United States
Environmental Protection
Agency
Office of Air Quality
Planning And Standards
Research Triangle Park, NC 27711
EPA-450/2-89-012a
July 1989
AIR
ANALYSIS OF AIR TOXICS
EMISSIONS, EXPOSURES,
CANCER RISKS AND
CONTROLLABILITY
IN FIVE URBAN AREAS
VOLUME I
BASE YEAR ANALYSIS AND RESULTS
V
\ ? /T*^ Vs-
n iMh
ol
Vi
^
*/....}*
-------
-------
EPA-450/2-89-012a
July 1989
ANALYSIS OF AIR TOXICS EMISSIONS,
EXPOSURES, CANCER RISKS AND
CONTROLLABILITY IN FIVE URBAN AREAS
VOLUME I
Base Year Analysis And Results
U. S. ENVIRONMENTAL PROTECTION AGENCY
Office Of Air and Radiation
Office Of Air Quality Planning And Standards
Research Triangle Park, North Carolina 27711
-------
-------
This report has been reviewed by the Office of Air Quality Planning
and Standards, U. S. Environmental Protection Agency, and has been
approved for publication. Mention of trade names or commercial
products does not constitute endorsement or recommendation for use.
EPA 450/2-89-012a
-------
-------
EXECUTIVE SUMMARY
BACKGROUND
The U.S. Environmental Protection Agency (EPA), under its National Air Toxics
Strategy (U.S. EPA, 1985a), has initiated a program to assess the nature and magnitude of
the air toxics problem in urban areas. This urban air toxics problem, sometimes called
"urban soup," is characterized by complex multi-source interactions and multi-pollutant
exposures. Numerous studies, including this study, suggest that areawide lifetime excess
cancer risks from urban soup may range from about 1 in 10,000 to 1 in 1,000, and that
cancer incidence may range from 1 to 23 excess cases per year per million population
(Lahre, 1988a).
EPA has been encouraging state and local air pollution control agencies to assess their
urban air toxics problems. Numerous assessment activities have been undertaken in the
last several years in the areas of ambient air monitoring, emission inventorying, exposure
and risk characterization, and mitigation analyses. At the same time, EPA is promoting the
coordination of air toxics control programs with ozone (Os) and paniculate matter (PMio)
control programs to assure that, whenever possible, State Implementation Plans (SIPs)
incorporate measures that reflect co-control of air toxics (52 FR 45072).
PURPOSE OF STUDY
The purpose of this study was as follows:
• define the multi-source, multi-pollutant nature of the urban air toxics problem, and
• discern what control measures (or combinations of measures) can best be employed
to mitigate the urban air toxics problem.
This study was conducted in two phases, each phase relating directly to the above-stated
goals. The first phase was the base year analysis, and involved dispersion modeling of
emissions data for five urban areas in the country. In this first phase, the latest available
emissions and source data were compiled and used as input to EPA's Human Exposure
Model (HEM) to estimate ambient air concentrations and population exposures to 25
known or suspected air carcinogens. From this modeling exercise, estimates were made of
the sources and pollutants contributing to additive (i.e., multi-pollutant) cancer risk
throughout each urban area. Emphasis was placed on estimating areawide population risk,
i.e., aggregate incidence, from multi-pollutant, multi-source exposures. The year 19X0 was
nominally defined as the base year in this analysis, although some later year data were
incorporated into the data base.
The second phase was the projection analysis, and included an evaluation of the
mitigation potential of various control measures when superimposed on the base year
emission inventory. One set of measures represented "baseline" controls, i.e., those
111
-------
-------
controls already in place as a result of current "on-the-books" measures. These "baseline"
measures include the Federal Motor Vehicle Control Program (FMVCP), New Source
Performance Standards (NSPS), and National Emission Standards for Hazardous Air
Pollutants (NESHAP). In addition, various alternative control measures were evaluated
over and above the baseline measures. These alternative measures included new measures
reasonably anticipated to be in effect in the near future as well as a maximum controls or
Best Available Control Technology (BACT) strategy. The year 1995 was defined as the
"projection year" in this analysis. Changes unrelated to emissions (e.g., meteorology) were
not considered in this analysis.
VOLUME I REPORT ORGANIZATION -
This study is documented in two major parts. Volume I reports on the methods and
results associated with the base year analysis. Volume II will be published under separate
cover as a companion document and reports on the methods and results associated with the
projection, or controllability, analyses.
Volume I is organized according to each major activity as follows:
• methodology overview
.• emission inventory enhancement
• dispersion modeling and exposure/risk characterization
• base year results
• assumptions/limitations
• conclusions
DISPERSION MODELING ANALYSES
Two major approaches are currently used for conducting urban air toxics assessments:
(I) ambient air monitoring and (2) dispersion modeling of emissions. In the former,
measured ambient levels of air toxics in an urban area are multiplied by cancer unit risk
factors and applied to population data to calculate individual risks and aggregate incidence.
In the dispersion modeling approach, an emission inventory is compiled for the air toxics of
concern and is modeled (using various long-term dispersion models) to predict ambient air
concentrations over the urban area. These modeled ambient air concentrations are then
used to estimate individual risks and aggregate incidence as if they were measured data.
Important advantages of the dispersion modeling approach are that it allows one to
predict risk reductions as a function of anticipated emission changes and it allows ambient
air levels to be projected at many more receptors than may be possible in air sampling
networks. It also allows one to handle pollutants (e.g., Cr+r>) for which ambient sampling
IV
-------
methods are not available or cost effective and to assign risk to different emitters of the
same pollutant.
This study employs the dispersion modeling approach for all pollutants with the
exception of secondary formaldehyde. No validated photochemical models were available
to predict secondary formaldehyde production via modeling; hence, ambient formaldehyde
measurements were used to estimate cancer incidence and risks associated with secondary
formaldehyde exposures.
EMPHASIS ON CANCER
This study estimated cancer risks from long-term (i.e., annual average) exposures to
multiple pollutants. This type of analysis has predominated urban soup assessments. To
date, limited work has been done to quantify noncancer risks associated with short-term
(i.e., acute and subchronic) exposures to air toxics.
The most commonly used measure of cancer risk in urban soup assessments is "aggregate
cancer incidence," which is a measure of the number of excess lifetime cancer cases over an
entire urban area associated with multi-source, multi-pollutant exposures to air toxics;
(This is also called "population risk.") Incidence is typically expressed as the number of
excess cancer cases expected in a single year, but is estimated based on an assumed 70 year
"lifetime" exposure. In this study, incidence is considered additive (incidence is estimated
for individual compounds and then summed) and. is population-normalized by adjusting per
million persons. As such, the measure of cancer risk most commonly expressed herein is
"excess aggregate additive incidence per year per million population." Normalization by
population allows incidence to be compared from one urban area to another.
SCREENING STUDY
This study, like most urban assessments to date, should be considered a screening (or
scoping) analysis, performed to yield an order-of-magnitude estimate of the relative nature
of the urban cancer problem rather than to provide an absolute prediction of cancer
incidence and individual risks. It is especially critical in this type of study to ensure that the
scope and use of the conclusions be kept compatible with one's confidence in the
underlying data base and analyses. Most studies to date of this type are acknowledged to be
screening or scoping studies whose results should be used only in a relative sense for
providing broad program direction to suggest where more detailed and focused follow-up
work is needed.
Not all of the data and procedures used in this analysis have been reviewed and
approved by the States or local air agencies whose jurisdictions encompass the study cities.
In many cases, especially with small point and area sources, EPA and its support contractor
made their own emission estimates based on national data and "top down" procedures. For
-------
these reasons, no results are associated with any particular city in this report. Instead, all
city-specific results are associated with "Study Area A," "Study Area B," etc.
Because of the many assumptions and limitations inherent in this type of assessment, and
because of different characteristics of each of the five cities, the composite results for ail
five cities may provide a better overall representation of the urban air toxics problem in the
United States than the results for any single city.
RELATIONSHIP TO PREVIOUS CONTROLLABILITY STUDY
EPA conducted a controllability study for these same five cities in 1985. The results
were not formally published, but were used to help formulate EPA's National Air Toxics
Strategy. Since 1985, many of the procedures and data used for conducting urban
assessments have changed significantly (especially the emission factors and cancer potency
factors for a number of important pollutants). Hence, this study builds on the previous data
base by (1) updating the source and emissions data with more recent national and local
information, (2) incorporating more recent cancer unit risk factors, and (3) reflecting a new
comparative potency approach (U.S. EPA, 1987b) for quantifying risks from polycyclic
organic matter (POM), previously defined as products of incomplete combustion
(Haemisegger, 1985). Nevertheless, for many point sources, the same data exist in the data
base as were present in the 1985 study.
Because the results from the 1985 controllability study are now considered outdated,
little emphasis is placed on them in this report. The major changes are highlighted briefly
in Chapter HI.
SUMMARY OF RESULTS
Aggregate cancer incidence across the five cities in this study averaged about 6 excess
annual cases per million persons, ranging from about 2 to 10 in individual cities. Area wide
individual lifetime cancer risks averaged about 4 x 10," ranging from about 1.5 x 10" to 7 x
10." Note that these risks are not maximum individual risks, which can be as high as 10"" or
even 10"'at specific receptor sites around some large point sources (Haemisegger, 1985).
Instead, these are individual risks averaged over entire urban populations. This phase of
this study did not estimate maximum individual risks.
Figures 1 and 2 show the major sources and pollutants contributing to urban cancer
incidence from air toxics, based on average results from the five cities. The major"
contributors to total aggregate incidence tend to be small point and area sources and road
vehicles, the last source category figuring importantly in most urban studies. Not
surprisingly, the pollutants of primary importance tend to be those associated with these
same source contributors. Total cancer incidence associated with POM in Figure 2 is
largely due to diesel participate (45 percent), gasoline paniculate (32 percent) and wood
smoke (17 percent), all area sources. Total cancer incidence associated with chromium is
VI
-------
05
o
o
O
H O
P5 <£
EH O
z w
O 05
o o
o
OH
o
£3
PL.
O
CL.
CO
-------
O
o
Q
I— H
o
PS
o w
EH O
0
CQ
PS
EH
o
o
O
w
PS
o
o
CO
EH
HH
O
o
OH
o
g
<
CO
CO
cj
00
II
Q
HH
O
O
viii
-------
predominantly due to hexavalent chromium (Cr ') emitted from industrial cooling towers
(19 percent), comfort cooling towers (28 percent), and chrome platers (51 percent). Cancer
incidence from benzene and 1,3-butadiene exposures is primarily due to road vehicles.
Risks from formaldehyde exposures are attributable both to secondary (or
photochemically produced) formaldehyde and to primary (or directly emitted)
formaldehyde. This study suggests that direct formaldehyde emissions account for about 40
percent of the total formaldehyde-related cancer risk whereas secondary formaldehyde
accounts for about 60 percent. The primary VOC sources contributing to secondary
formaldehyde production are road vehicles (35 percent), solvent use (29 percent), gasoline
marketing (8 percent), and refining (6 percent).
Detailed results are presented in Chapter IV.
ONGOING ASSESSMENT PROCESS
The methods, data, and assumptions reflected in any study such as this tend to change,
sometimes rapidly, as one's understanding of the urban air toxics problem evolves and
matures. For example, new sources and pollutants may be identified, new emission factors
may be developed, and new assumptions may be adopted concerning how exposure and risk
characterizations should be conducted. Indeed, numerous such changes have been
required in this study since the original controllability study concluded just 5 years ago. The
biggest change reflected updated emission factors for the major area sources.
Because continued changes are expected, EPA intends to continue this analysis to reflect
new trends in urban assessments. Hence, just as the original 1985 controllability study's
results are no longer considered current, it is expected that the results presented herein will
also become outdated. For this reason, all major data, procedures, and assumptions are
documented herein to help the reader judge the validity of the conclusions at some future
date.
IX
-------
-------
-------
CONTENTS
EXECUTIVE SUMMARY Hi
TABLES xiii
FIGURES xiv
I METHODOLOGY OVERVIEW 1
A. BASE YEAR INVENTORY 1
B. DISPERSION MODELING AND EXPOSURE/RISK
CHARACTERIZATION 3
II EMISSION INVENTORY ENHANCEMENT 5
A. IMPORTANT SOURCE CATEGORIES , 5
1. Road Vehicles 5
2. Chrome Platers ... 13
3. Cooling Towers 13
4. Wood Stoves 18
5. Coal/Oil Combustion 18
6. Other Hexavalent Chromium Emissions 18
7. Methylene Chloride Emissions 18
B. STATE AND LOCAL DATA INPUTS .'. 20
C. ORIGINAL EMISSIONS INVENTORY DEVELOPMENT 21
1. Introduction 21
2. Special Methods for Selected Source Categories 22
3. Source Assessment Documents 26
4. Emission Factors 26
III DISPERSION MODELING AND EXPOSURE/RISK ASSESSMENT 27
A. MODELING METHODS ...; 27
B. COMPARATIVE POTENCY APPROACH FOR ESTIMATING POM
RISKS 31
C. FORMALDEHYDE MODELING APPROACH 34
XI
-------
CONTENTS (continued)
Page
IV BASE YEAR RESULTS 39
A. AGGREGATE RESULTS 39
B. CITY-SPECIFIC RESULTS .44
V ASSUMPTIONS/LIMITATIONS , 54
VI CONCLUSIONS 56
GLOSSARY OF TERMS 58
REFERENCES.
.69
xn
-------
TABLES
Number Page
11.1 Mobile Source Air Toxic Emission Factors for 1980 6
II.2 Formaldehyde Motor Vehicle Emission Factor Calculations -1980 8
H.3 Benzene Motor Vehicle Emission Factor Calculations -1980 9
II.4 Diesel Paniculate Emission Factor Calculation 11
II.5 Calculation of 1980 Diesel Travel Fractions 12
II.6 Hexavalent and Total Chromium Emissions from Chrome Plating
Operations 14
II.7 Hexavalent Chromium Emissions From Comfort Cooling Towers 16
II.8 Industrial Process Cooling Towers - Hexavalent Chromium Emissions 17
II.9 Model Stack Parameters for Industrial Cooling Towers 19
11.10 Emission Estimates for Selected Toxic Emitting Source Categories 23
111.1 Compounds and Associated Unit Risk Values : 28
III.2 Emission Cutoff Levels Used for Identifying Major Point Sources 32
III.3 Particle Unit Risk Estimates Used for Estimating Comparative POM Risks 35
III.4 Application of the Comparative Potency Method to Estimation of Lung
Cancer Unit Risks for Combustion Emissions '....36
III.5 Base Year Formaldehyde Incidence Estimates. 38
IV. 1 VOC Related Annual Incidence by Source Category: Five Cities
Combined 49
IV.2 PM Related Annual Incidence by Source Category: Five Cities Combined ...50
Xlll
-------
-------
FIGURES
Number Page
1.1 Five City Controllability Study Data Flow.. 2
IV. 1 Pollutants Contributing to Rve City Average Aggregate Cancer Incidence ....40
IV.2 Sources Contributing to Five City Average Aggregate Cancer Incidence 41
IV.3 Sources of POM-related Incidence 42
IV.4 Sources Contributing to Secondary Formaldehyde-related Incidence 43
IV.5 Sources Contributing to Incidence from Directly Emitted Formaldehyde 45
IV.6 Sources of Benzene-related Incidence 46
IV.7 Sources of 1,3-Butadiene-related Incidence 47
IV.8 Sources of Hexavalent Chromium-related Incidence 48
IV.9 City-specific Aggregate Cancer Incidence by Source Category 51
IV.10 City-specific Aggregate Cancer Incidence by Pollutant 53
XIV
-------
-------
CHAPTER I
METHODOLOGY OVERVIEW
An overview of the methodology used in this study is shown in Figure 1.1. This chapter
briefly discusses the overall data flow from development of the base year emissions
inventory to estimation of cancer incidence. Details are given in subsequent chapters on
each of the major steps in this process.
