vehicle interiors. In closed vehicles under
high temperatures, vinyl chloride was present at levels ranging
from below 2 ppb to 7 ppb. A number of other carcinogenic
compounds were identified qualitatively. Due to the low
exposure level, no significant risk should be present from
vehicle interior emissions.
Ethylene Dibromide (EDB) and Cadmium - Updated emission
factors for EDB and cadmium were estimated. The risk estimates
for EDB and cadmium were then determined by adjusting the risk
obtained in the Six Month Study to account for the difference
in emission factors. As seen in Tables S-l and S-2, the risks
from these pollutants are negligible in comparison to other
mobile source pollutants.
xviii
-------
1.0 INTRODUCTION
Considerable effort is underway within the Environmental
Protection Agency (EPA) to determine the magnitude of the air
toxics problem in the United States. For the purposes of this
report, air toxics are defined as carcinogens associated with
air pollution. The overall objective is to develop
quantitative estimates of the upper-bound cancer risk
associated with air toxics emissions, and to determine the
relative contribution of specific pollutants and sources.
In 1985, EPA completed a report which is commonly referred
to as the Six Month Study. This study contains estimated
cancer risks due to 15 to 45 toxic air pollutants (the number
of pollutants examined varied with the three different analyses
that were performed).[1]* The estimates of cancer incidence
from these analyses ranged from 1300 to 1700 cases annually
nationwide for all pollutants combined. These estimates were
based on use of upper-bound unit risks. The analyses further
indicated that mobile sources may be responsible for a large
portion of the aggregate cancer incidence. The Six Month
Study, however, was broad in nature with the goal to obtain a
quick assessment of the air toxics problem in the United States
and to guide further studies.
The purpose of this report is to focus on air toxics
emissions from mobile sources. Specific pollutants or
pollutant categories which will be discussed include diesel
particulate, formaldehyde, benzene, gasoline vapors, gas phase
organics, organics associated with non-diesel particulate,
dioxins, asbestos, vehicle interior emissions, and metals. For
each pollutant or pollutant category, available information is
given on the formation and composition, control technology,
emissions, ambient concentrations, unit risk estimate, the
current and projected public health impact (or risk), and
current EPA regulatory and/or research activity.
The results obtained are then compared to the earlier EPA
report.[1] Following this is a summary of the total risk from
mobile source air toxics emissions, together with the
limitations inherent in the estimate.
Numbers in brackets designate references at the end of the
report.
-1-
-------
2.0 DIESEL PARTICULATE
It should be noted that, since there has not been a recent
analysis by EPA of diesel particulate emission factors and
exposure, a detailed analysis was performed during preparation
of this report. Other pollutants are not treated as thoroughly.
2.1 Formation and Composition
Particulate emissions from diesel-equipped vehicles are
formed as a result of incomplete combustion of the fuel. The
particles in diesel exhaust differ both in quantity and
composition from particles in gasoline engine exhaust.
Dies el-equipped vehicles can emit from 30 to 100 times more
particulate mass, on a grams per mile basis, than
gasoline-powered, catalyst-equipped vehicles.[2] Over 90
percent of diesel particulates are less than 1 micron in size
and are therefore small enough to be inhaled and deposited deep
within the lungs.[3] Virtually all are less than 10 microns.
Gasoline particulate emissions from catalyst-equipped
vehicles using unleaded fuel are very low in mass and consist
largely of sulfates and also some organics. Diesel particulate
emissions, on the other hand, are very complex, being composed
of carbonaceous matter with condensed and/or absorbed fuel and
lubricant components and other varied combustion products. The
soluble organic fraction (SOF) of the total diesel particulate
mass (i.e., organics extractable with methylene chloride)
varies significantly. Typically, diesel passenger car SOF
ranges between 5 and 50 percent of the total particulate
mass.[4]
The chemical composition of diesel SOF is complex.
Generally, diesel SOF ranges from Ci4 to about C40.[4]
Hundreds of compounds are present and the analytical capability
does not exist to identify every compound. Instead, effort has
been focused to identify the chemical classes and specific
compounds associated with the SOF that are mutagenic in the
Ames bioassay test.
The mutagenicity of diesel SOF decreases upon the .addition
of S9 activation, indicating direct-acting frameshift mutagenic
activity. The SOF was solvent-partitioned into organic acids,
bases, and neutral components; the neutral components were
further fractionated and the mutagenic activity of each
fraction was determined using the Ames Salmonella
typhimurium/microsome assay. The moderately and highly polar
neutral compounds account for 89-94 percent of the mutagenic
activity and only 32 percent of the mass. [5] Gas
chromatography/mass spectroscopy identified nonmutagenic
fluorenones and methylated fluorenones as major constituents of
-2-
-------
the moderately and highly polar neutral fractions but they did
not account for the direct-acting activity observed. Studies
with nitroreductase-deficient strains of Salmonella typhimurium
suggested that nitrated compounds are present. Subsequent
studies show that nitropyrenes, dinitropyrenes and
nitrohydroxypyrenes together account for much of the
mutagenicity observed.[5, 6]
EPA conducted a large research program to evaluate the
health effects associated with exposure to diesel emissions,
with particular emphasis on the organic extracts. The
culmination of this effort was an estimation of the unit risk*
for diesel particulate.[7]
2.2 Control Technology
A variety of approaches are being developed by
manufacturers to control diesel particulate emissions. One
approach is the use of a catalytic converter to oxidize ar.d
remove the soluble organics absorbed on the particulate. T .e
other, more popular approach is the use of a particulate filter
or trap. Reduction of particulate formation via combustion
chamber modifications is also being investigated. The primary
traps being evaluated are: 1) a catalyzed ceramic monolith
trap, 2) a ceramic monolith trap, and 3) a catalyst-coated wire
mesh trap. These traps are being evaluated for both light-duty
and heavy-duty applications.
For light-duty applications, Daimler-Benz introduced in
the 1985 model year a catalyzed ceramic monolith trap in some
of its turbocharged diesels. Most manufacturers, however, are
investigating the feasibility of non-catalyzed, ceramic
monolith traps.
A trap must be periodically regenerated by oxidizing the
collected particulates. Otherwise, particulates collected on a
trap can cause the exhaust back pressure to increase and
adversely affect fuel economy and vehicle performance. Active
and passive regeneration techniques have been assessed. An
active technique would use a diesel-fueled burner or electric
resistance heater to raise the temperature of the engine
*Unit risk is defined as the individual life time excess cancer
risk from continuous exposure to 1 ug carcinogen per m3
inhaled air. Assuming a life time is 70 years, the excess lung
cancer risk in 1 year is derived by simply dividing the unit
risk by 70. Using this approach, latency is ignored.
-3-
-------
exhaust gas flowing into the trap or to raise the temperature
of the trap itself to the ignition temperature of the collected
particulates, thus combusting the particulate which regenerates
the trap. It should be noted that the organic emissions from
trap regeneration have not been fully characterized.
A passive technique uses catalytic material to lower the
ignition temperature of the collected particulate. One method
is to apply the catalytic coating to the trap substrate
itself. Daimler-Benz introduced in the 1985 model year a
catalyzed ceramic monolith trap in some of its turbocharged
diesels. Johnson-Matthey, a catalyst manufacturer involved in
trap development, is investigating the merits of a
catalyst-coated stainless steel wire mesh trap.
The other passive method is the use of a metal fuel
additive in conjunction with a ceramic wall flow monolith
trap. This yields the lowest ignition temperature of any other
passive method. The most promising metal fuel additives to
date are manganese and copper compounds.
2.3 Emissions
In this section, the Federal diesel particulate standards
are presented. This is followed by an estimation of in-use
emission factors by model year for each vehicle class (i.e.,
light-duty car, light-duty truck and heavy-duty vehicle).
These model year emission factors are then used in conjunction
with information on diesel sales fractions and fraction of
diesel travel by model year to obtain calendar year emission
factors for each vehicle class. For this analysis, calendar
years 1986 and 1995 were selected. Nationwide diesel
particulate emissions (metric tons/year) for 1986 and 1995 are
then calculated by combining the emission factors with
projected vehicle miles traveled (VMT) data for 1986 and 1995.
The results are then compared to the most recent national
particulate emission estimate.
2.3.1 Emission Standards
Table 2-1 provides a summary of the Federal diesel
particulate emission standards for light-duty cars and trucks
and heavy-duty engines. The test procedure specified for
light-duty vehicles is the CVS-75 or FTP procedure, a constant
volume sample test which includes both cold and hot starts.
For heavy-duty vehicles, a transient test procedure is
currently used. This is an engine dynamometer procedure with
starts, stops, and speed/load changes. In order to meet the
future Federal standards (1987 and later), some light-duty and
heavy-duty vehicles will need to be equipped with a particulate
aftertreatment system, such as a trap.
-4-
-------
Table 2-1
Diesel Particulate Standards
Light-Duty Vehicles
Standard
(grams per mile)
Year Cars Trucks
1981 and prior
1982-1986 0.60 0.60
1987 and later 0.20* 0.26*
Heavy-Duty Engines
Standard
Year (grams per brake horsepower-hour)
1987 and prior
1988-1990 0.60
1991-1993 0.25**
0.10***
1994 and later 0.10**
* Emissions averaging may be used to meet this standard.
For trucks, some restrictions apply.
** Emissions averaging may be used to meet this standard. Some
restrictions apply.
*** For urban bus engines, the standard is 0.10 g/Bhp-hr
beginning in 1991. Particulate averaging is not allowed
with this standard.
NOTE: In an emissions averaging program, the manufacturer
determines emission limits for each vehicle/engine family.
Family emission limits are allowed to exceed the standards;
however, the weighted average of the family emission limits
myst be in compliance with the applicable standard. For cars
and trucks, the average emission level is based upon a
production-weighted average of the family emission limits. For
heavy-duty engines, the average emission level is determined by
calculating a production- and horsepower-weighted average of
the family emission limits.
-5-
-------
2.3.2 Emission Factors by Model Year
The light-duty and heavy-duty diesel particulate emission
factors are given in Table 2-2. The heavy-duty sub-groups are
defined by gross vehicle weight rating (GVWR) as follows:
Class 2B = 8,500 to 10,000 Ibs.
Classes 3-5 = 10,001 to 19,500 Ibs.
Classes 6 = 19,501 to 26,000 Ibs.
Classes 7 and 8 = 26,001 Ibs and up.
The emission factors for light-duty cars and trucks were
obtained from an EPA report titled, "Diesel Particulate
Study."[8] For 1987 and later, these emission factors are
considered to be, for the purpose of this report, the
respective emission standards. For heavy-duty vehicles, the
emission factor assumed for 1983 and prior is 1.0 grams per
brake horsepower-hour (g/Bhp-hr). For 1984-1987, the emission
factor assumed is 0.75 g/Bhg-hr. This is based on transient
test data for a variety of in-use heavy-duty vehicles.[9] For
1988-1993, the emission factor assumed is the standard. For
1994 and 1995, the standard was adjusted upward slightly to
account for projected trap failure. Within classes 7-8, the
emission factor for buses was adjusted upward to account for
projected trap failure beginning in 1991. The equation used to
make this adjustment is given and discussed in detail in
reference 8.
It was necessary to convert the heavy-duty emission
factors from g/Bhp-hr to g/mile. As seen in Table 2-2, the
g/mile emission factors for the heavy-duty subgroups differ
even though they are subject to the same g/Bhp-hr standard for
1988 and later and similar in-use g/Bhp-hr emissions prior to
1988. The conversion of g/Bhp-hr to grams/mile is dependent on
the engine brake-specific fuel consumption (BSFC, Ib/Bhp-hr),
the fuel density (Ib/gallon) and the fuel economy (mile/gallon)
of the particular vehicle/engine configuration. Since the BSFC
and fuel economy varies among the different heavy-duty
subgroups, estimated g/mile emissions vary as well. Conversion
factors for the heavy-duty subgroups by model year were based
on the information provided in reference 10. These conversion
factors are consistent with those used in the MOBILE3 emissions
model.[11]
2.3.3 Emission Factors for Calendar Years 1986 and 1995
The next step is to combine the model year emission
factors for each vehicle type into a single, weighted calendar
year emission factor for each vehicle type. Model year data
for the previous 20 years are used, i.e., for 1986, model year
data back to 1967 are used; for 1995, model year data back to
1976 are used. Each model year's emission factor is multiplied
-6-
-------
Table 2-2
Diesel Particulate Emission Factors
Light-Duty Vehicles
grams per mile
Model Year
1978 and prior
1979
1980
1981-1986
1987 and later
Model Year
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
Cars
0.70
0.80
0.50
0.27
0.20
Heavy-Duty Engines*
grams per mile
2B 3-5 6
Trucks
0.90
0.90
0.50
0.28
0.26
7-8
1
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.99
.98
.98
.97
.97
.73
.73
.73
.72
.58
.57
.57
.24
.24
.24
.09
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
.71
.71
.71
.71
.71
.71
.71
.71
.71
.71
.71
.71
.53
.34
. 16
.97
.97
.73
.73
.73
.72
.58
.57
.57
.24
.24
.24
.09
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
.71
.73
.75
.75
.79
.82
.84
.85
.86
.87
.87
.86
.86
.86
.86
.87
.85
.37
.36
.34
.33
.06
.06
.06
.44
.44
.44
.19
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
2
2
2
2
1
1
1
0
0
0
0
.99
.02
.06
.09
.14
.20
.24
.29
.33
.24
.28
.23
.18
. 16
.13
.09
.08
.30
.29
.28
.28
.81
.79
.77
.69
.69
.67
.30
0.09
0.09
0.18
0.29
See text for explanation of the g/Bhp-hr emission factors
used.
-7-
-------
by that model year's fraction of calendar year VMT and the
diesel sales fraction for that model year, and then summed
across all 20 model years.
The fraction of travel by model year for the light-duty
vehicles and trucks is given in Table 2-3. These travel
fractions are dependent on vehicle age rather than model year;
therefore, the same travel fractions were used to determine
both the 1986 and 1995 emission factors. These travel
fractions were obtained from the MOBILES emissions model.[11]
The fraction of travel by model year for the heavy-duty
subclasses for calendar years 1986 and 1995 are given in Tables
2-4 and 2-5, respectively. For 1986, the travel fractions for
heavy-duty gasoline-fueled trucks were used for subclasses 2B,
3-5 and 6. The travel fractions for heavy-duty diesels were
used for subclasses 7-8, since this combined subclass is
dominated by heavy-duty diesels. For 1995, the travel
fractions for heavy-duty gasoline-fueled trucks were used for
subclasses 2B and 3-5; travel fractions for heavy-duty diesels
were used for subclasses 6 and 7-8. Unlike the heavy-duty
gasoline-fueled trucks, the mileage distributions and resulting
travel fractions for heavy-duty diesels are dependent on model
year and reflect the increasing penetration of diesels in the
lower mileage, lighter weight classes of the heavy-duty truck
category. These travel fractions were calculated based on the
information provided in Reference 11.
Projecting future diesel sales is rather uncertain.
Light-duty diesel sales in particular have dropped off quite
dramatically and should fall well short of projections made
just a few years ago that light-duty diesels would account for
25 percent of light-duty vehicle sales in 1995. To account for
these uncertainties, low and high diesel sales scenarios were
developed. Both scenarios incorporate actual sales data to the
extent available.
For light-duty vehicles, actual diesel sales fractions up
to and including 1985 are available and were used. [12,13] For
1986 and beyond, the low diesel sales scenario assumes that the
light-duty vehicle diesel sales fraction will remain .constant
at the 1985 level (0.009). The high diesel sales scenario uses
MOBILES projections of diesel sales for 1995 (0.009 in 1986
with a gradual increase to 0.115 in 1995).
For light-duty trucks and heavy-duty subclass 2B, actual
diesel sales fractions through 1984 were used.[12,13] For 1985
and beyond, the low diesel- sales scenario assumes that the
diesel sales fraction will remain constant at the 1984 level.
The high diesel sales scenario uses the latest EPA/OMS
projections for 1985-1995.[14]
-8-
-------
Table 2-3
Fraction of Travel by Model Year for
Light-Duty Vehicles and Trucks
Vehicle LDV LDT
Age Travel Fraction Travel Fraction
1 .038 .035
2 .142 .129
3 .125 .114
4 .111 .101
5 .098 .088
6 .084 .078
7 .075 . .067
8 .065 .058
9 .055 .050
10 .047 , .043
11 .040 .037
12 .032 .031
13 .026 .026
14 .021 .022
15 .015 .018
16 .011 .015
17 .007 .012
18 .003 .009
19 .003 .006
20+ .004 .009
-9-
-------
Table 2-4
Fraction of Travel by Model Year for
Heavy-Duty Vehicles in Calendar Year 1986
Model
Year
1986
1985
1984
1983
1982
1981
1980
1979
1978
1977
1976
1975
1974
1973
1972
1971
1970
1969
1968
1967
2B, 3-5
and 6
.000
.227
.175
.134
.105
.080
.062
.049
.037
.028
.023
.017
.013
.010
.009
.006
. 005
.004
.003
.013
7-8
.000
.255
. 186
.136
.105
.078
.057
.043
.033
.026
.019
.014
.011
.008
.006
.005
.004
.003
.002
.009
-10-
-------
Table 2-5
Fraction of Travel by Model Year for
Heavy-Duty Vehicles in Calendar Year 1995
Model
Year
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
1978
1977
1976
2B, 3-5
.000
.227
.175
. 134
.105
. 080
.062
.049
.037
.028
.023
.017
.013
.010
. 009
.006
.005
.004
.003
.013
6
.000
.292
.208
.146
.105
.074
.052
.037
.026
.019
.013
.008
.006
.004
.002
.002
.001
.001
.001
.003
7-8
.000
.272
.199
.144
. 106
.076
.055
.040
.029
.021
.015
.011
.008
.006
.004
.003
.002
.002
.001
.006
-11-
-------
The diesel sales fractions for heavy-duty classes 3-5, 6
and 7-8 were based on actual data through 1985. For 1986 to
1995, the latest EPA/OMS projections were used.[14] These
sales fractions were used for both the low and high diesel
sales scenarios.
The diesel sales fractions for the light-duty vehicles and
trucks are presented in Table 2-6. The diesel sales fractions
for the heavy-duty vehicle classes are given in Table 2-7.
The resulting weighted emission factors for calendar years
1986 and 1995 are given below:
(g/mile)
HDV
LDV LPT 2B 3-5 6_ 7-8
Calendar Year 1986
Low Diesel Sales .008 .0105 .089 .040 .479 2.35
High Diesel Sales .008 .0105 .091 .040 .479 2.35
Calendar Year 1995
Low Diesel Sales .003 .008 .061 .082 .330 .844
High Diesel Sales .013 .041 .084 .082 .330 .844
2.3.4 Nationwide Diesel Particulate Emissions
In this section, nationwide diesel particulate emissions
(metric tons/year) for 1986 and 1995 are calculated by
combining the calendar year emission factors with projected
vehicle miles traveled (VMT) data for 1986 and 1995.
The VMT fractions and resulting projected VMT for 1986 and
1995 is given in Table 2-8. These data were obtained from the
MOBILE3 Fuel Consumption Model. [15] In this model, heavy-duty
subgroups 2B and 3-5 were treated as one subgroup. To split
the VMT fraction for these subgroups, data from EEA's Motor
Fuel Consumption, Tenth Periodical Report were used.[16] Total
VMT is projected to increase 19 percent from 1986 to 1995.
The projected nationwide diesel particulate emissions for
1986 and 1995 are given in Table 2-9. As seen in this table,
nationwide diesel particulate emissions are projected to
decrease roughly 44 to 54 percent from 1986 to 1995, despite
the 19 percent projected increase in VMT. This is due to the
rather small projected further infiltration of the fleet by
diesels (with the exception of the heavier heavy-duty
subclasses 7-8) together with increasingly stringent standards
in future years.
-12-
-------
Table 2-6
Low and High Diesel Sales Fractions by
Model Year for Light-Duty Vehicles and Trucks
Model
Year
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
1978
1977
1976
1975
1974
1973
1972
1971
1970 +
Light-Duty Vehicles*
Low High
.009 .115
.009 .102
.009 .089
.009 .076
.009 .063
.009 .050
.009 .044
.009 .034
.009 .019
.009 .009
.009
.023
.019
.039
.060
.045
.026
.009
.003
.003
.003
.003
.002
.002
.001
.000
Light-Duty Trucks**
Low High
.026 .339
.026 .300
.026 .263
.026 .226
.026 .188
.026 .150
.026 .119
.026 .088
.026 .057
.026 .026
.026 .026
.026
.042
.092
.082
.048
.015
.010
.005
.003
.002
.000
.000
.000
.000
.000
*1970-1985 sales fractions are based on actual data and are the
same for both scenarios. These fractions are not repeated in
the high sales column in order to highlight the differences
between scenarios.
*1970-1984 sales fractions are based on actual data and are the
same for both scenarios.
-13-
-------
Table 2-7
Low and High Diesel Sales Fractions by
Model Year for Heavy-Duty Vehicles
Model
Year
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
1978
1977
1976
1975
1974
1973
1972
1971
1970
1969
1968
1967
2B
Low
.190
.190
.190
.190
.190
.190
. 190
. 190
. 190
.190
. 190
. 190
.153
.127
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
High
.300
.300
.300
.300
.286
.273
.259
.245
.231
.218
.204
3-5
Best
.300
.300
.300
.300
.291
.282
.273
.265
.256
.247
.239
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
.000
6
Best
.712
.693
.-674
.655
.635
.616
.597
.578
.559
.540
.521
.382
.312
.289
.322
.183
. 114
.078
.070
.042
.032
.016
.016
.016
.015
.016
.000
.000
.000
7-8
Best
.883
.882
.879
.878
.877
.875
.874
.872
.871
.868
.867
.849
.862
.845
.881
.871
.820
.815
.878
.823
.733
.770
.780
.760
.750
.750
.750
.750
.750
-14-
-------
Table 2-8
VMT Fractions and VMT for 1986 and 1995
VMT Fractions
1986
1995
1986
1995
LDV
1098.
1298.
LDV
.675
.671
53
75
LDT
.225
.227
Projected
LDT
365.25 35
438.59 49
2B
.022
.025
VMT (
2B
. 16
. 19
HDV
3-5 6 7-8
.004 .009 .065
.003 .006 .068
9
10 miles)
HDV
3-5 6 7-8
6.11 15.34 106.48
5.68 10.64 131.48
TOTAL
1626.87
1934.33
-15-
-------
Table 2-9
Nationwide Diesel Particulate Emissions
(metric tons per year)
1986 1995
Low Sales High Sales Low Sales High Sales
LDV 8,788 8,788 3,896 16,884
LOT 3,835 3,835 3,509 17,982
HDV
2B 3,129 3,200 3,001 4,132
3-5 244 244 466 466
6 7,348 7,348 3,511 3,511
7-8 250,228 250,228 110,969 110,969
TOTAL 273,572 273,643 125,352 153,944
-16-
-------
2.3.5 Other Sources of Particulate
To put the preceding emission estimates into perspective
the results are compared to national particulate emission
estimates for 1984. [17] 1984 is the most recent year for which
data are available.
Total particulate emissions in the U.S. in 1984 were
projected to be 7.0 million metric tons. In comparison, diesel
particulate emissions in 1986 were projected to be roughly
274,000 metric tons, or 3.9 percent of the total 1984
emissions. Transportation sources in 1984 were estimated to
contribute 18.6 percent of the total, or 1.3 million metric
tons. Diesel particulate emissions in turn currently appear to
account for about 21 percent of the emissions from
transportation sources.
2.4 Ambient Concentrations of Diesel Particulate
In this section, urban and rural concentrations of diesel
particulate are estimated for 1986 and 1995, using a modified
version of the EPA NAAQS Exposure Model (NEM) for CO. These
concentrations are then compared to previous exposure estimates
and to monitoring data for total suspended particulate.
