EPA-450/4-88-013
November 1988
MULTIPLE AIR TOXICS EXPOSURE STUDY
WORKING PAPER NO. 3
URBAN AIR TOXICS EXPOSURE MODEL:
DEVELOPMENT AND APPLICATION
Prepared by
South Coast Air Quality Management District
Systems Applications, Incorporated
Reprinted with permission of the
South Coast Air Quality Management District
U.S. Environmental Protection Agency
Office Of Air And Radiation
Office of Air Quality Planning And Standards
Research Triangle Park, North Carolina 27711
-------
-------
INTRODUCTION/PURPOSE
The Environmental Protection Agency's Administrator, Lee Thomas, released
"A Strategy to Reduce Risks to Public Health from Air Toxics" on June 11, 1985.
This has later become known as EPA's National Air Toxics Strategy." One impor-
tant leg of this strategy focuses on the multi-pollutant/multi-source impacts
which have been characterized as urban air toxics. EPA has been working to
implement the strategy in several ways.
Efforts to date, in the urban air toxics area have been to:
°Assess the problem from a national perspective to develop better
evidence and documentation of it's magnitude and character.
"Promote State and local urban air assessment activities by State and
local agencies.
"Develop guidance and analytical tools needed for States and local
agencies to assess the problems.
"Encourage State and local agencies to evaluate options for mitigation
of problem areas.
"Encourage State and local' agencies to mitigate these situations where
warranted.
Urban air .toxics assessment efforts have begun to provide returns in
several areas^ especially where State and local agencies were already interested
and involved in examining the problem. One such area where advanced concern
and activities have occured is the South Coast Air Quality Management District
of California (Los Angeles area). The South Coast District, with financial
assistance from EPA and substantial funding of their own has carried out a
study addressing the "Magnitude of Ambient Air Toxics Impact form Existing
Sources in the South Coast Air Basin" (Also, known as "MATES"). The methods
employed by Los Angeles, and the general purposes of the study are very much in
line with EPA's urban air toxic program objectives, though much more extensive
and elaborate than might be needed in many smaller areas. Thus, EPA is making
this report available to other State and local agencies. The rest of this
document is a reproduction of Working Paper Number 3 from the South Coast study.
This report is reproduced and distributed with the permission of the South
Coast Air Quality Management District to be used as a basis for further study
by various interested State and local agencies who may be contemplating work on
their own.
For further information contact:
Pollutant Characterization Section (MD-15)
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
ii
-------
DISCLAIMER
This report is the product of the South Coast (California) Air Quality
Management District and is reproduced as received. Though the report has
been generally reviewed by the U. S. Environmental Protection Agency, and
approved for publication, approval does not signify that the contents
necessarily reflect the views and policies of the Agency, neither does
mention of trade names or coramerical products constitute endorsement or
recommendation for use.
EPA-450/4-88-013
iii
-------
SOUTH COAST AIR QUALITY MANAGEMENT DISTRICT
GOVERNING BOARD
NORTON YOUNGLOVE, Chairman
Supervisor
County of Riverside
MARVIN BRAUDE, Vice-Chairman
Councilman
City of Los Angeles
LARRY L. BERG
Speaker of the Assembly
Appointee
FAYE MYERS DASTRUP
Councilwoman
Cities Representative
County of San Bernardino
EDMUND D. EDELMAN
Supervisor .
County of Los Angeles
ROBERT L. HAMMOCK
Supervisor
County of San Bernardino
PATRICIA HERRON
Councilwoman, City of Hemet
Cities Representative
County of Riverside
THOMAS F. HEINSHEIMER
Councilman, City of Rolling Hills
Cities Representative
County of Los Angeles
PETER F. SCHABARUM
Supervisor
County of Los Angeles
SABRINA SCHILLER
Senate Rules Committee
Appointee
VACANT
Governor's Appointee
GADDI VASQUEZ
Supervisor
County of Orange
HARRIETT M. WIEDER
Supervisor
County of Orange
HENRY W. WEDAA
Councilman, City of Yorba Linda
Cities Representative
County of Orange
EXECUTIVE OFFICER
JAMES M. LENTS
iv
-------
-------
Jo Anne Aplet
Acting Deputy Executive Officer/Planning and Analysis
Prepared by Planning Division
John E. Grisinger, Acting Director of Planning
Authors
Ditas Shikiya
Chung Liu
Emily Nelson
Rich Rapoport
Typist: Ann Kypreos
-------
-------
South Coast
AIR QUALITY MANAGEMENT DISTRICT
9150 FLAIR DRIVE. EL MONTE, CA 91731
(818)572-6200
July 25, 1988
Edward J. Lillis, Chief
Noncriteria Pollutant Programs Branch
Air Quality Management Division
Office of Air Quality Planning and Standards
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dear Mr. Lillis:
The South Coast Air Quality Management District is pleased
to endorse your proposed publication of the District's
report entitled "The Magnitude of Ambient Air Toxics Impacts
From Existing Sources in the South Coast Air Basin" under
EPA cover with a title acknowledging our authorship. Please
send us a copy of this publication when completed.
The District continues to improve the regional exposure and
risk assessment model used in the MATES study. Although it
has proven to be a very useful tool in prioritizing air
toxic species, the model in its current fbrm assumes the
population stays in the home all the time and that the only
exposure route is inhalation. This model will be improved
to account for (1) mobility, (2) microenvironments, and (3)
multi-media exposure. Results of this effort will be
available by spring of next year.
If we can be of further assistance to you or any other state
and local agencies, please feel free to call Ditas Shikiya
at (818) 572-2119. We look forward to continued
communication regarding this common interest.
Sincerely,
Carolyn L. Green
Deputy Executive Officer
Office of Planning & Analysis
CLG:DS
vi
-------
-------
TABLE OF CONTENTS
Page
LIST OF TABLES v1
LIST OF FIGURES vi i
ACRONYMS USED IN WORKING PAPER NO.3 viii
EXECUTIVE SUMMARY ix
CHAPTERS
I. INTRODUCTION 1-1
I.I OBJECTIVES. 1-1
1.2 BACKGROUND, '. .1-2.
II. AMBIENT CARCINOGEN CHARACTERIZATION METHOD II-l
III. MODEL DEVELOPMENT III-l
III.l POPULATION DATA BASE III-l
III.2 METEOROLOGICAL DATA BASE III-3
III.3 EXPOSURE ASSESSMENT ........111-3
. II1.4 SUMMARY OF ENHANCEMENTS AND OUTPUT. III-4
IV. EMISSIONS DATA IV-1
IV.1 EMISSION INVENTORY METHODOLOGY IV-1
IV. 1.1 Stationary Source Emissions IV-1
IV.1.2 Mobile Source Emissions IV-1
IV.2 SUMMARY OF EMISSIONS IV-5
IV.3 SPATIAL ALLOCATION IV-7
V. AMBIENT DATA.... ....V-l
V.I CONTINUOUS MONITORING NETWORK V-l
V.I.I Annual Average Concentrations V-l
V.I.2 Population-Weighted Annual Average V-9
Concentrations
V.2 LITERATURE SURVEY OF AMBIENT DATA V-9
vii
-------
Paoe
VI. MODEL APPLICATION AND RESULTS VI-1
VI.1 RISK CHARACTERIZATION VI-1
VI.2 SOURCE APPORTIONMENT VI-5
VI.3 COMPARISON OF MEASURED AND MODEL-
PREDICTED AMBIENT CONCENTRATIONS
AND RISKS VI-7
VII. ASSUMPTIONS AND UNCERTAINTIES VII-1
VIII. CONCLUSIONS AND RECOMMENDATIONS VIII-1
REFERENCES R-1
APPENDICES
APPENDIX A SPATIAL DISTRIBUTIONS OF POINT SOURCE
AIR TOXICS EMISSION. A-l
APPENDIX B SPATIAL DISTRIBUTION OF MODEL-PREDICTED
AMBIENT CONCENTRATIONS, .B-l
APPENDIX C SPATIAL DISTRIBUTION OF INDIVIDUAL
CANCER RISKS AND NUMBER OF EXCESS
CANCER IN THE SOUTH COAST AIR BASIN ....C-l
viii
-------
LIST OF TABLES
Table Page
IV-1 TOXIC AIR POLLUTANTS STUDIED IN THE SOUTH
COAST AIR BASIN IV-2
IV-2 STACK PARAMETERS IV-3
IV-3 EMISSION FACTORS FOR MOTOR VEHICLES. IV-4
IV-4 TOXIC EMISSIONS OF TWENTY SPECIES IN THE
SOUTH COAST AIR BASIN IN 1984 IV-6
V-l 1985 ANNUAL AVERAGE AMBIENT AIR CONCENTRATIONS
OF VARIOUS TOXIC ORGANIC GASES IN THE SOUTH
COAST AIR BASIN V-4
V-2 1985 ANNUAL AVERAGE AMBIENT AIR CONCENTRATIONS OF
VARIOUS TOXIC METALS IN THE SOUTH COAST AIR BASIN V-7
V-3 POPULATION-WEIGHTED ANNUAL AVERAGE AMBIENT
CONCENTRATIONS AND INDIVIDUAL CANCER RISKS
IN THE SOUTH COAST AIR BASIN ...V-10
V-4 AMBIENT FORMALDEHYDE CONCENTRATIONS IN THE SOUTH
COAST AIR BASIN AS MEASURED BY VARIOUS INVESTIGATORS...V-12
VI-1 SOURCE APPORTIONMENT OF LIFETIME- (70 YEAR) CANCER :
CASES FOR BENZENE AND HEXAVALENT CHROMIUM IN THE
SOUTH COAST AIR BASIN. VI-7
Vl-2 COMPARISON OF MEASURED AND MODEL-PREDICTED
TOXIC AIR POLLUTANTS VI-9
VI-3 ESTIMATION OF LIFETIME (70 YEAR) UPPER-BOUND
CANCER CASES ASSOCIATED WITH AMBIENT CARCINOGENS
IN THE SOUTH COAST AIR BASIN V..VI-11
-------
LIST OF FIGURES
Figure
II-l AMBIENT CARCINOGEN RISK ASSESSMENT APPROACH II-2
IV-2 SPATIAL DISTRIBUTION OF POPULATION DENSITY IV-8
IV-3 DISTRIBUTION OF MOBILE SOURCE EMISSIONS OF
REACTIVE ORGANIC GASES IV-9
V-l AMBIENT AIR TOXIC MONITORING SITES IN THE
SOUTH COAST AIR BASIN V-2
VI-1 MODEL PREDICTED ANNUAL AVERAGE BENZENE
CONCENTRATIONS IN THE SOUTH COAST AIR BASIN VI-2
VI-2 MODEL PREDICTED ANNUAL UPPER-BOUND INDIVIDUAL RISK
ASSOCIATED WITH LIFETIME EXPOSURE TO AMBIENT
BENZENE IN THE SOUTH COAST AIR BASIN VI-3
VI-3 MODEL PREDICTED UPPER-BOUND EXCESS CANCER
CASES ASSOCIATED WITH LIFETIME EXPOSURE TO
BENZENE IN. THE SOUTH COAST AIR BASIN , VI -4
VI-4 UPPER-BOUND MODEL PREDICTED POPULATION
FREQUENCY DISTRIBUTION RISK PROFILE OF
LIFETIME (70 YEAR) EXPOSURE TO NINE .AMBIENT
CARCINOGENS IN THE SOUTH COAST AIR BASIN... VI-6
VI-5 MATRIX OF CANCER RISK FROM AMBIENT CARCINOGENS VI-8
x:
-------
ACRONYMS USED IN WORKING PAPER NO. 3
AB ASSEMBLY BILL
AEIS AUTOMATED EQUIPMENT INFORMATION SYSTEM
AQMP AIR QUALITY MANAGEMENT PLAN
ARB AIR RESOURCES BOARD
CAG CARCINOGEN ASSESSMENT GROUP
DIME DUAL INDEPENDENT MAP ENCODING
DOHS DEPARTMENT OF HEALTH SERVICES
EIS EMISSION INVENTORY SYSTEM
EPA ENVIRONMENTAL PROTECTION AGENCY
GBF GEOGRAPHIC BASIC FILE
HEM . HUMAN EXPOSURE MODEL
MATES MULTIPLE AIR TOXICS EXPOSURE STUDY
OAQPS OFFICE OF AIR QUALITY PLANNING AND STANDARDS .