A. BASE YEAR INVENTORY
A base year inventory was compiled for each of the five cities from a number of different
sources of information. Since considerable resources had gone into inventory development
in the prior controllability study done in 1985, the decision was made to use this previous
inventory as a starting point and use the limited resources available in this effort mainly to
upgrade those areas of most importance reflecting newer source and emissions data. The
base year nominally represents the year 1980 but, in fact, the data were not this well
resolved temporally. Thus, the base year should really be considered to represent
conditions generally existing between 1980 and 1985. Motor vehicle emissions were
estimated using 1980 emission factors, but emission estimates for other categories were
updated to current conditions when better information was available than existed in the
1980 emission inventory. These revisions included deleting plants from the data base which
were known to have shut down after 1980. The first reliable ambient formaldehyde data,
used to estimate total formaldehyde-related risk, were not available until 1987.
The pollutants covered in this study were those for which cancer unit risk factors exist
and which are known or reasonably suspected of contributing to aggregate cancer incidence
in urban areas. These pollutants are listed below:
arsenic ethylene oxide
asbestos formaldehyde
benzene gasoline paniculate
benzo(a)pyrene, or B(a)P* gasoline vapors
beryllium manganese
1,3-butadiene mercury
cadmium methylene chloride
carbon tetrachloride nickel
chloroform perchloroethylene
chromium (VI and total) trichloroethylene
diesel paniculate vinyl chloride
ethyfene dichloride
*B(a)P was used as a surrogate for all polycyclic organic matter for certain sources.
-------
>-
ID
h-
CO
DQ
LJ
GC:
ID
o
o
a:
o
o
H:
o
LJ
t
a:
LJ
X
00
LJ
X
O
O
1
CO
c
\ o
£"0'~0
t-t.®
o c o
»- O vl
OOO-
uO to
O"i O
CD'C
x— O
C
Jr CD
.LJ- ^
-------
The base year inventory was compiled from a number of sources. The starting point was
the 1980 National Acid Precipitation Assessment Program (NAPAP) emission inventory.
This inventory was improved by incorporation of the following:
• information from local surveys;
• comments from State and local-air agencies;
• information and methodologies from EPA's Integrated Environmental Management
Projects (IEMP) "geographic" studies;
• updated emission factors and emission estimates from EPA's National Emission
Standards for Hazardous Air Pollutants (NESHAP) program;
• updated emission factors and emission estimates from EPA's Office of Mobile
Sources (QMS);
• updated emission factors from EPA's emission factor documents; and
• special contractor studies of area source activity levels and emissions (hospital
sterilizers, waste oil combustion, dry cleaning, residential wood combustion, and
wastewater treatment).
B. DISPERSION MODELING AND EXPOSURE/RISK
CHARACTERIZATION
Figure I.I characterizes dispersion modeling, exposure assessment, and risk
characterization as three separate and sequential operations. Conceptually, they are
separate operations and are accomplished sequentially in some other studies. EPA's HEM,
used in this study, performs ail of these steps during the course of a single model run. In
HEM, there are two modeling options: Systems Applications Human Exposure and
Dosage Model (SHED) and Systems Applications Human Exposure and Risk Model
(SHEAR). (See Chapter III for further discussion.) HEM/SHEAR was used in this study.
In HEM, air toxics emissions data are first modeled to estimate annual average ambient
concentrations of each pollutant throughout each study area. (Point and area sources are
modeled in fundamentally different ways in HEM.) The resulting model-predicted
concentrations are then applied to block group/enumeration district (BG/ED) populations
to estimate the number of people exposed to different pollutant concentrations. These
exposures are then converted to estimates of cancer incidence by applying cancer unit risk
factors, which relate individual exposures to certain probabilities of contracting cancer.
The number of excess cancer cases in each BG/ED is then summed for all pollutants and
across the entire urban area to estimate aggregate cancer incidence from multi-source,
multi-pollutant exposures.
-------
In this study, secondary formaldehyde exposures and risks could not he assessed using
the emissions inventory/dispersion modeling approach. Instead, measured ambient air
levels of formaldehyde were used to calculate total formaldehyde-related risk for each
urban area after which secondary formaldehyde risks were estimated by subtracting primary
formaldehyde-related risk from total formaldehyde-related risk. Secondary
formaldehyde-related risks were apportioned to source categories according to their VOC
emission totals.
To evaluate exposures and cancer incidence in 1995, a module was used that
superimposes various growth factors and emission reductions onto the base year inventory
prior .to emissions modeling. This module is discussed only in the companion Volume II,
which addresses the controllability aspects of this study. Volume I reports only on the
exposures and cancer incidence associated with the base year inventory. The growth and
control projection module was not needed for the base year analysis.
-------
CHAPTER IS
EMISSION INVENTORY ENHANCEMENT
The emissions data base used in this study has benefitted from recent national studies of
toxic emissions from specific source types and from local inventory efforts in some of the
cities. The purpose of this chapter is to describe specific changes made to the emissions
data base since the original air toxics controllability study in 1985. This chapter is divided
into three sections. The first section presents revised toxic emission estimates for source
categories where new information has become available nationally. This is followed by a
description of how local survey information was incorporated into the inventory. A
description of how the original emission inventory was developed for the 1985 study is
presented in the third section for completeness.
A. IMPORTANT SOURCE CATEGORIES
1. Road Vehicles
Mobile source emission factors were completely revised from the 1985 study. This
affected mobile source emissions estimates for the following compounds: formaldehyde,
benzene, 1,3-butadiene, gaseous particle associated organics, and diese! participates. The
source of new motor vehicle emissions information (Carey, 1987c) did not include emission
factors for the base year (1980) and contained only composite emission factors. Separate
factors for each of the four motor vehicle categories in the National Emissions Data System
(NEDS) were needed. Hence, some additional calculations were required. This section of
the report describes how motor vehicle emission calculations were performed and how
these numbers were applied in the analysis. A summary of resulting emission factors by
vehicle type is shown in Table II. 1. The emission factor calculation process is described
separately for each pollutant below.
MOBILE3 was used to calculate 1980 total hydrocarbon emission factors by vehicle type.
This information was then the basis for estimating emission factors for all organic toxic
compounds. MOBILE3 was used to estimate emission factors for three different vehicle
speeds — 20,45, and 55 miles per hour -- which are used to represent urban, rural, and
expressway driving, respectively. AH other conditions are kept constant in the model: low
altitude. Federal test procedure conditions are simulated. Emission factors for the three
speeds are weighted according to the percentage of travel on each roadway type (urban -
16 percent, rural — 30 percent, expressway — 54 percent) to develop a composite number
for each of the four vehicle types. These composites are then used to estimate
formaldehyde, benzene, and 1,3-butadiene emission factors.
-------
Table II.1
Mobile Source Air Toxic Emission Factors for 1980
Pollutant
Formaldehyde
Benzene
1,3-butadiene
Particulates
Vehicle Type
LDGV
LDGT
HDGV
HDDV
LDGV
LDGT
HDGV
HDDV
LDGV
LDGT
HDGV
HDDV
LDGV
LDGT
HDGV
HDDV
Emission Factors
(grams/mile)
0.0497
0.0762
0.2908
0.1475
0.1686
0.2292
0.4124
0.0513
0.0127
0.0205
0.0328
0.0159
0.0149
0.0182
0.0313
1.9730
LDGV » Light-duty gasoline powered vehicle
LDGT » Light-duty gasoline powered truck
HDGV = Heavy-duty gasoline powered vehicle
HDDV = Heavy-duty diesel powered vehicle
-------
a. Formaldehyde
To perform the 1980 formaldehyde emission factor calculation for light-duty gasoline
powered vehicles, it was necessary to know (by technology type) the percentage of exhaust
VOC that was formaldehyde. For this, EPA's Office of Mobile Sources provided the
following numbers (Carey, 1987a):
Control Technology Formaldehyde Emissions
noncatalyst 1.66 percent of exhaust VOC
oxidation catalyst - * 1.11 percent of exhaust VOC
3-way catalyst 0.83 percent of exhaust VOC
Formaldehyde is emitted in the exhaust, not the evaporative, portion of the organic '
emissions, so the percentages identified above are applied to the exhaust hydrocarbon (HC)
values only. Table II.2 illustrates the calculations made to estimate 1980 motor vehicle
emissions by vehicle type. As Table II.2 shows, formaldehyde fractions are identified
separately for four different diesel vehicle types as well as for the three different technology
types for light-duty gas vehicles (LDGVs). Fractions of diesel travel by vehicle type were
estimated from MOBILES Fuel Consumption Model values. Exhaust HC emission factors
listed in Table 11.2 are from MOBILE3. The result of the weighting of vehicle class/control
technology fractions, formaldehyde fractions, and exhaust HC emission factors is the
formaldehyde emission factor shown in the right-most column of the table. Use of these
values to estimate 1980 motor vehicle formaldehyde emissions is described later in this
road vehicles section.
b. Benzene
Benzene is emitted in both exhaust and evaporative HC, and the percentage of HC that
is benzene varies by vehicle type. These differences were accounted for in estimating 1980
benzene emissions. Benzene emissions expressed as percentages of HC are taken from
Table 4-1 of an Office of Mobile Sources report (Carey, 1987c). Table 11.3 shows how these
percentages and MOBILE3-generated HC emission factors were used to estimate benzene
emission factors by vehicle type. For diesels, the calculation is simplest because there are
no evaporative emissions. The fractions listed under technology type are vehicle miles
traveled (VMT) fractions from MOBILE3 for diesels. Diesel-emitted benzene then is
estimated by multiplying exhaust HC emission factors by the benzene-in-exhaust fraction.
The VMT fractions are used to estimate the emission factors by vehicle type shown in the
right-most column of Table II.3.
For light-duty gasoline powered vehicles, evaporative emissions vary according to
whether vehicles have fuel injectors or carburetors. It was estimated that 90 percent of
LDGVs were carbureted in 1980 and that the remainder were fuel injected.
-------
»
G flJ H ^
Q t-t V
CO MtH
CO >» O S
*s "° "o "ti
w o e *~f
m I-H <-< CM
oo «-H o vo
CM CM O r—
o o o o
* • * •
o o o o
CO
o
O\
CM
vo oo en oo
~* CM CO sj-
O O CM -<
O O «-i O
• * • *
o o o o
rt
rH
3
o
o
o
&<
CM a
• o
l™1 *"
M W
CD >H
rH Q
e2 .
to
O
§
EQ
Q Q Q CO
Q Q Q 3
J J 33 03
i i i i
i—(
01
o>
8
-------
x™\
¥
o
Q
•f(
M
•K
^4
-Q
cu
* ^"> x- -iH
rH 6
• CJ x.
W iH bO
B 0)
W !>
01
x. CU
CO >>
W EH
C (T3
0 rH S>vx
•H O bO <1>
W O i-H
01 >»r-H tH
•H -Q O S
e ex.
W . J3 bO
CJ CJ xx
fa E-i
VD
CO
vo
i — i
•
O
r- CM
p** -
<3) fa
01
(TJ
js a
x o
-x
W bo
CM I-H O O
CO Si- .-H .-H
f—i O *~H i—<
o o o o
• * • •
o o o o
CO CO
o o
d o
sf CO
CM Sf
CO CO
O O
O O O
sr sf I-H
CM CM ^H
O O O
* * »
000
in m
o o
• *
CM CM
CO
m
CO
CM
co
ooo
O
4->
O
0)
to
CQ
en
(U
4-> 1-H
M tH
=3 e
a \
-C bo
CO CO
VO vo
CO CO
vo
co
co
CO
CO O CM
~5T O CM
o>
CJ
•H
-C
(U
>
0
IH
4->
O
c^
OT
01
«J
rH
O
^i^
O
rH
O
JS
o
>
c
0
1-t
4-J
o
m
fa
o
o
0
o
o
0
o
o
co
CO
co
o
*
o
^.
CO
o
d
PX^
^
co
d
*-> o
Cu
K*"»
H
QJ
r—l
O
•ri
J3
o> at
3 O
^ -rH
CtJ rH
CJ 0)
1 — J
t> M I P
0 • C
> EH >
o a a
Q Q a
•J J S3
i i i
t/i
rH
O)
H > OT
S 0
-------
Benzene emission factors for LDGVs and other gasoline-powered vehicles were then
estimated by multiplying exhaust HC by the benzene exhaust fraction, multiplying
evaporative HC by the benzene evaporative fraction, and summing the products. Where
Carey (1987c) provided a range of evaporative HC percentages, the midpoint was used in
the calculations.
c. 1.3-butadiene
Motor vehicle emissions of 1,3-butadiene are roughly 0.35 percent of the total exhaust
HC (Carey, 1988). The,1,3-butadiene emission factors provided by the Office of Mobile
Sources were shown earlier in Table II. 1.
'' d. Particulate Matter
Paniculate emission estimates have also been significantly changed from previous work
so that risk assessments could be made using the comparative potency approach. (See
Section B of Chapter III for more discussion of the comparative potency approach.) Motor
vehicle emission factors for gaseous particle associated organics were estimated as the
soluble organic fraction. Diesel particulate emission estimates were made using total
particulate. Emission estimation methods for gasoline and diesel participates are described
separately below.