2.4.1 Estimation of Urban and Rural Exposure
The model used provides an estimate of nationwide annual
person-hours of exposure to any non-reactive mobile source
pollutant of interest.[ 18] It is based upon the NEM developed
originally by EPA's Office of Air Quality Planning and
Standards (OAQPS). The NEM approach relies on an activity
pattern model that simulates a set of population groups called
cohorts as they go about their day-to-day activities. Each of
these cohorts are assigned to a specific location type during
each hour of the day. Each of several specific location types
in the urban area are assigned a particular ambient pollutant
concentration based on fixed site monitor data. The model
computes the hourly exposures for each cohort and then sums up
these values over the desired average time to arrive at average
population exposure and exposure distributions. Annual
averages are possible because a full year's data from fixed
site monitors is an input to the model.
The NEM approach was designed to determine an integrated
exposure to a pollutant by estimating both indoor and outdoor
exposures to the pollutant. It was not designed to determine
exposure from a particular source, such as mobile sources.
Also, because its basic time unit is an hour, it does not
account well for short periods spent in locations with high
exposure such as an on-road vehicle. Hence, Southwest Research
Institute (SwRI), under EPA contract, developed a new model
based on the NEM for CO which could be used to better determine
-17-
-------
exposures resulting specifically from mobile source
pollutants. [18] The CO NEM was used since outdoor CO is almost
exclusively mobile source related. Since the CO monitor data,
on which the CO NEM was based, can be assumed to be related to
mobile source emission rates, exposure to other non-reactive
mobile source pollutants can be modeled using this
relationship.
The CO monitor data are used to provide CO concentration
data for each neighborhood and most of the microenvironments.
In each neighborhood or microenvironment, CO emission factors
(in grams/minute) are chosen which are felt to best represent
vehicle conditions in that neighborhood/microenvironment. For
example, in an urban commercial neighborhood, an emission
factor at 10 mph steady state is chosen since this is thought
to best represent the vehicle conditions in this neighborhood.
The emission factor is a fleet average emission factor, thus
weighting emissions from both light-duty and heavy-duty
vehicles. The model ratios the CO concentrations and
appropriate CO emission factors for each
neighborhood/microenvironment, so each
neighborhood/microenvironment contains a ug/m3/grams/minute
factor. Emission factors in grams/minute for the pollutant of
interest for each neighborhood/microenvironment are input to
the model. The model simply multiplies the input emission
factor (grams/minute) by the factor (ug/m3/grams/minute) to
obtain concentrations in each neighborhood/microenvironment for
the pollutant of interest.
It should be noted that indoor concentrations (and
therefore, exposure) due to ambient mobile source pollutants
are also accounted for in the model. A scaling factor of 0.85
was applied to the appropriate neighborhood CO monitor data to
estimate indoor exposures to the pollutant of interest in each
neighborhood. The scaling factor was based on comparisons of
indoor and outdoor CO levels of homes with no indoor CO
sources (e.g., gas stove, smokers).
The model does not account for photochemical reactions.
The exposure levels predicted by the model are those resulting
from direct exhaust emissions, and do not account for either
the destruction or photochemical formation of the pollutant in
the atmosphere. The model also assumes that the pollutant of
interest has emission formation and dispersion characteristics
similar to that of CO.
In the SwRI model, the relatively insignificant indoor CO
exposures were set to zero. Exposures in three mobile source
microenvironments (street canyons, tunnels and parking
garages), where elevated concentrations of mobile source
pollutants could be experienced, were added to the OAQPS
version of the model. Finally, a national extrapolation
procedure designed expressly for mobile sources was devised.
-18-
-------
There are three required inputs. The first is the
nationwide urban and rural populations for the year of
interest. The second is a series of 33 emission factors which
are described in Table 2-10. The third input is a list of 25
ambient pollutant concentrations, specifying the 24
concentration intervals or bins for which cumulative
person-hours of exposure are to be calculated.
The output lists the total annual person-hours of exposure
found in each of the specified concentration intervals. Using
this information, a mean exposure level may be calculated.
Urban and rural populations for 1986 and 1995 were
estimated based on U.S. Department of Commerce data and are
given below.[19]
Urban Rural Total
1986 180,000,000 60,000,000 240,000,000
1995 195,000,000 65,000,000 260,000,000
As seen in Table 2-10, the emissions input required for
the model are FTP emission factors and an emission factor at an
average speed of 10 mph, expressed in grams/minute. For the
tunnel microenvironment and rural areas, a 35 mph steady state
emission factor will be used. For light-duty vehicles, an idle
emission factor is also required. As mentioned previously,
these emission factors were chosen to best represent the
vehicle operating conditions in each neighborhood/
microenvironment.
The weighted emission factors given in Section 2.3.3 are
considered FTP (or transient) emission factors. VMT fractions
are then used to determine FTP emission factors for each
vehicle class and a composite emission factor for the fleet.
The LDV/LDT FTP has an average speed of 19.6 mph; this average
speed was used for both LDV/LDT and HDV to convert the g/mile
factor to g/min, as required for the model. For LDV/LDT this
is certainly correct. For HDV it is less clearly so but is a
reasonable approximation.
The VMT fractions in Table 2-8 represent VMT fractions
nationwide. Urban and rural VMT fractions differ, particularly
for the heavy-duty vehicle classes 7 and 8. For these
heavy-duty subclasses, rural VMT is estimated to exceed urban
VMT by a factor of 2.7.[15] Since these heavy-duty subclasses
are responsible for the majority of diesel particulate
emissions, urban and rural VMT fractions will be used. The
urban and rural VMT fractions for each vehicle class are given
below.[15]
-19-
-------
Table 2-10
Model Emission Factor Inputs
Neighborhood or Microenvironment Assumed Emission Factor (g/min)
Weekday urban residential
Weekday urban commercial
Weekday urban industrial
Weekday suburban residential
Weekday suburban commercial
Weekday suburban industrial
Weekday street canyon
Weekday tunnel
Weekday parking garage
Dummy mobile source emission factor
Weekday rural
Saturday urban residential
Saturday urban commercial
Saturday urban industrial
Saturday suburban residential
Saturday suburban commercial
Saturday suburban industrial
Saturday street canyon
Saturday tunnel
Saturday parking garage
Dummy mobile source emission factor
Saturday rural
Sunday emission factors
FTP
10 mph
10 mph
FTP
FTP
FTP
10 mph
35 mph
0.5 times idle factor plus 0.5
times 10 mph (LDV only)
Use tunnel factor as dummy (to
allow for future expansion)
35 mph
0.72* times weekday FTP factor
0.72 times 10 mph factor
0.72 times 10 mph factor
0.72 times FTP factor
0.72 times FTP factor
0.72 times FTP factor
0.72 times 10 mph factor
35 mph
0.5 times idle plus 0.5 times
10 mph (LDV only)
Use tunnel factor as dummy
0.72 times 35 mph factor
In current version of program,
Sunday emission factors are
equal to Saturday factors.
* The factor of 0.72 is used to adjust for relative traffic volume.
-20-
-------
VMT Fractions
HDV
LDV LOT 2B 3-5 6 7-8
1986
Urban .721 .206 .026 .005 .010 .032
Rural .617 .247 .016 .003 .008 .109
1995
Urban .717 .209 .031 .004 .006 .033
Rural .613 .250 .018 .002 .005 .112
To determine the 10 mph and 35 mph emissions factors and
the idle emission factor for the light-duty diesel vehicle
fleet, it was necessary to calculate ratios of emissions at
these speeds to FTP emissions. For the light-duty vehicles,
data from the New York City Cycle (NYCC), with an average speed
of 7.07 mph, were used to represent the 10 mph cyclic emission
factor. The results of two studies were used which
collectively contained data at idle, steady state speeds of 31
mph, 50 mph, and 53 mph as well as FTP and NYCC data. [20,21]
Ratios of emissions at the steady state speeds (g/min) to FTP
emissions (g/min) were plotted and found to increase
exponentially. The ratio at 35 mph was estimated from the plot.
For heavy-duty vehicles, emissions data from three
in-service buses over various chassis dynamometer driving
cycles were used to calculate the 35 mph steady state ratio for
the tunnel microenvironment and urban areas.[22] Emissions
data were taken at idle, steady state speeds of 12.5 mph and 25
mph and over the heavy-duty chassis driving cycle. To obtain
the ratio at 35 mph, it was assumed that the ratio from 25 mph
to 35 mph would continue to increase linearly. Data were not
available at higher speeds to determine whether the ratios
would increase exponentially at some point; therefore, the
linear assumption may result in a slight underprediction of the
35 mph ratio.
Very little data exist to calculate cyclic emission
factors at 10 mph for heavy-duty vehicles. Recent EPA test
programs have focused on emissions characterization of in-use
transit buses. These programs involved buses which were
temporarily removed from operating service and which were
tested without additional maintenance in their chassis
configurations over test cycles designed specifically to
simulate transit bus operation. The test cycles have average
speeds ranging from 8.8 to 12.4 mph. Information regarding the
test programs can be found in reference 23. An overall average
transit bus emission factor was estimated to be 5.52 g/mile, or
roughly 0.92 g/min, based on an average speed of 10 mph. A 10
mph/FTP ratio was estimated such that the resulting emission
factor would equal 0.92 g/min, based on the 1986 FTP data.
-21-
-------
This ratio will be used for both 1986 and 1995 to calculate the
heavy-duty contribution in the street canyon neighborhood,
where public exposure to transit bus emissions is relatively
high.
For the remaining urban industrial and commercial
neighborhoods, a 10 mph cyclic emission factor for heavy-duty
trucks is required. For these neighborhoods, the composite
heavy-duty FTP g/mile emission factor will be used. The ratio
at 10 mph is thus equal to the ratio at this speed to the FTP
average speed of 19.6 mph. The resulting ratio at 10 mph is
0.51.
The resulting ratios for the light-duty and heavy-duty
vehicles are given in Table 2-11. These ratios were applied to
the FTP g/min calculated for each vehicle class and then summed
to obtain VMT-weighted 10 mph and 35 mph g/min composite
emission factors. The emission factor inputs for the 1986 low
and high sales scenarios are the same. The reason is that the
sales fractions used for the scenarios only differ for the 1985
and 1986 model years. This difference was found to be
negligible when computing the model emission factor inputs.
It should be noted that the absolute emission factors and
the assumptions used to estimate the off-FTP values directly
influence the estimated absolute particulate exposures in 1986
and 1995, but that the percent decrease in particulate exposure
from 1986 to 1995 is sensitive only to the more reliable FTP
emission factors for the two years.
The final model input required is a list of 25 ambient
pollutant concentrations defining the concentration intervals
for which cumulative person-hours of exposure are calculated.
The chosen concentrations are given in Table 2-11 and range
from 0.00 to 50.00 ug/m3 for 1986 and 0.00 to 31.47 ug/m3
for 1995.
The modified NEM does not take into account projected
increases in VMT. Since most of the modified NEM is based on
1981 monitoring data, projected VMT for 1986 and 1995 will be
compared to 1981 VMT and the exposures predicted by the
modified NEM adjusted upward accordingly. Since VMT from 1981
to 1986 is expected to increase 10 percent, the 1986 exposures
were adjusted upward by 1.10. Similarly, since VMT from 1981
to 1995 is expected to increase 31 percent, the 1995 exposures
were adjusted upward by a factor of 1.31. [15] This VMT
adjustment is likely to be somewhat conservative on top of the
adjustment that is made for population growth, since some of
both types of growth will be outward at urban fringes, rather
than upward in the areas of current population and VMT
concentration. The caveat applies to many of the other
pollutants, whose exposure is estimated in the same manner.
-22-
-------
Table 2-11
Diesel Particulate Ratios and Model Inputs
Ratios
g/min/g/min
Idle/FTP 10 mph/FTP 35 mph/FTP
LDV/LDT 0.33 0.64 0.79
HDV N/A* 0.51** 0.63
25 Pollutant Concentrations (ug/m3)
1986
0.00, 0.50, 0.61, 0.75, 0.91, 1.11, 1.36, 1.66, 2.00, 2.50,
3.00, 3.70, 4.50, 5.50, 6.80, 8.20, 10.10, 12.30, 15.00,
18.40, 22.00, 27.00, 33.00, 41.00, 50.00
1995
0.00, 0.13, 0.16, 0.21, 0.26, 0.33, 0.42, 0.54, 0.68,
0.86, 1.10, 1.40, 1.77, 2.25, 2.87, 3.64, 4.63, 5.88,
7.47, 9.50, 12.07, 15.34, 19.49, 24.77, 31.47
*Anidle emission factor is used to simulate emissions in a
parking garage. It is assumed there are no HDV in parking
garages; therefore, a HDV idle/FTP ratio is not applicable.
** For the street canyon microenvironment, a 10 mph/FTP ratio
of 2.50 is used. For 1986, this ratio results in an
emission factor of 0.92 g/min, equal to a current
estimated transit bus emission factor. The same ratio is
applied in 1995.
-23-
-------
The mean exposure levels predicted by the model, adjusted
to take into account projected increases in VMT, are given
below.
Exposure (ug/m3)
Urban Rural Nationwide
1986 2.63 2.38 2.56
1995 Low Sales 1.27 1.06 1.22
1995 High Sales 1.69 1.27 1.58
As seen above, the rural exposures are quite similar to
the urban exposures. This is due to the greater fraction of
heavy-duty vehicle classes 7-8 in rural areas, resulting in a
higher composite emission factor. In addition, the modified
NEM calculates rural exposures quite crudely. This is a
weakness of the model and should be taken into consideration
when reviewing the results.
2.4.2 Comparison to Other Exposure Estimates
This section compares the mean exposure levels predicted
by the modified NEM to three other diesel particulate exposure
estimates made previously. These are: 1) diesel particulate
exposures estimated by OMS in a somewhat similar manner in
1983, 2) an estimate made by the Lovelace Inhalation Toxicology
Research Institute (ITRI) in 1981, and 3) exposures estimated
by OMS in 1983 using a lead surrogate approach. The comparison
among these methods also holds for other pollutants in this
report for which the modified NEM was used.
2.4.2.1 1983 QMS Exposure Estimate
In 1983, OMS projected diesel particulate exposure in
urban areas for 1995. [8] OMS used the original CO NEM, with
the only modification being the removal of all indoor sources.
The annual average CO exposure was 2.12 ppm. An average annual
exposure for diesel particulate for 1995 was then estimated by
ratioing CO and diesel particulate emission factors and
multiplying the result by 2.12 ppm. The CO emission factor was
62.3 g/mile. This is the CO national average emission factor
for 1978, which is the same year as the CO NEM data base. The
composite diesel particulate emission factor used (for the best
estimate base sales scenario) was 0.0554 g/mile. Since VMT
from 1978 to 1995 was expected to increase 45 percent, the
diesel particulate emission factor was adjusted upward by 1.45.
-24-
-------
The projected exposure in 1995 for the best estimate base
sales scenario is 1.5-1.6 ug/m3 from LDDV and 1.6-2.1 ug/mj
for HDDV. The total projected exposure is 3.1-3.7 ug/m3.
This is considerably higher than the urban exposure range of
1.3-1.7 ug/m3 estimated in this study. The urban exposure of
1.7 ug/m3 estimated in this study corresponds to an FTP
composite diesel particulate emission factor of 0.053 g/mile.
This is quite similar to the composite emission factor
calculated in 1983. It appears that the modified NEM
calculates lower exposures per g/mile that the CO NEM used
previously. The modified NEM does include more recent
monitoring data, including an expanded number of monitors,
relative to the CO NEM, which may account for the difference.
2.4.2.2 1981 Lovelace Inhalation Toxicology Research Institute
(ITRI) Exposure Estimate
ITRI's exposure estimate was confined to LDDV. ITRI used
a particle dispersion model to project future atmospheric
concentrations of LDDV particulate in urban and rural areas of
the U.S. in 1995.[24,25] The atmosphere is represented by a
grid of cells with variable heights that move with the velocity
and direction of wind. Atmospheric concentrations of particles
were calculated by sweeping the grid of cells, beginning on the
upwind side of the urban area and passing them across the
city. The model accounts for emissions as well as turbulent
mixing, particle diffusion, particle settling and particle
resuspension. Assumptions include: 1) the population was
distributed with the same density as diesel vehicles, 2) the
typical U.S. city has 40 percent of the population in the city
center during work hours and has an average wind speed of 5
m/sec, and 3) particles entering a cell are assumed to be mixed
uniformly throughout the cell.
Information used included the current (1981) land areas,
populations and gasoline consumption rates for all major U.S.
cities and standard metropolitan statistical areas. These data
were scaled to represent conditions after 1995 when 20 percent
of all LDV were assumed to be diesel-powered. ITRI projected
average concentrations of LDDV particulate based on assumed
LDDV emission rates of 0.2 g/mile and 0.5 g/mile. This
comparison will be limited to the results obtained with the 0.2
g/mile emission rate, since this emission rate agrees with that
used in this study and the previous QMS study.
Average concentrations of LDDV particulate in urban and
rural areas were projected to be 0.2 ug/m3 and 0.02 ug/m3,
respectively, based on an emission rate of 0.2 g/mile. In
addition to projecting these "background" atmospheric
concentrations in urban and rural areas, calculations were also
made of the higher concentrations that may occur near urban
street canyons and expressways. The exposure estimates
developed by ITRI are given below:
-25-
-------
Particulate Exposure (ug/m3)
Based on 0.2 g/mile
Typical urban resident 0.2
Urban residents near freeway 2.0
Workers on urban freeway 15.0
Workers in urban street canyon 15.0
The exposures in urban areas and near street canyons and
expressways were weighted, with the bulk of the resulting total
exposure being due to the urban background, and used in ITRI's
risk assessment. The composite LDDV urban particulate exposure
was not given by ITRI but has been calculated by EPA to be
roughly 0.50 ug/m5.
The exposure estimate projected by ITRI is lower than the
estimates projected in this study; however, the different
assumptions used in both studies make a direct comparison
difficult. ITRI assumed 20 percent of all LDV were diesels.
In comparison, this study assumed 0.9-11.5 percent of LDV sales
were diesels. ITRI therefore assumed a much greater
penetration of LDDV. Heavy-duty vehicles were not considered
by ITRI. In this study, heavy-duty vehicles are responsible
for 66-90 percent of the 1995 urban exposure depending on
diesel sales. (These percentages were derived using the g/mile
emission factors in Section 2.3.3 together with the 1995 urban
VMT fractions in Section 2.4.1.)
2.4.2.3 Exposure Estimate Based on a Lead Surrogate Approach
For comparison, in 1983, EPA also used 1975 atmospheric
lead monitoring data as a surrogate to estimate atmospheric
levels of diesel particulate in 1995.[8] Estimates were
provided of ambient diesel particulate concentrations at one or
two particular monitor locations in a large number of U.S.
cities. The monitors were chosen in areas having no large
stationary sources of lead.
For this analysis, an estimate was made of the fleet's
automotive lead emission factor which caused the observed
ambient lead levels, and is compared to the expected diesel
particulate emission factor. The 1995 projected ambient diesel
particulate concentration in 1995 was set equal to the urban
ambient lead concentration in 1975 multiplied by the ratios of
1995 diesel particulate to 1975 lead emission factors,
dispersion factors and VMT. Even though the 1995 fleet average
(best estimate sales) diesel particulate emission factor is
less than half the 1975 fleet average lead emission factor,
ambient concentrations of diesel particulate in 1995 were
projected to be 1.63 times the urban ambient lead
concentrations in 1975. This is due to the projected increase
-26-
-------
in VMT (60 percent) in 1995 and the higher estimated dispersion
factor for diesel particulate relative to lead (1.0 versus
0.43). Resulting ambient diesel particulate concentrations
varied from 0.7 to 3.9 ug/m3, depending on the city and
particular monitor. The estimates projected in this study, as
well as the earlier QMS estimates, fall within this range.
2.4.3 Comparison to Particulate Monitoring Data
The annual geometric mean total suspended particulate
(TSP) concentration in 1984 was 50 ug/m3. [17] It has
remained fairly constant at this level since 1982. The
nationwide average diesel particulate concentration in 1986
calculated above is roughly 5.1 percent of the 1984 total.
This percentage is similar to that estimated for emissions (3.9
percent) in Section 2.3.5.
o denotes particulate matter less than 10 microns in
size. Since virtually all diesel particulate falls under
PM10, it would be useful to estimate the contribution of
diesel particulate to PM10 concentrations. A PMi0/TSP
ratio of 0.50 is used, based on an examination of 1982-1983
monitoring data. [26] Applying this ratio to the 1984 TSP
concentration results in an annual mean PMio concentration of
25 ug/m3. Thus, diesel particulate emissions appear to
account for roughly 11 percent of the annual mean PMio
concentrations. It should be pointed out, however, that this
comparison is probably misleading since the TSP and PM10 data
are all at fixed site monitors. At these sites, the diesel
levels are probably much higher than the mean personal
estimates given in this report.
2. 5 Health Effects of Diesel Particulate and Unit Risk
Estimates
The mutagenic activity of organic extracts from diesel
particulate was first reported in 1979. At that time, the
limited data available from animal and epidemiology studies
were not sufficient for a cancer risk assessment. As a result,
EPA conducted a large research program to evaluate the health
effects associated with exposure to diesel emissions.
A comparative potency method was used to determine a unit
risk estimate for diesel particulate. In this method, the
potency of the organic extracts from diesel particulate are
compared with the potencies of extracts from sources for which
epidemiological data are available. A large number of
mutagenesis and carcinogenesis studies were performed to
determine relative potencies. The methodology and results are
discussed in detail in reference 7. The unit risks for the
light-duty diesel particulate sources ranged from
-27-
-------
2. OxlO"s-3.5x10"s. These represent the means of the lower
and upper 95 percent confidence limits. The upper confidence
limit unit risks ranged from 3.3x10"s-6.3xlO~s.
Two other analyses of EPA's comparative potency data were
performed by Dr. Harris for NAS [27] and by the Lovelace
Inhalation Toxicology Research Institute (ITRI).[24] Harris1
mean estimate was a 0.0035 percent proportional increase in
risk per year per ug/m3 exposure assuming lifetime exposure.
Harris' relative risk was translated into an absolute measure
of lung cancer incidence using the methodology described in
reference 28. The result is a unit risk of IxlO"4. Harris'
upper confidence limit estimate (0.0252 percent proportional
increase in risk) is roughly equivalent to a unit risk of 7.5 x
lO'4.
ITRI calculated a range of unit risks which differ
depending on which comparative source was used. The resulting
range was 4.9xlO~s to 2.1x10"4. These risks appear to be
upper confidence limits. ITRI chose 7.OxlO"5 as being most
representative.
The range of unit risks that will be used in this paper is
2. OxlO"s to 1.0 x 10"4. Assuming an average lifetime of 70
years, the range of risk estimates, on an annual basis, is 0.29
x 10~6 to 1.4 x 10 "6 lung cancers per person per ug/m3
particulate. The lower end of the range is the lowest EPA
estimate. The upper end of the range is Harris' mean estimate
which is also in rough agreement with the ITRI estimate. The
range incorporates both EPA's and ITRI's upper confidence limit
estimates. Harris' upper confidence limit estimate was not
included due to the uncertainty of translating his relative
risk into an absolute measure of risk.
The comparative potency studies that are the basis of the
risk estimates used the organics extracted from diesel
particulate. Inhalation studies using whole diesel emissions
(i.e., particulate and gas phase emissions) were also performed
concurrently but the results were negative or inconclusive.[29]
Long-term animal inhalation studies are presently being
conducted by ITRI, Fraunhofer, Battelle-Geneva and
Battelle-Northwest. In contrast to the previous studies,
preliminary results indicate that lung tumors have been found
at concentrations no higher than those tested with negative
results in the previous studies. Reports are expected shortly.
2.6 Current and Projected Health Risk
The annual risk estimates are combined with the exposure
estimates (given in Section 2.4.1) and population estimates to
obtain estimates of lung cancer incidence for 1986 and 1995.