POTW. PUBLICLY OWNED TREATMENT WORKS
ROG REACTIVE ORGANIC GASES
SCAG SOUTHERN CALIFORNIA ASSOCIATION OF GOVERNMENTS
SCREAM SOUTH COAST RISK AND EXPOSURE ASSESSMENT MODEL
SHEAR SYSTEMS APPLICATIONS HUMAN EXPOSURE AND RISK MODEL
SMSA STANDARD METROPOLITAN STATISTICAL. AREA
UTM UNIVERSAL TRANSVERSE MERCATOR
xi.
-------
-------
EXECUTIVE SUMMARY
Control of toxic air pollutants (or air toxics) is currently achieved by
continued reduction of industrial and mobile source emissions through
control measures already in place or proposed for implementation. Such
measures include regulations for criteria pollutants which reduce
emissions for a broad range of particulate and gaseous compounds that
are toxic. However, some of these measures may at the same time result
in increased emissions of toxic air pollutants or precursors.
An analysis of potential positive or adverse effects on ambient air
toxics concentrations of AQMP control strategies for criteria air
pollutants is a necessary part of the AQMP development effort. Control
strategies specifically for air toxics will be developed through the
California Air Resources Board (ARB) afr toxics contaminant program (AB
1807, Health and Safety Code Section 39650, et.seo.) and by the District
as more comprehensive toxic emissions data become available.
In order to determine if existing or proposed control approaches for
criteria pollutants are adequate to protect public health from exposure
to air toxics, an understanding of the air toxics problem in the South
Coast Air Basin (Basin) is needed. However, standard techniques
currently available for assessing such risks cannot be specifically
applied to an urban area such as this.Basin. Through an Environmental
Protection Agency (EPA) funding for a Multiple Air Toxics Exposure Study
(MATES) in the Basin, a method to identify the magnitude, of the air
toxics impacts from individual chemical species and emissions source
categories was developed and applied to this Basin. This method
integrates ambient concentration, population distribution, and health
risk data for individual chemical species into regional estimates of
inhalation exposure, risk, and number of excess cancer cases.
The model used in this method was developed from the Human Exposure
Model (HEM) and includes features for: (1) defining the receptor
network, (2) conducting dispersion calculation, and (3) determining
population exposure and areawide risks. This model was developed with
the concept that all necessary input data on emissions, meteorology, and
population are readily available, such that it can be applied to other
urbanized areas in the United States.
The estimation of population exposure to one or more air toxics is
conducted by first using dispersion modeling to calculate the long-term
concentrations at centroids of census areas and then multiplying the
calculated concentration with the population that each centroid
represents. The areawide risks, in terms of incremental cancer cases,
are then calculated by multiplying the population exposure with the
chemical-specific unit risk factor. A linear response relationship is
assumed and the exposure/risks associated with multiple sources and
species of air toxics are considered additive.
xii
-------
The enhanced HEM model (called the South Coast Risk and Exposure
Assessment Model, SCREAM) can be used to apportion the number of excess
cancer cases by source category and by pollutant, and to identify high-
risk chemical species and source categories. It can also be used to
identify high-risk locations and to estimate control measure
effectiveness in reducing exposure, cancer risk, and number of cancer
cases.
Of the 20 air toxics studied, benzene and hexavalent chromium appear to
have the greatest impact on the Basin's population. Almost the entire
population is exposed to ambient benzene and hexavalent chromium
concentrations corresponding to an upper-bound risk of 1 x 10"4 or
higher.
The assumptions used in developing the model and those associated with
the quantification of cancer risk inject a considerable degree of
uncertainty into the analysis. Some assumptions lead to a potential
underestimation of the risk to the population, while others result in an
upper-bound estimate of the cancer risk. An understanding of these
assumptions is necessary to evaluate the uncertainty associated with the
estimated risks.
Results of this study indicate the relative importance of the individual
carcinogenic species and the relative contribution of individual source
categories to the total risk from a specific pollutant. This
information can be used in developing and prioritizing an air toxics
control program and in evaluating potential air toxics impacts from
existing or proposed control approaches and sources.
xiii
-------
CHAPTER I
INTRODUCTION
I.I OBJECTIVES
Current regulations for criteria air pollutants reduce emissions for a
broad range of particulate and gaseous compounds that are toxic. These
regulations have been adopted based on meeting air quality standards
rather than assessments of health risk due to toxicity. An example is
.the use of alternative solvents and surface coatings to reduce emissions
of reactive organic gases which are precursors to ozone formation.
Changing paint formulations or using alternative coatings and solvents
which are photochemically less reactive could produce other
environmental impacts including adverse toxic effects. In view of the
current concern for toxic health effects, an analysis of potential
adverse consequences related to air toxic emissions will be made prior
to a recommendation to implement Air Quality Management Plan (AQMP)
control strategies.
This report presents the results of a study quantifying the magnitude of
population exposure from existing point, area, and mobile source
emissions of 20 selected air toxics. This understanding of the existing
air toxics problem in the South Coast Air Basin (Basin) will be useful
in evaluating AQMP control, strategies for criteria pollutants for their
potential to reduce or increase emissions of toxic air pollutants. An
analys-is of potential positive or adverse changes to ambient air toxics
concentrations as a result of AQMP strategies will prevent the
inadvertent replacement of the health threat from criteria pollutants
with the health threat from ambient air toxics.
The method described in this report would establish a scheme to rate
toxic air pollutants according to a number of selected factors which
will be determined based on this study and on existing or proposed
control measures. Control strategies specifically for air toxics will
be developed as needed by the District (see state's program below) and
as more comprehensive, Basin-specific toxics emissions data bases become
available.
Under the state's toxic air contaminant program (Assembly Bill 1807,
Health and Safety Code Section 39650 el sea), the California Air
Resources Board (ARB), with the participation of the local air pollution
control districts, evaluates and develops any needed control measures
for air toxics. Measures for the control of benzene emissions have been
developed and control measures for chromium emissions are currently
being developed. The information from this report will also be useful
to the ARB and the California Department of Health Services (DOHS) in
their identification and assessment of potential health risks of air
toxics as required by AB 1807.
-------
1.2 BACKGROUND
The Environmental Protection Agency (EPA), using standard risk
assessment techniques, prepared the most comprehensive report to date to
characterize the risk associated with exposure to carcinogens (EPA,
1985). This study, known as the "Six-Month Study," was national in
scope and used existing data extrapolated to a much larger geographical
area to estimate exposure and risk. The limitation of the available
data precludes performing specific risk assessments for most urban
areas, for many substances, and for many large sources of air toxics
risk. EPA's objectives were not intended for regulatory support but
only for guidance "in policy decisions and further studies.
To satisfy the need of a local air regulatory agency, both in terms of
air quality planning and for permit application review, a method was
developed to characterize existing ambient concentrations of
carcinogenic air pollutants and to summarize the present understanding
of the magnitude of the air toxics problem in the Basin.
This study focuses on cancer risks only because the analysis techniques
for carcinogenic effects are sufficiently developed to allow a rational
and defensible basis for regulation. For example, use of a non-
threshold assumption in estimating cancer risk has broad scientific
support. It is also generally accepted that a substance that causes
cancer in test animals is likely to be carcinogenic to humans as well;
this has not been established for other health effects. There is also a
well-established mathematical model for estimating risk at low doses;
this is not the case for other effects. More work is needed to
establish models arid methods- for assessing quantitatively the risk of
other health effects.