I. Diesel Particulate
Diesel particulate emission factors by model year were taken from Table 2-2 of an Office
of Mobile Sources (Carey, 1987c) report. Carey does not estimate emissions for 1980, so
diesel particulate emission factors and VMT fractions by model year estimated from
MOBILE3 were used to estimate composite emission factors by vehicle type. These
calculations are shown in Table II.4. The calculation of 1980 diesel travel fractions used to
estimate the amount of travel by light-duty diesel vehicles (LDDV), light-duty diesel trucks
(LDDT), and heavy-duty diesel vehicles (HDDV) is illustrated in Table II.5.
li. Gasoline Particle-associated Organics
Gasoline particle-associated organic emission rates were estimated using methods
consistent with the unit risk factor that is applied for gasoline particle-associated organics
(see Section B of Chapter III on the comparative potency approach). The original estimate
of gasoline particle-associated organic emission factors for 1980 was made with the Office
of Mobile Sources particulate model, using the organic fraction of the particulate. The
organic fraction in the particulate model includes elemental carbon, however, and because
only the soluble organic fraction should be counted, the earlier values were revised. The
correct values were listed in Table II. 1 under particulates. These revised values were
provided by Penny Carey (1987b).
e. Application to Study Areas
Table II. 1 emission factors were used along with the NEDS county fuel consumption
values to estimate toxic compound emissions for each of the four vehicle types. Because
fuel consumption estimates were available for only three vehicle types, the gasoline
10
-------
Table II.4
Diesel Particulate Emission Factor Calculation
Calculation of 1980 Light-duty Diesel Vehicle Particulate
Emission Factor
Model
Year
1980
1979
1978 and
.earlier
(A)
PM Emission
Factor
(grains/mile)
0.5
0.8
0.7
(B)
VMT
Fraction
0.038
0.142
0.820
(A*B)
Composite
Emission
Factor
0.0190
0.1136
0.5740
0.7066
Calculation of 1980 Light-duty Diesel Truck Particulate
Emission Factor
Model
Year
1980
1979 and
earlier
(A)
PM Emission
Factor
(grams/milei
O.5
0.9
(B)
VMT
Fraction
O.O35
0.965
(A*B)
Composite
Emission
Factor
O.O175
0.8685
0.8860
Calculation of 1980 Diesel Composite Emission Factor
Model
Year
LDDV
LDDT
HDDV
(A) .
1980 Emission
Factor
(grams/mile)
0.7066
0.8860
2.1859
(B)
VMT
Fraction
.0838
.0684
.8-177
(A*B)
Composite
Emission
Factor
0.0592
O.0606
1.8530
I .9731
11
-------
Table II.5
Calculation of 1980 Diesel Travel Fractions
(A) (B) (A*B)
1980 Estimated 1980 Estimated
Diesel Vehicle Fuel Use Road Miles VMT VMT
Categories (billion gals) per Gallon (billions) Fractions
LDDV 0.462 15.78 7.290 0.0838
LDDT 0.541 . 11.00 5.951 0.0684
HDDV 13.989 5.27 73.722 0.8477
86.963
12
-------
light-duty vehicle estimates were dissaggregated by percentage fuel consumption into
LOG V and LDGT emissions. Weighting factors were 77.9 percent for LDG Vs and 22.1
percent for LDGTs.
2. Chrome Platers
The estimated chromium emissions from chrome plating operations were revised from
the previous study. Previously only total chromium emissions were estimated for this
source category. For the current effort, information was available to estimate hexavalent
and total chromium emissions. Emission estimates for each pollutant are described
separately below.
a. Hexavalent Chromium
The EPA Engineering Standards Division, as part of its NESHAPS program, provided
hexavalent chromium emissions data by process type (hard and decorative plating), state,
and county for the five study areas (U.S. EPA, 1987d). The emissions were summed by
county for Study Areas B, C, D, and E. Study Area A was reported to have no hexavalent
chromium emissions.
The local air quality management agency provided total chromium emissions for Study
Area B based on extensive local surveys and source testing. These local agency-supplied
emissions agreed to within 10 percent of the EPA/NESHAPS data and were used in this
study. Only 92 percent of the local agency totals were assumed to.be Cr,"1"6 based on
EPA/NESHAPS test results identifying the hexavalent-to-total chromium ratios.
The EPA/NESHAPS data suggested there are no hexavalent chromium emissions in
Study Area A. The local air toxic survey noted, however, that there are at least six
operating chrome platers in the area, but the survey did not provide emission estimates for
each chrome plater. Hence, as an approximation, the national hexavalent chromium
emissions were apportioned to the county level by population (Radian, 1987a). The
resulting estimates of hexavalent chromium emissions by study area are listed in Table II.6.
b. Total Chromium
According to EPA tests, on average about 92 percent of the chromium emitted from
platers is hexavalent (U.S. EPA, 1987a). Using this percentage and the hexavalent
chromium emissions listed in Table II.6, total chromium emissions were calculated for each
study area by county, except for Study Area B. The local agency provided total, not
hexavalent, chromium emissions for this study area. Hexavalent chromium emissions were
then estimated using the 92 percent assumption mentioned above.
3. Cooling Towers
Cooling tower chromium emissions were revised to include hexavalent chromium
emissions as well as total chromium emissions. To further enhance the emission inventory,
13
-------
Table II.6
Hexavalent and Total Chromium
Emissions from Chrome Plating Operations
Emissions (Tons/year)
Study Hexavalent Total
Area Chromium Chromium
A
B
C
D
E
0.024
13.8
0.025
0.6
0.806
0.026
15.0
0.027
0.64
0.876
Source: .U.S. EPA, 1987d
14
-------
cooling towers were divided into two categories: comfort and industrial cooling towers.
Comfort cooling towers are inventoried as area sources while industrial cooling towers are
inventoried as point sources.
It was assumed that total chromium emissions for cooling towers are equivalent to
hexavalent chromium, although EPA is still conducting source tests to ascertain if this
assumption is true. The following discussion describes the emission estimation method
used for each type of cooling tower.
a. Comfort Cooling Towers
EPA's Office of Air Quality Planning and Standards (OAQPS) provided current
documentation (U.S. EPA, 1987b) to estimate hexavalent chromium emissions from
comfort cooling towers. These emissions were estimated using a per capita emission factor
specific to each state. These emission factors and the calculated hexavalent chromium
emissions for the five study areas are presented in Table II.7.
b. Industrial Cooling Towers
EPA-OAQPS provided national hexavalent chromium emissions (U.S. EPA, 1988) from
industrial cooling towers in eight industrial categories. These were reduced to four
categories by aggregating five that individually account for low or negligible emissions into
an "other" category, as shown in Table II.8. The five categories aggregated into "other" are
tobacco, utilities, tire and rubber, textiles, and glass manufacturing. The emissions for each
source category were apportioned to the county level by employment in the related
Standard Industrial Classification (SIC) category, as reported to the Department of
Commerce. The calculated hexavalent chromium emissions by industry and study area are
presented in Table II.8.
In prior work, industrial cooling towers were included in the emission inventory as an
area source. This possibly overpredicted cancer incidence, as area sources are allocated
within the modeling region by population. Cooling towers are at industrial sites, which may
be removed from residential areas. To obtain a more realistic estimate of cancer incidence
from cooling towers, the county level emissions were inventoried as point sources.
Petroleum refining, chemical manufacturing, and primary metals account for more than 98
percent of the industrial cooling tower chromium emissions in each study area. Each study
area was examined to determine the location of the greatest concentration (or grouping) of
facilities in each of these three industries. Once a grouping of facilities was chosen to
represent each industry, the cooling tower emissions were assigned as a new point at one of
tlie facilities within the group.
The emissions from the industrial category labeled "other" were assigned to a single
cooling tower. The utility industry was chosen to represent this category because the
majority of the remaining emissions are from utilities and there are utility plants in each
study area. The emissions were assigned to the largest utility in the study area. The
NESHAP Background Information Document (BID) (U.S. EPA, 1988) for chromium
15
-------
Table II.7
Hexavalent Chromium Emissions From Comfort Cooling Towers
Hexavalent Chromium
Study Emission Factor Emissions
Area fIb/person-vr)* fTon/vr)
A 1.21 (10~3) 0.92
B 1.19 (1CT3) 6.5
C , 1.44 (10~3) 0.27
D 1.01 (10~3) 0.7
E 8.81 (10~4) ' 0.74
Source: U.S. EPA, 1987b.
* Average emission factor used from cited reference rather than
upper bound of range.
16
-------
CO
M
M
0)
rH
J3
0)
>
SO T3
•H >^- *-•
S « co
o rt
I-I (U
o <;
** >»
s*s
rH <->
at co
n) >>
>
at
s
n
M
3
-a
c
(U
co
ai
o
u
3
O
CO
CLi
17
-------
emissions from industrial cooling towers provided model stack parameters for each
industrial category. The stack parameters assigned to each cooling tower are presented in
Table II.9.
4. Wood Stoves
EPA-OAQPS provided current emission factor documentation describing participate
matter (PMio) emissions from wood stoves. Based on the information provided, the PM
emission factor for a conventional noncatalytic wood stove of 23.5 g/kg (47 Ibs/ton) of wood
burned was used (U.S. EPA, 1987c). Particulate matter was computed rather than B(a)P or
some other POM because the comparative potency factor used to estimate total POM
cancer risk is applied to paniculate matter. (See Chapter. HI, Section B for details.)
Based on the information in the revised VOC species data manual, benzene emissions
from wood stoves were estimated to be 6.3 percent of the total VOC emissions (Radian,
1987b). Therefore, it was assumed that 6.3 percent of the VOC emission factor would
represent the benzene emission factor. Using this assumption, the benzene emission factor
was estimated to be 126 pounds of benzene per ton of VOC emitted.
5. Coal/Oil Combustion
• • r"
Coal and oil combustion emission factors for arsenic, chromium, formaldehyde, and
nickel were revised from those of the original study using more recent test data (Radian,
' 1986). Radian (1986) also provided emission factors for beryllium, cadmium, manganese,
and mercury, which were compounds not included in the 1985 study. Hexavalent chromium
combustion emission factors were developed using the information provided in two EPA
chromium screening test report documents on coal fired boilers (U.S. EPA, 1985b and
1985e). According to these documents, approximately 0.41 percent of the uncontrolled
chromium emitted is hexavalent and 0.15 percent of the controlled chromium emitted is
hexavalent. Lacking data on chromium emissions from oil combustion, these factors were
also applied to oil fired boilers.
6. Other Hexavalent Chromium Emissions
Hexavalent chromium emission factors for refractories, electric arc furnaces, and
argon-oxygen decarburization (AOD) furnaces were obtained from EPA chromium
screening study test reports for these source categories (U.S. EPA, 1985c and 1985d).
7. Methylene Chloride Emissions
The EPA source assessment document for methylene chloride provided nationwide
emission estimates by source category (U.S. EPA, 1985f). Degreaser emission estimates
.were provided at the state level and were apportioned to the study areas by population.
According to the source assessment document, 67 percent of the degreasing emissions are
from cold cleaners and the remaining emissions are split equally between open top and
18
-------
Table II.9
Model Stack Parameters for Industrial Cooling Towers
Stack Stack Exit Gas
Height Diameter Temperature Flow
Industry (feet) (feet) (F) (cu ft/min)
Petroleum Refining
Chemical Manufacturing
Primary Metals
Utilities
46
34
38
80
22
18
19
32
100
100
iod
100
666,797
440,888
448,997
1,344,333
19
-------
-------
conveyorized vapor degreasers. Using this information, methylene chloride emissions from
degreasing operations were assigned to the source categories: cold cleaners, open top
vapor degreasers, and conveyorized vapor degreasers. All other emissions were
apportioned to the study areas by population and assigned to the miscellaneous solvent use
source category. No large point source emitters of methylene chloride were identified in
the five study areas of concern.
B. STATE AND LOCAL DATA INPUTS
Plant specific emissions were verified and updated using local air toxic surveys and plant
specific information from EPA. Local air toxic surveys were available for Study Areas A
and B for the year 1985. These surveys also provided area source emissions which were
compared with the emissions in the existing inventory. Changes in the area source
categories resulting from this comparison, which were discussed briefly above, are discussed
in more detail here. Baseline air toxic emissions for Study Area D were compared with the
emissions in the Program Integration Project — Queries Using Interactive Commands
(PIPQUIC) computer data base. PIPQUIC is a data handling system developed by EPA as
part of the IEMP to calculate exposures and risks from air toxics emissions data. PIPQUIC
is being used in various urban air toxics studies. The PIPQUIC data base for Study Area E
had been available in the previous air toxics controllability study and was incorporated in
the five city data file at that time. The EPA provided plant specific information for two
point sources of 1,3-butadiene in Study Area C. These were incorporated into the
emissions data base.
State and local air pollution control agencies were also given an opportunity to review
the plant specific emission estimates by compound during the course of the 1985 study.
Their comments resulted in some significant changes to the emission inventory data base,
particularly for Study Areas B and C. These changes have now been incorporated.
In general, if a study area had a small number of air toxic emitting sources and local air
toxic emission estimates were available, changes were made to individual records in the
point source emission inventory to reflect these data. Project resources did not allow this
detailed point-by-point matching between local air toxic survey data and the five city data
base for all cities and all compounds. For compounds where there were local survey .
estimates for hundreds of facilities (many of which were too small to be included in the
1980 NEDS/NAPAP Emissions Inventory), only the largest emitters were inventoried as
point sources. The remainder were included in the emissions data base as area sources so
that five city data base emission totals matched those determined from local surveys. This
inventorying procedure has implications for the conclusions that may be drawn about the
relative importance of point vs. area sources because it does not completely reflect all point
source emissions as such. The extent of this bias will vary by compound, with
perchloroethylene being a good example of a compound whose point source emissions are
underrepresented in this analysis.
20
-------
-------
C. ORIGINAL EMISSION INVENTORY DEVELOPMENT
1. Introduction
While the toxic emission inventory used in the 1985 study was updated for many source
categories as described in the previous sections, there were many parts of the original
emission inventory that were not changed. For completeness, this section provides a
summary of the methods that were used to develop the original inventory, particularly for
those categories for which special techniques were used.
The starting point in developing the baseline inventory was the National Acid
Precipitation Assessment Program (NAPAP) emissions inventory for 1980. The NAPAP
inventory was chosen because it was considered to represent the best available detailed
inventory of emissions on a national scale at that time. The NAPAP inventory is similar to
EPA's National Emissions Data System (NEDS) but it has been improved substantially by
incorporation of the latest available emission factors, substituting data from the Northeast
Corridor Regional Modeling Project (NECRMP) and recent state submittals, and
incorporating electric utility data compiled by Pechan.
To further improve the NAPAP inventory, for the purposes of the toxics study, specific
air toxic emissions information was obtained from EPA Emission Factor Documents,
Source Assessment Documents, and local air toxic surveys. If more than one data source
was available for a pollutant and source category, the following criteria, listed in order of
priority, were applied to estimate air toxic emissions:
(1) Plant specific emission information from local air toxic surveys or source assessment
documents.