The results are given in Table 2-12.
As seen in this table, the total risk in 1986 ranges from
178-860 cancer incidences and drops from 33-48 percent in 1995,
depending on projected diesel sales.
-28-
-------
Table 2-12
Annual Lung Cancer Risk from Diesel Particulate Exposure
1986
1995 Low Sales
1995 High Sales
Urban
137-661
72-346
95-461
Rural
41-199
20- 97
24-115
Total
178-860
92-443
119-576
-29-
-------
3.0 FORMALDEHYDE
3.1 Formation, Composition and Control Technology
Formaldehyde is the most prevalent aldehyde in vehicle
exhaust and is formed as a result of incomplete combustion of
the fuel. Formaldehyde is emitted in the exhaust of both
gasoline and diesel-fueled vehicles. It is not a component of
evaporative emissions. Formaldehyde is of interest due to its
photochemical reactivity and suspected carcinogenicity.
Formaldehyde has the chemical formula HCHO. Its structure
is shown below.
H
H - C = 0
Use of a catalyst has been found to be effective for
controlling formaldehyde emissions. Formaldehyde emissions are
controlled to roughly the same extent as total hydrocarbon
emissions with a catalyst.
3.2 Emissions
In this section, formaldehyde emission factors (g/mile)
for calendar years 1986 and 1995 are presented. Nationwide
mobile source formaldehyde emissions (metric tons/year) for
1986 and 1995 are then calculated by combining the emission
factors with VMT data for 1986 and 1995. The results are then
compared to estimates of nationwide total formaldehyde
emissions.
3.2.1. Emission Factors for Calendar Years 1986 and 1995
For this analysis, formaldehyde emissions for the various
classes are expressed as a percentage of total hydrocarbons.
These percentages were then applied to MOBILES exhaust
hydrocarbon (HC) output for 1986 and 1995 to obtain the
formaldehyde emission factors. This was done for two reasons:
1) formaldehyde emissions can vary considerably within a
vehicle class but are more consistent when expressed as a
percentage of total hydrocarbons, and 2) use of the MOBILE3 HC
output should more accurately represent in-use emissions since
deterioration and effects of malfunction and
tampering/misfuel ing are accounted for in MOBILES.
An important issue to address is whether the percentages
chosen are adequate to use for the excess hydrocarbons that
come from deterioration, malfunction and tampering/misfueling.
Effects of deterioration, malfunction and tampering/misfueling
have been studied to the greatest extent with LDGV; therefore,
LDGV data are examined here.
-30-
-------
The percentages for LDGV were based on data for both new
vehicles and in-use low and high mileage vehicles. The
percentages do, therefore, factor in the effects of
deterioration and minor malfunctions. The percentages are
remarkably similar for non-catalyst-equipped vehicles as well
as vehicles equipped with oxidation catalysts and 3-way plus
oxidation catalysts. Overall, formaldehyde emissions varied
from roughly 1-3 percent of total HC while HC emissions varied
from roughly 0.2 -6.0 g/mile. Even within a particular vehicle
category, wide variations in HC emissions generally have little
effect on the formaldehyde percentages. For example, HC
emissions from two in-use oxidation catalyst-equipped vehicles
varied from 0.4 g/mile to 4.7 g/mile, yet the formaldehyde
percentage for both vehicles was 1.5 percent. [303 It is true
that the formaldehyde percentage itself is subject to some
variation, but the variation is slight compared to the
variations in the absolute level of formaldehyde emissions.
Studies have specifically examined the effects of
misfueling and malfunctions on formaldehyde and other exhaust
emissions. In a misfueling study, exhaust emissions from a
catalyst-equipped vehicle were measured initially with unleaded
fuel.[31] Total aliphatic aldehydes were measured but can be
used as an indicator of formaldehyde emissions. The vehicle
was then driven 5000 miles on various commercial leaded
gasolines. Total aliphatic aldehyde emissions initially (with
unleaded fuel) were 2.1 percent of the total HC emissions.
Following 5000 miles of misfueling, the percentage increased
only slightly to 2.5 percent.
Malfunctions that have been evaluated include, but are not
limited to, 12 percent misfire, disabled EGR, rich best idle
and high oil consumption. Results with non-catalyst-equipped
vehicles indicate roughly similar formaldehyde percentages with
and without malfunctions.[32] With catalyst-equipped vehicles,
formaldehyde emissions tend to be reduced with malfunctions
while total hydrocarbon emissions increase, resulting in lower
formaldehyde percentages with malfunctions.[33-35]
In summary, formaldehyde percentages appear relatively
stable over a wide range of operating conditions and HC
emissions. The data support the use of expressing formaldehyde
emissions as a percentage of total hydrocarbons.
Formaldehyde emissions, expressed as a percentage of total
hydrocarbons, for the vehicle classes in 1986 and 1995 are
given below:
-31-
-------
% of Total Exhaust HC
Vehicle Class 1986 1995
LDGV 1.3 1.0
LDGT 1,2 1.3 1.1
LDDV 4.1 4.1
LDDT 4.1 4.1
HDGV 3.1 3.1
HDDV-TRUCK 3.0 3.0
HDDV-COMM. BUS 10.0 10.0
Data for the light-duty gasoline-fueled vehicles were
taken from references 30 and 32-36. These data were assumed to
apply to light-duty gasoline-fueled trucks as well. These
references include data for both new and in-use vehicles. The
decreasing percentages for these vehicles from 1986 to 1995
reflect the phase-out of non-catalyst-eguipped vehicles. Since
a greater percentage of non-catalyst-equipped trucks than
vehicles are projected in 1995, the percentages for trucks do
not decrease as much in 1995 as the vehicles.
Data for the light-duty diesel vehicles were taken from
references 37 and 38. Again, these data were assumed to apply
to light-duty diesel trucks as well. The formaldehyde
percentage of total HC is assumed to remain constant from 1986
to 1995.
Data for the heavy-duty gasoline- and diesel-fueled
engines were taken from references 22 and 38. Since the
heavy-duty diesel bus engines appeared to have greater
percentages than the other heavy-duty diesel engines, they were
treated separately, and assigned a VMT fraction of 0.002.[17]
Applying these percentages to the MOBILES exhaust HC
emission factors by vehicle class for 1986 and 1995 and using
MOBILES VMT fractions, resulting composite FTP formaldehyde
emission factors are given below.
FTP g/mile
1986 1995
with I/M 0.0418 0.0201
without I/M 0.0453 0.0224
Note that MOBILES runs were used which assumed both the
presence and absence of an Inspection/Maintenance (I/M)
program. An I/M p-ogram should have a beneficial impact on
formaldehyde emissions since formaldehyde emissions are
correlated with HC emissions. The I/M program selected has the
following characteristics:
-32-
-------
Start year (January 1): 1983
Pre 1981 MYR stringency rate: 20%
Mechanic training program?: No
First MYR covered: 1951
Last MYR covered: 2020
Vehicle types covered: LDGV
1981 and later MYR test type: Idle
1981 and later MYR test cutpoints: 1.2% ICO/220 ppm IHC
This represents the minimum EPA requires of I/M programs.
3.2.2 Nationwide Mobile Source Formaldehyde Emissions
Nationwide mobile source formaldehyde emissions (metric
tons/year) for 1986 and 1995 are calculated by combining the
calendar year emission factors in the previous section with
projected VMT data for 1986 and 1995. The VMT data are given
in Table 2-8.
The nationwide mobile source formaldehyde emissions for
1986 and 1995 are given below.
Mobile Source Formaldehyde Emissions
(metric tons per year)
1986 1995
with I/M 68,003 38,880
without I/M 73,697 43,329
As seen, formaldehyde emissions decrease from 1986 to 1995
due to the projected HC reductions.
It should be noted that these estimates only account for
formaldehyde directly emitted. Formaldehyde formed indirectly
from photooxidation of mobile source volatile organic compounds
(VOC) is not included but will be addressed in a later section.
3.2.3 Other Sources of Formaldehyde
Formaldehyde is produced in the U.S. by 14 chemical
companies in 48 locations encompassing 21 states. Formaldehyde
is used in the manufacture of four major types of resins:
urea-formaldehyde, melamine-formaldehyde, phenol-formaldehyde
and polyacetal resins. These resins are used in a wide variety
of products, such as plywood, particle board and counter tops.
Formaldehyde is also used as a raw material in several
synthetic organic chemical production processes.
In addition, formaldehyde is produced as a by-product in
the following types of processes: combustion (mobile,
stationary and natural sources), petroleum refinery catalytic
-33-
-------
cracking and coking, phthalic anhydride production and
atmospheric photooxidation of unburned hydrocarbons. Table 3-1
provides a summary of 1985 estimated formaldehyde emissions by
source category.[39] The contribution of atmospheric
photooxidation is extremely uncertain and therefore not
included in Table 3-1.
The estimates of mobile source formaldehyde emissions in
1986 with and without I/M were 68,003 and 73,697 metric
tons/year, respectively. This represents 26-29 percent of the
total formaldehyde emissions given in Table 3-1. Therefore,
roughly 26-29 percent of the formaldehyde emissions emitted
directly into the air (versus formed photochemically) appear to
be attributable to mobile sources.
3.3 Ambient Concentrations of Formaldehyde Emitted by Mobile
Sources
In this section, urban and rural concentrations of
formaldehyde emitted from vehicle exhaust are estimated for
1986 and 1995, using the modified version of the CO NEM
discussed previously. The contribution of mobile sources to
ambient formaldehyde levels is then discussed.
3.3.1 Estimation of Urban and Rural Exposure
The three required model inputs are: 1) the nationwide
urban and rural populations for the year(s) of interest, 2)
FTP, 10 mph, 35 mph, and LDGV idle emission factors (g/min),
and 3) a list of 25 ambient pollutant concentrations, defining
the concentration intervals for which cumulative person-hours
of exposure are to be calculated. The first input is given in
Section 2.4.1. Similarly, the FTP emission factors have also
already been provided and are simply converted from g/mile to
g/min using the average FTP speed of 19.6 mph. To calculate
emission factors at 10 mph and 35 mph, MOBILES runs were made
at these speeds and the ratios of the total exhaust HC
emissions at these speeds to those at the FTP average speed
were determined. These speed correction factors were then
multiplied by the composite FTP emission factors to obtain the
emission factors at 10 mph and 35 mph. The speed correction
factors were 1.75 for 10 mph and 0.54 for 35 mph.
The idle LDGV emission factors for 1986 and 1995 are
0.0009 and 0.0006 g/min, respectively. Idle emission factors
of both HC and formaldehyde have been measured less often than
those for the FTP, and the formaldehyde:HC ratio approach was
not used for idle. They were instead determined directly using
the formaldehyde emission rate data in reference 40 together
with projected LDGV VMT fractions of non-catalyst-eguipped and
catalyst-equipped vehicles in 1986 and 1995. In 1986, 12
-34-
-------
Table 3-1
Summary of Estimated 1985 Formaldehyde
Emissions by Source Category
Formaldehyde Emissions
Source (metric tons/year)
Direct Producers 910
Resin Manufacture 2,789
Synthetic Chemical Production 655
Indirect Producers*
Combustion (Mobile, Stationary, Natural) 250,000
Petroleum Catalytic Cracking 3,200
Phthalic Anhydride Production 1
TOTAL: 257,555
* Excludes atmospheric photooxidation due to the
uncertainties inherent in the estimate.
-35-
-------
percent of the LDGV VMT is projected to be due to non-catalyst-
eguipped vehicles; in 1995, the percentage drops to 0.2 percent.
The 25 ambient pollutant concentrations for 1986 ranged
from 0.0 to 24.0 ug/m3 . For 1995, they ranged from 0.0 to
10.0 ug/m1. Logarithmic intervals were chosen to maximize
resolution over the entire range.
The mean exposure levels predicted by the model, adjusted
to account for increased VMT, are given below. These exposure
levels account for direct emissions from mobile sources only.
Formaldehyde Exposure (ug/m3)
Urban Rural Nationwide
1986
With I/M 1.21 0.56 1.04
Without I/M 1.30 0.60 1.13
1995
With I/M 0.68 0.31 0.59
Without I/M 0.76 0.35 0.65
It should be noted that these model predictions were made
assuming that formaldehyde has emission formation and
dispersion characteristics similar to that of CO. The model
also does not account for photochemical reactions. The
exposure levels predicted by the model are those resulting from
direct exhaust emissions, and do not account for either the
destruction or photochemical formation of formaldehyde in the
atmosphere.
3.3.2 Contribution of Mobile Sources to Ambient Formaldehyde
Levels
Ambient formaldehyde levels are a result of formaldehyde
directly emitted by sources and formaldehyde formed from
photooxidation of VOC. The mobile source contribution to
ambient formaldehyde levels also contains both components. The
previous sections have attempted to quantitate the directly
emitted component. It appears that roughly 26-29 percent of
directly emitted formaldehyde may be attributable to mobile
sources.
Formaldehyde formed photochemically is much more difficult
to quantify. One approach being used by EPA is to determine
the relative contribution of various sources based on estimates
of annual U.S. VOC emissions for each source. These estimates
are given in Table 3-2.[41] Mobile sources account for 30
percent of the total VOC emissions. If it is assumed that the
VOC from all sources have the equivalent potential to form
formaldehyde, then 30 percent of the formaldehyde formed
photochemically is due to mobile sources. It should be noted,
-36-
-------
Table 3-2
Sources of VOC Emissions and 1985 National
VOC Air Emission Estimates*
VOC Emissions
Source (metric tons/year) % of Total
Mobile Sources 7,200,000 30
Misc. Solvent Uses 3,600,000 15
Hazardous Waste Treatment,
Storage and Disposal
Facilities 3,500,000 14
Surface Coating 3,160,000 13
Petroleum Marketing 2,230,000 9
Petroleum Refining 740,000 3
Chemical Manufacture 500,000 2
Industrial Processes 365,000 2
Miscellaneous Sources 3,020,000 12
TOTAL: 24,315,000
* Reference 41.
-37-
-------
however, that mobile sources are estimated to account for a
higher percentage of the total VOC in urban areas. Examination
of 1983 VOC emissions in 61 urban ozone non-attainment areas
showed the average mobile source VOC contribution to fall
between 40-50 percent. [42]
Since the mobile source contribution to both directly
emitted and photochemically produced formaldehyde nationwide is
roughly equivalent, as a rough approximation, 30 percent of the
ambient formaldehyde concentration could be due to mobile
sources on a national basis. The mobile source contribution
will vary depending on vehicle use in a particular area, season
and a variety of other meteorological factors. In the winter,
for example, formaldehyde emissions from mobile sources are
expected to increase whereas photochemical reactions are
expected to be minimized. The mobile source contribution to
ambient formaldehyde emissions in the winter could, therefore,
be greater.
Photochemical modeling is one approach to use in
attempting to determine the relative contributions of mobile
and stationary sources to ambient formaldehyde concentrations.
Photochemical modeling has been done to simulate meteorology
and photochemistry occurring during summer months due to the
interest in ozone. As part of one modeling study, designed to
determine the ozone impact of methanol-fueled vehicle
substitution in California's South Coast Air Basin (SCAB)
during a severe ozone episode, mobile source VOC emissions were
removed and the resulting formaldehyde concentrations
determined.[43] The result indicated that 23 percent of the
ambient formaldehyde concentration is due to direct
formaldehyde emissions from mobile sources and formaldehyde
formed photochemically from mobile source VOC. This result
lends some support to the previous estimate. Mobile source VOC
in the SCAB is roughly 50 percent of the total which is
somewhat higher than most areas of the country. Thirty
percent, therefore, can be considered an upper estimate.
3.4 Health Effects of Formaldehyde and Unit Risk Estimates
Formaldehyde can cause a number of acute adverse health
effects such as eye, nose, throat and skin irritation,
headaches and nausea, as well as death. Formaldehyde has also
been found to cause nasal cancer by inhalation in males and
females of one strain of rat and in males of another strain,
and there is evidence of its carcinogenicity in mice. Human
data are more limited. EPA has classified formaldehyde as a
probable (Bl) carcinogen in humans.
The unit risk estimates derived by EPA are based on a
single animal inhalation study conducted by the Chemical
Industry Institute of Toxicology (CUT).[44] In this study,
-38-
-------
statistically significant increased levels of squamous cell
carcinomas (malignant tumors) were found in the nasal cavities
of rats at 14.3 ppm. (This is roughly 1000 times ambient
levels.) The smaller increase in these carcinomas observed in
mice was not statistically significant. In addition to the
squamous cell carcinomas, small numbers of benign tumors
characterized as polypoid adenomas were observed in rats at
each dose level.
The upper confidence limit unit risk estimates based on
these data are 1.7 x 10"4 for benign tumors and 1.3 x 10"s
for malignant tumors.[45] The unit risk for malignant tumors
of 1.3xlO"s was used for this analysis. The current
consensus in EPA favors use of malignant data only. EPA's
Office of Toxic Substances is using the unit risk based on
malignant tumors in assessing the risk to garment workers and
home residents.[45] Assuming an average lifetime of 70 years,
the annual risk estimate is 1.9xlO~s.
3.5 Current and Projected Health Risk
The annual risk estimate is combined with the exposure
estimates (given in Section 3.3.1) and population estimates to
obtain estimates of cancer incidence for 1986 and 1995. The
results are given in Table 3-3.
As seen in this table, the total risk in 1986 ranges from
46-50 cancer incidences and drops to 29-31 cancer incidences in
1995. This decrease is accounted for by the anticipated
decrease in HC emissions.
These risk estimates are for formaldehyde emitted directly
from vehicle exhaust. Unfortunately, the model used to
determine ambient concentrations could not account for the
destruction and photochemical formation of formaldehyde in the
atmosphere.
An approach to include photochemistry is to use actual
ambient monitoring data and assign a mobile source fraction.
This approach accounts for formaldehyde directly emitted and
formed or destroyed photochemically. OAQPS calculated an urban
population weighted average of 12.71 ug/m1 based on data from
four cities.[46] It represents the average of the population
weighted summer average in Baltimore, Los Angeles and New
Jersey with the winter average in Philadelphia. The summer
concentrations used represent maximum rather than average
values. As a result, 12.71 ug/m3 is closer to an urban
population weighted maximum concentration. It raay also be
biased high since these cities are in ozone non-attainment
areas. It will be used here to represent a plausible upper
limit. A concentration of 1.50 ug/m3 was selected to
represent rural areas.[46]
-39-
-------
Table 3-3
Annual Cancer Risk From Direct Emissions
of Formaldehyde from Mobile Sources
Urban Rural Total
1986
With I/M 40 6 46
Without I/M 43 7 50
1995
With I/M 25 4 29
Without I/M 27 4 31
-40-
-------
In Section 3.3.2, it was indicated that mobile sources
constituted roughly 30 percent of formaldehyde emitted directly
as well as 30 percent of the total VOC emitted. Assuming the
VOC from all sources have the equivalent potential to form
formaldehyde, a mobile source fraction of 0.30 is selected and
applied to the estimated urban and rural concentrations above.
Resulting urban and rural formaldehyde concentrations due to
mobile sources are 3.81 and 0.45 ug/m , respectively.
In comparison, the urban and rural formaldehyde
concentrations obtained from modeling are 1.21-1.30 and
0.56-0.60 ug/m3, respectively. The urban concentration
estimated from ambient monitoring data is roughly three times
greater than that obtained by modeling. Two possible reasons
are: 1) the ambient monitoring data accounts for formaldehyde
formed photochemically, and 2) the ambient data may be from
fixed site monitors that overrepresent 24-hour exposures of the
population. In contrast, the rural concentration obtained from
modeling is slightly higher than that estimated from ambient
monitoring data. The rural concentrations were estimated
rather crudely with both approaches.
The exposure estimates based on ambient monitoring data
are then combined with the annual risk estimates and urban and
rural population estimates to obtain estimates of lung cancer
incidence from current mobile sources. The results are given
below:
Annual Lung Cancer Risk
(accounts for photochemical reactions)
Urban Rural Total
Current Mobile Sources 126 5 131
Mobile sources currently could be responsible for as many
as 131 cancer deat.is from formaldehyde exposure. Direct
emissions of formaldehyde were projected to decrease roughly 42
percent from 1986 to 1995. Mobile source VOC follows a similar
trend. If it is assumed that the photochemical component is
also reduced similarly, the mobile source risk in 1995 is
estimated be roughly 77 cancer deaths.
3.6 Current Activities
Research activity is planned or underway in three areas:
1) emissions characterization, 2) photochemistry, and 3)
ambient monitoring. Formaldehyde emissions from mobile sources
are being characterized under cold temperature conditions,
since it appears formaldehyde emissions could increase under
these conditions. In addition, work is planned in FY87 to
-41-
-------
investigate factors affecting formaldehyde formation in the
atmosphere. Specifically, smog chamber data will be analyzed
to rank order the various organic species based on their
ability to generate formaldehyde. Modeling simulations will be
conducted to assess the effect of varying light intensity,
temperature and other meteorological parameters on formaldehyde
yields. Smog chamber experiments will also be conducted to
determine how well two major chemical mechanisms predict
formaldehyde formation.
In the monitoring area, formaldehyde measurements will be
included in EPA's Toxic Air Monitoring System (TAMS) network.
Measurements will initially be made in three cities (Houston,
Boston and Chicago) with a single monitoring site in each
city. TAMS should eventually be operational in three to five
cities with up to three to four monitors in each city.
It should be noted that, in the case of formaldehyde,
formaldehyde levels inside many homes are significantly greater
than outside levels. Recent monitoring results indicate that
formaldehyde levels in new (less than one year old)
conventional homes generally fall within the range of 62
ug/m3 to 250 ug/m3; few measurements exceeded 375 ug/m3.
In new mobile homes, formaldehyde levels generally fall within
the range of 250 ug/m1 to 375 ug/m3 with the highest levels
measured near 500 ug/m3.[45] EPA has calculated expected
10-year averages for formaldehyde levels in homes built today,
acknowledging that there is a significant source of uncertainty
associated with the estimates. The calculated 10-year averages
are 88 ug/m3 for conventional homes built using significant
amounts of ureaformaldehyde pressed wood and 125 ug/m3 for
mobile homes.[45] These exposures are significantly higher than
current outdoor urban exposures.
-42-
-------
4.0 BENZENE
The Office of Mobile Sources has recently completed a
thorough analysis on the carcinogenic impact of benzene
emissions.[47] This section will summarize that analysis.
Also discussed is an alternative approach to estimating the
risk using ambient monitoring data in conjunction with
emissions data. A risk analysis performed by the California
Air Resources Board (CARB) is then summarized and compared with
the risk estimates presented in this report.
4.1 Formation, Composition, and Gontrol Technology
Benzene is an aromatic hydrocarbon with the formula
C6HS. It is present in both exhaust and evaporative
emissions. Current data show the benzene level of current
gasoline to be about 1.3%, with diesel fuel containing
insignificant levels of benzene. Very little exhaust benzene
is unburned fuel benzene. Some work indicates that non-benzene
aromatics in the fuels cause about 70-80% of the exhaust
benzene formed. Benzene also forms from engine combustion of
non-aromatic fuel hydrocarbons. The fraction of benzene in the
exhaust varies depending on control technology and fuel
composition but is generally about 3-5%. The fraction of
benzene in the evaporative emissions also depends on control
technology (e.g., whether the vehicle has fuel injection or a
carburetor) and fuel composition (e.g., benzene level and RVP)
and is generally about 1%. These data also show that diesel
vehicles account for only about 3% of the total mobile source
benzene emitted.
4.2 Emissions
4.2.1 Emission Factors for Calendar Years 1985 and 1995
An approach similar to that described earlier for
formaldehyde was employed for benzene. Benzene emissions were
expressed as a percentage of exhaust and evaporative
hydrocarbons. These percentages were then applied to the
hydrocarbon emissions data in MOBILES to obtain composite
emission factors for calendar years 1985 and 1995. For the
purpose of this report, it will be assumed that calendar year
emission factors for 1985 and 1986 will not differ
significantly.