The individual lifetime cancer risks reported in this study could be
viewed in the context of other cancer risks. The overall probability of
contracting cancer is approximately 250,000 cases per million population
over a lifetime (ARB and DOHS, 1986). Doll and Peto (1981) have
estimated that about 65 percent of annual cancer deaths appear to be
related to smoking or diet, each of which are predominantly affected by
personal choice. The number of cancer cases for all exposures to
environmental pollution is reported to be about two percent of total
cancer incidences and is generally due to involuntary exposure to air
toxics emissions. This translates into approximately 50,000 excess
cancer cases in the Basin over 70 years or about 700 cases annually
given the current population. The cancer cases calculated in this
report are only a small portion of the cancer risks from all
environmental pollutants but are those over which the District or ARB
has regulatory authority for protecting public health.
1-2
-------
CHAPTER II
AMBIENT CARCINOGEN CHARACTERIZATION METHOD
Estimating the cancer risks from exposure to an environmental pollutant
requires the following information:
o An estimate of the carcinogenic potency of the pollutant;
o An estimate of the ambient concentrations that individuals or
groups of individuals may inhale; and
o An estimate of the number of individuals that are exposed to those
concentrations.
The method discussed in this report uses the above information and is
consistent with the EPA-proposed guidelines for air toxics assessment
(Federal Register, 1986). It utilizes an urban air toxics exposure and
risk model developed specifically for the Basin to determine the risks
associated with exposure to ambient toxics emitted from both stationary
and mobile sources in the Basin.
The District's method integrates ambient concentrations, population
distribution, and carcinogenic potency data for individual species (in
the form of unit risk factors) into regional estimates of exposure,
risk, and cancer cases to provide the following:
o Estimates of regional impacts from existing, sources of carcinogens
quantified in terms of population exposure, individual cancer
risk, and number of excess cancer cases;
o Apportionment of the number of excess cancer cases by source
category, including identification of high-risk species and source
categories;
o Identification of high-risk locations due to specific sources or
groupings of multiple sources;
o Estimates of the effectiveness of control measures in reducing
exposure, risk, and number of excess cancer cases;
Figure II-l shows the flow of information in the characterization
method. As shown, ambient concentrations of carcinogens were estimated
using: (1) ambient measurements, (2) modeling, and (3) a literature
survey. Annual average ambient concentrations determined by
measurements in the Basin or predicted by regional modeling of emissions
sources are considered as primary sources of information. A literature
survey of ambient data available for areas in or outside this Basin, but
which are not representing an annual average, are used for analysis as
secondary sources of information.
II-l
-------
cjce
cor:
=1
LU a
ceu.
_L,
25IS
eouj2-
es=J_i
UJ UJ UJ
ca:
CuJ
Sg
COuJ
UJ
Q.
Q.
I
UJ
C/5
UJ
t-i ^
i t-t
OL
LU
0£ Z
2 LU
Cfl C3
1-1 O
U- Z
II-2
-------
Ambient measurement and modeling are complementary in this process.
Ambient measurements of carcinogens quantify the total impact of
emissions from all sources. If the deviation between monitoring results
and model-predicted concentrations is larger than those that can be
explained by the uncertainties of modeling techniques, the accuracy of
the emissions data might be in question. Modification of the emissions
data may be needed as shown by the feed-back loop in Figure II-l.
An advantage of modeling techniques over ambient measurements is that
additional information such as apportioning the number of cancer cases
in the region by source category or by individual chemical species are
provided. This apportionment identifies high-risk chemical species and
source categories and ultimately can be used to estimate the
effectiveness of control measures in reducing cancer impacts.
The cancer potencies or unit risk factors of most of the pollutants
covered in this report were those developed by EPA's Carcinogen
Assessment Group (CAG) and by the DOHS pursuant to AB 1807. The unit
risk factor represents the probability of cancer cases, not deaths, and
is defined by CAG as the chance of contracting cancer from a 70-year
lifetime exposure to a concentration of 1 ug/nr of a given substance.
The two measures of risks calculated in this report are the lifetime
individual risk and the estimated number of excess cancer cases.
Lifetime individual risk is a measure of the probability of an
individual contracting cancer as a result of exposure to an ambient
concentration of an air pollutant or several air pollutants over a 70-
year period. The number of excess cancer cases is the estimate for the
entire affected population and is calculated by multiplying the
individual cancer risk in a receptor area by the number of people
exposed in that receptor area.
II-3
-------
-------
CHAPTER III
MODEL DEVELOPMENT
Estimating population exposure to emitted air toxics has been
investigated previously through dispersion modeling and concentration-
population integration. Anderson, et al. (1980) developed and applied
the Human Exposure Model (HEM) under contract to EPA's Office of Air
Quality Planning and Standards (OAQPS). The analyses were national in
scope and were conducted under the OAQPS mandate to review chemicals in
use for potential regulation for National Emissions Standards for
Hazardous Air Pollutants (NESHAPs) development under Section 112 of the
Clean Air Act. The analyses were intended as a preliminary screening
for scoping and prioritizing regulatory attention among the 35 chemicals
studied.
HEM numerically combines the results of modeled concentration
distributions and population data files to quantify population exposure
to air toxics. However, the additive impacts of multiple chemicals
emitted by a source and the additive impacts of multiple nearby sources
are not calculated. Anderson and Lunberg (1983) enhanced the HEM
package by adding the capability to produce combined exposure and risk
estimates from multiple sources of all studied toxic species. This
enhanced model is called Systems Applications Human Exposure and Risk
(SHEAR).
Several characteristics of HEM and SHEAR are not adequate for
.applications in the Basin either for air quality planning or permit
processing purposes. For area sources, both models use calculated city-
wide concentrations and city-wide average population density. For point
sources, population density data from the United States Census Bureau
files are incorporated only to the Block Group and Enumeration District
level. The meteorological data base for HEM'and SHEAR dispersion models
includes STAR tabulation for only three sites in the Basin, even though
detailed localized meteorological data may be available for a large
number of sites. Emission inventory for air toxics in the Basin
contains more detailed information on specific point sources and general
prototype sources than can be handled by HEM and SHEAR.
For these reasons HEM was further enhanced and tailored to the Basin
for air quality planning purposes. This effort requires the use of the
most detailed population data available and a more refined treatment of
temporal and spatial emission patterns. This second enhancement of HEM
(Liu, et al, 1986) is called the South Coast Risk and Exposure
Assessment Model (SCREAM).
III.l POPULATION DATA BASE
The location and population of all census 'areas are required input for
SCREAM. The following four levels of population units were selected:
III-l
-------
o Place (Incorporated or Census Designated)
o Census Tract and Minor Civic Divisions
o Block Group (B6) and Enumeration District (ED)
o Street Block
The 1980 United States census population of all these levels of
disaggregation are available at the Bureau of Census' Summary Tape File
1 (USBC, 1981). In 1980, there were 245 places, 3,198 census tracts,
and 93,630 street blocks in the Basin.
The locations of centroids are available at the BG/ED level. In order
to obtain the same information for street blocks, the United States
Bureau of Census' Geographic Basic File/Dual Independent Map Encoding
GBF/DIME files (USBC, 1980) were used. A geographic base file is a map
in a form that is computer readable. Dual Independent Map Encoding is a
method of representing map features numerically. GBF/DIME files are
organized by Standard Metropolitan Statistical Areas (SMSAs). .The
following three SMSAs cover the entire Basin and the mountain and desert
areas to the east and north of the Basin:
Los Angeles - Long Beach
Anaheim - Santa Ana - Garden Grove
Riverside - San Bernardino - Ontario
There are'a total of 357,714 records in the GBF/DIME files for the three
SMSAs. Each record identifies a segment of a feature on the map by its
node points (a "from" node and a "to" node), address range (for both the
odd and even sides), segment type (street, political boundary, streams,
etc.), and other data (census tract and left and right block numbers) on
each sides.
The following steps were followed to determine the centroid for each
street block based on the information contained in the GBF/DIME records:
o Specific data items such as census tract code, block numbers (both
sides), the ID number, latitude and longitude for both "from" and
"to" node were extracted from the original file for each SMSA. The
geodetic coordinates were then converted into UTM coordinates.
o Each record was then split into two separate records; one for
each block on each side of the segment with all other information
attached.
o The resulting files were sorted by census tract and street block
numbers and contained records of all blocks in a sequential
manner. Duplicate records were eliminated.
III-2
-------
o Records related to the same block were processed to enclose the
boundary surrounding the specific block by linking "from" and "to"
nodes on each record. The centroids for the block were then
calculated based on the nodes on the enclosed boundary.
Four files were created containing data on both the 1980 population and
the UTM coordinates of the centroids of each: (1) block, (2) BG/ED, (3)
census tract, and (4) place. An additional file was created containing
similar information for the population in GBF/DIME areas.
Annual growth factors derived from the projection compiled by the
Southern California Association of Governments (SCAG) at the place level
were used to forecast population.
III.2 METEOROLOGICAL DATA BASE
HEM and SHEAR use meteorological data in the form of STAR tabulation
(frequency of occurrences of various meteorological conditions) for only
three sites in the Basin. Because of the abundance of the long-term
meteorological data collected in this Basin and the importance of
terrain on meteorology, the meteorological data base was expanded to 16
sites, each representing a specific source-receptor area, to cover the
entire Basin.
Another enhancement made was the creation of seasonal- and diurnal-
specific STAR tabulation to include certain source types which have
emissions in the winter periods only and where solvent usage occurs
mostly during working hours.
III.3 EXPOSURE ASSESSMENT '
The receptors used in dispersion modeling of emissions from specific and
prototype sources include centroids of all four levels of population
disaggregation. The model, by default, dynamically locates the
population centroids within a pre-determined radius from the source
studied. Centroids for all street blocks within a. radius of 2.5 km of
the sources, all BG/ED's within 10 km, and all census tracts within 20
km, were used as model receptor areas. If desired, the model can
specify the radii for population unit transition or specific
concentration thresholds or risk levels for determining the transition
from a finer level of population to a coarser one. The assessment of
impacts from mobile sources uses the centroids of all census tracts as
the common basis for .matching the concentration estimates and the
population distribution.