(2) Emission factors.
(3) National or statewide emission estimates apportioned to counties.
The inventories developed using these procedures were reviewed by the appropriate
local agency for each study area. Corrections were then made to the air toxic inventories
based on comments received from these local agencies. The plant specific information
needed was facility name, control equipment, control effectiveness, location (coordinates),
pollutant, emissions, process (by Source Classification Code - SCC), and stack parameters
(height, diameter, temperature, and flow).
All of these data items were necessary in order to meet the multiple needs of emission
inventory users. Locations, emissions, and stack data were needed for dispersion modeling.
Controlled and uncontrolled emissions and current control equipment data were needed to
21
-------
estimate the effect of future regulations and to estimate future control costs for the
controllability portion of this study.
2. Special Methods for Selected Source Categories
A few source categories were identified for independent analyses of their toxic emissions
because it was felt that using NEDS data and emission factors (or species profiles) would be
inadequate. The following subsections describe the methods used to make emission
estimates for these selected source categories.
a. Data and Sourcss
• Radian Corporation provided Pechan with baseline emission estimates for dry cleaners
(VOC), refractory manufacturing (TSP), residential wood combustion (TSP and VOC), and
sterilizers (ethylene oxide) (Hall, 1985). These emission estimates are listed in Table 11.10.
The dry cleaning VOC estimates were obtained directly from state agencies. The data for
refractory manufacturing were extracted from source assessment documents arid represent
two model plants located in Study Area D. Sterilizer estimates include only ethylene oxide
emissions from hospital usage. Actual values were obtained for one study area and a
coefficient was developed; other areas are based on this value with adjustments made for
the number of hospitals and admissions.
b. Residential Wood Combustion
Residential wood combustion emissions were prepared by OMNI Corporation for
Radian Corporation. OMNI's estimates of residential wood combustion emissions
(Simons, 1984) for each of the five study areas were divided into three categories: (1)
fireplaces used primarily for ornamental or aesthetic wood burning, (2) primary heating,
done primarily with wood stoves, and (3) secondary heating, done primarily with modified
fireplaces (inserts) and wood stoves.
j. Fireplace Emission Estimate Methodology
Using the 1980 census data (U.S. DOC, 1982a) in conjunction with U.S. Bureau, of
Census data (U.S. DOC, 1975, 1980, and 1982b), OMNI estimated at the county level the
number of single family households having one or more fireplaces. OMNI used a
conservative approach, however, in that single family houses estimated to have two or more ,
fireplaces were assumed to have only one operating fireplace. No attempt was made to
estimate the amount of wood burned in fireplaces in multi-family dwelling units or mobile
homes because residential wood fuel use studies indicate that more than % percent of all
wood used for residential wood combustion purposes occurs in single family dwelling units
(Skog and Watterson, 1983).
OMNI applied a utilization factor to each of the calculated number of installed
fireplaces to estimate the number of operating fireplaces. OMNI estimated only those
fireplaces installed as part of the original construction of single family homes even though
22
-------
Table 11.10
Emission Estimates for Selected Toxic Emitting Source Categories
VOC Emissions (TPY)
Category
PCE Dry Cleaning
Sterilizers
(Ethylene Oxide)
Residential Wood
Combustion:
Fireplaces
Woodstoves
A
911
7.8
Study Areas
B C
0 *
44 2.8
0
443
9.7
E
338.4
11.4
923
6,177
6,968
32,832
461
1,639
820 243
5,180 2,457
Category
Refractory
Manufacturing
TSP Emissions (TPY)
B
Study Areas
C D
uncont
cont
5,192.1 236.06
E
Residential Wood
Combustion:
Fireplaces
Wood Stoves
1,008 7,295 475
2,592 13,771 685
866 260
2,234 1,040
* not available
** original OMNI Corporation emission estimates listed for fireplaces and wood stoves
for Study Area B were later adjusted downward from the estimates listed above per
comments from the local agency.
cont=controlled
uncont=uncontrolled
Source: Hall, 1985
23
-------
•national fireplace sales data (HID, 1984) indicate that 36 percent to 43 percent of recent
fireplace installations are associated with residential repair and remodeling projects.
OMNI estimated the number of operating fireplaces used for ornamental or aesthetic
purposes by applying a conversion factor that assumes a certain percentage of existing
operating fireplaces were modified (i.e., inserts installed) to be used for secondary heating
rather than for aesthetic wood burning purposes. Fireplace conversion factors were
estimated from residential fuel wood use studies (Skog and Watterson, 1983; EIA, 1982b
and 1982c) and a review of regional marketing data of residential wood burning devices.
The average amount of wood burned per dwelling unit having operating fireplaces was
based on a number of studies, discussions with state and regional foresters, state energy
officials, and commercial/retail wood dealers. OMNI calculated the total amount of wood
(in cords) burned in fireplaces used primarily for ornamental or aesthetic purposes in a
county. AP-42 fireplace emission factors (U.S. EPA, 1985g) were applied to the estimated
amount of wood burned to estimate TSP and VOC emissions.
H. Primary Heating Emission Estimate Methodology
To estimate primary wood heating emissions, OMNI used 1980 census data to determine
the number of households using wood as a primary heating fuel. The households answering
"wood" are summarized at the county level in the 1980 Census Survey. Other studies (Skog
and Watterson, 1983; EIA, 1982b, 1982c) have indicated that the 1980 Census Survey may
have underestimated the number of households using wood as a primary heating fuel by as
much as a factor of 2.5 on a nationwide basis. These studies do not provide data at a county
level, however, and it was OMNI's opinion that in lieu of more detailed site-specific data
the 1980 Census Survey Data should be used and would represent a conservative estimate
of households using wood as a primary heating fuel.
OMNI estimated the average amount of wood consumed per primary wood burning
household using a variety of residential fuel wood use studies (Skog and Watterson, 1983;
EIA, I982a, 1982b, 1982c; Marshall, 1981) as well as information derived from discussions
with other experts in the wood energy field. Since there is very little data regarding the
average amount of wood consumed per primary wood burning household for the areas
evaluated, OMNI estimates are to be considered conservative since the values selected
were generally on the low end of data ranges available from these studies.
For each study area, total amount of wood consumed (in cords) was estimated for all
households using wood as a primary heating source. It was assumed that all households
using wood as a primary heating fuel used space heating devices that had the emission
performance characteristics of an "average" wood stove. Therefore, AP-42 wood stove
emission factors were applied to this residential wood combustion class.
Hi. Secondary Heating Emission Estimate Methodology
As noted above, it was assumed that most of the secondary or supplemental heating was
done primarily with modified fireplaces or wood stoves. It was recognized that ordinary
24
-------
fireplaces are occasionally used to provide supplemental heat (primarily for emergencies),
but it was OMNI's view that the emission contribution of this source was insignificant.
The average amount of wood burned per modified fireplace was based on national and
regional residential firewood use studies' (Skog and Watterson, 1983; EIA, 1982b; 1982c;
DER, 1982; NRBP, 1984; USD A, 1980) evaluations of heating demand characteristics of
single family homes by geographic area and discussions with regional foresters and
commercial firewood dealers.
OMNI assumed, based on its emission testing experience, that fireplace inserts
essentially have the same emissions characteristics as the average population of both
airtight and nonairtight wood stoves. Therefore, AP-42 wood stove emission factors were
applied to the amount of wood burned in modified fireplaces for secondary heating
purposes to estimate TSP and VOC emissions.
The number of wood stoves used for secondary heating purposes was estimated from
data in the 1980 Annual Housing Survey (U.S. DOC, 1983) and data regarding sales of
wood stoves in the United States (EIA, 1982b). Sales data indicate that the secondary wood
heating device market did not evolve until after the 1973 Arab oil embargo. Therefore,
OMNI estimates reflect wood stove installations occurring between 1975 and 1980.
The average amount of wood used per secondary heating wood stove was determined
using the same data sources and methodology as used for. modified fireplaces. AP-42 wood
stove emission factors were used to estimate TSP and VOC emissions.
Note that there is considerable uncertainty in the residential wood combustion emission
estimates. Comments received from one local air quality agency indicated that the OMNI
emission estimates were much higher than its own estimates. Therefore, OMNI's estimates
of the amount of wood burned per wood burning device were adjusted downward to try to
better account for the local conditions in that study area.
c. Toxic Emissions from POTWs
The 35 County study methodology (Versar, 1984) was adopted for calculating emissions
from POTWs. This method involves a multistep process. First, EPA's Industrial Facilities
Discharge (IFD) file was used to identify POTWs known to handle industrial discharges.
The following plant specific data were extracted from the IFD file for each POTW: county,
state, total flow, level of treatment (primary, secondary, or tertiary), percentage industrial
contribution, and type of industry discharging into the POTW (by SIC code). The POTWs
were next sorted into 13 categories based on industrial dischargers and level of treatment.
The 35 County study then derived seven prototype plants and two equations to be used to
calculate annual POTW air toxic emissions. Industrial discharges and level of treatment
were used to determine the appropriate equation and prototype plant to be applied.
25
-------
d. Toxic Emissions from Waste Oil Burning
The method used to calculate toxic emissions from waste oil burning was adopted from
the 35 County study (Versar, 1984). Each state's total residual oil consumption was
obtained from NEDS. If this value was not zero, equations were used to calculate waste oil
burned based on residential, commercial/institutional, and industrial residual oil
consumption.
3. Source Assessment Documents
Source assessment documents have been developed by EPA to provide information on
significant emitters of an air pollutant to help determine whether regulation is needed and
how effective controls might be. In these documents each source category is assessed
separately and emission information is provided either in the form of plant specific
emissions or nationwide emission rates. Plant specific information usually includes plant
name, location (longitude-latitude or UTM coordinates), stack parameters, pollutants, and
emissions. Plant specific data obtained from these documents were used in preference to
those in the NAPAP file.
Nationwide emission rates were used for source categories for which plant specific or
area specific emissions could not be found. These source categories were usually area
sources (grain fumigation, unidentified, or miscellaneous sources). The nationwide average
emission estimates were apportioned to counties by population. If regional or state
emission estimates were provided, these data were used instead of apportioning the
nationwide average estimate.
4. Emission Factors
Emission factors for air toxic compounds were developed from a number of different
sources including EPA emission factor documents, VOC species profiles, and air pollution
periodicals. EPA emission factor documents are for individual pollutants and contain
uncontrolled emission factor estimates for numerous sources. Unfortunately, emission
factor documents were not available for all of the compounds included in this study.
26
-------
CHAPTER
DISPERSION MODELING AND EXPOSURE/RISK
ASSESSMENT
A. MODELING METHODS
Assessing the cancer risks posed by a suspected airborne carcinogen requires three basic
pieces of information: (1) an estimate of the carcinogenic potency of the pollutant in
question, (2) an estimate of the ambient concentration to which an individual or a
population is exposed, and (3) an estimate of the exposed population. Described below are
the methods for developing and combining the above three pieces of information in order
to estimate cancer incidence in the five study areas.
The carcinogenic potency of a compound represents the probability of contracting
cancer from a lifetime (70 years) exposure to a 1 microgram per cubic meter concentration
of that pollutant. The carcinogenic potency estimates used In this study were developed by
EPA's Carcinogen Assessment Group. They are listed in Table III. 1.
EPA's Human Exposure Model was used in this analysis. HEM is a general model
designed to estimate the population exposed to air pollutants emitted from point and area
sources and the carcinogenic risk associated with this exposure (U.S. EPA, 1986). The
HEM comprises (1) an atmospheric dispersion model, with included meteorological data,
(2) a population data base of Bureau of Census data, and (3) a procedure for estimating
risks resulting from the predicted exposure. Using emission source data as input, HEM
estimates the magnitude and distribution of ambient air concentrations of the pollutant in
the vicinity of the source. These concentration estimates are coupled with population to
estimate public exposure to the pollutant. The HEM then estimates population risks if a
unit risk number determined from health data is input for the pollutant.
As discussed in Chapter I, within HEM there are two models, SHED and SHEAR. The
SHED model was developed to assess sources on a nationwide source category basis, e.g.,
all primary copper smelters. The SHEAR model was developed from SHED to model hot
spots, or smaller regions with more than one source. In addition to modeling major point
sources, SHEAR can also be used to model prototype sources (model plants and area
sources). Thus, SHEAR was appropriate for the urban scale modeling being performed in
this study.
In SHEAR, two basic dispersion algorithms are used: a plume algorithm for computing
concentration patterns resulting from a single source (point sources) and a box algorithm
for computing concentration patterns resulting from an urban-wide distribution of a single
source type (area sources). The dispersion zone for each modeled point source is limited to
27
-------
Table III.l
Compounds and Associated Unit
Pollutant
1. Formaldehyde
2. Ethylene Oxide
3. Chloroform
4. Carbon Tetrachloride
5. Ethylene Dichloride
6. Perchloroethylene
7. Trichloroethylene
8. Vinyl Chloride
9. Benzene •-,—
10. Gas Vapors
11. Arsenic
12. Chromium (Total)
13. Nickel'
14. PIC (BaP)
15. Lead
16. 1/3-Butadiene
17. Diesel Particulates
18. Gasoline Particle Associated prganics
A. Catalyst
B. Non catalyst
19. Asbestos
20. Chromium (Hexavalent)
21. Beryllium
Risk Values
Unit Risk*
1.3 x 10~5
1.0 x 10"4
2.3 x 10"5
1.5 X 10"5
2.6 x 10~5
5.8 X 10~7
1.3 X 10~6
4.1 X 10"6
8.0 X 10~6
6.6 X 10~7.
4.3 X 10""3
see Hexavalent
Chromium
3.3 x 10~4
4.2 x 10"1
npt available
2.8 x 10~4
3.0 x 10~5
-4
•1.2 x 10
7.9 X 10~4
7.6 X 10
-3
1.2 x 10~2
2.4 x 10~3
28
-------
Table III.l (continued)
Pollutant . Unit Risk*
22. Cadmium 1.8 x 10~3
23. Mercury • not available
24. Manganese not available
25. Methylene Chloride 4.1 x 10~6
*The unit risk value is the estimated probability of contracting
cancer as the result of a constant exposure over 70 years to an
ambient concentration of 1 microgram per cubic meter. Values for
some compounds have changed since the modeling portion of this
study was performed. If the revised unit risk factors were used
in this study, incidence estimates would differ by about 2
percent.
29
-------
a circle of a 20 kilometer radius about the source location. This limitation should have
minimal effect on study results because concentrations from most sources will be small
beyond 20 kilometers.
Exposure to air toxics in SHEAR is computed for residential population patterns for
annual-average concentrations. No account is taken of varying source strengths or varying
population patterns within a year. In this study, atmospheric transformation of toxic
compounds has been'ignored. (The assumptions and limitations of this study are delineated
in Chapter V.)