Benzene emissions, expressed as a percentage of exhaust
and evaporative emissions for the various vehicle classes, are
given in Table 4-1.
The test vehicles used to determine the benzene
percentages were low mileage, well maintained vehicles. Like
formaldehyde, it is important to address whether the
percentages chosen are adequate to use for the excess
hydrocarbons that come from deterioration, malfunction and
tampering/misfueling.
-43-
-------
Table 4-1
Benzene Emissions Expressed as Percentages of HC
- Vehicle Class % of Exhaust HC % of Evap. HC
LDGV 0.35-1.53*
3 Way Cat 5.12
3 Way + Ox Cat 2.78
Non-Cat or Ox Cat 3.95
LDGT1,2 3.24 1.1
LDDV 2.40
LDDT 2.40
HDGV 3.48 1.1
HDDV 1.10
l.ll-l.53% for carbureted LDGV.
0.35-0.46% for fuel injected vehicles.
-44-
-------
For non-malfunction conditions, there did not appear to be
any consistent differences in benzene percentages among the
various control technologies (non-catalyst, oxidation catalyst,
3-way catalyst, and 3-way plus oxidation catalyst), regardless
of whether the vehicles were tested new or in-use. There were
some differences in benzene percentages found between certain
malfunction modes and the unmodified test conditions. The most
significant of these were 1) the 12 percent misfire mode which
consistently decreased the benzene percentage (while greatly
increasing total HO and 2) the rich best idle mode which
consistently increased the benzene percentage as well as the
total HC (but not as much as the total HC increase in the 12%
misfire mode). Due to the lack of differences among control
technologies, the offsetting nature of these two malfunctions,
and the lesser effects of the other malfunctions, no
adjustments were made to the benzene percentages for
malmaintenance/tampering.
Resulting composite FTP benzene emissions for 1985 and
1995 are 0.128-0.135 g/mile and 0.055-0.057 g/mile,
respectively. The ranges result from the ranges given for the
percentage of evaporative hydrocarbons for LDGV. The MOBILES
runs assumed the pr5sence of a standard, minimum I/M program.
As stated previously, diesel vehicles account for a small
percentage (3%) of the total mobile source benzene emissions.
Based on the QMS analysis, RVP control, which would be
accompanied by a small increase in both benzene content and
total aromatic content of gasoline, would have little or no
effect on overall fleet emissions or on the number of cancer
incidences.[47]
4.2.2 Contribution of Mobile Sources to Nationwide Benzene
Emissions
In 1982, total benzene emissions were roughly 293,000
metric tons.[47] Mobile sources account for 250,000 metric
tons, or 85 percent of the total. Of the mobile source
contribution, 70% comes from exhaust of motor vehicles while
14% of the benzene emissions are motor vehicle evaporative
emissions. About 1% of the total benzene emissions occur
during motor vehicle refueling. The remaining benzene
emissions come from stationary sources with coke ovens being
responsible for 10% of the total. Obviously, in regions
without coke ovens virtually all benzene is from mobile sources.
4.3 Health Effects of Benzene and the Unit Risk Estimate
Several epidemiology studies on workers exposed to benzene
have identified benzene as a carcinogen causing leukemia in
humans. The upper confidence limit unit risk estimate has been
determined from these studies to be 8.0 x 10"6.[47,48]
-45-
-------
4.4 Current and Projected Health Risk
Nationwide exposure levels from both exhaust and
evaporative emissions were estimated using the modified NEM
exposure model. The speed correction factors at 10 mph and 35
mph required for the NEM inputs were those used for
formaldehyde. The nationwide exposure levels predicted by NEM,
adjusted upward to account for increased VMT, are 3.09-3.25
ug/mj and 1.68-1.77 ug/m3 for 1985 and 1995, respectively.
The range is due to consideration of both a low and high range
evaporative emissions estimate for light-duty gasoline-fueled
vehicles. Annual cancer incidences from exhaust and
evaporative emissions are estimated to be 84-89 in 1985 and
50-52 in 1995. The reason for this marked decrease is the
decrease in projected HC in 1995 and, thus, benzene emissions.
Exposure to benzene during refueling includes self-service
refueling, occupational exposure (service station attendants)
and community exposure in an urban area. Exposure to
self-service refueling and occupational exposure was determined
by measuring benzene levels in the region of the face of a
person refueling a vehicle tank. The exposure in a typical
urban area was estimated using a dispersion model.[48] Annual
cancer incidences from benzene refueling are estimated to be 8
in 1985 and 7 in 1995.
The total estimated cancer incidences due to mobile source
benzene and the contribution of evaporative, exhaust and
refueling emissions are given below.
Annual Cancer Incidences due to Mobile Source Benzene
Year Evaporative* Exhaust Refueling TOTAL
1985 17-22 67 8 92-97
1995 8-10 42 7 57-59
* Includes low and high range evaporative emissions estimate.
These numbers indicate that, as vehicle hydrocarbon
evaporative and exhaust emissions are controlled, the estimated
carcinogenic impact due to motor vehicle benzene emissions will
decrease. They also indicate that the impact is somewhat less
than that predicted for diesel particulates or formaldehyde.
However, it should be pointed out that the potency on which
these estimates are based is derived from several epidemiology
studies and is thus a much firmer potency than that derived for
diesel particulates or f or.naldehyde. The potency for diesel
particulates is based on several animal studies rather than
human data. For formaldehyde, it is based on a single animal
study.
-46-
-------
An alternative approach for estimating the exposure and
resulting risk of mobile source benzene emissions is to use
available ambient monitoring data and assign a mobile source
fraction. This approach was used for formaldehyde in order to
account for photochemistry.
An urban population weighted average of 10.24 ug/m3 was
calculated based on data from six metropolitan areas. Ambient
data for Baltimore, Los Angeles, Northern New Jersey,
Philadelphia and Chicago were taken from reference 49. Ambient
data for Houston were obtained from reference 46. Their
estimated metropolitan area populations in 1983 were taken from
reference 15 and used to calculate a population weighted
average. Since these cities are in ozone non-attainment areas,
the resulting urban concentration may represent more of an
upper-bound. A concentration of 7.52 ug/m3 was selected to
represent rural areas.[46]
Since mobile sources appear to be responsible for roughly
85 percent of the total benzene emitted, a mobile source
fraction of 0.85 was selected and applied to the estimated
urban and rural concentrations above. Resulting urban and
rural benzene concentrations due to mobile sources are 8.70 and
6.39 ug/m3, respectively.
These exposure estimates are then combined with the annual
risk estimates to obtain estimates of cancer incidence from
current mobile sources. In 1986, the urban and rural risks,
using this alternative approach, are estimated to be 179 and 44
cancer incidences, respectively, for a total of 223 cancer
incidences. Based on the NEM modeling, emissions of benzene
from mobile sources are projected to decrease roughly 40
percent from 1985 to 1995. The mobile source urban, rural and
total risk in 1995, accounting for this decrease and the
projected population increase, is 116, 29, and 145 cancer
incidences, respectively.
These risk estimates are somewhat higher than those
calculated with the NEM approach. Both approaches contain
uncertainties. For example, temperature effects are only
partially accounted for. With the NEM approach, it is assumed
that CO and benzene have similar dispersion characteristics.
This may not be true. With the ambient apportionment approach,
the ambient data may be from fixed site monitors that
overrepresent 24-hour exposures of the population. It is also
not certain whether the cities chosen are representative of the
entire urban population. For the purposes of this report, a
range of risk estimates for benzene will be reported, using the
results of both approaches. The resulting range of cancer
incidences is 92-223 in 1985 and 57-145 in 1995.
-47-
-------
4.5 GARB Analysis of Current and Projected Health Risk
In January, 1985, the California Air Resources Board
(GARB) identified benzene as a toxic air contaminant. As a
result, CARB released a draft report which includes information
regarding present (1984) and future (2000) benzene emissions
and levels of benzene in California as well as the magnitude of
risk posed by benzene emissions from various sources.[50] An
addendum to the report was recently prepared.[51] The
information in these reports will be briefly summarized in this
section. The following section provides a comparison of the
EPA and CARB risk estimates. For reference, California
contains 11 percent of the U.S. population and 11 percent of
the motor vehicles.[19,52]
Total benzene emissions in California were roughly 18,400
tons/year in 1984 and projected to decrease to 13,000 tons/year
in the year 2000. The vehicular contribution, which includes
both on-road and off-road vehicles as well as other mobile
sources, constitute roughly 91 percent of the total emissions
in 1984 and 84 percent of the total emissions in 2000. The
statewide population in 1984 was 25.8 million and is projected
to increase to 31.4 million in 2000.
The population exposure to benzene in California for 1984
and 2000 was estimated using benzene emissions data and
population and monitoring data. The estimated statewide annual
average exposure to benzene in 1984, 8.4 x 107 ppb-persons,
is equal to 25.8 million, the number of people exposed to
benzene, times 3.3 ppb, the population weighted annual average
benzene concentration. The vehicular contribution (68.1xl06
ppb-persons) was calculated based on the emissions data. This
is similar to the alternative approach described in the
previous section. Based on exposure of 25.8 million people,
the annual average benzene concentration due to vehicular
sources in 1984 is 2.64 ppb (8.5 ug/m1) . This accounts for
both urban and rural exposure.
The estimated statewide annual average exposure to benzene
in 2000 is 74 x 10s ppb-persons. This was calculated by
adjusting the 1984 value to account for the projected reduction
in total benzene emissions and the projected increase in
population. The vehicular contribution of 63.5 x 106
ppb-persons is again based on the emissions contribution.
Based on exposure of 31.4 million people, the annual average
benzene concentration due to vehicular sources in 2000 is 2.0
ppb (6.4 ug/m3). This is a 25 percent reduction relative to
1984.
-48-
-------
CARB used a range of risk estimates of 22-170 excess
cancers per million people exposed per ppb per 70 years. The
range is based on an examination of both mouse and human data.
Using this range together with the estimated statewide annual
average benzene exposures in 1984 and 2000, the resulting range
of risks in 1984 and 2000 are 1,900-14,500 and 1,630-12,600
excess cancer cases in California per 70 years. On an annual
basis, this translates to 27-207 cancer cases in 1984 and
23-180 cancer cases in 2000. The vehicular contribution is
21-166 cancer cases in California in 1984 and 20-154 cancer
cases in 2000. Expressed as individual risk, the vehicular
contribution is 8.11 - 64.3 x 10"7 in 1984 and 6.4 - 49.0 x
10"7 in 2000.
4.6 Comparison of EPA and CARB Health Risks
The range of cancer incidences given in this report is
92-223 in 1986 and 57-145 in 1995. In order to compare these
risk estimates to those of CARB, they are expressed as
individual risks. The resulting individual risk estimates are
3.8 - 9.3 x 10"7 in 1986 and 2.2 - 5.6 x 10~7 in 1995. The
upper end of these risk estimates are roughly equivalent to the
lower end of the CARB estimates.
CARB used a range of risk estimates of 22-170 excess
cancers per million people exposed per ppb per 70 years. In
terms of unit risks, this translates to 6.81 - 52.6 x 10"6.
The upper confidence limit unit risk used in this study is 8 x
10"s; this is near the low end of the range of CARB unit
risks. CARB's high end risk estimate is based on animal
studies, whereas EPA's risk estimate is based on human studies.
The CARB estimate of the benzene concentration due to
vehicular sources in 1984 is 8.5 ug/m3 . In comparison, the
range of estimated exposures in this study for 1985 is 3.2 -
8.1 ug/m3 (accounting for both urban and rural exposure, as
CARB has done). Clearly, the CARB estimate exceeds the
exposure estimates used in this study. This is not surprising
since monitoring data indicate that benzene concentrations in
urban areas of California are higher than most other areas of
the country. (The high benzene level in Los Angeles was
factored into the national estimate used in this study.)
The CARB projections do not show benzene emissions
dropping as rapidly as predicted in this report. CARB predicts
that vehicular benzene emissions will decrease 35 percent from
1984 to 2000. This can be compared to a 40 percent decrease
from 1985 to 1995 given in this report. Also, when the CARB
risk is expressed in terms of total cancer cases, the number of
cancer cases only decreases by 13 percent from 1984 to 2000.
-49-
-------
This is because the decrease in emissions is slightly offset by
the projected increase in population. The population in
California is projected to increase 22 percent from 1984 to
2000. In comparison, the U.S. population is projected to
increase 8 percent from 1986 to 1995.
When reviewing the GARB risk estimates, it should also be
remembered that the vehicular contribution includes on-road
vehicles, off-road vehicles, trains, ships, aircraft, mobile
equipment and utility equipment. The EPA estimates only
include emissions from on-road vehicles. Most of the
difference seen in the two estimates, however, is attributed to
the wide range of unit risks used by CARB.
4.7 Current Activities
The California Air Resources Board (CARB) is considering
implementing regulations requiring control of motor vehicle
benzene emissions. This may include a limit on the benzene
content of gasoline and/or stricter light-duty exhaust HC
standards. The first measures being considered are a more
stringent exhaust HC standard by the end of 1987, and possible
changes in the evaporative test procedure, e.g., multiple
diurnal tests, longer soak times, and higher soak temperatures.
The EPA Office of Air Quality Planning and Standards
(OAQPS) has designated benzene as a hazardous air pollutant
under Section 112 of the Clean Air Act and is implementing
necessary controls for stationary sources. The EPA Office of
Mobile Sources (QMS) is determining if any motor vehicle
controls specific to benzene are needed. Also, QMS is
determining how other regulations that may be proposed for
additional hydrocarbon control would affect benzene.
-50-
-------
5.0 GASOLINE VAPORS
Several years ago, the American Petroleum Institute
released the results of a lengthy animal inhalation study
showing that totally vaporized gasoline resulted in an increase
in kidney tumors in male rats and liver tumors in female
mice.[53] EPA evaluated these data in depth to determine the
potential carcinogenic impact.[48,54] From these studies, EPA
calculated the following plausible upper limit carcinogenic
potencies for exposure to 1 ppm for a year. The maximum
likelihood potency is lower by almost a factor of two for both
data sets.
Gasoline Carcinogenic Potencies Based on API Study
Rat data 4.9 x 10~s/ppm
Mice data 2.9 x 10"5/ppm
Exposure to gasoline vapors and benzene from gasoline
during refueling was estimated based on an American Petroleum
Institute study that involved measuring gasoline and vapor
levels in the region of the face of a person refueling a
vehicle tank. The exposure in a typical urban area for these
refueling emissions was also estimated by using the Industrial
Source Complex (ISC) dispersion model to calculate annual
concentrations.[48]
The exposure numbers determined were combined with the rat
potency data to calculate the potential carcinogenic impact due
to refueling emissions. These numbers for both benzene and
gasoline vapors are as below.
Potential Carcinogenic Impact of Refueling Loss Emissions
Annual Incidences
Exposure Benzene Gasoline Vapors
Self-service refueling 5 35
Occupational exposure
(service station attendants) 2 17
Community exposure in
an urban area 0.5 13
TOTAL EXPOSURE 7.5 65
There are several limitations to the use of the API
bioassay, however. First, totally vaporized gasoline was used
which includes higher molecular weight components as well as
the lighter components. Some short term animal tests done by
API indicate that the branched chain paraffins (e.g., C-6, C-7,
C-8) are the ones that would cause kidney damage. It is not
certain what relation the kidney damage has to the kidney
-51-
-------
tumors in the rats. However, it is known that most (i.e.,
80-90%) of the gasoline vapors in a realistic refueling episode
are below C-6 while only about 23% of the vaporized gasoline is
below C-6.[55,56]
Another limitation to the API work is the apparent
involvement of a low molecular weight protein (alpha
2-globulin) that is present only in rats. It is thought that
this protein, which is synthesized in the liver of mature, male
rats, accumulates in the kidney following hydrocarbon
exposure. The effect results in kidney damage (nephropathy, a
chronic inflammation and vascular thickening of portions of the
kidney) similar to that known as "old rat nephropathy". This
condition frequently occurs spontaneously in older rats; the
presence of hydrocarbons accelerates this change. There is
some concern that this kidney condition is related to the small
cancerous growths found. Alpha 2-globulin is not present in
humans which are not affected by this kidney condition. Female
mice have a large and variable rate of spontaneous liver
tumors. Like the case of the rats, some questions have been
raised about extrapolating the results with the mice to
humans. These issues have been discussed in a report by the
Health Effects Institute (HEI) which concluded the following.
...the usefulness of available animal and human data in
helping to determine health risks is quite limited.
Unburnt gasoline vapors may, upon further investigation,
prove to present significant carcinogenic risks for
humans. The evidence is not available to make that
statement today. Significant additional research would
have to be undertaken to understand important mechanisms
of action, physiological differences between test animals
and people and the extent and nature of exposures.[57]
It should be noted that HEI has since decided to perform no
research in the gasoline vapor area. Also, aside from some
limited work being done by API, EPA is aware of no additional
health research being done on gasoline vapors.
In an internal EPA memo, EPA has concluded the following
about the potential carcinogenicity of gasoline vapors and the
EPA risk assessment discussed above.
All issues raised by HEI were known to EPA and thoroughly
assessed in EPA's analyses of the same issues. EPA's
quantitative risk assessment for carcinogenicity of
unleaded gasoline vapors was prepared according to EPA
Guidelines and expresses its uncertainties, which are not
greater than EPA routinely addresses. Further research
would not change the underlying evidence of the
carcinogenicity of gasoline vapors but may shed light on
which components are responsible for the carcinogenic
response and this could alter the risk calculation.[58]
-52-
-------
On an overall basis then, it appears that gasoline vapors
must be regarded as a potential human carcinogen. Also,
gasoline vapors represent a hydrocarbon source which may need
to be controlled to help attain the ozone National Ambient Air
Quality Standard. EPA has not yet made a decision to propose
controls for gasoline vapors or a decision on the form of the
control.
-53-
-------
6.0 GAS PHASE ORGANICS
6.1 Formation and Control Technology
Gas phase organics, or volatile organic compounds (VOC),
are present in both exhaust and evaporative emissions. Over
300 gas phase organics have been identified.[59] The majority
of VOC are formed from incomplete combustion of the fuel while
others are simply evaporated fuel components. The emphasis of
this section will be on exhaust compounds; evaporative
compounds are covered in Section 5 discussing gasoline vapors.
The total mass of VOC has decreased since 1975 due to
introduction of the oxidation catalyst (caused by more
stringent HC standards). More advanced control technology
(i.e., 3-way and 3-way plus oxidation catalysts) introduced in
the early 1980's to allow further simultaneous control of NOx
and HC has reduced the total mass even further.
6.2 Composition
The majority of gas phase organics consist of unsaturated
and saturated hydrocarbons along with benzene, alkyl benzenes,
aliphatic aldehydes and a variety of polycyclic aromatic
hydrocarbons (PAH) including nitro-PAH. A list of the VOC
measured in ECTD contractor studies (that include unregulated
emissions characterization) is given in Table 6-1. It should
be noted that this is not meant to be a complete list; instead,
it represents those compounds for which the most emissions data
are available.
Studies have been conducted to determine the mass and the
detailed hydrocarbon composition of vehicle exhaust.[30,60]
Gas chromatographic (GO analysis was used in these studies to
identify individual hydrocarbons of carbon numbers 1 through 10
(Ci - Cio). In a recent study, 82 individual hydrocarbons
and 10 aldehydes were measured in the exhaust of 46 in-use
1975-1982 gasoline-fueled vehicles.[30] The gas
chromatographic conditions employed did not permit
identification of each individual compound above Cio-
Fortunately, the hydrocarbon composition of gasoline engine
exhaust consists primarily of components with carbon numbers 1
through 10 so a fairly complete description of VOC emissions
from gasoline-fueled vehicles is possible with GC analysis.
The exception is the small quantity of PAH which is also
present in gasoline exhaust. A more detailed discussion of gas
phase PAH will be presented later in this section.
Unlike gasoline-fueled vehicles, the VOC emitted from
diesel-fueled vehicles range from Ci to about C40, with the
majority being below C2S. The C^Cio hydrocarbons result
almost entirely from the combustion process, which involves
-54-
-------
Table 6-1
VOC Unregulated Emissions Most Commonly Characterized
Total Hydrocarbons
Individual Hydrocarbons
Methane
Ethylene
Ethane
Acetylene
Propane
Propylene
Benzene
Toluene
Aldehydes and Ketones
Formaldehyde
Acetaldehyde
Acetone
Isobutyraldehyde
Methyl Ethyl Ketone
Crotonaldehyde
Hexanaldehyde
Benzaldehyde
-55-
-------
cracking, and possibly subsequent polymerization, of higher
molecular weight materials. It is postulated that the
Cio-Czs hydrocarbons result, to a large extent, from
uncombusted fuel, and the Cis-C40 hydrocarbons from
lubricants.[4]
Gas phase PAH have been characterized in the exhaust of
both gasoline- and diesel-fueled vehicles.[61,6] Some of the
PAH identified included anthracene and phenanthrene,
fluoranthene, pyrene, benzo(a)anthracene and isomers, chrysene,
indenopyrene, benzo(ghi)perylene, benzo(e)pyrene,
benzo(a)pyrene and coronene. These compounds, which contain
two to six benzene rings, have also been identified in diesel
particulate extracts. With the exception of anthracene,
phenanthrene and possibly pyrene (the data for pyrene appear
conflicting), the majority of these PAH are emitted in the
particulate phase. This is particularly true for
benzo(a)pyrene. Particulate/gas phase mass ratios for
benzo(a)pyrene range from roughly 4:1 [61] to 15-27:1.[6]
6.3 Mutagenicity of VOC
Two studies were conducted to determine the relative
mutagenicity of the gas and particulate phases in the exhaust
of gasoline and diesel-fueled vehicles. A XAD-2 trapping
system was used in both studies to collect the gas phase
compounds (greater than C7). The highly volatile compounds,
i.e., less than C7, can not be collected on a XAD-2 trap.
Mutagenicity of the trapped compounds was determined with the
Ames test (Salmonella typhimurium).
In the first study, conducted with gasoline-fueled
stratified charge engine, only 3 percent of the total
direct-acting (i.e., mutagenicity decreases with the addition
of S9 activation) and 5 percent of the total indirect-acting
mutagenicity was found in the gas phase sample. [6] Most of the
gas phase indirect-acting mutagenicity was found in the
non-polar fraction which may be due to the presence of PAH.[6]
In the second study, similar results were found.[62] The
mutagenic activity of the gas phase emissions from the three
gasoline-fueled vehicles tested were at or near background
levels. For the single diesel-fueled vehicle tested, the
mutagenic activity of the gas phase emissions, expressed as
revertants per mile, was less than 11 percent of the
mutagenicity of the particle-bound organics.
6.4 Risk Associated with Individual VOC
For the purpose of this report, available data on detailed
hydrocarbon emissions were reviewed, and individual compounds
were selected which have unit risks associated with them. Of
-56-
-------
the gas phase organics emitted in motor vehicle exhaust,
benzene, formaldehyde, 1,3-butadiene, ethylene and
benzo(a)pyrene have unit risks. Benzene and formaldehyde have
been discussed in previous sections.
1,3-Butadiene
It is difficult due to the limited data available to
develop accurate NEM inputs for 1,3-butadiene. Only data for
light-duty gasoline-fueled vehicles exist. 1,3-Butadiene poses
an additional problem because 1,3-butadiene and n-butane
coelute and thus have the same retention point on the gas
chromatograph. Emission characterization studies to date have
not attempted to determine the percentage of the peak due to
1,3-butadiene. Therefore, assumptions must be made about the
percentage each compound contributes to this peak. In a
previous study, it was assumed that 20 percent of the peak was
due to 1,3-butadiene, for the purpose of grouping the detailed
HC data into various compound classes.[63] Based on data
collected in the Lincoln Tunnel, 1,3-butadiene constitutes 13.9
percent of the total peak, although unknown dilution with air
containing n-butane and 1,3-butadiene complicates the
analysis.[64] Morning ambient samples, probably daily
maximums, collected in 1984 and 1985 show much lower
percentages, with overall averages of 5.08 and 4.21 percent,
respectively.[64] For this report, fifteen percent was chosen
as an upper limit.