III-3
-------
III.4 SUMMARY OF ENHANCEMENTS AND OUTPUT
SCREAM contains the following enhancements:
o Census data at the street block rather than block group level;
o City-specific growth projections;
o Extensive site-specific meteorological data for dispersion
modeling;
o Special gravity treatment to locate prototype sources such as
service stations.
The types of output generated by SCREAM consist of:
o Population distribution in close proximity of specific point
sources;
o Tabulation of the number of people exposed to specific ambient
carcinogens above a pre-defined concentration;
o Tabulation of exposure, cancer risk, and number of cancer cases
with different re-defined thresholds by source category and
chemical species;
o Isopleth plots of concentrations, 'exposure, and cancer risk
patterns related to emission of specific chemicals from specific
point sources;
o Isopleth plots of concentration, exposure, and cancer risk
patterns on a regional basis for specific chemicals or for all
defined species.
III-4
-------
CHAPTER IV
EMISSIONS DATA
A detailed emissions data base was developed as input data for SCREAM
and discussed In the following sections.
IV.1 EMISSION INVENTORY METHODOLOGY
The District's first toxic air pollutant emissions inventory was
compiled for 30 toxic air pollutants for the year 1982 (Zwiacher, et
a!., 1983) for stationary sources only. For ten of the pollutants, the
data were generated from emissions compiled from the District's computer
data bases, including the Automated Equipment Information System (AEIS)
and Emission Inventory System (EIS) files and 1982 Emission Fee Reports.
For the remaining 20, emissions data were obtained from a mail survey of
1606 companies in the Basin and followed by literature searches, and
letter and telephone inquiries.
For the MATES, 20 of the pollutants (shown in Table IV-1) were updated
to 1984 (Zwiacher, et a!., 1985). In addition, mobile source emissions
data for 12 of the 20 toxics under study were compiled.
IV.1.1 Stationary Source Emissions .
The data for 1244 point sources were entered into a computer file along
with company name, address, AEIS identification number, and Universal
Transverse Mercator (UTM) coordinates. Over the last year additional
corrections have been made to the inventory. Specific stack parameters
for individual power plants and a generic set of parameters used for all
refineries are presented in Table IV-2.
IV.1.2 Mobile Source: Emissions
Emissions of potentially toxic air pollutants from mobile sources were
estimated for on-road motor vehicles only. Emission rate data were
unavailable for other mobile sources (i.e., aircraft, locomotives,
ships, and off-road vehicles). Estimated motor vehicle emission factors
for 12 of the 20 pollutants under study were provided by ARB and are
shown in Table IV-3. Emissions of these compounds result from
combustion and evaporation of motor vehicle fuels. The emission factors
were provided as a weight percent of total hydrocarbons (THC) except as
indicated in Table IV-3.
ARB's program BURDEN calculates estimates of motor vehicle emissions by
vehicle type (automobiles, trucks, and motorcycles), fuel type (gasoline
and diesel), emission type (exhaust, hot soak evaporation, and diurnal
evaporation), control technology (catalyst and non-catalyst), and
IV-1
-------
TABLE IV-1
TOXIC AIR POLLUTANTS STUDIED
IN THE SOUTH COAST AIR BASIN
METALS
ORGANICS
Arsenic
Beryl1i urn
Cadmiurn
Chromi urn
Lead
Mercury
Nickel
Benzene
Carbon Tetrachloride
Chloroform
Ethylene Dibromide
Ethylene Dichloride
Methyl Bromide
Methylene Chloride
Perch!oroethylene
To!uene
1,1,1-Tri chloroethane
Trichloroethylene
Vinyl Chloride
Xylenes
IV-2
-------
CM g
> uj
UJUJ
3EUJ
Su.
ta uj
ta*LU
UJ U.
4. co in
evt
CM CM CM
o
in
CM
O
o»
CM
o
0
o
tsj
g 2 S S3 £ §
5- i * * O» °*
*r ui co « co
o
o
o
c\i
O o
« CM
o
o
in
i*I
o o co «J» o o
«*l C3 O f*» O O
2 CM S - CM CM
a
0
tsj tnin
oa.ro i*» *
s
s~"-l:
CO CM
o in ^
oo o w
CM T CO
oo
CM
CO CM
10 r^
5 S
SSS" § §§
to
UJ
g
i
^_y >^
co co
UJ
o
UJ
fc
IV-3
-------
o
41
U
8.
ui " g
-J «/» J3
i s«
>- f- o
S2
"1
111 *.
o
5
?
UJ
f
15 S
s - * « - g
«- S S _ o S gaa^^54,c
^ * *S ^SSeo^gg^
j 5s
"t S a
4> ' -S °1S
^, S g CM 1 I -B g ^if^-S
1 i-u;tZ.
go j ! iiffit:2SI45i
s
M
go | I S^ooa^^^^
* W M
cr tf M. r-
*» T> o» I. a
o** «>
CO Z «- *"
a »- s e
Ul ^ 3 O ;
C O E
^0 VI 41 C
£*=*- «
4> < <
5 ..
C» C 4* V ^^
^3 o o ^^ etf en
as "«Ti g o
i» If 5 "5 42 eno
« *» a S E
^ « 4* Z B <3°
**
_
iis
O 4J (. « 41
4i M o.u c e
otlu o o.t-= o t?
T» 4t «S e
>» »f - I
^Se""lo<£
IggglTaBB 1
L***v***M"*>>cn33 f\
M r- « « * '
-------
enrittant (THC, PM, CO, NOX, and SOX) for any given year. BURDEN was run
for 1984 using emission factors from ARB's emission factor program
EMFAC6D. The emissions data from the BURDEN output for the appropriate
vehicle types, fuel types, emission types, and control technologies were
applied to the emission rates in Table IV-2 to calculate the ratio of
emissions to THC emissions for motor vehicles. The resulting fractions
applicable to the Basin in 1984 are shown below:
Benzene 3.4 x 10
Cadmium 3.4 x 10
Chloroform 2.9 x 10
Chromium 6.5 x 10
Ethylene Dibromide 5.9 x 10
Ethylene Dichloride
Nickel
Toluene
Xylenes 4.4 x
5.9 x
2.1 x 10
1.2 x 10
7.0 x 10
-2
-5
-9
-5
-5
-4
-5
2
10-2
Data in the 1979 emissions inventory and 1987 forecast for lead in the
Basin were interpolated to estimate 1984 lead emissions from motor
vehicles. The ratio of lead to THC was calculated to be 0.010.
The above factors were applied to the Basin total THC on-road motor
vehicle emissions (556 tons/day) and to. the gridded ROG emissions.
IV.2 SUMMARY OF EMISSIONS .
A summary of the 1984 toxics emissions inventory is presented in Table
IV-4. Area sources, and those small point sources which are too numerous
to individually spatially allocate are combined under the area source
emissions column.
Point sources are primarily associated with emissions of arsenic,
beryllium, mercury, methyl bromide, nickel, and vinyl chloride.
Emissions of metal species were contributed by combustion, plating, and
other processes. Methyl bromide is used as a soil and space fumigant
and in organic synthesis. Vinyl chloride is emitted from polyvinyl
chloride production in addition to a small, but poorly quantified,
contribution from municipal and hazardous waste landfills.
Area sources contributed the majority of methylene chloride,
perch!oroethylene, and trichloroethylene emissions. These substances
are predominantly used in metal degreasing, solvent extractions, and dry
cleaning.
Motor vehicles comprised the major sources of cadmium, ethylene
dibromide, ethylene dichloride, lead, toluene, and xylene emissions.
Each of these substances is a constituent of gasoline and diesel fuels.
Benzene emissions are split almost evenly between mobile sources and
IV-5
-------
TABLE IV-4
TOXICS EMISSIONS OF TWEttTY SPECIES
IN THE SOUTH COAST AIR BASIN
IN 1984
Emissions (tons/year)
Arsenic
Benzene
Beryllium
Cadmium
Carbon Tetrachloride
Chloroform
Chromium
Ethyl ene Dibromide
Ethylene Bichloride
Lead
Mercury
Methyl Bromide
Methyl ene Chloride
Nickel
Perch! oroethyl ene
Toluene
1,1,1-Trichloroethane
Trichl oroethyl ene
Vinyl Chloride
Xylenes
Point Area
0.047
118. 7870.
0.037
1.12
3.20
0
16.0
1.09
3.53
14.5
0.13
24.4 '
4780. 10,200.
5.40
3970. 8850.
714. 276.
8590. 6150.
9.52 546.
1.37
230. 185.
Mobile lota]
0.047
6910. 14,898.
0.037
6.91 8.03
3.20
0.0006 0.0006
13.2 29.2
12.0 13.1
42.7 46.2
2030. 2045.
0.13
24.4
14,980.
2.44 7.84
12,820.
14,200. 15,190.
14,740.
556.
1.37
8950. 9365.
no data available
IV-6
-------
area sources such as gasoline marketing, stationary gasoline engines,
crude oil production, and agricultural burning.
Some emissions sources have not yet been adequately assessed, including
municipal and hazardous waste landfills, while other sources have not
been incorporated. Although stationary source emissions of chloroform
are listed as zero in Table IV-4, several sources may have significant
emissions of chloroform. These sources include publicly owned treatment
works, sewer lines, swimming pools, showers, laundry machines, and power
plant cooling towers. Publicly owned treatment works (POTWs) are also
thought to be a significant source of vinyl chloride nationally (Versar,
1984); however, knowledge of existing industrial sources of discharge
into local sewer systems and recent downwind testing do not indicate
substantial vinyl chloride emissions from POTWs in this Basin (Roberts,
1985). Further source testing would need to be conducted to develop
emission factors for these and other sources and to estimate routine
emissions of chloroform and vinyl chloride as well as other air toxics.
IV..3 SPATIAL ALLOCATION OF EMISSIONS
Regional modeling analysis requires spatially resolved emissions data by
grid cell. A 5 km by 5 km grid cell system is generally used for this
purpose in the Basin. Maps showing the emissions by grid cell are also
useful in characterizing the spatial pattern of emissions.