In the point source model in SHEAR, the concentrations from each source are
computed on a separate polar grid centered on that source. Population is defined on an
irregular grid of block group/enumeration district centroids. To define all concentrations at
a common set of points with a minimum of interpolation, SHEAR adds interpolated values
of concentration from each polar grid to the grid of centroids.
SHEAR was applied in the five study areas to estimate expected increases in cancer
incidence resulting from air toxics exposure as it exists in the base year (1980),. All point
sources for which Pechan had coordinates and stack parameters (from NEDS) and which
exceeded a specified emission cutoff (see discussion on following page) were modeled
individually in SHEAR as point sources. Area sources, point sources for which Pechan had
no stack or locational information, and point sources below the emission cutoff levels, were
all modeled collectively as area sources. .All area source emissions were subcounty
apportioned by the population density in each block group/enumeration district by HEM.
(This is a limitation of HEM and is handled in other ways in other studies, through the use
of other surrogate parameters.)
A sensitivity analysis of the difference between apportioning area source emissions by
the population density in each block group/enumeration district versus modeling area
source emissions as if they were constant throughout the study area, showed a fairly
dramatic increase in area source related incidence in each study area in the population
weighted approach. Depending on the characteristics of the urban area being modeled,
results varied by a factor of 2 to 4 between the two cases.
Because the SHEAR was not designed to be run for large urban areas with many point
sources, a large number of pollutants, and many scenarios, some changes were made to the
model to make it more efficient for the purposes of the present analysis. Rather than
running the SHEAR separately for each scenario and pollutant, it was run once assuming
100 tons per year of pollutant emissions from each major point source. For each point, the
output of SHEAR was saved in an intermediate file for later use. This intermediate file
contained the cumulative population exposure (micrograms/nv -persons/year) for each
modeled point. These values were then used to estimate population exposure to individual
pollutants by multiplying them by the ratio of actual emissions to modeled emissions.( 100
TPY). This normalized modeling approach saved considerable computer resources.
30
-------
Because the cost and execution time of SHEAR is linearly related to the number of
point sources modeled, it was necessary to establish a point source emissions cutoff level for
each pollutant in each study area. These cutoff levels were set to minimize the number of
sources to be modeled individually as point sources while maximizing the percentage of
emissions so treated. For some pollutants, there were so few emitters that ail sources
within a study area could be modeled (e.g., carbon tetrachloride). Other pollutants had so
many sources that a cutoff level was needed to avoid having hundreds of sources included in
the modeling of that pollutant (e.g., formaldehyde).
Table III.2 shows the point source cutoff levels defined for each pollutant by study area.
Point sources with emissions below these levels were modeled as area sources.
Because accuracy in the stack data used in the dispersion modeling is as critical as the
emission rates in estimating ground-level concentrations, considerable effort was taken in
performing quality control checks on the point source stack parameters. Steps taken to
ensure reasonable values for stack parameters were as follows:
(1) UTM coordinates that were widely different for the same plant were corrected.
(2) The stack heights and diameters were reviewed for compatibility.
(3) Sources with stacks taller than 100 feet should be fuel combustion sources and have a
stack gas temperature of more than 200 degrees F.
(4) If temperature, height, and diameter are zero or missing, values were assigned.
(5) Default values for exhaust gas flow rate and stack temperature for utility and industrial
boilers are based on the type of fuel burned.
(6) In cases where plume height was estimated but needed stack data were missing, plume
height was used to estimate stack height. Other stack parameters were then estimated
using plume rise formulas.
B. COMPARATIVE POTENCY APPROACH FOR ESTIMATING POM
RISKS
The approach used for estimating cancer incidence from exposure to POM in this study
differs markedly from previous approaches that used B(a)P as a surrogate for POM. Where
possible, this study adopted the comparative potency approach which allows published
cancer unit-risk estimates to be applied to particle emission estimates directly (Lewtas, :
1987), thus directly accounting for all POM emitted by each source category. (For sources
where comparative potency factors were not available, the "old" approach of using B(a)P as
a surrogate of POM was still used.)
In this comparative potency method, the risk of a suspect human carcinogen for which
there are no epidemiologic cancer data (e.g., diesel particulate) is estimated by comparison
31
-------
Table III.2
Emission Cutoff Levels Used for
Identifying Major Point Sources for
Modeling Purposes* (TPY)
Study Areas
Pollutant
Arsenic
Benzene
Carbon Tetrachloride 0.0
Chloroform
Chromium
Ethylene Dichloride
Ethylene Oxide
Formaldehyde
Gasoline Vapors
Lead
Nickel
Perchloroethylene
PIC
Trichloroethylene
Vinyl Chloride
1,3-Butadiene
Diesel PM
Gasoline PM .
Asbestos
* As defined in the text, sources having emissions above these
levels were modeled individually as point sources, whereas
sources with lower emissions were modeled collectively as area
sources.
32
A
0.0
0.0
0.0
0.0
0.0
O.Q
0.0
1.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
B
0.2
1.0
0.0
0.4
0.1
0.04
0.0
70.0
40.0
10.0
0.1
50.0
0.0
1.0
0.0
0.0
0.0
. 0.0
0.0
c
0.0
0.0
0.0
0.0
0.0
0.0
0.0
100.0
0.0
0.0
0.01
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
D
0.05
0.0
0.0
0.0
0.05
0.0
0.0
20.0
100.0
1.0
0.1
0.0
0.0
0.
0.0
0.0
0.0
0.0
0.0
E
0.1
0.0
0.0
0.0
0.05
0.0
0.2
1.0
0.0
1.0
0.1
1.0
0.0
1.0
0.0
0.0
0.0
0.0
0.0
-------
Table III.2 (continued)
Pollutant
Chromium (Hex)
Beryllium
Cadmium
Mercury
Manganese
Methylene Chloride
A
1X10~5
ixicr3
0.0
0.0
0.0
0.0
B
1X10~5
1X10~3
2X10"3
IxlO"3
5xlO~3
0.0
C
IxlO"5
0.0
0.0
0.0
0.0
0.0
D
1X10~5
1X10~3
2xlO~3
IxlO"3
5xlO~3
0.0
E
ixio~5
1X10~3
2X10"3
1X10~3
5X10~3
0.0
* As defined in the text, sources having emissions above these
levels were modeled individually as point sources, whereas
sources with lower emissions were modeled collectively as area
sources.
33
-------
with a known human carcinogen (e.g., coke oven emissions) according to the following
equation:
estimated
human risk
(diesel)
human
risk
(coke oven)
bioassay
potency (diesel)
bioassay potency
(coke oven)
The relative bioassay potency is obtained by the ratio of the slopes of the dose responses
from the same in vitro or in vim bioassays. . .
The comparative potency method described here estimates the risk of human lung
cancer from diesel paniculate (or other) emissions based upon lung cancer data obtained
from humans exposed to other organic combustion and. pyrolysis products. Epidemiologic
studies have shown an increased incidence of lung cancer in humans exposed to coke oven
emissions, roofing tar emissions, and cigarette smoke. The estimation of cancer risk from
diesel particles was, therefore, determined by comparing the carcinogenic and mutagenic
potencies of diesel exhaust paniculate extracts with those of coke oven emissions, roofing
tar emissions, and cigarette smoke tars in a battery of inyjUQ mutagenicity and mouse skin
tumor initiation and carcinogenicity bioassays. Particle unit risk estimates recommended
for estimating comparative POM risks in an Integrated. Environmental Management
Project (IEMP) are shown in Table III.3. These Table III.3 unit risk factors were applied to
the paniculate emissions by source category to estimate POM related cancer incidence.
Essentially, this is the same as treating each source category as if it emits a unique pollutant.
The incidence for all categories is then summed to estimate paniculate based incidence.
Table III.4 presents some of the combustion source test results that were considered in
selecting the Table III.3 unit risk values. EPA's Office of Mobile Sources recommended
using the Mustang II unit-risk data across the board for gasoline particulates. Because a
significant portion of the light-duty vehicle fleet in 1980 was noncatalyst (and most trucks
would be noncatalyst), a separate factor was applied to the noncatalyst portion of the
gasoline-powered fleet. The data for the Chevrolet 366 were used for this case (see Table
III.4).
For diesels, the only heavy-duty vehicle tested was the Caterpillar. This Caterpillar was
tested to represent the lower end of the potency range and, therefore, was not considered
representative of the diesel fleet. (Caterpillar tests were performed under low speed and
load conditions and the sample was stored for one year.) Thus, as was done by Carey
(1987c), the light-duty vehicle test data were used to represent diesel particulate unit-risk.
C. FORMALDEHYDE MODELING APPROACH
This study used (1) the emissions data bases and (2) modeling and exposure capabilities
built into HEM/SHEAR to estimate excess cancer incidence associated with all pollutants
except secondary formaldehyde. Because the majority of ambient formaldehyde is believed
34
-------
Table III.3
Particle Unit Risk Estimates Used for Estimating
Comparative POM Risks
Particle Unit-Risk Estimates*
Emission Source (Lifetime risk/ug/cubic meter)
VEHICLES
Gasoline-Catalyst 1.2 x 10~4
Gasoline-Noncatalyst 7.9 x 10~4
Diesel 3.0 x 10"5 •
RESIDENTIAL HEATING
Oil 0.9 X 10"5
Coal 1.0 x 10~5
Wood 1.0 x 10~5
INDUSTRIAL AND UTILITY POWER PLANTS
Oil 0.030 x 10~5
Coal 0.008 X 10~5
MUNICIPAL INCINERATION 0.008 X 10~5
COKE OVEN EMISSIONS 6.5 X 10"5
Source: Lewtas, 1987.
* These factors have been adjusted such that they are applied to
the total particle concentration to estimate risk from the POM
fraction of the particulate matter.
. 35
-------
Table III.4
Application of the Comparative Potency Method to Estimation of Lung Cancer
Unit Risks for Combustion Emissions
Unit-Risk Estimates
(Lifetime risk per million persons
per microgram per cubic meter)
Organics
Emission Sources
DIESELS
Nissan
Volkswagen Rabbit
Oldsmobile
Caterpillar
Relative
Potency in
Skin Tumor
Initiation
440
170
240
Relative
Potency in
Short -Term
Bioassays
-__
130
120
6.6
Particles
Relative
Potency in
Skin Tumor
Initiation
35
30
39
...
Relative
Potency in
Short -Term
Bioassays
23
20
2
GASOLINE CATALYST
Mustang II
GASOLINE NONCATLYST
Chevrolet 366
Ford Van
RESIDENTIAL HEATERS
Oil
Wood
UTILITY POWER PLANTS
Coal
120
95
140
790
1600
240
4-800
10-4000
51
9.1
60
16
300
23
4-800
0.002-0.8
Source: Lewtas, 1987
36
-------
to he formed secondarily, it was felt that the previously used technique of estimating
ambient formaldehyde concentrations exclusively from primary emissions was
inappropriate. Therefore, in this study, total formaldehyde exposure in the base year was
estimated from observed ambient concentrations. (No validated photochemical models
were available to estimate secondary formaldehyde production from VOC and other
precursors.)
The method used to estimate incidence from formaldehyde exposure was quite simple.
A single representative annual average formaldehyde concentration was selected for each
city; this value was multiplied by the number of people exposed; and this product was then
multiplied by the formaldehyde unit risk factor of 1.3 x 10. Table 1II.5 shows the data used
to make base year formaldehyde incidence estimates for each study area. The considerable
uncertainty in these incidence estimates should be recognized.
Once total formaldehyde-related cancer incidence was
estimated, secondary formaldehyde-related cancer incidence was
estimated by subtracting, from total incidence, that incidence
associated only with directly emitted formaldehyde (also termed
"primary formaldehyde risk"). Primary formaldehyde incidence was
determined from dispersion modeling of emissions just as all
other incidence figures were in this study.
37
-------
Table III.5
Base Year Total Formaldehyde Incidence Estimates
Annual Average
Formaldehyde Ambient Estimated
Study
Area
A
B
C
D
E
Total
Unit Risk
(per million)
13
13
13
13
13
Concentration*
(ug/nT3)
3
6.7
3.1
3.7
3.7
Population
(millions)
1.509
10.867
0.383
1.438
1.688
70 Year
Incidence
59
947
15
69
81
1,171
* These concentrations were based on ambient measurements (Lahre, 1988b)
and are assumed in this analysis, to account for total (direct and
secondarily formed) formaldehyde exposures and incidence.
38
-------
CHAPTER IV
BASE YEAR RESULTS
A. AGGREGATE RESULTS
Aggregate cancer incidence (or population risk) across the .5 cities in this study averaged
about 6 excess annual cases per million persons, ranging from about 2 to 10 in individual
cities. Areawide individual lifetime cancer risks averaged about 4 x 1()~4 and ranged from
about 1.5 x 10"4 to 7 x 10."4 (Note that these risks are not maximum individual risks, which
can be as high as 10~2 near some large point sources. Instead, these are individual risks
averaged for entire urban populations. This study did not estimate maximum individual
risks.)
Figure I V.I shows the pollutants and their contribution to aggregate incidence for the
five cities combined. Of the 18 compounds listed in Figure I V.I, only 4 contribute to more
than 10 percent of the aggregate incidence. These four, in order of importance, are POM,
1,3-butadiene, formaldehyde, and hexavalent chromium. Compounds contributing more
than 1 percent of the five city incidence but less than 10 percent are benzene (9 percent),
methylene chloride (3 percent), and ethylene oxide (2 percent). All of the other studied
compounds combined account for only 5 percent of the aggregate incidence. (These other
pollutants may be responsible for high risk point source-related problems, though;)
Figure IV.2 provides further information about the sources responsible for the incidence
estimates shown in Figure I V.I. The most important source category contributors to
aggregate incidence are road vehicles, comfort and industrial cooling towers, chrome
platers, solvent use, and fuel combustion, with road vehicles contributing more than all
other source types combined.
A detailed breakdown of the sources of POM risk is shown in Figure I V.3. About 77
percent of the POM risk is associated with road vehicle emissions, be they diesel participate
or gaseous particle associated organics. POM emitted by residential wood combustors -
primarily wood stoves - is the source of 17 percent of the POM risk. The only other
sources with 1 percent or more of the POM risk are external fuel combustors and iron and
steel facilities. Thus, area sources dominate estimated incidence from POM exposure.
Contributors to secondary formaldehyde risk are detailed in Figure IV.4. Because
contributions to secondary formaldehyde risk are assumed to be in proportion to VOC
emissions, Figure IV.4 also depicts the distribution of VOC emissions by source type in the
base year. Road vehicles and solvent use are the primary VOC emitting categories. There
are other significant VOC emitting source types, but none with more than 10 percent of
total emissions.