1,3-Butadiene emissions were expressed as a percentage of
the total exhaust HC predicted by MOBILES. Based on the data
from 46 in-use gasoline-fueled vehicles provided in reference
30 (and the fifteen percent assumption), 1,3-butadiene is
roughly 0.94 percent of the total FID exhaust HC. Due to the
lack of data for the other vehicle classes, this percentage was
simply applied to the MOBILES composite exhaust HC emission
factor. It was further assumed that the percentage would
remain the same from 1986 to 1995. Composite 1,3-butadiene
emission factors for 1986 and 1995 are 0.0238-0.0263 g/mile and
0.0121-0.0143 g/mile, respectively. The range accounts for
both the presence and absence of an Inspection/Maintenance
program. The speed correction factors were those used for
formaldehyde.
The modified NEM model was used to estimate exposures.
Nationwide urban and rural exposure in 1986 is estimated to be
0.69-0.76 ug/m3 and 0.32-0.35 ug/m1, respectively. In
1995, nationwide urban and rural exposure is estimated to be
0.42-0.49 ug/mj and 0.19-0.23 ug/mj, respectively.
These exposure estimates are for direct emissions of
1,3-butadiene and do not account for reactions of 1,3-butadiene
in the atmosphere. Available ambient monitoring data were
-57-
-------
reviewed and compared to the exposure estimates. A condensed
summary of 1,3-butadiene monitoring data is given in Table
6-2. [64] As seen in the table, average mean values in urban
settings range from 0.77-24.23 ug/m3. The 1986 NEM estimate
of urban exposure from motor vehicles (0.69-0.76 ug/m3) lies
near the low end of the range. The monitoring data indicate
that 1,3-butadiene is, in fact, emitted from motor vehicles as
evidenced by the elevated levels in tunnels. The average mean
value in a rural setting (0.67 ug/m3) is also within the same
order of magnitude as the NEM rural estimate (0.32-0.35
ug/m3).
More recent 1986 1,3-butadiene monitoring data are now
available for 18 cities.[65] These data represent 6-9 a.m.
averages of about 10-15 samples for each city, measured during
the summer months. The average concentrations range from
0.24-1.98 ug/m3 (Data for Los Angeles were not obtained).
Again, the 1986 NEM estimate of urban exposure from motor
vehicles lies within this range.
The EPA estimate for the upper confidence limit unit risk
for 1,3-butadiene has changed substantially. Based on
inhalation studies of 1,3-butadiene in mice, a 95 percent upper
confidence limit unit risk for 1,3-butaciiene is estimated to be
2.8 x 10"4.[66] This is much greater than the previous
estimated unit risk of 4.6 x 10~7.
Estimates of cancer incidence for 1986 and 1995 are given
in Table 6-3. The total risk in 1986 (with the fifteen percent
assumption) ranges from 593-656 cancer incidences and drops to
391-460 cancer incidences in 1995. Preliminary emission
characterization results indicate the presence of
1,3-butadiene, but the amount has not yet been quantified.
Therefore, a lower risk estimate of zero will also be used.
The resulting ranges of cancer incidences for 1986 and 1995
given in Table 6-3 are 0-656 and 0-460, respectively.
Ethylene
The upper confidence limit unit risk of 2.7 x 10"6 for
ethylene was provided in the Six Month Study although it was
not developed by EPA. Unlike 1,3-butadiene, more extensive and
reliable emissions data exist to construct NEM inputs and
predict resulting exposure and risk from direct emissions of
ethylene. Ethylene was handled in much the same way as
formaldehyde. Emissions were expressed as a percentage of the
total exhaust HC predicted by MOBILES. Ethylene in evaporative
emissions is negligible and was not considered.
Ethylene emissions, expressed as a percentage of total
hydrocarbons, for the vehicle classes in 1986 and 1995 are
given below:
-58-
-------
Table 6-2
1,3-Butadiene Monitoring Data; Condensed Summary
ug/m^
Average Mean Maximum
Location Value Value
Houston, TX area*
Los Angeles, CA area*
Riverside, CA
Atlanta, GA
Lincoln Tunnel, NY
Columbus, OH
Denver , CO
5.22
2.81
24.23
3.98
5.57
20.11
1.64
0.77
1.7
88.61
33.76
88.39
6.19
24.09
2.71
5.47
7.6
Comments
incl . tunnels
excl. tunnels
in tunnel
outside air
only one
Jones State Forest
0.67
2.41
sample set
non-urban
setting
All locations within approximately 15 miles of the
metropolitan area.
-59-
-------
Table 6-3
Annual Cancer Risk from Direct Emissions
of 1,3-Butadiene from Mobile Sources
Urban Rural Total
1986
With I/M 0-514 0-79 0-593
Without I/M 0-568 0-88 0-656
1995
With I/M 0-339 0-52 0-391
Without I/M 0-399 0-61 0-460
-60-
-------
% of Total Exhaust HC
Vehicle Class 1986 1995
LDGV 9.0 6.4
LDGT 1,2 9.0 6.4
LDDV 11.0 11.0
LDDT 11.0 11.0
HDGV 13.0 13.0
HDDV 9.0 9.0
Data were taken from references 30, 32-34, 36, 38, and
40. The percentages of exhaust HC for the light-duty gasoline
classes were projected to decrease from 1986 to 1995 due to the
increasing penetration of 3-way and 3-way plus oxidation
catalyst-equipped vehicles and trucks. The speed correction
factors used to compute 10 mph and 35 mph emission factors were
those used for formaldehyde. Composite FTP emission factors
for 1986 and 1995, with and without I/M, are 0.2367-0.2607
g/mile and 0.0946-0.1092 g/mile, respectively.
Estimates of cancer incidence for 1986 and 1995 are given
in Table 6-4. As seen in this table, the total risk in 1986
ranges from 55-60 cancer incidences and drops to 29-31 cancer
incidences in 1995.
Two important limitations need to be mentioned when
discussing the risk estimates for ethylene. The first is the
fact that ethylene is photochemically reactive. The risk
estimates are for direct emissions of ethylene and do not
account for reactions of ethylene in the atmosphere.
The second, and most important, limitation is the unit
risk estimate. The unit risk quoted in the Six Month Study was
estimated by Clement Associates, Inc. for EPA but has not been
endorsed by EPA.[67] There is no available evidence that
ethylene is carcinogenic although ethylene oxide, a metabolite
of ethylene, has been shown to be an animal carcinogen. EPA
has calculated a unit risk for ethylene oxide. Clement assumed
a downward difference in potency of 100 to calculate a unit
risk for ethylene. The basis for this assumed difference in
potency was not given. Since there is no available evidence
that ethylene is carcinogenic, the risk estimates must be
regarded as extremely tentative. For this reason, a lower risk
estimate of zero will also be used.
Benzo (a) Pyrene (B(a)P)
Gas phase B(a)P is emitted in small quantities. The risk
posed by gas phase B(a)P will be considered together with
particle-associated B(a)P. It will be assumed that the
particle-associated B(a)P emission factor used will adequately
represent what little gas phase B(a)P may also be present. The
risk from particle-associated B(a)P is discussed in Section 7.
-61-
-------
Table 6-4
Annual Cancer Risk from Direct Emissions
Of Ethylene from Mobile Sources*
Urban Rural Total
1986
With I/M 0-47 0-8 0-55
Without I/M 0-52 0-8 0-60
1985
With I/M 0-25 0-4 0-29
Without I/M 0-27 0-4 0-31
Lower limit of zero used due to the uncertainty of the unit
risk estimate.
-62-
-------
6.5 Reactivity of VQC
The atmospheric photochemical reaction products of mobile
source volatile organic compounds (VOC) are largely unknown.
It is likely that carcinogens may be formed after emissions
leave the vehicle as well as some carcinogens degrading to
non-carcinogenic compounds. A study was conducted by Calspan
Corporation for the Coordinating Research Council, the
objective being to investigate the fate of diesel exhaust in
the atmosphere.[68] Smog chamber experiments with diluted
diesel exhaust were conducted to investigate parameters such as
the presence or absence of UV irradiation, dilution ratio,
aging and exposure to ozone and N02. The presence of ozone
increased the mutagenic response of the collected
particle-bound organics, in some cases by as much as an order
of magnitude. Similar results were obtained when NO2 and/or
irradiation is present in the chamber. Thus, it appears that
the mutagenicity of diesel particle-bound organics is affected
by ambient conditions.
It also appears likely that irradiation of even innocuous
VOC compounds may lead to the formation of mutagens. In one
study, irradiation of propylene, S02 and NOX did
demonstrate a mutagenic response. [69]
Mobile source VOC is known to contain photochemically
reactive compounds. The percent composition of exhaust
hydrocarbons for pre-1975, 1975-1980 and 1981-1982 model year
gasoline-fueled vehicles is given in Table 6-5.[30,63] The
table provides a general indication of changes in reactivity
with advancing control technology. Pre-1975 vehicles are not
catalyst-equipped. Model year 1975-1980 vehicles are primarily
equipped with oxidation catalysts whereas 1981-1982 vehicles
are primarily equipped with 3-way and 3-way plus oxidation
catalysts.
Olefins and aromatics are all fairly reactive; paraffins
and acetylene are less reactive. In particular, methane is
essentially non-reactive. From Table 6-5, it can be seen that
vehicles equipped with catalysts, particularly 3-way and 3-way
plus oxidation catalysts, emit a higher percentage of- methane
in their exhaust than do non-catalyst-equipped vehicles. Most
catalytic converter systems preferentially oxidize non-methane
hydrocarbons because methane is harder for the catalytic
converter to oxidize. Since methane is essentially
non-reactive, the total photochemical reactivity of the HC
mixture tends to be reduced by the catalyst. In addition, the
catalyst reduces the total hydrocarbon mass and generally
oxidizes the unsaturated HC compounds to a greater extent than
the saturated compounds. This is evidenced by the decreasing
percentages of olefins and aromatics coupled with the
increasing percentage of paraffins in the exhaust.
-63-
-------
Table 6-5
Percent Composition of Exhaust Hydrocarbons
Paraffins
(Methane)
Olefins
Acetylene
Aromatics
% of Total
Pre-1975*
41.0
(9.2)
30.0
8.8
20.4
Exhaust HC
1975-1980**
49.2
(9.3)
23.2
3.4
25.1
by Model Year
1981-1982**
66.5
(21.8)
15.2
0.9
17.4
* Reference 63
** Reference 30
-64-
-------
Aldehydes were not included .in Table 6-5. Aldehydes,
particularly formaldehyde, are photochemically reactive. The
catalyst has also been found to reduce formaldehyde and total
aldehyde emissions. Thus, it appears that, as
non-catalyst-equipped vehicles are phased out of the fleet, the
reactivity of vehicle exhaust will decrease. The reaction
products of mobile source VOC to which people are exposed
remains to be determined.
6.6 Current Activities
EPA-ORD is conducting a long-term research project
referred to as the Integrated Air Cancer Project (IACP). The
goal of the IACP is to identify the principal airborne
carcinogens and their sources. Initially, airsheds with cne or
two emission sources will be examined, followed by areas of
increasing complexity.
In Phase I, initial field measurements were conducted in
Albuquerque, New Mexico and Raleigh, North Carolina. The
purpose of Phase I was to evaluate and select the sampling,
analytical and bioassay methodologies, as well as other
approaches for an integrated field study.
In Phase II, which will be conducted in FY86 and FY87, an
integrated field study will be conducted in Boise, Idaho. The
two major emission sources will be mobile sources and emissions
from wood combustion for residential heating. Some of the
objectives are to: 1) identify and quantify classes of
compounds in the ambient air resulting from residential wood
combustion and motor vehicles, 2) quantify the relative
contributions of these sources to the mutagenic activity,
organic and fine particulate mass of ambient airborne
pollutants, and 3) characterize the chemical changes which may
occur to the source emissions in the atmosphere and assess the
resultant changes in mutagenic response.
In conjunction with the integrated field study in Boise,
separate smog chamber studies will be conducted under IACP to
assess the mutagenicity of vehicle exhaust before and after
irradiation. These results could provide at least an initial
indication of the photochemical transformation of mobile source
emissions, both gas phase and particle-associated, and the
resulting effect on mutagenicity.
-65-
-------
7.0 ORGANICS ASSOCIATED WITH NON-DIESEL PARTICIPATE
7.1 Emission Rates and Composition
This section will deal with the organic compounds
associated with gasoline particulate emissions.
Table 7-1 presents total particulate and associated
soluble organic fraction (SOF) emissions for the different
vehicle classes.[2,70] Gasoline-fueled vehicles emit far less
particulate than their diesel counterparts. As discussed in
Section 2.1, it is thought that nitro-polycyclic aromatic
hydrocarbons (PAH) such as nitropyrenes, dinitropyrenes and
nitrohydroxypyrenes together account for much of the
mutagenicity of diesel particulate emissions. Particulate
emissions from gasoline-fueled vehicles contain significantly
less of these nitro-PAH's; however, as seen in Table 71, the
mutagenicity of gasoline SOF, expressed as reverents/ug SOF, is
greater than diesel SOF. Also, unlike diesel SOF, the
mutagenic activity of gasoline SOF increases with the addition
of S9 activation, indicating indirect-acting activity. This
suggests that the classical PAH's may be responsible for the
mutagenicity of gasoline SOF, rather than the nitro-PAH's.
7.2 Risk from Non-Diesel Particle-Associated Organics
For gasoline-fueled vehicles, three different approaches
were taken to estimate the risk from non-diesel
particle-associated organics, referred to here as gasoline
particle-bound PIC (products of incomplete combustion). The
first approach estimates the risk of B(a)P emissions from
gasoline-fueled vehicles and assumes no risk from the remaining
gasoline particle-bound PIC emissions. The second approach
uses B(a)P emissions as a surrograte for gasoline
particle-bound PIC emissions. Unlike the first approach, which
uses the unit risk for B(a)P, the second approach uses the PIC
unit risk presented in the Six Month Study. The third approach
uses estimated gasoline particle-associated organic emission
rates together with a unit risk for gasoline
particle-associated organics. All three approaches and
resulting risk estimates will be described in this section.
The first approach uses B(a)P emission factors from
gasoline-fueled vehicles together with the B(a)P unit risk.
The annual cancer risk of B(a)P from gasoline-fueled vehicles
was determined by multiplying the B(a)P risk obtained in the
Six Month Study by the ratio of the emission factors (this
-66-
-------
Table 7-1
Particulate Emissions and Mutaqenicity*
Total particulate,
mg/mile
SOF, mg/mile
SOF, as % of total
particulate
Nitropyrene, ug/mi
TA98, +S9, rev/ug SOF
Leaded
LDGV
102.5
21.1
1
20.6
g/mile 14.6
i 0.20
SOF** 7.3
SOF 12.5
(x!03) 152
(xlQ3) 258
Unleaded
LDGV
31.7
14.4
45.4
3.2
0.24
7.6
13.4
42.1
79.3
Leaded
HDG
735
27.6
3.8
39.5
5.3
16.1
110.5
428.3
LDP
606.8
124.1
20.5
4.5
7.4
4.1
509
HDD
1948
385.2
19.8
2.3
0.88
0.84
287.4
279.4
**
For the light-duty vehicles, emissions were collected during the
FTP; for heavy-duty vehicles, the transient test procedure was used,
Denotes revertants per microgram of the soluble organic fraction
(SOF), using Salmonella typhimurium strain TA98 with or without
metabolic activation (S9) .
-67-
-------
study/Six Month Study).* This approach assumes that there is
no risk from the remaining gasoline particle-bound PIC.
The B(a)P emission factors were taken from references 2
and 70 and summarized in Table 7-1. B(a)P emission factors can
vary substantially from one reference to another. The emission
factors used represent the higher end of the range. Since
gas phase B(a)P is emitted in such small quantities, it is
assumed that the emission factors used adequately represent
what little gas phase B(a)P may also be present. The unleaded
and leaded LDGV weightings were 0.88 and 0.12, respectively to
compute a composite LDGV emission factor. These weightings are
based on MOBILES estimates for 1986. The resulting overall
composite emission factor is 5.56 x 10 ~6 g/mile. The
exclusion of diesel-fueled vehicles does not have much impact
on the composite emission factor since most of the B(a)P is
emitted from gasoline-fueled vehicles.
The composite emission factor of 5.56 x 10~6 g/mile is
roughly one-third of the emission factor used in the Six Month
Study (1.66 x 10"sg/mile). The primary reason for this
difference is that the Six Month Study assumed a 50/50 split
for leaded and unleaded gasoline-fueled vehicles. A B(a)P unit
risk of 3.3 x 10~J was used for both the Six Month Study and
this approach.
The B(a)P risk from motor vehicles estimated in the Six
Month Study is 0.02 cancer incidences per urban million. The
current (1986) cancer risk of B(a)P from gasoline-fueled
vehicles using the first approach was determined by simply
multiplying the risk obtained in the Six Month Study by the
ratio of the composite emission factors (5.56 x 10"6/1.66 x
10~s = 0.33). The resulting annual cancer risk is 0.007 per
urban million, or 1.3 cancer incidences, assuming an urban
population of 180 million in 1986.
In 1995, virtually all LDGV and HDGV will be using
unleaded fuel. The composite B(a)P emission factor will
decrease as a result to roughly 2.95 x 10~6 g/mile. When
this emission factor is used, the resulting cancer risk is
0.004 per urban million, or 0.78 cancer incidences, assuming an
urban population of 195 million in 1995.
* This is the first place in the report where the Six Month
Study exposures are used in place of NEM-predicted
exposures. The Six Month Study used the GAMS dispersion
model developed by EPA's Office of Toxic Substances to
calculate exposures. Area sources were generally assumed
to be distributed equally throughout each county. As a
very crude comparison, assuming the same FTP emission
factor, the NEM-predicted exposure is roughly 1.6 times as
high as that predicted by GAMS.
-68-
-------
The second approach uses B(a)P emission factors from
gasoline-fueled vehicles together with the PIC unit risk (which
is expressed per unit of exposure of B(a)P) to estimate the
annual cancer risk of PIC from gasoline-fueled vehicles. The
PIC unit risk is 4.2 x 10"l. The Six Month Study also used
B(a)P emission factors together with the PIC unit risk to
estimate the PIC risk from motor vehicles. The annual cancer
risk of PIC from gasoline-fueled vehicles was determined by
multiplying the PIC risk obtained in the Six Month Study by the
ratio of the B(a)P emission factors.
The PIC risk from motor vehicles estimated in the Six
Month Study is 2.07 per urban million. When this is multiplied
by the ratio of the B(a)P emission factors (5.56 x 10"s /1.66
x 10"s = 0.33), the resulting annual cancer risk is 0.68 per
urban million, or 122 cancer incidences in 1986. When the
estimated 1995 B(a)P emission factor is used (2.95 x 10"6),
the resulting cancer risk is 0.37 per urban million, or 72
cancer incidences in 1995.
The third approach uses estimated emission rates of
gasoline particle-associated organics as an unspeciated mixture
together with a unit risk for these organics. Exposures were
estimated using the modified NEM model.
Emission rates of gasoline particle-associated organics in
1986 and 1995 were estimated using the methodology in reference
71. Unfortunately, the organic emission factors contained in
reference 71 include both soluble organics and elemental
carbon. The organic emission factors in reference 71 were
revised to include only the soluble organic fraction (SOF).
The SOF emission factors used for the various vehicle classes
are given in Table 7-2. SOF emission factors for light-duty
gasoline-fueled vehicles and trucks were taken from references
70 and 72. SOF emission factors for the heavy-duty
gasoline-fueled trucks were taken from references 73, 74 and
75. The pre-1987 heavy-duty data were obtained using EPA's
transient chassis test procedure. It is assumed that the more
stringent emission standards for 1987 and later heavy-duty
gasoline vehicles in the 8,501-14,000 Ibs range (classes 2B and
3) will require the use of catalysts. The SOF emission-factors
in Table 7-2 differ somewhat from the emission factors in Table
7-1. The reasons for this difference are: 1) the emission
factors in Table 7-2 are specific by model year unlike those in
Table 7-1, and 2) more references were used in compiling Table
7-2.
MOBILES VMT fractions for 1986 and 1995 for the various
vehicle classes were used to calculate composite emission
factors. Composite emission factors were estimated assuming
both the presence and absence of an Inspection/Maintenance
(I/M) program. (The effect of the I/M program is to reduce the
rates of misfueling and catalyst removal, which affect
particle-bound organic emission rates.) An FTP average speed
-69-
-------
Table 7-2
Gasoline Particle-Associated Organic Emission Factors
Light-Duty Gasoline Vehicles and Trucks
Model Year
Pre-1970
1970-1974
1975+
1975+
1975+
1975+
Control
System
NOCAT
NOCAT
NOCAT
CAT
CAT
NOCAT
Fuel
Used
Leaded
Leaded
Leaded
Unleaded
Leaded
Unleaded
Emission Factor
grams per mile
0.0873
0.023"
0.019b
0.005b
0.023C
0.019d
Heavy-Duty Gasoline Trucks
Model Year
Pre-1987
1987+
1987+
1987+
Control
System
NOCAT
CAT
CAT
NOCAT
Fuel
Used
Leaded
Unleaded
Leaded
Leaded
Emission Factor
grams per mile
0.072"
0.013f
0.0509
0.105h
a From reference 72
b From reference 70
c Same as 1970-1974 no catalyst leaded value.
d Same as 1975+ no catalyst leaded value.
e From references 73, 74 and 75. The average class 2B truck
value was assigned a weighting of 60% based on VMT.
f Average class 2B no catalyst leaded truck value multiplied
by ratio of light-duty 1975+ catalyst unleaded and 1975+ no
catalyst leaded values (0.050 x 0.005/0.019).
g Average class 2B no catalyst leaded truck value.
h Average no catalyst leaded truck value for classes 5 and 6.
-70-
-------
of 19.6 mph was assumed. The estimated fleet composite
emission factors are given below.
Gasoline Particle-Associated Organic Emissions (q/mile)
With I/M Without I/M
1986 0.0075 0.0082
1995 0.0048 0.0058
Based on data obtained over a number of different driving
cycles, no trends are apparent with regard to speed.[70]
Therefore, the g/mile emission factors above were used to
calculate g/min emission factors at 10 mph, 19.6 mph and 35
mph, as required by the model. Idle emission factors were also
determined based on data for non-catalyst-equipped and
oxidation catalyst-equipped vehicles.[40] Idle emission
factors for 1986 and 1995 are 0.0196 g/min and 0.0138 g/min,
respectively.
A unit risk estimate for gasoline particle-associated
organics was estimated by EPA scientists in 1983. [7] Like
diesel particulate, the unit risk estimate has not been
reviewed by the Science Advisory Board (SAB) and an official
EPA risk assessment has not been done. It is based on data for
only one catalyst-equipped vehicle. Furthermore, the vehicle
had exceptionally high exhaust emissions, comparable to those
from a non-catalyst-equipped vehicle. It was originally chosen
on this basis since it was easier to collect enough extractable
organics for analysis. The mutagenic activity of the
particle-associated organics from this vehicle, as indicated by
the Ames Salmonella strain TA-98 bioassay, is on the low end of
the range, when compared with other catalyst-equipped
vehicles. As a result, the vehicle should be considered to be
of uncertain representativeness. An upper confidence limit
unit risk estimate based on this vehicle is 2.5 x 10"4. The
bioassays used to estimate the unit risk were the same as those
used to estimate the unit risk for diesel particulate, and the
same approach was used.
Nationwide urban and rural exposure in 1986, using the
modified NEM, is estimated to be 0.20-0.21 ug/m3 and
0.17-0.19 ug/m3, respectively. The range accounts for both
the presence and absence of an Inspection/Maintenance program.