Emissions from 1244 individual point sources were spatially located by
UTM coordinates to within .0.1 kilometer. The spatial distributions by
grid cell of point source emissions of each chemical species under study
are included in Appendix A. Most of the emissions from point sources
are spread throughout the coastal and metropolitan portions of the
Basin.
Area source emissions were spatially allocated to grid cells by
population using population distribution data provided by the California
State Census Data Center. The emissions from these sources were assumed
to be linearly proportional to the population. Figure IV-1 presents the
population distribution in the Basin and thus the relative distribution
of the area source emissions by grid cell.
Exceptions were made for gasoline stations which are often clustered at
street intersections. The degree of clustering of these sources was
determined through a telephone survey and was incorporated in the
modeling algorithm using a weighting method. Emissions of these sources
were assumed to be in proportion to the survey reported throughputs.
Annual average emissions of reactive organic gases (ROG) from motor
vehicle sources in the Basin for 1984 have been gridded into 5 km by 5
km cells for the entire Basin. These emissions are presented by grid
cell in Figure IV-2. These emissions were further divided into 1 km by
1 km cells by assuming that each of the 25, 1 km by 1 km cells within
IV-7
-------
C«0 3NIH1HON Win
IV-8
-------
13
illfi
E >
(M) 3NIH1UON
IV-9
-------
any 5 km by 5 km cell has equal emissions. The emissions of individual
toxic species in each cell were calculated by multiplying THC emissions
in any cell by the ratios listed in Section IV.1.2.
IV-10
-------
CHAPTER V
AMBIENT DATA
The District and ARB have been conducting continuous long-term ambient
air monitoring programs for several toxic air pollutants in the Basin.
EPA, through its National Air Surveillance Network, also monitors
continuously for several airborne metals in the Basin. Using these
data, annual average ambient concentrations for several organic and
metal air pollutants at several locations in the Basin have been
estimated for 1985. District-wide population-weighted average
concentrations were also estimated for most of the compounds.
Since none of the three toxic air pollutant monitoring networks measured
ambient formaldehyde, a brief discussion of selected short-term air
sampling studies conducted in the Basin for this compound are summarized
to provide an indication of representative ambient concentrations in the
region.
V.I CONTINUOUS MONITORING NETWORK
Figure V-l shows the location of each continuous toxic air pollutant
monitoring station in the Basin. All monitoring sites are located at
existing District criteria pollutant stations except for El Monte. The
District collects samples for 11 organic gases about once every two
weeks at four of these stations. ARB collects samples for eight organic
gases at about the same rate as the District and for six metals about
once per week'at the other five sites.. EPA samples for 14 metals and
,benzo(a)pyrene about every 10 to 12 days at two sites.
Data gaps were present in all three network data bases. These gaps were
more extensive within the District and ARB data bases. For example, in
1985, the District did not start collecting data until the. beginning of
March. Gaps in the ARB data base were also present during a three-month
period starting in June for seven organics, and during a 22-week period
starting in July for three metal pollutants.
V.I.I Annual Average Ambient Concentrations
A range of arithmetic annual average values was calculated when one or
more observations for a particular pollutant were measured below minimum
detection limits. That is, low and high average values were calculated
assuming that all below-detection-limit values were equal to zero and
equal to the detection-limit value, respectively.
ARB and EPA reported one detection limit value for each pollutant;
whereas, the District reported several because the limit changed from
observation to observation depending on laboratory conditions existing
at the time of sample analysis. Standard deviations were calculated for
V-l
-------
i §
S fn
I » M «
O^O 3NIH1HON Wif!
co
K S
crj
«
r
CO
en
CD
en
o
CO
en
Lu
CO
cc
X
o
cc
CD
V-2
-------
all pollutants with at least 10 percent or more of their observations
above detection limits. For pollutants with less than 10 percent of
their observations above detection limits, standard deviations were
calculated using only those values above the detection limit. Data
category codes are given based on the percentage of observations
measured above detection limits: Agreater than 90 percent, B--less
than 90 but greater than 10, and C--less than 10 percent. These
categories provide a crude measure of confidence associated with the
data.
Table V-l and Table V-2 summarize' the annual average ambient organic
gases and ambient metals concentrations at several monitoring stations
throughout the Basin in 1985. For organic gases the range of
concentrations appears to vary from station to station. Ethylene
dibromide concentrations, for example, vary by a factor of 20 between
stations. ARB stations consistently measured lower levels than District
stations for ethylene dibromide and perch!oroethylene. Stations
measuring benzene and carbon tetrachloride showed relatively comparable
average concentrations. Concentrations of vinyl chloride, chloroform,
ethylene dibromide, and ethylene dichloride were predominantly below
detection limits for most stations.
Unlike organic gases concentrations, estimated average concentrations
for metals did not vary by more than about a factor of three from
station, to station, except for arsenic and beryllium. EPA stations
measured the highest concentrations for arsenic and the lowest for
beryllium compared to ARB stations. The large variation in beryllium
concentrations can be explained, in part, by the different detection
limits'reported by ARB and EPA. Most ARB samples were reported below
detection limits which changed from 0.5 to 0.02 ng/nr after July 1.
Therefore, since the former value was over 3.00 times greater than the
EPA detection limit, estimated annual average values for ARB data are
much higher than would otherwise be expected based on the more recent
ARB detection limit. The quality of the annual average concentrations
can be judged, in part, by looking at the number of detection limit
observations for the various pollutant/station combinations shown in
Table V-2. Arsenic, beryllium, and cadmium were observed at detection
limit concentrations at many stations. Actual annual averages are
probably somewhat lower than estimated.
Discerning spatial trends throughout the Basin for organic gases or
metals is difficult because of the problem in determining if the spatial
variations observed from station to station are due to either actual
ambient conditions or differences in sampling and analytical procedures.
The degree of comparability between the District and ARB sampling and
analytical procedures for individual pollutants has not been clarified
yet. Therefore, any conclusions based on an analysis of combined data
from various data bases must be made with care.
V-3
-------
TABLE . V-l
1985 ANNUAL AVERAGE AMBIENT AIR CONCENTRATIONS OF
VARIOUS TOXIC ORGANIC GASES IN THE SOUTH COAST AIR BASIN*
(concentration in ppbv)
Pollutant
BENZENE
Ave. Cone.
Std. D«v.
Detection Limitb
Sample SIM/# < DLC
Data. Category
S
Ana-
heim
1.7-2.8
1.6
' 2.0-4.0
24/10
B
C A Q M D
Azusa
1.0-2.6
.92
1.0-4.0
21/13
B
Burbank
2.0-3.0
1.3
1.0-3.0
23/9
B
Lennox
1.7-2.8
1.5
2.0-3.0
23/10
B
El
Monte
4.9
2.6
.5
39/0
A
A R B
Long
Beach
4.1
1.9
.5
25/0
A
LA
4.2
2.2
.5
23/0
A
Rubi-
doux
2.5
1.3
.5
24/0
A
Upland
3.4
1.4
.5
22/0
A
CARBON TETRACHLORtoB
Avt. Cone.
Std. D«v.
Detection Limit
Sample SIM/# < DL
Data Category
CHLOROFORM
Ave. Cone.
Std. Dev.
Detection Limit
Sample Size/# < DL
Data Category
.12
.038 .
.016
22/0
.04S-.30
*«
.02-1.0
22/21
C
.12
.036 '
.016
19/0
.OS3-.29
»
.02-1.0
19/18
C
.10
.023
.016
20/0
. .013-24
*
.077-.5
20/20
C
.11
.033
.016
20/0
.063-.27
*
.077-89
20/18
C
.095
.024
.004
36/0
.063
.023
.02
36/0
A
.10
.014
.004
22/0
.082
.025
.02
22/0
A
.11
.016
.004
21/0
.11
.090
,02
21/0
A
.099
.015
.004
20/0
.053
.028
.02
20/0
A
.12
.073
.004
20/0
.071
.045
.02
20/0
A
BTHYLENB DIBROMIDB
Ave. Cone, (ppt)
Std. Dev. (ppt)
Detection Limit (ppt)
Sample Si«e/# < DL
Data Category
5-100
*
100
22/21
C
0-100
*
100
19/19
C
0-100
'*
100
20/20
C
0-100
*
100
20/20
C
3-6
4
5
36/26
B
4-8
9.0
5
22/16
B
2-6
2.0
5
21/14
B
2-6
1.4
5
20/16
B
3-5
1.0
5
20/9
B
ETHYLENE DICHLORIDE
Ave. Cone.
Std. Dev.
Detection Limit
Sample Sira/# < DL
Data Category
0-17
*
2.1-28
22/22
C
0-14
*
4.0-28
19/19
C
0-18
*
4.0-28
20/20
C
1.0-18
*
4.0-28
20/19
C
V-4
-------
TABLE V-l (continued)
S C A Q M D
Pollutant
Ana- Azuaa Burbank Lennox
heim
A R B
El Long LA
Monte Beach
Rubi- Upland
doux
METHYLENE CHLORIDE
Ave. Cone.
Std. Dev.
Detection Limit
Sample Size/# < DL
Data Category
5.1 4.6-4.7 3.5 2.2-2.3 2.7-2.8
2.6 2.9 2.0 1.9 1.7
.006 .006 .006 .006 .006
36/0 22/1 21/0 20/4 20/3
A A A B B
PERCHLOROETHYLENE
Ave. Cone.
Std. Dev.
Detection Limit
Sample SUe/# < DL
Data Category
3.1
2.4
.2
22/0
A
2.0
1.8
.2
19/0
A
2.7
1.6
.2
20/0
A
2.3
2.5
.2
20/0
A
1.6 1.0 1.2 .45 .70
.86 .56 .91 .32 .45
.01 .01 - .01 .01 .01
36/0 22/0 21/0 20/0 20/0
A A A A A
TOLUENE
Ave. Cone.
Std. Dev.
Detection Limit
Sample Si»e/# < DL
Data Category
4.0-5.6
2.9
3.0-6.0
24/8
B .
2.5-4.9
2.3 .
3.0-7.0
21/11
B
5.6-6.7
3.0
5.0
23/5
B
3.5-5.3
4.2
.40-5.0
23/11.