39
-------
10
w
EH
O
J«
00
o
fr5
N
C3
K
10
!.It,i!l!i
"»"*«['
OCQUQF3.-U = i.
p*i y\ Q» Qj ^3 S A ffi CC
S^aQpqnp^nH
ttoSOHw^w
o o
^1!
-ii
o
1—<
3
&
Pu,
O
O
ox
CO
^ FW'
^ ^-a
^ g§
•« |g
CO O O
w
PH
&
O
KH
PH
CQ
H
OT
-------
O
O
Q
O
2
o w
EH O
O
2 O
& EH
OP «
I li! Jill
Il
O
I-H
3
O
PH
o
I—I
d
CO
w
CO
-------
.0
iz;
»•— i
Q
CO
o a
o
o
o
CO
PH
CO
II
5i
42
-------
PS
o
P3
O
CO
o
iz:
o w
£>S
CO
o
EH Q
O
iz:
HH
EH
K
o g |
EH P
o ffi
o w
CO0
o
z
w
Q
>—i
O CO
fe H
CO
>H -H
O CO
w o
I"
&q
O
43
-------
Sources contributing to incidence from directly emitted formaldehyde are shown in
Figure IV.5. Direct emitted formaldehyde is estimated (via HEM) to contribute 40 percent
of total formal'dehyde related incidence. Road vehicles are the predominant contributor to
direct formaldehyde related incidence, with 63 percent of the total. Incineration,
off-highway vehicles, and external combustors are the next most prominent contributors to
direct formaldehyde related incidence, in order of importance.
Sources of benzene related incidence are illustrated in Figure IV.6. This pollutant's
incidence is dominated by motor vehicles, which account for 76 percent of the total.
Residential wood combustion is the only other category with more than 10 percent of the
benzene contributed incidence. As Figure IV.7 shows, road vehicles contribute 98 percent
of the 1,3-butadiene related incidence in the five study areas. Chemical manufacturing
facilities that emit 1,3-butadiene are found in one of the study areas and contribute 2
percent of the incidence estimated for this compound for the five areas. Primary sources of
hexavalent chromium-related incidence are chrome platers and industrial and comfort
cooling towers as shown in Figure IV.8.
Tables IV.l and IV.2 provide source category specific information broken down by VOC
or PM related compounds. Table IV.l results combine the secondary formaldehyde related
incidence with the HEM model results for the other organic compounds. The two tables
are important when evaluating future control options for reducing incidence because VOC
control options would be expected to lower organic toxic compound emissions and PM
controls would be expected to reduce particulate toxic emissions. Expected differences
between 1980 emissions and incidence and the same values in 1995 under different control
scenarios for the Table IV.l and IV.2 source categories will be presented in Volume II of
this study report.
B. CITY-SPECIFIC RESULTS
To accompany the five city aggregate results presented thus far, it was also of interest to
see if the same source categories and pollutants were important in the individual city
results. Figure IV.9 shows that expected incidence per million people varies from a low of
2.1 cases in Study Area A to 10 cases in Study Area D. In general, the city-specific results
mirror the five city aggregate results, with the primary source categories of road vehicles,
cooling towers, chrome platers, residential wood combustion, and solvent use being
important contributors to incidence in all the study areas. The "other" category in Figure
IV.9 has varying degrees of importance in the five cities, with Study Area C having the
greatest percentage of its total incidence contributed by source categories other than the
primary five. Study Area C was selected for inclusion in this study because of its chemical
manufacturing facilities and petroleum refineries, so it is not surprising that point sources
contribute more to the incidence in that area than they do in the other cities. (Although not
identified in Figure IV.9, point sources contribute a majority of what is shown as "other"
incidence for Study Area C in that figure. Point sources contribute less to the "other"
incidence in Study Areas A, B, D, and E than they do in Study Area C.)
44
-------
i
O
fe
o
W S
ON
^«^
5P
55
O
O JD H
CQ
O
O
O
O
CO
I - 8 CLj£ t
rtW*
Iff1*1
|i
es
5 S'wg:
HWJ
•Will
ili'Sl
1*1*41
o> • c o a
Iflfil
giigli-
|gB§£gc
"a a 5 •!>*'- n
I»«i1*i
1I1§!|J
ill|l|!
2: "o o « w *5 2
Q
t—i
O
^
H-(
Q ^
fvi ^
W 55
P=3 O
CO
CO
o
W
45
-------
o
Q
H— I
O
CO
N
g
o
o
CO
46
-------
w
O
o
55
O
jz:
s
W
I
w
w
PQ
O
C/3
I??*
!f«|
IIFII
^ in
•gi, S-g ° —
IpJi!
I'jlsf
SJB|.B!
3 |
Mi
iS !
UEl
I!!?!
3i
:j
i§
02
Q
t—t
O
§2
CQ
CO
CO
>H CO
f—I -^J
« O
47
-------
O
g
a
§3 §
g 5
0
w
o
;=>
o
CO
o
jz;
O
12:
K
o
>—•
fe
O
48
-------
Table IV.1
VOC Related Annual Incidence by Source Category*
Five Cities Combined
Source Category
1980
Incidence
Incidence
per Million
People
Light-duty Gas Vehicles
Light-duty Gas Trucks
Miscellaneous Solvent Use
Heavy-duty Gas Trucks
Surface Coating
Hospital Sterilizers
Heavy-duty Diesel Vehicles
Woodstoves
Residential Incineration
Cold Cleaners
POTWs
Dry Cleaning
Miscellaneous EDC
20.0
6.4
2.8
1.9
1.6
1.5
1.3
0.9
0.7
0.5
0.5
0.5
0.4
1.26
0.40
0.18
0.12
0.10
0.09
0.08
0.06
0.04
0.03
0.03
0.03
0.03
Total - All Categories
50.2
3.16
* VOC related compounds include formaldehyde, ethylene oxide,
chloroform, carbon tetrachloride, ethylene dichloride,
perchloroethylene, trichloroethylene, vinyl chloride, benzene,
gasoline vapors, 1,3-butadiene, and methylene chloride.
Note: No incidence values in this table should be considered to
represent absolute predictions of cancer occurrence. They should
be used in a relative sense only for broad screening activiies
and to provide more focused program direction. The dose-response
relationships and exposure assumptions inherent in this type of
cancer assessment are generally quite conservative. Errors
resulting from missing pollutants (either directly emitted or
secondarily formed), uncharacterized sources, and the lack of any
mechanism to simulate synergism among pollutants may potentially
offset this conservative bias, but to an unknown extent.
49
-------
Table IV.2
PM Related Annual Incidence by Source Category*
Five Cities Combined
Incidence Per
1980 Million
Source Category Incidence —People
Heavy-duty Diesel Vehicles 11.2 0.71
Chrome Plating 8.0 0.50
Light-duty Gas Vehicles 6.2 0.39
Comfort Cooling Towers 4.4 0.28
Industrial Cooling Towers 3.0 0.19
Woodstoves 2.9 0.: 18
Light-duty Gas Trucks 1.4 0.09
Fireplaces 1.4 O..Q9
Glass Mfg. - Melting Furnace 0.7 0,04
Total - All Categories 42.5 2.,68
* PM related compounds include arsenic, chromium (hexavalent),
nickel, POM, diesel particulates, gasoline particle associated
organics, asbestos, beryllium, and cadmium.
Note: No incidence values in this table should be considered to
represent absolute predictions of cancer occurrence. They should
be used in a relative sense only for broad screening activiies
and to provide more focused program direction.' The dose-response
relationships and exposure assumptions inherent in this type of
cancer assessment are generally quite conservative. Errors
resulting, from missing pollutants (either directly emitted or
secondarily formed), uncharacterized sources, and the lack of any
mechanism to simulate synergism among pollutants may potentially
offset this conservative bias, but to an unknown extent.
50
-------
o
Q
H—I
2
Oi
2
O
w
H
PS
o
o
o
C3 °
W w
PH U
O g
^§
5. co
CQ
CJ
O
CO
I
OT
P
•E
O
O
S
C/3
Q
O
O
Gft
12
(V
a
CJ
C/3
O
E-i
O
o
O
O
i—i
a
1W
*l1l:
S 6ST5!
||I|S||t|
-•2iiS.3i
8-al!ls
OS
•
CO
//
/
/
/
/
\/
:/
CO
cvi
\
\
CV2
C\2
O
o
O
m
52:
o
>—I
S
CQ
H
CO
<3
O
CO
id
II
W
O
o
iz;
o
NOLLVindOd
51
-------
City-specific results are presented by pollutant in Figure IV. 10. Benzene, hexavalent
chromium, formaldehyde, 1,3-hutadiene, and POM contribute to more than 80 percent of
the incidence estimated for each city. These top five pollutants constitute as much as 93
percent of the incidence in some study areas. POM related incidence seems to vary quite a
bit by city, and this variation is probably related to the percentage of emissions in each
study area that are from mobile sources. Incidence attributable to 1,3-butadiene exposure
is a larger portion of the total in Study Area C than it is elsewhere. This occurs because
there were some point source emitters of 1,3-butadiene identified for Study Area C, but not
in the other areas.
52
-------
o
•«-<•
•
>
PS
o
Q
t—t
o
O
o
PH
CO
I
CJ
H
d
w
CO
+
Pi
Q
>-
ffl
w
q
k—
<
is
K
o
CO
00
o
PH
CD
\
\
\
\
CO
CM
o
o
fc
o
o
25
O
25
O
CO
CO
CO
II
o
25
O'
25
OS
e
1—4
o
IS
53
-------
-------
CHAPTER V
ASSUMPTIONS/LIMITATIONS
Some of the more important assumptions and caveats associated with this study are as
follows:
(1) Personal, exposure to air toxics was estimated using annual-average concentration
estimates and it was assumed that exposures occur where people reside. In addition,
only outdoor exposures were modeled. Thus, this methodology ignores people's
movements throughout the urban area, travel outside the urban area, indoor exposures
and seasonal or diurnal variations in emissions. Because exposures were simulated
over a 70-year period, it is unclear how much this restricted modeling methodology
affects the study results.
(2) The study relied solely on quantitative estimates of cancer risk associated with
inhalation of ambient air over long periods. Acute and subchronic effects were not
included, and cancer cases associated with exposure routes other than inhalation of
ambient air were not quantified.
(3) Only 25 compounds were explicitly included in this study, although monitoring studies
have shown that urban atmospheres typically contain additional carcinogenic
pollutants. The compounds selected for study were chosen because they were
estimated to be the most important contributors to excess cancer incidence.
(4) This study focused on routine, continuous emissions. Accidental releases were not
modeled.
(5) Unit risk factors employed in this study represent the chance of contracting cancer
from a lifetime (70 years) exposure to a given concentration of that pollutant. It was
assumed that the resulting lifetime incidence levels could be divided by 70 to represent
annual incidence levels. The carcinogenic potency estimates used in this study were
developed by EPA's Carcinogen Assessment Group, and generally represent
conservative (upper bound) dose/response relationships.
(6) Cancer incidence estimates are presented for "existing" conditions (1980). These
incidence estimates are based on the assumption that emission levels for each scenario
remain constant for a 70-year period. In reality, emissions will vary from year to year.
(7) In assessing cancer risk within an urban area, each of the compounds under study has
been analyzed individually and the effects considered additive. No synergistic or
antagonistic health effects of these compounds have been assumed.
54
-------
(8) Sources included in the exposure modeling data set for each study area were limited to
those in the counties understudy. Therefore, while contributions from these sources
to areas outside the county boundaries were considered, contributions of sources
outside the county boundaries to air toxic concentrations within the study areas were
not.
(9) Modeling results presented in this report for exposure to gasoline vapors from
marketing operations do not include self-service exposures at service stations.
(10) Except for secondary formaldehyde formation, atmospheric transformation of toxic
compounds and precursors has been ignored. Both secondary formation and
scavenging may occur for the compounds included in this study. Thus, it is difficult to
quantify how neglecting transformation might affect the final results.
(11) Incidence from formaldehyde exposure was estimated using ambient monitoring data
and assuming that everyone within an urban area is exposed to the same
concentration. This is a relatively crude technique, but because so much of
formaldehyde is formed secondarily, this procedure was judged to be preferable to
modeling direct formaldehyde concentrations and ignoring secondary formation.
(12) It has been suggested that global background concentrations of some toxic compounds,
notably carbon tetrachloride, may be contributing significantly to observed ambient
readings. No attempt has been made in the dispersion modeling performed for this
study to account for background concentrations.
(13) While the comparative potency approach used to estimate POM risks for several
important sources is judged to be an improvement, conceptually, over previous
techniques which used B(a)P as a surrogate for POM, available comparative potency
factors are, in fact, based on few measurements, especially for the important motor
vehicle categories, and the uncertainty in these values should be recognized.
(14) The handling of some point sources as area sources for modeling purposes may
introduce some upward bias in the resulting exposure/risk estimates since HEM
distributes area source emissions by population and since area sources are emitted
closer to ground level.
(15) Caution is urged when applying the study results to other cities since the sources and
pollutants may not be representative of some other areas. For example, no study areas
include the heavy concentration of wood stoves that characterize some northern cities;
hence, the relative importance of wood smoke may be understated in this report.
(16) Any study such as this represents a "snapshot in time" of one's collective understanding
of the urban air toxics problem. In fact, the emission estimates and dose-response
relationships used in this study are subject to frequent revision as newer data become
available. Hence, care should be taken when interpreting any results from this study
or comparing these results to those from other studies where different data have been
used.
55
-------
CHAPTER VI
CONCLUSIONS
There are three primary conclusions that can be drawn from the exposure analyses
conducted to date. The first is that small point and area sources dominate point sources in
contributing to area wide cancer incidence in the five urban areas studied. The second
finding is, that a few pollutants dominate cancer incidence (POM, formaldehyde,
1,3-butadiene, hexavalent chromium, and benzene). In fact, the same sources and
pollutants are important in all the cities analyzed in this study. The third finding is that
there is a high level of uncertainty in the results because of the rapidly changing data bases
and because of the many assumptions made, which continually evolve. Cancer incidence
estimates can change by a large magnitude and the ranking in importance of various sources
and pollutants can change depending on the emission factors, unit risk factors, and
assumptions made in the analysis, and these assumptions will continue to change as new
information becomes available.
An unpublished modeling study conducted for the same five cities in 1985 produced
incidence estimates that were 800 percent higher than the current results. This dramatic
decrease in incidence occurred despite adding a number of compounds to the study that
were not analyzed previously and changing the approach for estimating
formaldehyde-related incidence to capture secondary formation. One major reason for this
decrease was the change in emission estimates for all important source categories reflecting
newer emission factors and activity levels. A second reason is that many of the unit risk
values listed in Table III.l have been changed in the past 5 years as new information has
become available. Another major change affecting analysis results has been the method for
estimating ROM-related incidence. The comparative potency approach used here produces
much lower POM incidence estimates for some important source categories than the
previous approach of using B(a)P as a surrogate for POM.