In 1995, nationwide urban and rural exposure is estimated to be
0.13-0.15 ug/m3 and 0.11-0.13 ug/m3, respectively.
Estimates of lung cancer incidence for 1986 and 1995 are
given in Table 7-3. The total risk in 1986 ranges from 163-176
cancer incidences and drops to 115-136 cancer incidences in
1995.
For this report, a range of risk estimates for gasoline
PIC will be reported, which encompasses the results of all
three approaches. The resulting range of cancer incidences is
1.3-176 in 1986 and 0.78-136 in 1995.
-71-
-------
Table 7-3
Annual Lung Cancer Risk from Gasoline Particle-
Associated Orqanics
Urban Rural Total
1986
With I/M 1.3-127 0-36 1.3-163
Without I/M 1.3-136 0-40 1.3-176
1995
With I/M 0.78-90 0-25 0.78-115
Without I/M 0.78-106 0-30 0.78-136
-72-
-------
8.0 DIOXINS
8.1 Composition
Over 75 different chlorinated dioxin isomers have been
identified. One of the 22 isomers with four chlorine atoms is
2,3,7,8-tetrachlorodibenzo-p-dioxin (2378-TCDD). This dioxin
compound has a high molecular weight and exists in the
particulate state or is adsorbed onto particulates. It is the
dioxin compound of most interest since it is thought to be the
most toxic of the chlorinated dioxins and is most often
associated with exposure and potential health risks to humans
based on available data.
8.2 Emissions
Some qualitative analytical measurements have found dioxin
to be present in the muffler scrapings of vehicles using either
leaded or unleaded gasoline. Also, some measurements have
shown this compound to be present in diesel particulate. Some
work has been done attempting to measure this compound
quantitatively in vehicle exhaust. Since it is apparently
emitted in only trace quantities (e.g., 10"9 g/mile), it is
very difficult to collect enough particulate sample for
analysis.[76] Also, a radioactive tracer compound must be used
to correct for losses in sample work-up.
8.3 Concentrations of Dioxins
Most reports of environmental contamination from dioxins
have concerned non-air releases of dioxin from industrial or
chemical processes (e.g., pesticide production) or measurement
of dioxin levels in fish in waters that could have been
contaminated by non-air routes. At this point, there are no
accurate measurements of ambient levels of dioxin since it
would be present in such low levels in the ambient air. Thus,
it is not possible to say for certain whether the known sources
of dioxin such as municipal and industrial incinerators account
for most of the dioxin found in the ambient air although it is
thought that such sources account for a significant fraction of
dioxin emitted. A recent EPA OAQPS report states that, due to
the low chlorine level (0-100 ppm) in gasoline and diesel fuel,
mobile sources may not be a significant source of dioxin.[76]
8.4 Current Activities
OAQPS plans to make a decision within the coming year on
whether to list dioxin as a hazardous air pollutant under
Section 112 of the Clean Air Act. Also, OAQPS has recommended
that QMS consider a program to quantitatively measure dioxin
emissions from mobile sources. QMS is presently considering
whether work in this area would be useful and has asked for
some additional OAQPS input on the relative priority of work on
dioxins from mobile sources versus other sources.
-73-
-------
9.0 ASBESTOS
9.1 Emissions and Ambient Concentrations
Asbestos, due to its high friction and heat resistance
characteristics, is used in brake linings, clutch facings and
automatic transmissions. About 22% of the total asbestos used
in 1984 was used in motor vehicles. Previous EPA work involved
measurement of asbestos emissions from vehicles during typical
vehicle operations for a vehicle with front disc brakes and
rear drum brakes. [77,78] This work showed that the asbestos
emissions were 4 ug/mile and possibly as high as 28 ug/mile.
It is estimated that the 4 ug/mile emission figure would result
in a maximum annual average asbestos level in a central city
area of about 0.25 ng/m3.[79] This is in agreement with what
would be predicted in a typical street canyon, using a method
developed by Southwest Research Institute under EPA
contract.[80] Urban asbestos levels from all sources in large
cities show average readings of 2.6-5 ng/m3 but average
levels in New York range from 8 to 30 ng/m3; localized
asbestos levels in dense traffic can be somewhat higher. Thus,
asbestos from mobile sources is responsible for 1-10% of the
total asbestos although a higher fraction (7-70%) could result
with the high emission factor.
9.2 Cancer Risk
The National Academy of Sciences (NAS) has estimated
lifetime risks for persons in urban areas.[81] Based on the
data in the NAS report, unit risk estimates range from 6.6 x
10~7 to 2.6 x 10~5 per ng/m3 exposure (6.6 x 10~4 - 2.6
x 10~2 per ug/m3) . On an annual basis, this translates to
9 x 10~9 to 3.6 x 10~7 per ng/m3 exposure (NAS assumed an
average lifetime of 73 years).
Maximum annual average asbestos levels in urban areas due
to motor vehicle emissions are estimated to range from 0.25 to
1.75 ng/m3, based on estimated emission rates of 4 to 28
ug/mile. Using these exposure estimates in conjunction with
the range of unit risks, and assuming an urban population of
180,000,000, the resulting cancer risk is estimated to range
from 0.405-113.4 cancer incidences per year.
9.3 Current Activities
Due to the risk to the general population as well as the
risk to workers exposed to asbestos, the EPA Office of
Pesticides and Toxic Substances, in January 1986, has proposed
regulations under Section 6 of TSCA to ban certain uses of
asbestos (i.e., asbestos-cement pipes, flooring tiles, and
asbestos clothing) and to allocate permits to mine and import
asbestos which would restrict its remaining uses.[79] The
-74-
-------
National Resources Defense Council petitioned EPA in 1984 to
ban the use of asbestos in motor vehicles. However, EPA feels
that effective substitutes are still not available for many
applications of asbestos in brakes. EPA feels that the
restrictions on use of asbestos will encourage the development
of suitable replacements for brake linings. EPA is also
considering a ban on asbestos friction products about 5 years
after the final rules are promulgated. Presently,
semi-metallic disc brake pads made without asbestos are used in
about 85% of new domestic cars with front wheel drive. Also,
some manufacturers are introducing aramid fiber instead of
asbestos for disc brakes. Development of non-asbestos
substitutes for drum brakes has not yet been as successful as
for disc brakes due to concerns about durability and heat
resistance. Still, one automobile manufacturer is using
semi-metallic drum brake linings in its new minivans. Another
has reported progress in developing a drum brake lining with
aramid fiber.
The proposed regulations would involve EPA issuing permits
for mining of asbestos in the U.S. and importation of asbestos
and its products. The total amount of asbestos to be permitted
under these permits for the first year after the regulation is
promulgated would be 30% of the average amount of asbestos
mined or imported yearly during the base period of 1981, 1982,
and 1983. The amount of the asbestos used in the following
years would decrease annually by 3% from the 30% level until it
reached a 3% level in the tenth year of the regulation after
which no asbestos could be mined or imported. Recent findings
that there are not yet available good substitutes to replace
asbestos in certain automobile and truck brakes, however, may
push back EPA's goal for banning and phasing out asbestos for
such uses.
-75-
-------
10.0 VEHICLE INTERIOR EMISSIONS
10.1 Composition and Concentrations
Two different EPA projects were conducted measuring
different compounds in vehicle interiors.[82,83]
One project showed vinyl chloride to be present at levels
ranging from below 2 ppb to 7 ppb in several closed vehicles
under high temperatures as might exist on a summer day.[82]
Under similar conditions other carcinogenic compounds were
identified qualitatively. The carcinogenic compounds detected
are listed below.
aniline
biphenyl
1,2-dibromoethane
dichlorobenzene
dimethylphenol
isobutyl alcohol
maleic anhydride
naphthalene
benzene
carbon tetrachloride
chloroform
phenol
A total of 147 compounds were identified.
In another EPA project, 58 vehicle interiors were sampled
for nitrosamine which is a potent carcinogen.[83] The main
compound measured is N-nitrosodimethylamine which was found at
levels of about 0.05 ug/m3. This exposure level for 3
hours/day which would represent a long commute is less than
that from a can of beer or a strip of bacon a day.
Also, from time to time, complaints have been received by
EPA on the formation of a white semi-opaque film on the
interior of a vehicle windshield. This film very slightly
obstructs visibility and can be somewhat difficult though not
impossible to remove. EPA ORD has analyzed samples of this
film and finds it to consist of a phthlate ester; dioctyl
phthlate is used as a plastisizer inside vehicles. The
exposure level to and the health effects of this compound are
not known.
10.2 Cancer Risk and Current Activities
At this point, no risk assessment calculations have been
made for these vehicle interior emissions (with the exception
of some simple calculations for the nitrosamines). Since the
exposure level is low, it is thought that no significant risk
could be occurring due to exposure to these substances. No
further work is planned to identify other substances present in
vehicle interiors at this time.
-76-
-------
11.0 METALS
This section discusses metals which are emitted in vehicle
exhaust. Lead, manganese , platinum and cadmium are
considered. For each metal, available information regarding
the source of the metal, emission rates and health effects are
given. In addition, for cadmium, a risk estimate is presented.
11.1 Lead
11.1.1 Source and Emission Factors
Lead, in the form of tetraethyl lead and tetramethyl lead,
is used in gasoline fuel to increase the octane number of
gasoline and suppress knock. Knock is the premature
autoignition and very rapid combustion of the fuel-air charge
in the engine combustion chambers. Knock is frequently audible
as a sharp metallic rap, and it may cause damage to engines.
Lead alkyl compounds were first added to gasoline in 1923 as a
means of suppressing engine knock by promoting uniform burning
of the fuel-air mixture in the engine combustion chambers.
When lead alkyl compounds are burned, lead oxide is
formed. To reduce the tendency of lead oxide to build deposits
in automobile engines, halogen compounds are included in the
fuel to scavenge the lead deposits from the engine. Ethylene
dichloride and ethylene dibromide are commonly used. Analysis
of lead particulate indicates that most of the exhausted lead
appears as PbClBr.[71]
Vehicles equipped with catalytic converters are required
to use unleaded gasoline since the presence of lead poisons the
catalytic converter. Catalyst-equipped vehicles were
introduced in 1975. Unleaded gasoline is subject to the
requirement that it not contain any lead additives and that it
not include more than 0.05 grams per gallon (gpg) lead.
Due to the widespread and persistent misuse of leaded
gasoline in vehicles designed for unleaded gasoline (termed
misfueling or fuel switching) and the adverse health effects of
lead in gasoline, EPA promulgated regulations to reduce the
lead content of gasoline.
In 1973, EPA required refiners to meet a 0.5 gpg standard
for the average lead content of all gasoline. EPA later
replaced this standard with a standard for the lead content of
leaded gasoline only. Effective November 1, 1982, large
refineries were required to meet a standard of 1.10 grams per
leaded gallon (gplg). Certain smaller refineries were subject
to a 1.90 gplg standard until July 1, 1983, at which time they
would also be subject to the 1.10 gplg standard. [84] EPA
-77-
-------
further reduced the standard to 0.10 gplg effective on
January 1, 1986, with an interim standard of 0.5 gplg effective
on July 1, 1985. [85] EPA is now investigating the possibility
of regulatory action banning the use of leaded gasoline. A
range of alternatives is being considered, ranging from no ban
to a ban in 1995 to a ban effective as early as January 1,
1988.[86]
Lead emissions from mobile sources are calculated based on
the percentage of burned lead exhausted at different speeds,
the lead content of gasoline, vehicle fuel economy and the
model year mix of vehicles on the road. A report containing
guidelines for predicting lead emission factors is
available.[87] Based on the information in this report,
simplified lead emission factors have been estimated for
light-duty and heavy-duty model year 1986 (MY1986) vehicles.
These emission factors are summarized in Table 11-1. Note that
these estimates are for the mass of lead alone. To account for
the halogens, these lead emission estimates should be
multiplied by 1.557.[71]
Composite calendar year 1986 emission factors will vary
from area to area depending on the model year mix of vehicles
on the road, the driving conditions (which will affect fuel
economy) and the misfueling rates in a particular area. Lead
emission factors for heavy-duty gasoline-fueled vehicles are
projected to decrease in later years since it is assumed that
emission standards effective in 1987 will require virtually all
new heavy-duty gasoline-powered vehicles under 14,001 pounds
gross vehicle weight to use catalytic converters and thereby
burn unleaded fuel.
11.1.2 Health Effects
A strong correlation has been demonstrated between
gasoline lead usage and blood lead levels. The Centers for
Disease Control (CDC) has defined 25 ug/dl as an elevated blood
lead level. The list of demonstrated health effects at blood
lead levels exceeding 30 ug/dl is well established.[88] Such
effects include: l) death at blood levels of 80+ ug/dl; 2)
frank anemia, anorexia, abdominal pain, and vomiting at 70
ug/dl; 3) reduced hemoglobin at 40 ug/dl and interference in
heme synthesis at levels down to 15-20 ug/dl; and 4) vitamin D
metabolism interference at PbBs possibly as low as 12 ug/dl.
In addition, there is a great deal of evidence on
neurological effects in children. At levels of 30+ ug/dl,
studies have found significantly slowed nerve conduction, fine
motor dysfunctions, impaired concept formation, lower IQ, and
altered behavior among pre-school children. At PbBs as low as
15-30 ug/dl, a number of studies suggest possible small effects
on IQ, behavioral dysfunctions (attentional deficits, poor
classroom behavior), and changes in electrical brain wave
activity and hearing function among children. Children appear
-78-
-------
Vehicle
LDGV catalyst
LDGV cat (misfueled)
HDGV
LDGV non-cat**
Table 11-1
Lead Emission Factors for MY1986 Vehicles
1986 Fuel
Economy mpg*
23.8
23.8
9
13
1986 Lead
Fuel Content qpg
0.014
0.10
0.10
0.10
Fraction of Lead
Burned That is
Exhausted
0.75
0.44
0.75
0.75
MY1986 Lead
Emissions gpm
0.0004
0.002
0.008
0.006
Combined city/highway fuel economy.
This estimate is for a non-catalyst-equipped vehicle for comparison to the 1986
fleet. Fuel economy was based on data for pre-1975 light-duty gasoline-fueled
vehicles.
-7Q-
-------
to be particularly susceptible to adverse health effects from
lead exposure. Recently emerging data from several ongoing
longitudinal studies consistently indicate that fetal exposures
to lead, at maternal or umbilical cord PbB levels as low as
10-15 ug/dl are associated with reduced birth weight and early
growth, and delays in early mental, motor, and emotional
development.
In addition, more recent data provide convincing evidence
for strong associations between blood pressure increases and
blood lead levels, even at blood lead levels below 30 ug/dl.
It is difficult to conclude what role lead may play in the
induction of human tumors. Epidemiological studies of
lead-exposed workers provide no definitive findings. Lead
acetate has produced renal tumors in some experimental animals
but does not seem to be a potent carcinogen.[89]
11.2 Manganese
11.2.1 Source and Emission Factors
A nonleaded organometallic additive which was used widely
in the middle 1970's as an octane improver in unleaded gasoline
is methycyclopentadienyl manganese tricarbonyl (MMT).
In September, 1978, EPA banned its use in unleaded
gasoline because of evidence which showed that MMT increased
hydrocarbon emissions and plugged catalysts.[90] This ban was
temporarily suspended for four months (June through September)
in 1979 because of concerns over potentially short supplies of
unleaded gasoline.[91] MMT has always been allowed in leaded
gasoline, but has been used at only low levels due to its high
cost compared to lead additives.
More recently, a particulate trap regeneration system
based on use of a manganese (Mn) containing fuel additive has
been selected by Volkswagen as a means of meeting the 1986
California and 1987 Federal light-duty diesel particulate
standard of 0.2 g/mile. EPA has given Volkswagen temporary
approval to use this additive for the 1986-1988 model years.
Volkswagen estimates that for highway driving, emission
rates will range from an expected value of 1.38 mg Mn/mile to a
worst-case value of 4 mg Mn/mile. For street driving, expected
emissions are 3.13 mg Mn/mile with a worst-case value of 10 mg
Mn/mile.[92] The worst-case numbers are based on the
assumption that all of the manganese is emitted in the exhaust
while the expected numbers reflect expected manganese retention
in the engine and particle trap. Data are presently lacking on
the size distribution of particles emitted in the exhaust and
the amount and species of manganese they contain.
-80-
-------
The California Air Resources Board (GARB) estimates that,
by 1990, between 2.6 percent and 6.6 percent of the California
vehicle fleet will have traps. Using this estimate of trap
usage, ambient Mn concentrations corresponding to highway
emissions of 1.3 - 4 mg Mn/mile are estimated to be 0.03 - 0.24
ug Mn/m3. Ambient concentrations corresponding to street
canyon emissions of 3.13 - 10 mg' Mn/mile are estimated to be
0.06 - 0.5 ug Mn/m3. [92]
11.2.2 Health Effects
In June, 1984, Volkswagen requested that the Health
Effects Institute (HEI) undertake an evaluation of potential
health issues related to emissions from diesel automobiles
using manganese-containing fuel additives for particle trap
regeneration. This section briefly summarizes the highlights
of HEI's October, 1985 response to Volkswagen.[92]
The effects of greatest concern are neurotoxic and
respiratory. The neurotoxic effects in humans reguire at least
several months of exposure and progress from an early
reversible stage to a more advanced and irreversible stage.
The human clinical and epidemiologic literature suggests that
neurotoxicity is not strongly indicated until exposure exceeds
5 mg Mn/m3, but that neurological symptoms may occur at
levels as low as 0.3 mg Mn/m3.
The respiratory effects of manganese are identical to
those associated with exposure to fine or respirable
particulate matter in general, and involve an inflammatory
effect or pneumonitis, which may lead to diminished pulmonary
function, bronchitis, or altered susceptibility to infection.
EPA considers respiratory symptoms to be the critical effect
because respiratory effects are reported at levels lower than
those reported for neurotoxicity. EPA derived human No
Observed Effect Levels (NOEL) of 5 ug Mn/m3 based on rat data
and 8.7 ug Mn/m3 based on monkey data. [93]
Occupational studies generally show that respiratory
effects follow exposures in excess of 5 mg Mn/m3. However,
data from a Japanese study of junior high school students
exposed to manganese emissions from a ferromanganese plant
apparently associates increased respiratory symptoms and
diminished pulmonary function with exposures below 5 mg/m3.
Exposure was determined by the amount of manganese in the
dustfall. EPA estimated, on the basis of analyses of dustfalls
near a ferromanganese plant in the Kanawha Valley in West
Virginia, that the dustfall in the Japanese study may translate
to between 3 and 11 ug Mn/m3 . These are the lowest levels
associated with adverse health effects; however, the
correlation of settled dust with suspended particulate matter
is subject to broad variability, even in the same geographic
-81-
-------
locations, as a result/of air turbulence, humidity, topography,
windspeed, etc. As a result, there is a great deal of
uncertainty regarding the level of exposure to manganese
concentrations associated with the health effects in this
study. HEI is presently evaluating the study in detail.
There is some evidence of carcinogenic activity of
manganese in laboratory animals, although the value of these
studies is questionable. There is no epidemiologic information
relating manganese exposure to cancer occurrence in humans.
11.3 Platinum
11.3.1 Source and Emission Factors
Platinum is a catalyst attrition product. It is emitted
in small quantities. Most studies that have attempted to
characterize platinum emissions report no detectable levels.
The measurement methods used are often not sensitive enough to
detect the small quantities of platinum emitted.
Limited quantitative measurements of platinum have been
made. In 1978, Ford measured FTP particulate platinum
emissions from two vehicles (1976 LTD with monolithic three-way
catalyst, oxidation catalyst and air pump and a 1978 Pinto with
monolithic three-way catalyst, oxidation catalyst and air
pump). Platinum emissions ranged from 0.4 to 1.4 micrograms
per mile (ug/mile).[94] In 1977, General Motors reported that
platinum emissions from pelletized catalysts ranged from 1 to 3
ug/mile.[95]
It has been thought for some time that much of the
platinum emitted may accumulate alongside roadways. This was
confirmed in a recent study.[96] Dust samples collected from
the leaves of roadside plants contained as high as 0.7 ppm of
platinum and 0.3 ppm of palladium (another catalyst attrition
product). The highest concentrations of both metals were found
in dust collected from plants growing at the edge of heavily
trafficked streets and highways and the lowest in samples
collected from plants growing in the yards of houses located on
lightly trafficked streets. The concentrations of both metals
in the dust samples are much higher than the reported natural
abundances of these elements.
11-3.2 Health Effects
In 1977, the National Academy of Sciences prepared a study
which examined the uses, sources of supply, and potential
health effects of noble metals. With regard to platinum and
palladium emissions from catalyst-equipped vehicles, the HAS
study concluded:
-82-
-------
"Minute quantities of platinum and palladium (about 1-3
ug/mile) are emitted from the exhaust systems of
automobiles equipped with catalytic converters; much of
this material may accumulate alongside roadways. However,
this material is in a chemical form that is
physiologically innocuous (no detectable soluble salts),
and it is concluded that such emission poses no threat to
the environment. Because there is no evidence that
platinum metal can be methylated by microorganisms and
solubilized in the same way that mercury is methylated,
this deposited material should not have an adverse effect
on the environment."[97]
Recently, Dr. Hans A. Nieper of the Paracelsus Clinic in
West Germany has claimed that platinum emissions from
catalyst-equipped vehicles are associated with rising
incidences of AIDS, cancer and leukemia.[98] He also states
that slight traces of platinum are extraordinarily toxic
against the genetic defense systems. These claims are not
currently scientifically supportable.
11.4 Cadmium
11.4.1 Source and Emission Factors
It is postulated that cadmium is present in fuel as a lead
contaminant and therefore emitted by vehicles burning leaded
fuel. Cadmium emissions from non-catalyst-equipped vehicles
are roughly 1.6 x 10~s g/mile.[32] Assuming 12 percent of
the current light-duty fleet are non-catalyst-equipped
vehicles, the resulting light-duty fleet emission factor is 1.9
x 10~6 .
11.4.2 Health Effects and Risk Estimate
A unit risk estimate for cadmium is estimated by EPA to be
1.8 x 10"3. The risk from cadmium was determined by
multiplying the risk determined in the Six Month Study by
ratios of the updated to the Six Month Study emission factors
and unit risks to obtain a crude estimate of the risk. The
resulting risk of 0.001 per urban million appears negligible.
This risk is further projected to decrease to zero by 1995 as
non-catalyst-equipped vehicles are phased out of the fleet.
-83-
-------
12.0 SIX MONTH STUDY: SUMMARY AND COMPARISON OF RESULTS
12.1 Purpose of Six Month Study and Summary of Results
The final study is formally titled, "The Air Toxics
Problem in the United States: An Analysis of Cancer Risks for
Selected Pollutants."[1] The EPA study attempted to assess the
magnitude and nature of the air toxics problem by developing
estimates of the cancer risks posed by selected air pollutants
and their sources.
Three major analyses were used to estimate excess cancer
risk from exposure to 15-45 toxic air pollutants. The Ambient
Air Quality study used ambient data for five metals, 11 organic
compounds, and benzo (a) pyrene (B(a)P) together with unit
risks for these compounds to estimate national excess cancer
incidence. The other two analyses (the NESHAPS study and the
35 County study) used exposure models together with unit risks
to estimate excess cancer incidence. The NESHAPS study
'provides national estimates for about 40 compounds. The 35
County study was limited to 22 compounds and 35 counties but
was designed to allow more detailed assessment of source
contributions. Since the 35 County study was used to estimate
the mobile source contribution to the air toxics risk, it will
be discussed in more detail. The other two studies do not
contradict any relevant findings from the 35 County study.
The 35 counties were chosen from counties with the highest
expected ambient exposures, based on high populations and large
aggregate emissions from all source categories. Also, some
counties were included simply because they contain large
industrial point sources of potential interest. The 35
counties contain roughly 20 percent of the total U.S.
population (45 million people), 20 percent of total releases of
VOC, and 10 percent of total particulate matter loading, based
on 1982 data. They represent a variety of industrial and
population distributions, but are not considered a
statistically representative sample of the country.[99] When
comparing the results of the 35 County study with the results
of this study, cancer incidences will be expressed per million
people exposed.