B
1.1.1-TRICHLOROETHANE
Ave. Cone.
Std. Dev.
Detection Limit
Sample S»e/# < DL
Data Category
2.3
1.4
.23
22/0
A
2.6
1.7
.23
19/0
A
3.3
1.7
.23
.20/0
A
2.5
1.5
.23
20/0
A
7.1
4.7
.02
36/0
A
3.0
2.1
.02
22/0
A
2.4
2.6
102
21/0
A
1-1
.73
.02
20/0
A
1.6
1.1
.02
20/0
A
TRICHLOROETHYLENB
Ave. Cone.
Std. Dev.
Detection Limit
Sample Size/# < DL
Data Category
.23:35
.23
.20-.90
22/7
B
.16-.33
.26
.11-.90
19/8
B
.39-45
.72
.12-.22
20/7
B
.20-.23
.22
.12-.15
20/15
B
.40
.25
.02
36/0
A '
.29
.20
.02
22/0
A
.34
.22
.02
21/0
A
.10
.06
.02
20/0
A
.37
.17
.02
20/0
A
VINYL CHLORIDE
Ave. Cone. 0-2.0 0-2.0
Std. Dev. * *
Detection Limit 2.0 2.0
Sample Size/# < DL 24/24 21/21
Data Category C C
0-2.0
*
2.0
24/24
C
0-2.0
*
2.0
24/23
C
V-5
-------
TABLE V-l (continued)
S C A Q M D
A R B
Pollutant
Ana-
heim
Azusa Burbank Lennox
El Long LA Rubi- Upland
Monte Beaeh doux
BENZOfAlPYRENB (from EPA monitoring network)
Ave. Cone, (ng/m3)
Std. D«v.
Det. Lim. (ng/m3)
Simple S!M/# < DL
Data Category
.75
.17
.33
21/0
A
.75
.0079
.33
18/0
A
Blanks indicate no data available. Ranges of arithmetic annual averages defined as follows: first estimate is the
average assuming all sub-detection limit observations equal zero; second estimate is the average assuming all sub-
detection limit observations are equal to the detection limit concentration.
Standard deviations were calculated using only the observations above detection limits; if more than 90 percent
of the observations wer« below detection limits, standard deviations were not calculated.
Detection limit* for some SCAQMD-measured pollutants reported as range because limits changed from-sampie
to sample depending on analytical conditions. See text for further explanation.
"Sample Sii«/# < DL" = (the total number of samples taken pver the year) / (total number of these samples
with concentrations below minimum detectable limits).
Data Category codes for SCAQMD and ARB'data are defined as: A - Most of the data above detection limits
(>SO%), C - Very few of the data points are above detection limits (<10%), and B - Several data points fall above
'and below detection limits.
* standard deviation not calculated for Data Category C.
V-6
-------
TABLE V-2
1985 ANNUAL AVERAGE AMHIENT AIR CONCENTRATIONS OF VARIOUS
TOXIC METALS IN THE SOUTH COAST AIR BASIN'1
(concentration In nj/m )
E P
Pollutant Anaheim
ARSENIC
AT*. Cone.
3ld. DOT.
Detection Limit
Sample Siu/# < DLe
Data. Category
BARIUM
A»e. Cone.
Sid. DOT.
Detection Limit
Sample SI*e/# < DL
Data Category
BERYLLIUM
A»e. Cone.
Std. Dm.
Detection Limit*
Sample She/* < DL
Data Category
CADMIUM
Av«. Cone.
Std. DOT.
Detection Limit
Sample 3l»/# < DL
Data Category
CHROMIUM
Av«. Cone.
Std. Dm.
Detection Limit
Sample Shw/# < DL
Data Category '
COBALT
Ave. Gone.
Std. DOT.
Detection Limit
Sample Siae/# < DL
Data Category
COPPER
Ave. Cone.
Std. DOT.
Detection Limit
Sample 3Iie/# < DL
Data Category
0-6.7
.b
6.7
21/21
C
0-110
110
27/27
C
0-.OOI6
.0016
27/27
C
0-1.0
1.0
27/2G
C
1.8-5.1
1.2
4.5
SI/15
B
1.1
.70
.37
21/0
A
170
71
4.8
27/0
A
A
Loe El
Angale* Monte
5.1-8.8 2.7
S.3 1.S
6.7 .4
18/10 15/0
B A
0-110
«
110
23/22
C
0-.OOIC 0-.20
«
.0016 .S-.02
23/23 IS/IS
C C
2.0-2.2 4.1
2.0 2.0
1.0 .3
23/5 15/0
B A
10-11
5.0
4.5
18/2
B
1.0
.40
.37
18/0
A
180
46
4.8
23/0
A
Lnn(
Oaaeh
2.2-2.3
1.9
.4
S8/3
A
.007-.J8
.008
.5-.02
58/40
B
.72-1.1
.59
4
S3/23
B
4.7
1.8
1.0
31/0
A
A a D
Loo Pico River-
Angflt* lliver.i «id«
2.2-2.3 3.4-3.S 2.0
1.7 2.7 .98
,4 .4 .4
S7/2 57/1 29/0
A A A
.008-.28 .000-.28 .046
.007 .011 .02
.5-.02 .S-.02 .£-.02 .
§7/39 S7/37 29/0
BOA
1.0-2.2 1.1-1.5 .8G-.87
3.7 1.4 .51
.3 .3 . .3
S7/17 57/20 29/1
B B A
7.0 S.O
2.6 1.5
1.0 1.0
31/0 31/0
A A
Rubi-
doux
1.7-2.0
1.3
.4
31/10
B
0-.50
.5-.02
31/31
C
.16-1.0
.50
.3
31/27
B
3.S
1.8
1.0
31/0
A
Upland
1.8-1.0
1.1
.4
57/8
B
.OI4-.J8
.018
.S-.02
57/35
B
.73-1.1
.55
.3
57/21
B
3.2'
1.3
1.0
30/0
A
V-7
-------
TABLE V-2 (continued)
B P
Pollutant Anaheim
AT*. Cone.
Sld.Dev.
Defection Limit
SampU SIie/iK DL
Data Category
LEAD
AT*. Cone.
Std. Dtv.
Detection Limit
Sampk 3tt*/# < DL
Data Category
MANCANE3B
AT*. Cone.
Std. D*r.
DefeetSoa Limit
Sampk Site/* < DL
Data Cafegory
MOLYBDENUM
AT*. Cone.
314. Dev.
Defection Limit-
Sample 31n/# < DL
Data Cafegory
NTCKBtr
AT*. Cone.
Std. Dvr.
Defection Limit
Sample S!t»/# < DL
Data Category
VANADIUM
AT*. Cone.
Std. Dev.
Defection Limit
Samp)* 3b*/# < DL
Data Cafegory
tTNC
AT*. Cone.
SW-Der.
Defection Limit
Sample Slu/£ < DL
Data Cafegory
1300
680
22
27/0
A
180
82
8.8
27/0
A
33
15
3.6
27/0
A
1.9
.5 .
I.S
27/0
A
3.7-7.2
1.2
S.9
27/16
B
4.7-7.9
1.S
6.2
27/14
B
37-100
,17
93
27/19
B
A A R U
Lo« El Long Lo* I'ico River-
AngelM Monte Beach Angele* Rivrro tid*
1800
760
22
23/0
A
280
140
8.8
23/0
A
44 18 2E 2 above deleclion limili; if more thnn 00 percent of th* obMrratlont nere
below detectkm limit*, itandard delation* were not ealcnlnted.
* (tandard dariatlon not calculated for Data Category C.
e -SampI* SIi*/# < DL- » (th* totnl number of anmplra taken orer th* y«ar) / (total nninl.rr of thne nample* with concentratioiu below
minimum delectable limili).
Data ClUjory code« for EPA and CARD data are rfefinwl as: A - Meat of th* data above detection limit. (>9055). C - Very few of the
data polnla an above detection limit. (<10%), and B - Several data point* fall above nnd below detection limit*.
Detection limit* for beryllium chanted from O.S to O.02 tig/in3 after July 1.
V-8
-------
V.I.2 Population-Weighted Annual Average Ambient Concentrations
For purpose of calculating cancer risk using ambient air quality data,
the basinwide population-weighted annual average concentration for each
pollutant was estimated as follows. First, each station was spatially
located with UTM coordinates and plotted onto a gridded map of the
Basin. Average concentrations for each individual grid cell were
interpolated based on their proximity to surrounding monitoring
stations. This was done by weighting concentration data from each
station by 1/R.j , where R.J is the distance of monitoring station i from
the particular grid cell, and then calculating the average over all
stations. Therefore, stations farther away from a grid cell have less
influence on the estimated grid cell average compared to those that are
closer.
Finally, gridded population data for 1985, obtained from the ARB, were
superimposed over the gridded concentration data. Basinwide population-
weighted averages were calculated by summing the products of the
population and concentration for each grid cell and then dividing by the
total Basin population. Low and high averages were estimated for
pollutants that were observed at sub-detection limit concentrations.
Table V-3 is a list of population weighted annual average concentrations
for selected organic gases and metal pollutants for 1985. The following
important factors must be noted before proper interpretation of these
estimates can be made. First, pollutants that were measured at only a
few stations were not included on this list because it would not be
appropriate to calculate a basinwide average from only a few data
points. Second, weighted averages appearing, in the list should be used
with caution because of questions regarding the sampling and analytical
comparability of data from different monitoring networks. Ethylene
dibromide and beryllium are two examples where caution should be used
because of the wide range of concentrations observed by the respective
monitoring networks. Also, averages for these two compounds, plus vinyl
chloride, ethylene dichloride, arsenic, and cadmium, should be
interpreted carefully because the majority of the observations were
below detection limits.
The only way to ensure truly accurate and precise results from an
analysis using combined data bases is to require that all monitoring
networks conform to the same sampling, analytical, and quality assurance
and control procedures. Although such requirement is not currently
possible, these estimates provide a reasonable approximation of
basinwide annual average ambient concentrations.