Of the compounds modeled in this study but not included in the previous study,
1,3-butadiene had the greatest effect on results. In fact, incidence associated with
1,3-butadiene exposure was estimated to be higher than that of any other compound except
POM. The uncertainty in these 1,3-butadiene related incidence results is pointed out by
revisions made within the last year by the Office of Mobile Sources to its motor vehicle
emission factors for 1,3-butadiene that lowered the previous estimates by 63 percent. Thus,
readers of this report should use caution in making judgments from results which may
change significantly in the future as more information becomes available on air toxic
emissions, atmospheric reactions, and risks associated with exposure.
One significant conclusion of the previous study which has not changed here is the
dominance of mobile, small point, and area sources in contributing to incidence. Large
56
-------
point sources are not in the list of top five source category-pollutant contributors to
incidence in any of the study areas. This suggests that either large point sources are
removed from where most people live or they are already well controlled, or both. An
important measure of the impact of point sources is maximum individual risk (MIR), a
distinctly different measure from the aggregate incidence measure that was the focus of this
phase of this study. Maximum individual risk and the potential for reducing this risk
through alternative control strategies is addressed in Volume II of this report.
57
-------
GLOSSARY OF TERMS
Acute exposure - One or a series of short-term exposures generally less than 24 hours.
Additive risk/incidence - Risk/incidence due to the interaction of two or more chemicals in
which the combined health effect is equal to the sum of the effect of each chemical
alone.
Aggregate incidence - In an urban air toxics context, this term generally refers to areawide,
additive incidence. This term sometimes refers to just areawide incidence or additive
incidence, so one should pay attention to the contextual use for the proper meaning.
Ambient (air) monitoring - The collection of ambient air samples and the analysis thereof
for air pollutant concentrations.
Annual incidence - Lifetime incidence adjusted to a yearly basis, typically by dividing.
lifetime incidence by 70.
Area source - Any source too small and/or numerous to consider individually as a point
source in an emissions inventory.
Areawide average individual risk - Average individual risk to everyone in an area (but not
necessarily the actual "'sk to anyone). May be computed by dividing lifetime
aggregate incidence by the population within the area.
Areawide incidence - Incidence over a broad area, such as a city or county, rather than at a
particular location, such as an individual grid cell. .
B(a)P - Benzo(a)pyrene. One of a group of compounds called polycyclic organic matter
(POM). B(a)P is sometimes used as a surrogate for all POM in computing emissions,
exposures and risks.
Background - A term used in dispersion modeling representing the contribution to ambient
concentrations from sources not specifically modeled in the analysis, including natural
and manmade sources.
BACT- Best Available Control Technology.
BG/ED - Block group/enumeration district. Designated by the Bureau of Census. A block
group is an area representing a combination of contiguous blocks having an average
58
-------
population of about 1100. An enumeration district is an area containing an average of
about 800 people and is designated when block groups are not defined.
Bioassay-A test in living organisms, e.g., a test for carcinogenicity in laboratory animals,
generally rats and mice, which includes a near-lifelong exposure to the agent under
test.
Box model - A simplified modeling technique that assumes uniform emissions within an
urban area and uniformly mixed concentrations within a specified mixing depth.
CAG_- Cancer assessment group. EPA group that prepares qualitative and quantitative
carcinogenic risk assessments.
Cancer- A cellular tumor, the natural course of which is fatal. Cancer cells, unlike benign
cells, exhibit properties of invasion and metastasis (malignancy). Cancers are divided
into two broad categories: carcinoma and sarcoma.
Csircinogenfcitv - The extent to which a substance is able to induce a cancer response.
Catalyst - A substance which promotes a chemical reaction. In the context of this report, a
device installed on the tailpipe of a motor vehicle to control exhaust emissions.
£D_M.- Climatological Dispersion Model. A Gaussian dispersion model whose particular
strength is its detailed area source treatment. Can also handle point sources, but not
in as detailed a manner as the Industrial Source Complex (ISC) model.
Centroi'd (population, source) - A single point whose coordinates represent the location of
•• a BG/ED, in the case of a population centroid, or the location of an emission point, in
the case of a complex source.
Chronic exposure - Long-term exposure usually lasting 6 months to a lifetime.
Co-control-In the context of air toxics, co-control represents the simultaneous control or
mitigation of air toxics and criteria pollutants via the same control measure. For
example, a motor vehicle catalyst would reduce VOC and CO emissions and also
reduce benzene and other gaseous toxics.
Comparative potency factor - A cancer unit risk factor for a complex substance or mixture
that is extrapolated from human risk data for a reference substance and the ratio of
short term bioassay responses of the complex substance to the reference substance.
EPA is developing comparative potency factors for various classes of POM.
59
-------
Coverage (pollutant, source, spatial) - The extent of inclusion of pollutants or sources in an
emission inventory or risk analysis or the amount of space or area represented in a
monitoring program or inventory or risk analysis.
Cr+ * - Hexavalent chromium, i.e., chromium in the +6 valence state..
Criteria pollutants - Pollutants defined pursuant to Section 108 of the Clean Air Act and for
which national ambient air quality standards are prescribed. Current criteria
pollutants include paniculate matter, SOx, NOx, ozone, CO and lead.
Culpability - The extent to which something (generally a pollutant or source) is responsible
for some effect (such as exposure or risk).
Decay - A term that represents pollutant removal by physical or chemical processes.
Deposition - The removal of paniculate matter and gases, at a land or water body surface,
by precipitation or dry removal mechanisms, including surface reactions and filtering.
Dispersion coefficients - Parameters used in Gaussian dispersion modeling to estimate
plume growth through dispersion along the horizontal and vertical axes. These are
computed as a function of downwind distance and atmospheric stability.
Dispersion modeling - A means of estimating ambient concentrations at locations
(receptors) downwind of a source, or an array of sources, based on emission rates,
release specifications and meteorological factors such as wind speed, wind direction,
atmospheric stability, mixing height and ambient temperature.
DNPH- Dinitrophenylhydrazine. Material used in special cartridges for monitoring of
formaldehyde'and other aldehydes.
Effective stack height - The height above ground level of the centerline of a plume. It is the
sum of the physical stack height, plume and stack-tip downwash (as applicable).
EPA - U.S. Environmental Protection Agency.
Excess cancer risk - An increased risk of cancer above the normal background.
Exposure - An event in which an organism comes into contact with a chemical or physical .
agent.
60
-------
Exposure assessment - Measurement or estimation of the magnitude, frequency, duration
and route of exposure to substances in the environment. The exposure assessment
also describes the nature of exposure and the size and nature of the exposed
populations.
FM VCP - Federal Motor Vehicle Control Program. EPA's program to control motor
vehicle emissions.
Fugitive emission/release - Emissions unconfined to a stack or duct, such as equipment
leaks from valves, flanges, etc., or open spills.
Gaussian model - A Gaussian dispersion model represents the distribution of
concentrations within a plume by assuming a normal distribution along the horizontal
and vertical axes. In the basic form, predicted concentrations are estimated as a
function of emission rate, horizontal and vertical dispersion coefficients, and vertical
and horizontal distance from the plume centerline.
Global buildup -The widespread accumulation of pollutants in the atmosphere over the
years. Global buildup is generally associated with more inert compounds such as
halogens (e.g., carbon tetrachloride).
. Grid - A network of rectilinear or polar grid cells superimposed over an area, generally for
modeling analyses. A rectilinear grid is defined by a series of perpendicular lines
defining rectangular or square grid cells whereas a polar grid is defined by a series of
concentric circles and straight lines radiating from the center of the circles.
Grid cell-The smallest area resolved within a modeling grid.
Grid spacing-The dimensions of the grid cells within a grid.
Grid square - A grid cell whose sides are equal.
*
HC- Hydrocarbon.
HDDV- Heavy duty diesel vehicles.
HEM- Human Exposure Model. EPA model used for exposure and risk analysis, which
defines polar receptor grids around each point source. Can also model area sources
by apportioning county level emissions to each BG/ED and running a simple box
model. HEM contains two component modules: SHED and SHEAR.
61
-------
Highway vehicle source - Car, truck or motorcycle. Also called road vehicles or motor
vehicles.
Hotspot - A particular receptor, grid cell or localized area wherein exposure or risk is high.
IACP - Integrated Air Cancer Program. EPA long-term research and development
program to develop methods and conduct field and lab tests to learn what causes
cancer in complex urban air mixtures and what sources are contributing to this cancer
burden.
1EMP- Integrated Environmental Management Project. A series of studies conducted in
Philadelphia, Baltimore, Santa Clara, Kanawha Valley and Denver to evaluate
multi-media contributions to various health risks, with emphasis on cancer.
Incidence - The frequency of occurrence of a certain event or conditions, such as the
number of new cases of a specific disease or tumor occurring during a certain period.
The incidence rate is the number of new cases during a certain period divided by the
population size (e.g., 10 cases per 100,000 exposed persons).
Individual risk - The increased risk for a person exposed to a specific concentration of a
toxicant. May be expressed as a lifetime individual risk or as an annual individual risk,
the latter usually computed as 1/70 of the lifetime risk.
LDDT - Light-duty diesel truck.
LDDV - Light-duty diesel vehicle.
LDGV - Light-duty gasoline vehicle.
Lifetime - Considered to be 70 years in EPA health risk assessments.
MEI - Maximum exposed individual.
Microenvironment - Localized environment in which one may be exposed to pollutant
concentrations considerably different than in ambient (outdoor) air, e.g., indoor
household air, occupational exposures, air within automobiles. EPA's Total Exposure
Assessment Methodology (TEAM) studies evaluate personal exposures as individuals
are exposed to air in different microenvironments during each day.
Mkroji- One millionth of a meter. A dimensional unit used to measure the diameters of
particles.
62
-------
MIR-Maximum individual risk, i.e., risk to the most exposed individual.
Mitigation-The reduction or control of emissions, exposures or risks due to air toxics.
Mobile source - Any motorized vehicle, such as cars, trucks, airplanes, trains. Sometimes
refers specifically to highway vehicle sources.
Modeling! - See dispersion modeling or receptor modeling. This term can also be used in
the context of emissions modeling, referring to the prediction of emissions from a
source or source category.
Monitoring - The collection and analysis of ambient air samples. Sometimes refers
specifically to just sampling and not to analysis. Can also refer to source (stack)
sampling.
Motor vehicle - On-road or off-road cars, trucks or motorcycles.
Mntngenicitv - The extent to which a chemical or physical agent interacts with DN A to
cause a permanent, transmissible change in the genetic material of a cell.
NAAOS - National Ambient Air Quality Standard. Set by EPA for criteria pollutants
under the Clean Air Act.
NAPAP - National Acid Precipitation Assessment Program..
HEQS- National Emissions Data System. EPA's centralized emission inventory of criteria
pollutant emissions.
NESHAP- National Emission Standards for Hazardous Air Pollutant. Standards set by
EPA for hazardous air pollutants under Section 112 of the Clean Air Act.
Noncancer risk - Risk of a health effect other than cancer.
Nontradi'tional sources - Sources not usually included in an emission inventory, such as
wastewater treatment plants, groundwater aeration facilities, hazardous waste
combustors, landfills, which are air emitters due to intermedia transfer from water or
solid waste. .
Normalized modeling - Modeling of unit weights (e.g., 1 Mg/yr) of emissions from each
source, rather than modeling of actual emissions, and displaying incremental receptor
concentrations or receptor coefficients. Thereafter, the resulting normalized receptor
63
-------
coefficients are adjusted by actual emission rates to simulate different emission
scenarios rather than re-running the model over and over with different emissions
totals. This process assumes linearity between emissions and modeled ambient air
concentrations, which does not always hold if stack and exhaust parameters change.
NSPS - New Source Performance Standards.
NSR - New Source Review. Permit process for evaluating emissions and need for controls
before construction and operation of a proposed facility. EPA as well as many States
and local agencies have NSR requirements for air toxics sources.
OAQPS - EPA's Office of Air Quality Planning and Standards.
QMS - EPA's Office of Mobile Sources.
OPPE - EPA's Office of Policy, Planning and Evaluation. Initiator of the Integrated
Environmental Management Project, a series of geographic, multi-media studies in
various cities (see IEMP).
P.AM - Peroxyacetyl Nitrate. A photochemical oxidant formed in urban atmospheres along
with ozone.
PhotochemicaHy formed pollutant - A secondarily formed pollutant due to atmospheric
photochemistry. Some exampels are formaldehyde and PAN.
PJC - Products of incomplete combustion. A term used somewhat loosely in various studies
referring generally to polycyclic organic matter.
PIPOUIC - Program Integration Project -- Queries Using Interactive Commands. A data
handling system developed by EPA as part of the IEMP to calculate exposures and
risks from air toxics emissions data. PIPQUIC is being used in various urban air
toxics studies.
PM - Participate matter.
PMin - PM less than 10 microns in diameter.
Point source - A source large enough to keep an individual record on in an emission
inventory, often emitting above a certain cutoff level or threshold.
Polar grid - See grid.
64
-------
EQM- Polycyclic organic matter. A broad class of compounds which generally includes all
organic structures having two or more fused aromatic rings (i.e., rings sharing a
common border). POM includes polynuclear aromatic hydrocarbons (PAH or PNA).
Primary pollutant - One emitted directly from an emission source, prior to any secondary
physical or chemical reaction.
Receptor - A particular point in space where a monitor is located or where an exposure or
risk is modeled.
Receptor grid - An array of receptors. Generally synonymous with network.
Receptor modeling - A technique for inferring source culpability at a receptor(s) by
analysis of the ambient sample composition. There are various receptor models
employing microscopic and chemical methods for analysis.
Release specifications - Release specifications are used as model inputs to characterize the
location, release height, and buoyant and momentum fluxes of each source. Required
terms include stack height, exit velocity, inner stack diameter, exhaust temperature
and the dimensions of nearby structures.
Risk - The probability of injury, disease, or death under specific circumstances., In
quantitative terms, risk is expressed in values ranging from zero (representing the
certainty that harm will not occur) to one (representing the certainty that harm will
occur).
assessment -The use of the factual base to define the health effects of exposure of
individuals or populations to hazardous materials and situations. May contain some
or all of the following four steps:
Hazard identification -The determination of whether a particular chemical is or
is not causally linked to particular health effects.
Dose-response assessment - The determination of the relation between the
magnitude of exposure and the probability of occurrence of the health effects in
question.
Exposure assessment - The determination of the extent of human exposure.