The aggregate cancer risk from all sources in the 35
County study is 207, or 4.6 cancer incidences per million, when
products of incomplete combustion (PIC) are included. Mobile
sources were estimated to account for 60 percent of the total
incidence when PIC are included and 23 percent of the total
incidence when PIC are excluded. A breakdown of the specific
mobile source pollutants responsible and the risk associated
with each pollutant is given in Table 12-1.
-84-
-------
Table 12-1
Six-Month Study Results for Mobile Sources*
Pollutant
PIC
Benzene
Formaldehyde
Ethylene
B(a)P
Cadmium
Total
93.2
14.0
le 3.1
ibromide 0 . 9
0.8
0.4
Annual Cancer Incidence
Per Million** %
2.07
0.31
0.07
0.02
0.02
0.01
Contribution
82.9
12.5
2.7
0.8
0.7
0.4
TOTAL: 112.4 2.50 100.0
* Gasoline vapors were considered separately. The total
risk was 6.8 or 0.15 per million.
** Based on 45 million people in 35 County study.
-85-
-------
PIC refers to a large number of hydrocarbon compounds
(mainly polynuclear organics). The PIC unit risk value is
derived from various older epidemiological data and is
controversial. Epidemiological studies of the general
population exposed to ambient air and studies of workers
occupationally exposed to PIC (e.g., coke oven emissions, hot
pitch fumes) were used to estimate the PIC unit risk. B(a)P
was used as an index of exposure to PIC. As a result, the PIC
unit risk is expressed per unit of exposure of B(a)P. The PIC
unit risk used in the Six Month Study is 4.2X10'1. The
uncertainties associated with this unit risk are detailed in
the Six Month Study.
The B(a)P emission factor for mobile sources was used to
calculate an annual average exposure. The B(a)P exposure was
then multiplied by the PIC unit risk to estimate annual cancer
incidence due to PIC from mobile sources. The same process was
performed for all other sources emitting B(a)P to determine
cumulative annual cancer incidence from PIC.
In the 35 County study, roughly 75 percent of the total
B(a)P emissions was said to be attributable to mobile sources.
Sources emitting the remaining 25 percent are residential
fireplaces, woodstoves and waste oil burning, all considered
area sources. Point sources were assumed to emit no B(a)P,
since coke oven emissions were dealt with separately in the
report. As a result, 75 percent of the total PIC cancer
incidence is due to mobile sources. PIC cancer incidence
accounts for a large portion of the aggregate cancer incidence
because the unit risk for PIC is orders of magnitude greater
than the unit risks for the other toxic compounds examined in
this study. This explains why, when PIC are included, the
fraction of total cancer incidence related to mobile sources
increases from 23 to 60 percent.
From Table 12-1, the pollutants responsible for the total
estimated cancer incidence from motor vehicles in the 35 County
study, in order of importance, are PIC, benzene, formaldehyde,
ethylene dibromide (EDB), B(a)P and cadmium. Gasoline vapors
from motor vehicles were not included in the motor vehicle
category. Instead, gasoline vapors from service stations,
which includes vehicle refueling and delivery of gasoline, were
considered separately.
12.2 Comparison of Six Month Study Results with Results of This
Study
For benzene, formaldehyde, gasoline vapors and
1,3-butadiene, the annual cancer incidence per million
estimated in both studies can be directly compared. For PIC,
the PIC risk estimated in the Six Month Study will be compared
to the combined risk from diesel particulate and gasoline
POM/PIC estimated in this study. For EDB and cadmium, risks
have not been estimated in this report due to the rather low
-86-
-------
emissions and the lack of data on emissions at various speeds
required as inputs to the NEM model. For these pollutants,
rough estimates of the risk will be made by comparing the
emission factors used in the Six Month Study with those which
will be developed in the subsequent sections, and adjusting the
risks obtained in the Six Month Study accordingly. This
approach assumes that the emissions are directly associated
with ambient concentrations. This appears to have some merit,
considering that EDB and cadmium are not considered
photochemically reactive.
In addition to emission factors, the unit risks used in
both studies should be compared. In some cases, the unit risks
have changed substantially since release of the Six Month
Study. 1,3-Butadiene is one notable example.
In the following sections, the risks from each pollutant
will be compared. The emission factors and unit risks used to
calculate the annual cancer incidences will also be compared in
order to pinpoint the discrepancies. Finally, the aggregate
risks will be compared. Table 12-2 presents a summary of this
comparison that will be referred to in the following
subsections.
12.2.1 Formaldehyde
Before comparing the composite emission factors for
formaldehyde, the vehicle class VMT fractions used will be
briefly discussed. For this study, the VMT fractions are those
used in MOBILES for calendar year 1986. For the Six Month
Study, the supporting document prepared by Versar for the 35
County study included VMT data for only one of the 35
counties. [99] Also, the emissions and VMT data in the 35
County study were grouped into three major vehicle categories
(LDG, HDG and HDD). The MOBILES VMT fractions were grouped
into these three vehicle categories and compared to the VMT
fractions for the single county. The VMT fractions were
generally similar; the MOBILES heavy-duty VMT fractions were
somewhat higher. As a result, the MOBILES grouped VMT
fractions were applied to the emissions data for LDG, HDG and
HDD given in the Versar report to calculate composite emission
factors for the 35 County study.
As seen in Table 12-2, both the composite emission factor
and the unit risk for formaldehyde used in this report are
higher than those used in the 35 County study (referred to as
the Six Month Study in Table 12-2). The emission factors used
in the 35 County study were those derived from studies of low
mileage vehicles. In this study, the emission factors were
percentages of total exhaust HC, as predicted by MOBILES for
1986, and were designed to account for in-use deterioration.
Also, since the release of the Six Month Study, the unit risk
-87-
-------
0)
O
10
O
0)
u
u
3
O
01
0)
rH tl
H 3
T3
C 3
O 4-»
i-H (Q
H -H
E x:
CH
(o
o
i
r
vo
ro
I
vO
vO
t-
O
1
O
O
VD
ro
vO
ro
I
o
Cn
CNJ
O
I
O
ro
o
i
O
o
o
o
CO
in
I
o
oo
(M
i-H
0)
X)
01
01
0)
M.
u
a
o
Oi C
rH *- O
3*S
C 01 vO
o
4->
00
01
r-
o
CO
o
CM
O
in
i-H
o o
p
2
o
o
(M
o
in
vO
u
4->
rl
c
X
o
CO
o
rH
X
VO
X
CO
I
o
(M
X
CO
T o 7
O vO O
X
CO
rg
I
vO
vO
X
00
X
CM
X
ro
X
in
0) 3
rH 4J
H ryj
^ w
*- !c
H
u
o
4-1
o
0) IT]
rl XI
M C 4-1
o o c
Ql-H O
E en S
O CO
U -H VO
o
I
o
o
CO
CO
o
o
o
X
VO
vO
eg
o
o
I
p
2
VO
rg
o
CO
X
oo
X
ro
(Nl
o
rH
X
o
i-H
X
in
o
rH
X
rH
in
o
i-H
x
3
CO
rH
O
U
P
O
i-H
X
VO
T)
X
VO
vO
4J
C
(0
o
O4
0)
x;
0)
T3
U
O
0)
C
0)
N
0)
CQ
U
M
a,
m
p
0)
4-1
(0
i-H
3
U
m
[0
0)
X
O
p
p
X
in
00
01
a
(0
0)
o
01
(0
o
0)
0)
H
TJ
(0
4J
CQ
I
CO
O)
C
o>
01
O
01
0)
XI
0}
(0
U
x
CO
O)
-O
O
XI
H
P
0)
C
0)
rH
x;
u
o
vV
01
01
g
c
0
H
1 1
1 1
H
S
o
00
rH
MH
O
C
o
H
) 1
m
rH
3
O4
O
O4
c
1C
^J
3
fj
<
«
anics from gasoline-fueled vehicles were considered to
risk was not determined in this study.
?0
*O
0) «
4-1
H
1 X
U
if
s w
* .H
X
(U i)
4-)
(0 c
y o
rl r-l
4J p,
U
(0 4->
ft C
0)
a>
GO
3
I
oo
OO
I
-------
for formaldehyde has increased substantially. The unit risk
used in this report is an upper 95 percent confidence limit and
includes only malignant tumors.
The annual cancer risk for both studies assume an average
lifetime of 70 years. The risk per million for the 35 County
study is based on a population of 45 million. This population
was assumed to be urban. The calendar year 1986 annual cancer
risks in this study presented in Table 12-2 are based on an
estimated 1986 urban population of 180 million.
When the annual cancer risk for formaldehyde given in the
35 County study is adjusted to account for the updated unit
risk estimate, the resulting annual cancer risk is 0.15 per
million. This still falls below the low end of the range given
in this study. The range (0.22-0.70 per urban million)
attempts to account for formaldehyde photochemistry. The range
is large because 1) it is difficult to determine what the
annual average formaldehyde concentration is for any particular
area, much less the entire urban population, and 2) the
relative contribution of mobile and stationary sources is not
known. For example, the low estimate (0.22 per urban million)
is based on a NEM-predicted concentration of 1.21 ug/m3 for
direct emissions of formaldehyde from mobile sources. The
highest estimate (0.70 per urban million) assumes a maximum
annual average concentration of 12.7 ug/m3 based on
monitoring data, with 30 percent attributable to mobile
sources, and attempts to account for formaldehyde
photochemistry.
From comparing the two risk estimates (the 35 County
estimate and the low estimate in this study) and accounting for
differences in the emission factors and unit risks, it appears
that the NEM used in this study predicts slightly higher
exposures per g/mile emitted than the dispersion modeling used
in the 35 County study. This should be kept in mind when
comparing the results for benzene, diesel particulate, gasoline
POM/PIC and 1,3-butadiene as well since the modified NEM was
used to estimate exposures for these pollutants.
12.2.2 Benzene
As seen in Table 12-2, the risks for benzene are more
similar; however, the risk obtained in the Six Month Study
still falls below the range of risk estimates determined in
this study. The low estimate in this study is based on a
NEM-predicted concentration whereas the high estimate is based
on available ambient monitoring data, assuming 85 percent is
attributable to mobile sources. The 35 County study used
greater emission factors but a lower unit risk. Like
formaldehyde, the unit risk for benzene has increased since
release of the Six Month Study.
-89-
-------
12.2.3 PIC, B(a)P, Diesel Particulate and Gasoline PIC/POM
The Six Month Study calculated a risk estimate for
products of incomplete combustion (PIC). The approach used has
been previously described. Benzo (a) pyrene (B(a)P) was
treated separately. In this study, information existed to
enable separate treatment of diesel and gasoline PIC. Diesel
particulate was considered to represent diesel PIC while
particle-associated organic emissions from gasoline-fueled
vehicles were considered to represent gasoline PIC. The sum of
these risks will be compared to the PIC risk obtained in the
Six Month Study.
The risk from diesel particulate ranges from 0.76 - 3.67
per urban million. The range is due to the range of unit risks
chosen.
For gasoline-fueled vehicles, three different approaches
were taken to estimate the risk from gasoline PIC. These
approaches are discussed in detail in Section 7. The first
approach estimates the risk of B(a)P from gasoline-fueled
vehicles. B(a)P emission factors are used together with the
B(a)P unit risk. The annual cancer risk of B(a)P from
gasoline-fueled vehicles was determined by multiplying the
B(a)P risk obtained in the Six Month Study by the ratio of the
emission factors (this study / Six Month Study). This approach
assumes that there is no risk from the remaining gasoline PIC
emissions.
The composite B(a)P emission factor of 5.56 x 10"6
g/mile used in this study is roughly one-third of the emission
factor used in the Six Month Study (1.66 x 10"s). The main
reason for this difference is that the Six Month Study assumed
50 percent of the gasoline-fueled vehicles were leaded while
this study used a more realistic percentage of 12 percent.
When the B(a)P risk estimated in the Six Month Study (0.02
cancer incidences per urban million) is multiplied by the ratio
of the composite emission factors (5.56 x 10"6/1.66 x
10~s), the resulting annual cancer risk is 0.007 per urban
million.
The second approach uses B(a)P emission factors from
gasoline-fueled vehicles together with the PIC unit risk (which
is expressed per unit of exposure of B(a)P) to estimate the
annual cancer risk of PIC from gasoline-fueled vehicles. This
is the approach used in the Six Month Study. The annual cancer
risk of PIC from gasoline-fueled vehicles was determined by
multiplying the PIC risk obtained in the Six Month Study by the
ratio of the B(a)P emission factors.
-90-
-------
When the PIC risk estimated in the Six Month Study (2.07
per urban million) is multiplied by the ratio of the composite
emission factors (5.56 x 10"6/l-66 x 10~5), the resulting
annual cancer risk is 0.68 per urban million.
Unlike the second approach, the third approach does not
use B(a)P as a surrogate for PIC emissions. The third approach
uses estimated emission rates of gasoline particle-associated
organics (as an unspeciated mixture) together with a unit risk
for these mixed organics. Exposures were estimated using the
modified NEM.
Estimated composite emission factors for gasoline
particle-associated organics in 1986 are estimated to be 0.0075
- 0.0082 g/mile, depending on whether a minimum I/M program was
assumed to exist. An upper confidence limit unit risk is 2.5 x
io-4.
Nationwide urban exposure in 1986, using the modified NEM
is estimated to be 0.20-0.21 ug/m3. Estimates of urban lung
cancer incidence in 1986 range from 127-136, or 0.71-0.76 per
urban million.
The resulting range of risk estimates for gasoline PIC,
using the results of all three approaches, is 0.007-0.76 per
urban million.
The sum of diesel particulate and gasoline POM/PIC risks,
or equivalent "PIC" risks, range from 0.77-4.43 per urban
million. The PIC risk obtained in the 35 County study (2.07
per urban million) lies within this range.
12.2.4 Gasoline Vapors
The annual cancer risk from gasoline vapors reported
(based on other documents) in this study is 0.36 per urban
million. In the Six Month Study, the risk from gasoline
marketing was considered. Gasoline marketing in the Six Month
Study refers to emissions from service stations. It includes
emissions from vehicle refueling and from the delivery of
gasoline to the service station. The resulting risk was 0.15
per million. This is less than half the risk presented in this
study. It is mentioned, however, in the supporting Versar
report that high exposures near the pump from self-service
refueling were excluded. The vehicle refueling in the Six
Month Study then refers to occupational exposure (service
station attendants). The risk presented in the Six Month Study
is, as a result, more in line with the risk presented in this
study, since the risk from self-service refueling was projected
to be 54 percent of the total.
-91-
-------
12.2.5 1,3-Butadiene
The 35 County study found no risk associated with
emissions of 1,3-butadiene from mobile sources. A cancer risk
estimate for 1,3-butadiene was estimated in this study using an
updated emission factor and unit risk. The risk estimate is
discussed in detail in Section 6.
Determination of an emission factor for 1,3-butadiene is
difficult. This is because 1,3-butadiene and n-butane coelute
and thus have the same retention point on the gas
chromatograph. Emission characterization studies to date have
not attempted to determine the percentage of the peak due to
1,3-butadiene. Therefore, assumptions must be made about the
percentage each compound contributes to this peak. It will be
assumed that 15 percent of the peak is due to
1,3-butadiene.[64] 1,3-Butadiene emissions were expressed as a
percentage of the total exhaust HC predicted by MOBILES. Based
on the data from 46 in-use gasoline-fueled vehicles provided in
reference 30, 1,3-butadiene is roughly 0.94 percent of the
total FID exhaust HC. Due to the lack of data for the other
vehicle classes, this percentage was simply applied to the
MOBILES composite exhaust HC emission factor.
The modified NEM model was used to estimate exposures.
Nationwide urban exposure in 1986 is estimated to be 0.69-0.76
ug/m3. The range accounts for both the presence and absence
of an Inspection/Maintenance program. These exposure estimates
are for direct emissions of 1,3-butadiene and do not account
for reactions of 1,3-butadiene in the atmosphere. Available
ambient monitoring data were reviewed and compared to the
exposure estimates. Average mean values in urban settings
range from 0.77-24.23 ug/m3. [64] The NEM estimate of urban
exposure from motor vehicles lies near the low end of this
range.
The unit risk for 1,3-butadiene has changed
substantially. Based on inhalation studies of 1,3-butadiene in
mice, a 95 percent upper confidence limit unit risk for
1,3-butadiene is estimated to be 2.8 x 10~4.[66] This is
much greater than the previous estimated unit risk of 4.6 x
10~7 used in the Six Month Study.
Estimates of urban cancer incidence in 1986 range from
514-568, or 2.86-3.16 per urban million. Due to the
uncertainty associated with the emissions estimate, a lower
risk estimate of zero will also be used.
12.2.6 Ethylene
Ethylene from motor vehicles was not considered in the 35
County study. Based on the unit risk for ethylene provided in
the Six Month Study, risk estimates for ethylene were computed
in this study (Section 6.4). The urban risk in 1986 was
estimated to be 47-52 cancer incidences, or 0.26-0.29 per urban
-92-
-------
million. As mentioned in Section 6.4, however, the unit risk
and resulting risk estimates must be regarded as extremely
tentative, since there is no available evidence that ethylene
is carcinogenic. The unit risk for ethylene was estimated
based on assumptions regarding its potency relative to ethylene
oxide, a metabolite of ethylene and an animal carcinogen. As a
result, a lower risk estimate of zero is also used.
12.2.7 Asbestos
Asbestos emissions from motor vehicles were not considered
in the Six Month Study. Based on available emission factors
and resulting expected ambient concentrations, it is estimated
in this study that asbestos emissions from motor vehicles could
currently be responsible for as many as 0.002-0.63 cancer
incidences per urban million.
12.2.8 Ethylene Dibromide (EPS) and Cadmium
EDB and cadmium were handled similarly. Updated emission
factors and unit risks were used in this study. For each
pollutant, the risk determined in the 35 County study was
multiplied by ratios of the updated to the 35 County emission
factors and unit risks to obtain a crude estimate of the risk.
EDB is used as a lead scavenger. It is therefore emitted
from vehicles using leaded gasoline. £DB data came from
reference 100. The light-duty vehicle fleet was assumed to be
88 percent catalyst-equipped and 12 percent
non-catalyst-equipped based on MOBILES data. Of the
catalyst-equipped vehicles, a misfueling rate of 14.5 percent
was used.[87] The composite emission factor for light-duty
vehicles is 7.1x10"5 g/mile. Heavy-duty gasoline-fueled
vehicles were not considered. The impact of this vehicle class
is not expected to be great due to its small contribution to
total VMT (4 percent).
The composite emission factor and resulting annual cancer
risk is roughly half that calculated in the 35 County study.
This is because a much higher percentage of
non-catalyst-equipped vehicles was assumed in the 35 County
study (50 versus 12 percent). The resulting risk of 0.01 per
urban million appears negligible.
In 1995, with the phase-out of leaded-fueled vehicles, the
emission factor is expected to decrease to 1.8 x 10"5
g/mile. The resulting risk is 0.003 per urban million.
Data for cadmium were obtained from references 32-35.
Cadmium does not appear to be emitted from catalyst-equipped
vehicles or it is emitted below the detection limit. Cadmium
emissions from non-catalyst-equipped vehicles are roughly
. -93-
-------
1.6xlO~s g/mile. It appears that cadmium may be a lead
contaminant. Assuming 12 percent of the light-duty fleet are
non-catalyst-equipped vehicles, the resulting light-duty fleet
emission factor is 1.9xlO~6. Again, this is less than what
was used in the 35 County study because of the lower percentage
of non-catalyst-equipped vehicles expected. The resulting risk
of 0.001 per million is negligible compared to the other risks.
12.2.9 Total Aggregate Risk
The sum of all the individual annual cancer risk estimates
in the 35 County study (including gasoline vapors) is 2.65 per
million. In this study, the aggregate risk ranges from 1.80 to
10.58 per urban million. If the formaldehyde risk in the 35
County study is increased to reflect the updated unit risk, the
aggregate risk in the 35 County study would increase to 2.73
per million. The range in this study is due to a number of
factors including: 1) the uncertainty of annual average
formaldehyde and benzene concentrations and the contribution of
mobile sources, 2) the range of unit risks used for diesel
particulate, and 3) the different approaches used to determine
the contribution of gasoline POM/PIC.
It should be noted when reviewing Table 12-2 that the
pollutants given do not represent a complete list. This list
does not include pollutants which are formed photochemically
from mobile source emissions. This category of pollutants
could have considerable impact but not enough is known to make
a quantitative estimate.
-94-
-------
13.0 SUMMARY AND LIMITATIONS
The aggregate risk from mobile source pollutants in
calendar year 1986 was estimated to range from 1.80-10.58 per
urban million. This translates into roughly 325-1900 urban
cancer incidences. The majority of risk is attributed to
formaldehyde, diesel particulate, and benzene. Due to
increasing use of advanced control technology, the risk in 1995
is projected to decrease to roughly 60 percent the risk in 1986.
When reviewing these estimates, the following important
limitations should be considered.
o This report only accounts for a small number of
mobile source pollutants known to be emitted. In
reality, mobile sources emit hundreds of compounds.
A combination of health and/or exposure and/or
emissions data are lacking for many of these
compounds.
o With the exception of formaldehyde, this report does
not consider reactions of mobile source pollutants
in the atmosphere. Little data are available.
Resulting pollutants may be more or less
carcinogenic than what was originally emitted.
o Exposures for many of the pollutants were estimated
using a modified version of the NAAQS Exposure Model
(NEM) for CO. In order to apply this model to
other mobile source pollutants, it was assumed that
the other pollutants have the same 'dispersion
characteristics as CO. This is probably not true
for every pollutant; however, this was determined to
be the best approach currently available to
determine exposure to mobile source pollutants on a
national scale.
An alternative is to use ambient data.
Unfortunately, ambient data for many mobile source
pollutants, particularly VOC, are scarce and not of
high quality. In addition, the relative
contributions of mobile and stationary sources are
not known.
o This report does not fully take into account
seasonal variations in emissions. Hydrocarbon
emissions from mobile sources are known to increase
as the temperature decreases. Work is being done to
characterize emissions of formaldehyde under cold
temperature conditions. This will be followed by
further characterization of individual HC emissions
from mobile sources under cold temperature
conditions.
-95-
-------
The unit risk estimates' are always subject to some
uncertainty. With the exception of benzene, they
are generally based on the results of animal rather
than human studies.
The risks are assumed to be additive. It may be
that certain combinations of exposures have
synergistic or antagonistic effects.
The risk projections for 1995 are based on the
emission standards currently in place. Changes in
fuel composition are not considered.
-96-
-------
REFERENCES
1. Elaine Haemisegger, Alan Jones, Bern Steigerwald and
Vivian Thomson, "The Air Toxics Problem in the United
States: An Analysis of Cancer Risks for Selected
Pollutants," EPA Report, Office of Air and Radiation, May
1985.
2. Roy B. Zweidinger, "Emission Factors from Diesel and
Gasoline Powered Vehicles: Correlation with the Ames
Test," In: Toxicological Effects of Emissions from
Diesel Engines, Joellen Lewtas, ed., Elsevier Science
Publishing Co., Inc., 1982.
3. Ronald L. Williams, "Diesel Particulate Emissions:
Composition, Concentration, and Control," In:
Toxicological Effects of Emissions from Diesel Engines,
Joellen Lewtas,ed., Elsevier Science Publishing Co., Inc.,
1982.
4. Frank Black and Larry High, "Methodology for Determining
Particulate and Gaseous Diesel Hydrocarbon Emissions," SAE
Paper 790422, February 1979.
5. Joellen Lewtas, Marcia G. Nishioka and Bruce A. Peterson,
"Identification and Comparative Risk Assessment of
Airborne Carcinogens from Combustion Sources," U.S. EPA,
Office of Research and Development, EPA/600/D-86/013,
January 1986.