V.2 LITERATURE SURVEY OF AMBIENT DATA
A literature survey was conducted as a secondary source of information
for estimating the impact of pollutants considered significant but for
which Basin-specific monitoring and emissions data are not yet
V-9
-------
TABLE V-3
POPULATION-WEIOIITKI) ANNUAL AVERAGE AMBIBNT CONCENTRATIONS
AND INDIVIDUAL CANCER RISKS IN THE SOUTH COAST AIR BASIN1
Unltb
Riik Factor
Pollutant (uf/m )"
Qrranie Gw(
Benxen«
Benxo(a)Pyrene
Carbon Tetraehloride
Chloroform
Ethylene Dibromide
Methylene Chloride
Perchloroethylene
Trlchloroethylene
TfKt Mtf *'*
Anente (Inorganic)
Beryllium
Cadmium
Chromium
Nickel
SUSxlO-5
-3
3.3x10 3
LSxlO'5
2JxlO'S
7.2xlO'S
4.1X10-6
8.8xlO"7
a^xio'8
4.3X10-3
2.4X10'3
1.SX10"3
1.5x10^ «
3.2x10"* h
1985 Population0
Weighted Annual
Average Ambient
Concentration
(ug/m )
SCAQMP
5.1-8.9
_-4 e
7.5x10
0.9x1 0"1
.f
17
1.2-1.8
BCA
(2.7-7.8)xlO"3
>
. (l.l-1.6)xlO"3
(0.3-8.1 )xlO"3
(S.6-8.1)xlO'3
MS
12
K
6.3X10*1
3.8X10*1
(2.1-4.8)xW2
13-13
6.8
i.y
ARB
(2.3-2.4)xlO*3
(.U-2.6jxlO~*
(1.4-I.8)xIO'S
4.8X10"3
7.9X10-3
Population
Weighted
Lifetime
Individual
Risk
SCAQMP
(2.7-4.7)x)o"*
, n-6 *
l.OxlO"5
*
9.8X10'8
(l.S-2.4)xlO"6
EPA
(I.2-3.3)xIO"S
»
(2.0-2.9)xlO~8
(.94-I.2)xlO"3
(l.S-2.6)xlO'6
ARB
6.6x10"*
9.4X10'8
8.7X10'8
(1.5-3.5)xl6-8
(5.3-S.4)xlO-S
3.9xlO"6
2.2xlO"5
ARB
(.99-1.0)xlO'S
(.26-6.2)xlO-7
(2.S-3.2)xW8
7.2x10"*
2.SX10"8
Reference
ARB/DOHS, 1984
EPA, 1984b
EPA, 1984c
EPA, 198Gd
ARB/DOIIS, 198Sa
EPA, 1985e
EPA, 1985g
EPA, 1985c
EPA, 1984a.
EPA, 1986a
EPA, 1085b
ARB/DOHS, 198Se
EPA, 1986b
Bl«nlc» indicate no data available. popiiinuon-wrigniTO II...HH. «n«w.»-
pollutant, with a Data CaUgory of A or B. (»« Table, IH-1. IIM) and a published «..it ri»k factor.
b Unit ri.k factor r.p««nU carcinogenic rtt f»r a ,««on brtathin, 1 ti»/m3 of a ,«ll.,tant over a 70-year lifetime.
lation-wdght^ av.rage. are defined a. follow: fi«t -Hmata b the average a-uming all .ub^.tection limit
^uJ to »ro; ,«ond estimate i, the average a«umin« M .ub-deteclio,, Umit ob.ervat.on. are «,ual to the
detection limit concentration.
Population-righted aver.g« for organic ga«,. were convert* from ppbr unit, to ug/m3 turning .t»,dard condition.
(temp.« 2S°C and pmeura s 1 atmoiphere).
d A rUk of 1.0X10-8 mean, that an individual ha, a one in a million chance of contracting cancer. Reader .hould understand
thai th«« ralue. are probably accompanied by .ignifieant, but unquantified, uncertainty.
Data for ben*o(a)pyrene obtained from EPA monitoring ttatiotu.
f indicate that the pollutant powe.Md a Data Category cod. of C and w». therefore, excluded. Pollutant. aMigned Data
Catefory C were thow in which Ie« than 10 percent of the obMrvation. were above detection lim.te.
« Unit ri.k factor for hexavalent chromium; it i. not knnwn what fraction of the annual average concentration i. chromium (VI).
h The potency of nickel varie. by .pecies; unit ri.k factor reprewnta .ubwlfide .pecies.
V-10
-------
available. The types of data derived from a literature survey include
short-term (less than one year) ambient data in the Basin and monitoring
data collected in other major urban areas.
A number of short-term studies to measure ambient formaldehyde have been
conducted in the Basin over the past several years. Table V-4 presents
a summary of several representative studies referencing nine specific
monitoring projects in this Basin with up to three hundred samples
collected and analyzed. Formaldehyde concentrations appear to be
somewhat higher inland compared to coastal areas, although the
statistical significance of this relatively, weak trend cannot be
verified with these data because sampling dates, methods, and analytical
procedures vary.
Diurnal and seasonal concentration patterns have been reported by a few
investigators. Grosjean (1982) found that formaldehyde concentrations
peaked in late afternoon and dipped to a minimum during the early
morning hours. On a seasonal basis, Salas and Singh (1986) found that
at one site in the Basin (Downey), concentrations appeared to be
somewhat lower in the winter months compared to other times of the year.
Formaldehyde is emitted directly into the atmosphere from mobile and
stationary sources and indirectly through photochemical formation.
Grosjean, et al. (1983) suggested that formaldehyde concentrations
nearer to the coast are influenced primarily by direct emission sources;
whereas, inland concentrations are . largely due to secondary
photochemical formation processes in the atmosphere. The investigators
estimated that at Azusa and Claremont, photochemical formation accounted
for, on average, 44 and 78 percent, respectively, of the observed
formaldehyde concentration relative to the Lennox site which was fixed
at zero percent. Lennox was chosen as the reference because it was
assumed to represent a site dominated by direct source emissions.
There are currently no available data on ambient concentrations of
chlorinated dioxins and dibenzofurans in the Basin (ARB and DOHS, 1986).
Ambient particulate samples collected in St. Louis, Missouri, and
Washington, D.C., had average concentrations of dioxins and furans of
200 ppb (Czuczwa and Hites, 1984). Octachlorodioxin accounted for
almost 90 percent of the total dioxins and furans collected at each
location. These data have limited usefulness because the vapor phase
concentrations of these pollutants were not measured. In addition,
these measurements may not be representative of background
concentrations in the Basin. ARB is currently conducting a special
monitoring study in the Basin to determine existing dioxin and furan
concentrations.
For purposes of estimating exposure and risk, these ambient
concentrations were assumed to be representative of those experienced on
an annual average basis in the Basin. Since octachlorinated dioxin is
thought to be relatively non-carcinogenic (ARB and DOHS, 1986), it was
subtracted from the average concentration of total dioxins and furans.
V-ll
-------
TABLE V-4
AMBIENT FORMALDEHYDE CONCENTRATIONS IN THE SOUTH COAST AIR BASIN
AS MEASURED BY VARIOUS INVESTIGATORS
Location
Riverside
Downey
Downtown Los Angele*
East Loi Angeles
Claremont
Asusa
Ltnnox
Various So. Calif.
Locations (mobile lab)
Lennox
Pico Rivera
Pico Rivera
Azuia
Riverside
Date Number of
Samples
Jul 8-10, 1980
Feb 28-Mar 1, 1984
Sept 29-Nov 13, 1981
May 19-Jun 20, 1980
Sept 19-Oct 8, 1980
Jul 30-Oct 24, 1980
Jul 30-Oct 24, 1980
Jul 30-Oct 24, 1980
Jan 13-19, 1983
Jan 13-19, 1983
May 26-Jun 16, 1983
May 26-Jun 16, 1983
Jul 2-12, 1980
18
48.
23
36
70
18
18
20
9
8
12
12
not reported
Concentration1 Reference
(ppb)
19 ±7.6 (41.0 max)
15.5 ±11.9 (67.7 max)
4-86
2-40
3-48
0.7-35.4 (1S.5±9.26)C
0.5-39.6 (8.94 ± 9.68)C
4.6 - 65.9 (45.0 ± 17.3)c
7.3 - 18.2 (12.3 ± 3.51)
4.3 - 33.3 (13.6 ± 9.20)
2:0 - 17 (7.8 ± 4.15)
5.6-23.3 (13.5 ±4.80)
10.4 - 41 (19 ± 7.6)
1
1
2
3
3
4
4
4
5
5
5
5
6
* Data presented either as ranges of concentration or mean ± standard deviation unless otherwise noted.
References:
1 Salas and Singh (1986).
2 Grosjean and Fung (1984).
S Grosjean (1982).
4 Grosjean et al. (1983).
6 RogoMn et al. (1984).
6 Singh et al. (1982).
c Mean and standard deviation values are derived from authors' data.
V-12
-------
CHAPTER VI
MODEL APPLICATION AND RESULTS
The enhanced model SCREAM was applied to the Basin using the detailed
emissions, meteorology, and population data bases previously described.
Annual average concentrations predicted by the model and those obtained
from monitoring data were combined with unit risk factors to
characterize individual cancer risk. For pollutants already identified
as toxic and listed for regulation under AB 1807, unit risk factors
developed by DOHS pursuant to this legislation were used for risk
calculations. To date, unit risk factors have been developed only for
benzene and hexavalent chromium. EPA values were used in all other
cases. The DOHS benzene and hexavalent chromium unit risk factors are
approximately 7 and 12 times higher than EPA's values and represent
upper-bound estimates of these substances' lifetime carcinogenic
potencies.
VI.1 RISK CHARACTERIZATION
To characterize risk from existing sources, the two measures of risk
were generated for each of the 20 pollutants with unit risk factors.
The results are included in the Appendices. To illustrate these
results, the risk characterization results for benzene are discussed in
.this section. Figure VI-1 displays the spatial distribution of model-
predicted ambient concentrations of benzene in the Basin. The highest
concentrations are located in the metropolitan Los Angeles area where
population density is greatest. Mobile source and gasoline marketing
dominate the benzene emissions in this area. These model-predicted
concentrations compare well with ambient measurements as discussed in
Section VI.3.