Risk characterization -The description of the nature and often the magnitude of
human risk, including attendant uncertainty.
65
-------
Risk management -The decision-making process that uses the results of risk assessment to
produce a decision about environmental action. Risk management includes
consideration of technical, social, economic and political information.
Road vehicle source - See highway vehicle source.
Sampling - See monitoring or ambient air monitoring.
Scoping study - Also known as screening study. An assessment or analysis using tentative or
preliminary data whose results are not accepted as absolute indicators of risk or
exposures, but rather, are taken as an indication of the relative importance of various
sources, pollutants and control measures. Most urban air toxics assessments
conducted to date have been considered to be scoping studies, useful for pointing out
where more detailed work is needed prior to regulation.
Screening study - See scoping study.
Secondary pollutant - Also, "secondarily formed pollutant." A pollutant formed in the
atmosphere as a result of chemical reaction and/or condensation, such as PAN. Some
pollutants (e.g., formaldehyde) are both primary and secondary pollutants.
SHEAR-Systems Application Human Exposure and Risk. A module within EPA's
Human Exposure Model designed to focus on multiple pollutant, multiple source
exposures, including area source analyses. SHEAR uses a Gaussian dispersion model
for point sources and a box model for area sources.
SJIEQ- Systems Applications Human Exposure and Dosage Model.
SJ£- Standard Industrial Classification. A series of codes or classifications to categorize
industry, published regularly by the Office of Management and Budget.
S_l£- State Implementation Plan. Required by States under the Clean Air Act to indicate
plan of action to meet National Ambient Air Quality Standards for criteria! pollutants.
Source apportionment - See receptor modeling.
Source grid - A grid defined to encompass all emission sources that one wants to model.
The source grid is sometimes defined bigger than a corresponding receptor grid so
that all local sources impacting on the receptor grid will be considered. More
typically, the source grid and receptor grid coincide in most studies.
66
-------
Spatial coverage - The area included or covered by a sampling network or a source/receptor
grid.
gpatial resolution - The extent to which emissions, monitoring or any other data are
subdivided or resolved in space, generally across a geographical area. For example,
emissions data may be spatially resolved to 1 km by 1 km squares within an urban area.
Species profile - A set of apportioning factors that allow one to subdivide VOC or PM
emission totals into individual chemicals or chemical classes. Generally, species
profiles are multiplicative in nature.
iily.- A parameter to describe the degree of turbulence in the atmosphere ranging
from unstable (vigorous mixing) to stable (suppressed mixing).
gubchronic exposure - Exposure to a substance spanning approximately 10 percent of the
lifetime of an organism.
Surrogate indicator - A variable whose spatial or temporal distribution is assumed to
behave in the same manner as some variable of interest. Surrogate indicators are
used for spatial and temporal apportionment of emissions data, especially for area
sources.
Temporal resolution - The extent to which some variable, typically emissions and
monitoring data, is subdivided or resolved in time. Data, for example, may be
resolved hourly or seasonally.
Transformation - The conversion, through chemical or physical processes, of one
compound or several compounds into other compounds as a result of aging and
irradiation in the atmosphere.
Transport -The movement of pollutants by wind flow. Transport is characterized for
modeling purposes by wind speed and wind direction.
Unit cancer risk factors - The incremental upper bound lifetime risk estimated to result
from a lifetime exposure to an agent if it is in the air at a concentration of I
microgram per cubic meter.
Urban soup - An expression referring to the multi-source, multi-pollutant urban air toxics
problem resulting from the complex interaction of many pollutants, sources and
atmospheric transformation.
67
-------
VMT - Vehicle miles traveled. Mobile source emission factors are typically expressed in
terms of grams per VMT.
VQC - Volatile organic compounds..
68
-------
-------
REFERENCES
Carey, 1987a: Penny M. Carey, personal communication, July 6,1987, U.S. Environmental
Protection Agency, Ann Arbor, MI.
Carey, 1987b: Penny M. Carey, personal communication, September 15,1987, U.S.
Environmental Protection Agency, Ann Arbor, MI.
Carey, 1987c: Penny M. Carey, "Air Toxics Emissions from Motor Vehicles,"
EPA-AA-TSS-PA-86-5, U.S. Environmental Protection Agency, Office of Mobile
Sources, Ann Arbor, MI, September 1987.
Carey, 1988: Penny M. Carey, U.S. Environmental Protection Agency, Ann Arbor, MI,
"1,3-Butadiene Emission Factors," memorandum to T.F. Lahre, May 24,1988.
DER, 1982: Department of Environmental Resources, "Pennsylvania Residential
Fuelwood Use Assessment, 1980-81," Office of Resources Management, Bureau of
Forestry, State of Pennsylvania, Harrisburg, Pennsylvania, 1982.
EIA, 1982a: Energy Information Administration, "Estimates of U.S. Wood Energy
Consumption from 1949 to 1981," DOE/ELA-0341, U.S. Department of Energy,
Washington, DC, 1982.
EIA, 1982b: Energy Information Administration, "Residential Energy Consumption
Survey: Consumption and Expenditures, April 1980-March 1981," DOE/EIA-0321/1,
U.S. Department of Energy, Washington, DC, 1982.
EIA, 1982c: Energy Information Administration, "Residential Energy Consumption
Survey: Housing Characteristics 1980," DOE/EIA-0314, U.S. Department of Energy,
Washington, DC, 1982.
Haemisegger et al., 1985: E. Haemisegger et al., 'The Air Toxics Problem in the United
States: An Assessment of Cancer Risks for Selected Pollutants," U.S. Environmental
Protection Agency, Washington, DC, May 1985.
Hall, 1985: Hall,T.A., Radian Corporation, Research Triangle Park, NC, "Phase II Work -
Air Toxics Controllability Study" memorandum to J. Weigold, January 17, 1985.
HID, 1984: Housing Industry Dynamics, "Sales and Marketing Information for Members of
the Wood Heating Alliance," Crofton, MD, 1984.
69
-------
Lahre, 1988a: Thomas Lahre, "Cancer Risks From Air Toxics In Urban Areas." Paper No.
88-127.6, presented at 81st Annual Meeting of APCA, Dallas, Texas, June 19-24, 1988.
Lahre, 1988b: Thomas Lahre, U.S. Environmental Protection Agency, Research Triangle
Park, NC, letter to Jim Wilson, E.H. Pechan & Associates, Inc., July 5,1988.
Lewtas, 1987: J. Lewtas, "POM Risk Assessment Using Comparative Potency Factors,"
memorandum to A. Manale and T. Lahre, U.S. Environmental Protection Agency,
RTP, NC, August 12,1987.
Marshall, 1981: Norman L. Marshall, The Dynamics of Residential Wood Energy Use in
New England, 1970-2000," Policy"Resource Center, Thayer School of Engineering,
Dartmouth College. Hanover, New Hampshire, October 1981.
NRBP, 1984: Northeast Regional Biomass Program, "Particulate Emissions from
Residential Wood Combustion - Summary Report," CONEG Policy Research Center,
Inc., Washington, DC, May 1984.
Radian, 1986: Radian Corporation, "Summary of Trace Emissions from and
Recommendations of Risk-Assessment Methodologies for Coal and Oil Combustion
Sources, External Review Draft," July 1986.
Radian, 1987a: Radian Corporation, "Phoenix and Tucson Air Toxics Emission Inventory
Study, Final Report," April 1987.
Radian, 1987b: Radian Corporation, "Update of Volatile Organic Compound (VOC)
Species Data Manual (Second Edition) Interim Document," prepared foir U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards and
Air and Energy Engineering Research Laboratory, April 1987.
Simons, "1984: Carl Simons, OMNI, "Caveats to Information Provided on Emission
Constraint Coding and Cost Data Forms and Growth Factors," memorandum report
to Leigh Hayes, Radian Corporation, December 5, 1984.
Skog and Watterson, 1983: Kenneth E. Skog and Irene Watterson, "Residential Fuelwood
in the United States: 1980-1981," Forest Products Laboratory, U.S. Forest Service,
Madison, WI, July 1983.
USDA, 1980: U.S. Department of Agriculture, "Prospectus: Firewood Manufacturing and
Marketing," U.S. Forest Service NA-FR-17, Madison, WI, February 1980.
70
-------
U.S. DOC, 1975: U.S. Department of Commerce, Bureau of the Census, "Construction
Reports: Characteristics of Nw Housing," C 25-74-13, Washington, DC, August 1975.
U.S. DOC, 1980: U.S. Department of Commerce, Bureau of the Census, "Construction
Reports: Characteristics of New Housing," C 25-79-13, Washington, DC, August 1980.
U.S. DOC, 1982a: U.S. Department of Commerce, Bureau of the Census. "1980 Decennial
Census, Detailed Housing Characteristics," Washington, DC 1982.
U.S. DOC, 1982b: U.S. Department of Commerce, Bureau of the Census, "Construction
Reports: Characteristics of New Housing," C 25-81-13, Washington, DC, August 1982.
U.S. DOC, 1983: U.S. Department of Commerce, Bureau of the Census, "1980 Annual
Housing Survey: Energy Related Housing Characteristics Park F," H-150-80,
Washington, DC: 1983.
U.S. EPA, 1985a: U.S. Environmental Protection Agency, "A Strategy to Reduce Risks to
Public Health from Air Toxics," Washington, DC, June 1985.
U.S. EPA, 1985b: U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards, "Chromium Screening Test Report, Coal Fired Boiler, Adolph Coors,
Golden, Colorado," EMB Report 85-CHM-6, April 1985.
U.S. EPA, 1985c: U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards, "Chromium Screening Test Report, Electric Arc and AOD Furnaces,
Carpenter Technology Inc., Reading, Pennsylvania," EMB Report 85-CHM-9, April
1985.
U.S. EPA, 1985d: U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards, "Chromium Screening Test Report, Harbison Walker Refractories,
Baltimore, Maryland," EMB Report 85-CHM-12, June 1985.
U.S. EPA, 1985e: U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards, "NESHAP Screening Study Chromium Emission Test Report,
Burlington Industries, Raleigh, North Carolina. "EMB Report 85-CHM-8, August,
1985.
U.S. EPA, 1985f: U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards, "Survey of Methylene Chloride Emission Sources," EPA-450/3-85-OI5,
June 1985.
71
-------
U.S. EPA, 1985g: U.S. Environmental Protection Agency, "Compilation of Air Pollutant
Emission Factors, Volume I: Stationary Point and Area Sources," AP-42, Fourth
Edition, Office of Air and Radiation, Research Triangle Park, NC, September 1985.
U.S. EPA, 1986: U.S. Environmental Protection Agency, "User's Manual for the Human
Exposure Model (HEM)," EPA-450/5-86-001, Office of Air Quality Planning and
Standards, Research Triangle Park, NC, June 1986.
U.S. EPA, 1987a: U.S. Environmental Protection Agency, Industrial Standards Branch,
"Chromium Electroplating NESHAP - Background Information Document Chapters
3 through 5," February 1987.
U.S. EPA, 1987b: U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards, "Regulatory Impact Analysis for Chromium Emissions from Comfort
Cooling Towers, Draft," February 1987.
U.S. EPA, 1987c: U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards, "Emission Factor Documentation for AP-42: Section -1.10, Residential
Wood Stoves, Draft," prepared by Engineering-Science, Inc., May 1987.
U.S. EPA, 1987d: U.S. Environmental Protection Agency, Strategies and Air Standards
Division, printout of hexavalent chromium emissions from chromeplating operations
by state and county, July 13,1987.
U.S. EPA, 1988: U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards, "Background Information Document for Chromium Emissions from
Industrial Process Cooling Towers," February 1986 report updated to 1988 by
personal communication with Thomas F. Lahre.
Versar, 1984: Versar Inc., "Hazardous Air Pollutants, A Preliminary Exposure and Risk
Appraisal for 35 U.S. Counties," Springfield, VA, September 1984.
52 FR 45072: "State Implementation Plans; Approval of Post-1987 Ozone and Carbon
Monoxide Plan Revisions for Areas Not Attaining the National Air Quality
Standards; Notice," Federal Register. Vol. 52, No. 226, part II, Tuesday, November
24, 1987, p. 45072.
72
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing/
REPORT NO.
EPA-450/2-89-012a
2.
3. RECIPIENT'S ACCESSION NO.
TITLE AND SUBTITLE
Analysis of Air Toxics Emissions, Exposures, Cancer
Risks and Controllability in Five Urban Areas
Volume I - Base Year Analysis and Results
5. REPORT DATE
July, 1989
6. PERFORMING ORGANIZATION CODE
OAQPS
AUTHOR(S)
Jim Wilson, Deborah Istvan, and Erica Laich of E. H.
Pechan; Tom Lahre, EPA, OAQPS
8. PERFORMING ORGANIZATION REPORT NO
PERFORMING ORGANIZATION NAME ANO ADDRESS
E. H. Pechan and Associates, Inc.
5537 Hempstead Way
Springfield, VA 22151
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
2. SPONSORING AGENCY NAME ANO ADDRESS
Noncriteria Pollutant Programs Branch
Air Quality Management Division
Office of Air Quality Planning and Standards
Research Triangle Park, N.C. 27711
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
5. SUPPLEMENTARY NOTES
6. ABSTRACT
This report (Volume I) is the first phase of a study to (1)
define the multiple source, multiple pollutant nature of the ;
urban air toxics problem (also known as "urban soup"), and (2) to
discern what control measures (or combinations of measures) can ;
best be employed to mitigate the urban air toxics problem. This
report documents the base year analysis, involving dispersion
modeling of emissions data for 25 carcinogenic air toxics in five
U. S. urban areas and a subsequent exposure/risk assessment to
estimate aggregate cancer incidence.
Aggregate (multi-source, multi-pollutant) cancer incidence
(or population risk) across the 5 cities in this study averaged
about 6 excess cases per million persons, ranging from about 2 to
10 in individual cities. The most important pollutants
contributing to aggregate incidence are polycyclic organic matter
(POM), 1,3-butadiene, formaldehyde and hexavalent chromium. The
most important sources are road vehicles, comfort and industrial
cooling towers, chrome platers, solvent use and fuel combustion,
including woodstoves. . . ..
7.
KEY WORDS ANO DOCUMENT ANALYSIS
DESCRIPTORS
b.lOENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Air toxics
Cancer incidence, air toxics
Controllability, air toxics
Hazardous air pollutants
Mitigation, air toxics
Risk assessment, air toxics
Urban air toxics (urban soup)
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS I This Report}
21. NO. OF PAGES
90
20. SECURITY CLASS /This page I
22. PRICE
EPA Form 2220-1 (R«». 4—77) PREVIOUS EDITION is OBSOLETE
-------
-------
-------
------- |