6. Dennis Schuetzle, "Sampling of Vehicle Emissions for
Chemical Analysis and Biological Testing," Environmental
Health Perspectives, Vol. 47, pp. 65-80, 1983.
7. Roy E. Albert, Joellen Lewtas, Stephen Nesnow, Todd W.
Thorslund, and Elizabeth Anderson, "Comparative Potency
Method for Cancer Risk Assessment: Application to Diesel
Particulate Emissions," Risk Analysis, Vol. 3, No. 2, 1983.
8. "Diesel Particulate Study," EPA Report, Office of Mobile
Sources, October 1983.
9. EPA Emission Factors Program Data Base. Report in
preparation.
10. Mahlon C. Smith, IV, "Heavy-Duty Vehicle Emission
Conversion Factors, 1962-1997," EPA Technical Report
EPA-AA-SDSB-84-1, August 1984.
11. "Users Guide to Mobiles (Mobile Source Emissions Model),"
Office of Mobile Sources, EPA-460/3-84-002, June 1984.
12. MVMA Facts and Figures and Fact Sheets.
-97-
-------
13. EPA Memo from Paul Laing, Standards Development and Support
Branch (SDSB) to Chester J. France, Chief, SDSB,
"Historical and Projected Light-Duty Truck and Heavy-Duty
Vehicle Sales Data," February 1986.
14. EPA Memo from John W. Mueller, Standards Development and
Support Branch to Charles L. Gray, Jr., Director, Emission
Control Technology Division, "Revised Diesel Sales Fraction
Projections," January 9, 1987.
15. Draft MOBILES Fuel Consumption Model, February, 1985.
16. "The Motor Fuel Consumption Model, Tenth Periodical
Report," prepared by Energy and Environmental Analysis,
Inc. for U.S. EPA, Office of Mobile Sources, November 1983.
17. "National Air Quality and Emissions Trends Report, 1984,"
U.S. EPA, Office of Air Quality Planning and Standards,
EPA-450/4-86-001, 1986.
18. Melvin N. Ingalls, Southwest Research Institute, "Improved
Mobile Source Exposure Estimation," EPA-460/3-85-002, March
1985.
19. "Statistical Abstract of the United States," 105th Edition,
U.S. Department of Commerce, Bureau of the Census, 1985.
20. Charles T. Hare and Thomas M. Baines, "Characterization of
Particulate and Gaseous Emissions from Two Diesel
Automobiles as Functions of Fuel and Driving Cycle," SAE
Paper 790424, February 1979.
21. James D. Hyde, Richard E. Gibbs, Robert A. Whitby, Stanley
M. Byer, Benjamin J. Hill, Thomas E. Hoffman, Robert E.
Johnson, and Paul L. Werner, "Analysis of Particulate and
Gaseous Emissions Data from In-Use Diesel Passenger Cars,"
SAE Paper 820772, June 1982.
22. Terry L. Ullman and Charles T. Hare, Southwest Research
Institute, "Emissions Characterization of Heavy-Duty Diesel
and Gasoline Engines and Vehicles," EPA-460/3-85-001, May
1985.
23. "Compilation of Air Pollutant Emission Factors, Volume II:
Mobile Sources," AP-42, Fourth Edition, September 1985,
Appendix N.
24. R.G. Cuddihy, W.C. Griffith, C.R. Clark and R.O. McClellan,
"Potential Health and Environmental Effects of Light-Duty
Diesel Vehicles II," Inhalation Toxicology Research
Institute, Lovelace Biomedical and Environmental Research
Institute, prepared for U.S. Department of Energy, LMF-89,
UC-48, October 1981.
-98-
-------
25. Roger 0. McClellan, "Health Effects of Diesel Exhaust: A
Case Study in Risk Assessment," Am. Ind. Hyg. Assoc. J.,
47(1): 1-13, 1986.
26. A.K. Pollack, A.B. Hudischewskyj and A.D. Thrall, Systems
Applications, Inc., "An Examination of 1982-83 Particulate
Matter Ratios and Their Use in the Estimation of PM10
NAAQS Attainment Status, "EPA-450/4-85-010, August 1985.
27. Jeffrey E. Harris, "Potential Risk of Lung Cancer from
Diesel Engine Emissions," report to the Diesel Impacts
Study Committee, National Research Council, National
Academy Press, Washington, D.C., 1981.
28. Todd W. Thorslund, "A Suggested Approach for the
Calculation of the Respiratory Cancer Risk Due to Diesel
Engine Exhaust," presented at the EPA Workshop on the
Evaluation of Research in Support of the Carcinogenic Risk
Assessment for Diesel Engine Exhaust, February 24-25, 1981.
29. W.E. Pepelko and W.B. Peirano, "Health Effects of Exposure
to Diesel Engine Emissions," Journal of the American
College of Toxicology, Volume 2, No. 4, 1983.
30. John E. Sigsby, Jr., Silvestre Tejada and William Ray,
Mobile Sources Emission Research Branch, U.S. EPA,
"Volatile Organic Compound Emissions from 46 In-Use
Passenger Cars," Environ. Sci. Technol., Volume 21, 1987.
31. George J. Nebel, "The Effect of Misfueling on Aldehyde and
Other Auto Exhaust Emissions," APCA Journal, Vol. 31, No.
8, August 1981.
32. Charles Urban, Southwest Research Institute, "Unregulated
Exhaust Emissions from Non-Catalyst Baseline Cars Under
Malfunction Conditions," EPA-460/3-81-020, May 1981.
33. Charles Urban, Southwest Research Institute, "Regulated and
Unregulated Exhaust Emissions from Malfunctioning
Non-Catalyst and Oxidation Catalyst Gasoline Automobiles,"
EPA-460/3-80-003, January 1980.
34. Charles Urban, Southwest Research Institute, "Regulated and
Unregulated Exhaust Emissions from Malfunctioning Three-Way
Catalyst Gasoline Automobiles," EPA-460/3-80-004, January
1980.
35. Charles Urban, Southwest Research Institute, "Regulated and
Unregulated Exhaust Emissions from a Malfunctioning
Three-Way Catalyst Gasoline Automobile, EPA-460/3-80-005,
January 1980.
-99-
-------
36. Lawrence R. Smith, "Characterization of Exhaust Emissions
from High Mileage Catalyst-Equipped Automobiles,"
EPA-460/3-81-024, September 1981.
37. "Investigation of Diesel-Powered Vehicle Emissions VII,"
EPA-460/3-76-034, February 1977.
38. Karl J. Springer,"Characterization of Sulfates, Odor,
Smoke, POM and Particulates from Light and Heavy-Duty
Engines - Part IX," EPA-460/3-79-007, June 1979.
39. M.W. Phillips and G.E. Wilkins, Radian Corporation,
"Preliminary Source Assessment for Formaldehyde, Final
Report," prepared for U.S. EPA, Office of Air Quality
Planning and Standards, September 3, 1985.
40. Lawrence R. Smith, Southwest Research Institute,
"Unregulated Emissions for Vehicles Operated Under Low
Speed Conditions," EPA-460/3-83-006, May 1983.
41. "Control Techniques for VOC Emissions from Stationary
Sources," U.S. EPA, Draft Report, March 1986, Table 2-5.
42. 1983 NMHC Emissions Inventory, Office of Mobile Sources,
Technology Evaluation Branch, Draft, November, 1986.
43. Gary Z. Whitten and Henry Hogo, Systems Applications, Inc.,
"Impact of Methanol on Smog: A Preliminary Estimate,"
prepared for ARCO Petroleum Products Company, February 1983.
44. W.D. Kerns, et al. , "Carcinogenicity of Formaldehyde in
Rats and Mice after Long-Term Inhalation Exposure," Cancer
Research, 43: 4382-4392.
45. "Assessment of Health Risks to Garment Workers and Certain
Home Residents from Exposure to Formaldehyde," U.S. EPA,
Office of Pesticides and Toxic Substances, Draft Final
Report, March 1987.
46. William F. Hunt, Jr., Robert B. Faoro, Thomas C. Curran and
Jena Muntz, U.S. EPA, Office of Air Quality Planning and
Standards, "Estimated Cancer Incidence Rates for Selected
Toxic Air Pollutants Using Ambient Air Pollution Data,"
April 23, 1985.
47. Memo from Charles L. Gray, Jr., Director, Emission Control
Technology Division to Richard D. Wilson, Director, Office
of Mobile Sources, "Mobile Source Benzene Emissions and a
Preliminary Estimate of Their Health Impacts," May 15, 1986.
-100-
-------
48. "Evaluation of Air Pollution Regulatory Strategies for
Gasoline Marketing Industry," EPA Report 450/3-84-012b,
July 1984.
49. Memo from William F. Hunt, Jr., Chief, Data Analysis
Section, OAQPS to Stan Meiburg, Special Assistant for
Program Development, OAQPS, "Comparison of Additive
Individual Lifetime Risks for Five Cities," April 3, 1986.
50. Donald J. Ames, Barbara Fry and Richard Vincent, "Proposed
Benzene Control Plan," State of California Air Resources
Board Draft Staff Report, May 1986.
51. "Addendum to Proposed Benzene Control Plan," State of
California Air Resource Board Staff Report, July 1986.
52. Ward's Automotive Yearbook 1986, Forty-Eighth Edition.
53. H.N. MacFarland, C.E. Ulrich, C.E. Holdsworth, D.N.
Kitchen, W.H. Halliwell and S.C. Blum, Journal of the
American College of Toxicology, 3, 4, 1984.
54. "Level I Options Package for a Gasoline Marketing
Decision," Office of Air Quality Planning and Standards and
Office of Mobile Sources Report, July 11, 1985.
55. Andrew E. Bond, Vinson L. Thompson, and Gordon C. Ortman,
Environmental Monitoring Systems Laboratory, EPA, and
Francis M. Black and John E. Sigsby, Jr., Atmospheric
Sciences Research Laboratory, EPA, "Self Service Station
Vehicle Refueling Exposure Study," no date given.
56. Memorandum from Charles L. Gray, Jr., Director, Emission
Control Technology Division to James B. Weigold, Office of
Air Quality Planning and Standards, "Lower Limit on
Gasoline Vapor Risk," April 19, 1985.
57. "Gasoline Vapor Exposure and Human Cancer: Evaluation of
Existing Scientific Information and Recommendations for
Future Research," Health Effects Institute, September 1985.
58. Memo from Donald J. Ehreth, Acting Assistant Administrator
for Research and Development to Lee Thomas, EPA
Administrator, "Brief Statement Concerning Health Effects
Institute (HEI) Report," October 8, 1985.
59. "Feasibility of Assessment of Health Risks From Vapor-Phase
Organic Chemicals in Gasoline and Diesel Exhaust," prepared
by the Committee on Vapor-Phase Organics, National Research
Council, National Academy Press, Washington, D.C., 1983.
-101-
-------
60. Penny Carey and Janet Cohen, U.S. EPA, Office of Mobile
Sources, "Comparison of Gas Phase Hydrocarbon Emissions
From Light-Duty Gasoline Vehicles and Light-Duty Vehicles
Equipped with Diesel Engines," CTAB/PA/80-5, September
1980.
61. J. Kraft and K.-H. Lies, "Polycyclic Aromatic Hydrocarbons
in the Exhaust of Gasoline and Diesel Vehicles," SAE Paper
810082, February 1981.
62. Fred Stump, Ronald Bradow, William Ray, David Dropkin, Roy
Zweidinger, and John Sigsby, "Trapping Gaseous Hydrocarbons
for Mutagenic Testing," SAE Paper 820776, June 1982.
63. Frank Black and Larry High, "Automotive Hydrocarbon
Emission Patterns in the Measurement of Nonmethane
Hydrocarbon Emission Rates," SAE 770144, February 1977.
64. Summary of 1,3-butadiene monitoring data provided by Ila
Cote, EPA, Office of Air Quality Planning and Standards,
1986.
65. EPA Memo from Edward J. Li 11 is, Chief, Air Management
Technology Branch to List of Addressees, "Urban Toxics,"
June 12, 1987.
66. "Mutagenicity and Carcinogenicity Assessment of
1,3-Butadiene," final report prepared by U.S. EPA, Office
of Health and Environmental Assessment, EPA/600/8-85/004F,
September, 1985.
67. Carl O. Schulz and David M. Siegel, Clement Associates,
Inc., "The Relative Carcinogenic Potential of 50 Chemicals
That May Be Air Pollutants," final report prepared for U.S.
EPA, Office of Air Quality Planning and Standards, March
30, 1984.
68. T.M. Albrechcinski, J.G. Michalovic, B.J. Wattle, and E.P.
Wilkinson, Calspan Corp., "Laboratory Investigation of the
Fate of Diesel Emissions in the Atmosphere. Task 2. Final
Report," CRC-APRAC-CAPA-13-76-05, December 1985.
69. Larry D. Claxton and H.M. Barnes, "The Mutagenicity of
Diesel-Exhaust Particle Extracts Collected Under
Smog-Chamber Conditions Using the Salmonella typhimurium
Test System," Mutation Research, 88, 255-272, 1981.
70. John M. Lang, et al. , "Characterization of Particulate
Emissions from In-Use Gasoline-Fueled Motor Vehicles," SAE
Paper 811186, October 1981.
71. Energy and Environmental Analysis, Inc., "Size Specific
Total Particulate Emission Factors for Mobile Sources,"
final report prepared for U.S. EPA, Office of Mobile
Sources, EPA 460/3-85-005, August 1985.
-102-
-------
72. Charles T. Hare and Frank M. Black, "Motor Vehicle
Particulate Emission Factors," APCA Paper 81-56.5, June
1981.
73. James N. Braddock and Ned K. Perry, Jr., "Gaseous and
Particulate Emissions from Gasoline - and Diesel-Powered
Heavy-Duty Trucks," SAE Paper 860617, February 1986.
74. Frank Black et al. , "Emission from In-Use Heavy-Duty
Gasoline Trucks," SAE Paper 841356, October 1984.
75. H.E. Dietzman et al. , "Emissions from Gasoline and Diesel
Delivery Trucks by Chassis Transient Cycle," ASME Paper
81-DGP-6, January 1981.
76. "National Dioxin Study Tier 4-Combustion Sources,"
EPA-450/4-84-014a, February 1985.
77. M.G. Jacko, R.T. DuCharme, and J.H. Somers, "Brake and
Clutch Emissions Generated During Vehicle Operation," SAE
Paper 730548, 1973.
78. S. Cha, P. Carter, and R.L. Bradow, "Simulation of
Automobile Brake Wear Dynamics and Estimation of
Emissions," SAE Paper 831036, 1983.
79. "Asbestos; Proposed Mining and Import Restrictions and
Proposed Manufacturing Importation and Processing
Prohibitions," Federal Register, Vol. 51, No. 19, 3738,
January 29, 1986.
80. Melvin N. Ingalls and Robert J. Garbe, "Ambient Pollutant
Concentrations from Mobile Sources in Microscale
Situations," SAE 820787, June 1982.
81. National Research Council, "Nonoccupational Health Risks of
Asbestiform Fibers," National Academy Press, Washington,
D.C., 1984.
82. Ruth A. Zweidinger, Joan T. Bursey, Nora C. Castillo,
Ronald Keefe and Doris Smith, "Organic Emissions from
Automobile Interiors," SAE Paper 820784, June 1982.
83. Lawrence R. Smith and Thomas M. Baines, "Nitrosamines in
Vehicle Interiors," SAE Paper 820785, June 1982.
84. "Regulation of Fuels and Fuel Additives," Federal Register,
Vol. 47, No. 210, 49322, October 29, 1982.
85. "Regulation of Fuels and Fuel Additives; Gasoline Lead
Content," Federal Register, Vol. 50, No. 45, 9385, March 7,
1985.
86. Regulation of Fuels and Fuel Additives; Gasoline Lead
Content," Federal Register, Vol. 50, No. 45, 9400, March 7,
1985.
-103-
-------
87. Energy and Environmental Analysis, Inc., "Supplementary
Guidelines for Lead Implementation Plans - Updated
Projections for Motor Vehicle Lead Emissions," final report
prepared for U.S. EPA, Office of Mobile Sources,
EPA-460/3-85-006, August 1985.
88. "Regulation of Fuels and Fuel Additives; Lead Phase Down,"
Federal Register, Vol. 49, No. 150, 31032, August 2, 1984.
89. "Air Quality Criteria for Lead," External Review Draft,
EPA-600/8-83-028B, September 1984.
90. "In Re Applications for MMT Waiver," Federal Register, Vol.
43, No. 181, 41424, September 18, 1978.
91. "Regulation of Fuel and Fuel Additives: MMT - Suspension
of Enforcement," Federal Register, Vol. 44, No. 109, 32281,
June 5, 1979.
92. Letter from Thomas P. Crumbly, Executive Director, Health
Effects Institute to Wolfgang Groth, Manager, Emissions and
Fuel Economy, Volkswagen of America, Inc., October 7, 1985.
93. "Health Assessment Document for Manganese,"
EPA-600/8-83-031F, Final report, August 1984.
94. Memo from Lawrence E. Slimak, Director, Planning and
Industrial Department, Motor Vehicle Manufacturers
Association to Phil Lorang, Chief, Technical Support Staff,
EPA, May 6, 1986.
95. R.F. Hill and W.J. Mayer, IEEE Trans. Nucl. Sci., NS-24,
2549-2554, 1977.
96. Vernon F. Hodge and Martha O. Stallard, "Platinum and
Palladium in Roadside Dust," Environ. Sci. Technol., Vol.
20, No. 10, 1986.
97. "Medical and Biologic Effects of Environmental Pollutants,
Platinum-Group Metals," National Academy of Sciences
Report, 1977.
98. Open Letter from Dr. Hans A, Nieper, Medical Department,
Paracelsus Clinic to the Board of Directors of a
Technological Firm, published in Space & Time, 1985.
99. Versar Inc. and American Management Systems, "Hazardous Air
Pollutants, A Preliminary Exposure and Risk Appraisal for
35 U.S. Counties," draft report prepared for U.S. EPA,
Office of Policy Analysis, September 1984.
100. John E. Sigsby, Jr., David L. Dropkin, Ronald L. Bradow and
John M. Lang, "Automotive Emissions of Ethylene Dibromide,"
SAE Paper 820786, June 1982.
-104-
-------
GLOSSARY OF TERMS
Airshed: A geographical area which, because of topography,
meteorology, and climate, shares the same air mass.
Air Toxic: A compound in the air capable of causing adverse
health effects. For the purpose of this report, the air
toxics examined were limited to known or suspected
carcinogens.
Aldehydes: A class of fast-reacting organic compounds
containing oxygen, hydrogen, and carbon. They contain the
group -CHO.
Aliphatic: A class of hydrocarbon compounds which are open
chained and fully saturated (e.g., no double bonds).
Ambient Air: That portion of the atmosphere, external to
buildings, to which the general public has access.
Ames Test: A mutagenicity bioassay using the bacteria.
Salmonella typhimurium.
Aromatic: A class of hydrocarbon compounds originating from
benzene, C6H6, or containing at least
one benzene ring or similar unsaturated heterocyclic ring.
Benzo(a)Pyrene (B(a)P): A polycyclic aromatic hydrocarbon with
the molecular formula C20H12.
Bioassay: Using living organisms to measure the effect of a
substance, factor, or condition.
Catalyst: Used in this report to denote a catalytic converter,
a chamber in the exhaust system of vehicles containing a
catalyst system which aids in oxidizing the carbon monoxide
and unburned hydrocarbons in the exhaust gases or in reducing
nitrogen oxides in the exhaust gases to innocuous products
(carbon dioxide, N2, O2, and water).
Control Technology: A combination of measures designated to
achieve the aggregate reduction of emissions.
Diesel Particulate Trap: A device located in the exhaust stream
of a diesel vehicle that filters a certain percentage of
exhaust particulate. This device must include some means by
which accumulated particulate can be burned, thus
regenerating the trap and making the trap available for
continued particulate filtration.
-105-
-------
Dynamometer: A device that is used to measure or to simulate
loads, engine torque, and driving forces on vehicles or
engines.
Exhaust Gas Recirculation (EGR): A system or device (such as
modification of the engine's carburetor or positive crankcase
ventilation system) that results in engine operation at an
increased air-fuel ratio so as to achieve reductions in
exhaust emissions of nitrogen oxides.
Emission Factor: For motor vehicles, an emission factor is the
amount of pollutant emitted per unit of distance. In this
report, emission factors are expressed in units of grams of
pollutant per mile travelled.
Epidemiology: The study of diseases as they affect populations
rather than individuals, including the distribution and
incidence of a disease; mortality and morbidity rates; and
the relationship of climate, age, sex, race and other factors.
Evaporative Emissions: Hydrocarbons emitted into the atmosphere
from a motor vehicle through fuel evaporation.
Exhaust Emissions: Substances emitted to the atmosphere from
any opening downstream from the exhaust port of a motor
vehicle engine.
Federal Test Procedure (FTP): A multistage (multimodal) test
procedure for new car certification by the Environmental
Protection Agency.
Halogen: Any one of the nonmetallic elements chlorine, iodine,
bromine, and fluorine.
Heavy-Duty Vehicle: Any motor vehicle rated at more than 8,500
pounds gross weight or that has a vehicle curb weight of more
than 6,000 pounds or that has a basic vehicle frontal area in
excess of 45 square feet.
Hydrocarbon: Any of a vast family of compounds containing
carbon and hydrogen in various combinations: found especially
in fossil fuels.
Ketone: An organic compound derived by oxidation from a
secondary alcohol; it contains the carbonyl group (= CO).
Light-Duty Vehicle: A passenger car or passenger car derivative
capable of seating 12 passengers or less.
-106-
-------
MOBILE3: A computer program that calculates emission factors
for hydrocarbons (HC), carbon monoxide (CO), and oxides of
nitrogen (NOx) from highway motor vehicles.
Mutagenic: The property of a substance or mixture of substances
to induce changes in the genetic structure in subsequent
generations.
NAAQS Exposure Model (NEM): An exposure model suitable for
evaluating alternative ambient air standards.
Neurotoxic: Harmful to nerve tissue.
Noble Metal: Metal such as platinum or palladium that is
non-reactive to most chemical substances.
Organic: In chemistry, any compound containing carbon.
Oxidation Catalyst: A catalytic converter used to oxidize the
carbon monoxide and unburned hydrocarbons in the exhaust
gases to innocuous products.
PAH: Polynuclear aromatic hydrocarbons.
Particulate: A particle of solid or liquid matter.
Photochemistry: Chemical changes brought about by the radiant
energy of the sun acting upon various polluting substances.
The products are known as photochemical smog.
POM: Polycyclic organic matter. Many POM compounds are also PAH
compounds.
PIC: Products of Incomplete Combustion.
Soluble Organic Fraction (SOF): In this report, defined as the
organic fraction of particulate soluble with dichloromethane.
Steady-State: Constant operating conditions with no variation
in fuel supply or load.
Street Canyon: A street lined with buildings. In general, the
streets are less than seven lanes wide and the buildings
under 26 stories. The minimum canyon height to width ratio
for a building-lined street to be considered a street canyon
is approximately 0.3.
Three-Way Catalyst: A catalytic converter that is capable of
both oxidizing carbon monoxide and unburned hydrocarbons and
reducing nitrogen oxides in the exhaust.
-107-
-------
TSCA: Toxic Substances Control Act.
Unit Risk: The individual life time excess cancer risk from
continuous exposure to 1 ug carcinogen per m3 inhaled air.
VMT: Vehicle miles travelled.
Volatile Organic Compounds (VOC): Any compound containing
carbon and hydrogen or containing carbon and hydrogen in
combination with any other element which has a vapor pressure
of 1.5 pounds per square inch absolute or greater under
actual storage conditions.
-108-
;,. Environmental Protection Agenc.
ion 5, Library (5PL-16)
j S, Dearborn Street, Room 1670
IL 60604
------- |