The spatial distribution of the benzene individual cancer risks are
presented in Figure VI-2. The highest grid-cell average upper-bound
individual lifetime cancer risks for benzene are greater than IxlO"3.
There may be receptors with higher individual risks than those shown.
Population risk is estimated by interpolating individual lifetime cancer
risks with population data for the Basin. The upper-bound number of
excess cancer cases associated with lifetime (70-year) exposure to
model-predicted ambient concentrations of benzene are illustrated in
Figure VI-3. The highest estimates are in the grid cells with both the
highest population density and highest model-predicted ambient
concentrations. Again, these .estimates are based on upper-bound 95
percent confidence limit estimates of carcinogenic potency. The true
risk values may be considerably lower than those estimated.
VI-1
-------
QJ CO
tvl
z. ce
>-«:
I 0£
U) :
O Z
Lul O
O£ -
Q. H-
_i at
LU I
a z
O LU
S O
VI-2
-------
1
*-
Si
a
("H)
Hin
VI-3
-------
ai
3 8
i
a.5
8 8
I O
3.S
ffl
(M) 9NIH1MON NJJ1
' VI-4
-------
The spatial concentration patterns and the spatial distributions of
individual and population cancer risks for the other potential
carcinogens studied are presented in Appendix B and C, respectively.
Also included and presented in-Appendix C are the risk estimates for
formaldehyde and dioxins based on ambient concentrations obtained from
the literature.
This model can also generate data on the population-risk profile for the
Basin. Figure VI-4 shows the number of people exposed to various risks
from nine gaseous and trace metal species on a log-log scale.
Population frequency distribution risk profiles are presented using both
EPA and DOHS potency estimates for those substances for which DOHS has
developed unit risk values. Almost the entire population in the Basin
is exposed to ambient benzene concentrations corresponding to an upper-
bound risk of 10"4 or higher; whereas, a small portion of.the population
is exposed to an upper-bound lifetime risk as high as 10"3. Figure VI-4
illustrates the magnitude of risks and relative importance of the
individual carcinogenic species. Of the nine species evaluated, ambient
concentrations of benzene and hexavalent chromium appear to have the
greatest impact on this Basin's population. These results are specific
to this Basin because the estimated risks from benzene consider the
existing control requirement of Phase I and Phase II vapor recovery for
gasoline marketing. Risks and cancer cases would be higher in areas
which do not employ these emissions controls.
VI.2 SOURCE APPORTIONMENT
The. model can also be used to conduct source apportionment of excess
cancer cases associated with each individual source category.
Table VI-1 breaks down basinwide lifetime excess cancer cases for
benzene and hexavalent chromium by mobile and stationary sources.
Again, benzene cancer cases reflect the District's requirement for Phase
I and Phase II vapor recovery for gasoline marketing.
Chromium emissions from mobile sources were assumed to be 10 percent
hexavalent, while stationary sources were assumed to be 100 percent
hexavalent. .These assumptions are health protective yet plausible
since stationary sources of chromium in the Basin are predominantly hard
chrome platers.
VI-5
-------
Sal
.
^? *^ t
§
O
VI-6
-------
TABLE VI-1
SOURCE APPORTIONMENT OF LIFETIME (70-YEAR) CANCER
CASES FOR BENZENE AND HEXAVALENT CHROMIUM
IN THE SOUTH COAST AIR BASIN
BENZENE
HEXAVALENT
CHROMIUM
MOBILE
STATIONARY
TOTAL
248-2,110
199-1,690
477-3,800
85-1,020
508-6,100
593-7,120
Source apportionment is an effective method of prioritizing allocation
of resources to reduce risks. Table VI-1 indicates that stationary
source emissions of hexavalent chromium should be considered a high
priority for control.
Given sufficient data, a matrix of cancer risk from ambient carcinogens
could, be .developed, from model output. Figure VI-5 depicts a cancer case
matrix for ambient carcinogens and: source categories. This matrix
contains estimates of the number of excess cancer cases from exposure
to: (1) individual chemical species emitted from a single source
category-, (2) individual chemical species emitted from all source
categories, and (3) all chemical species emitted from an individual
source category. The total number of excess cancer cases for the whole
region is also included in this matrix. Emissions data are not yet
available to complete this matrix.
VI.3 COMPARISON OF MEASURED AND MODEL-PREDICTED AMBIENT CONCENTRATIONS
AND RISKS'
Annual average ambient concentrations obtained from monitoring data were
compared with annual average model-predicted concentrations at the same
receptors to identify problems in the modeling approach and the input
data to the model. Table VI-2 compares the measured and model-predicted
annual average concentrations in the Basin for both carcinogenic organic
gases and metals. Also shown in this table are the range of ratios of
the measured to model-predicted concentrations. The further the ratio
is from one, the greater is the discrepancy between measured and modeled
concentrations, indicating problems with model input data or
assumptions.
VI-7
-------
Source - Cace
Speciea > Total
Source I
Cacesory 1
««^bM
-
^^^^
^^M^H^^
^MWMM*'
^^^MIB^
»^«i^^
9 «f
o£
-8
^B^***
tn
o _
'5 ra
^ : u u
* Q 0
(flh
Ui
I
CO
o
o
O
Lul
ee.
u,
tu :*£
S «2
-a
C3 c*
H *
* i . t
5
b
X
0£
VI-8
-------
TABLE VI-2
COMPARISON OF MEASURED AND MODEL-PREDICTED TOXIC AIR POLLUTANTS
AIR TOXICS
MEASURED
MODEL-
PREDICTED
PREDICTED/
MEASURED RATIO
ORGANIC GASES (ppb)
Benzene
1.0-4.9 0.56-5.0
Carbon Tetrachloride 0.10-0.12 l«l-24 x 10'5
0.22-1.8
1.0-25 x 10"4
Chioroform
Ethylene Dibromide 0-100
Ethylene Dichloride 0-18
Perch!oroethylene 0.5-3.1
0.02-0.30 2.0-17 x 10'8 0.68-49 x 10'7
1.1-22 x 10-4
1.7-10 x 10"3
0.28-2.4
Toluene
Trichl oroethylene
Vinyl Chloride
TRACE METALS (ng/m3)
Arsenic
Beryllium
Cadmi urn
Chromi urn
Lead
Nickel
2.5-6.7
1.1-7.1
0-2.0
0-8.8
0-0.5
0-4.1
1.8-11
180-280
3.7-8.9
0.80-3.7
0.33-2.9
5.1 x 10"5
5.0-10 x 10"4
0-5.4 x 10'3
1.1-9.6
3.6-60
1100-1700
0.7-5.6
0.11-110 x 10'5
1.2-52 x 10"4
0.22-1.5
0.16-0.66
0.13-54
2.6 x 10
-5.
1.5-2.3 x 10'4
0.003-3.4
0.71-1200
1.06-8.6
3.9-9.4
0.08-7.3
VI-9
-------
There is good agreement between the measured and model predicted
concentrations for several organic gases and metals. Ratios for
benzene, perch!oroethylene, toluene, 1,1,1-trichloroethane, beryllium,
cadmium, chromium, lead, and nickel are all very close to 1.0. These
results give a great deal of confidence to the model-predicted
concentrations. The ratios for carbon tetrachloride, chloroform,
ethylene dibromide, ethylene dichloride, vinyl chloride, and arsenic
range between 10"' to 10"4 which indicates that ambient concentrations
were under-predicted by the model.
The following may explain at least part of the discrepancies found.
Carbon tetr'achloride is extremely persistent in the atmosphere, with a
half-life of approximately 40 years, and has globally accumulated in the
ambient air. Thus, ambient concentrations are much greater than can be
accounted for by present emissions data used as input to the model.
Chloroform is thought to be emitted in large quantities from such non-
traditional sources as swimming pools and sewage treatment plants. The
District's toxics emissions data does not currently include these types
of sources. Likewise, vinyl chloride emissions from landfills have not
been adequately quantified and a default value of one pound per year was
assumed for modeling purposes. The discrepancies between ethylene
dibromide and ethylene dichloride ambient measurements and model
predicted concentrations may be due to the increases in measured
concentrations resulting from entrainment and out-gassing from the
ground. Concentrations of arsenic were also predicted to be lower than
measured results and may be a result of the contribution from soil dust
or that other sources of arsenic emissions may not have been included in
the emissions data. .
Additional emissions inventory efforts may resolve many discrepancies
between measured and model-predicted concentrations. The model's
treatment of carbon tetrachloride's persistence in the ambient air could
also be revised.
Estimates of the upper-bound lifetime number of cancer cases in the
Basin based on the measured, model-predicted, and literature survey
annual average concentrations are presented in Table VI-3. The
differences in concentrations should be considered when comparing the
three sets of cancer estimates. Since the model under-predicted ambient
concentrations for several pollutants, the risk estimates based on
measured concentrations may be more representative upper-bound
estimates.
In either case, the relative risks of the different pollutants are
easily discerned. Again, benzene and hexavalent chromium contribute the
greatest number of cancer cases to the total estimate. In addition,
existing ambient concentrations of formaldehyde may result in a
relatively large number of excess cancer cases in the Basin. The
relative importance of the other pollutants is apparent from Table VI-3.
VI-10
-------
TABLE VI-3
ESTIMATION OF LIFETIME (70 YEAR) UPPER-BOUND CANCER
CASES ASSOCIATED WITH AMBIENT CARCINOGENS IN
THE SOUTH COAST AIR BASIN
AIR TOXICS
ORGANIC GASES
Benzene
Carbon Tetrachloride
Chloroform
Dioxins and Furans
Ethyl ene Di bromide
Ethyl ene Di chloride
Formaldehyde
Methyl ene Chloride
Perch! oroethyl ene
Tri chl oroethyl ene
TRACE METALS
Arsenic
Beryl 1 i urn
Cadmi urn
Chromium
Nickel
TOTAL
AMBIENT
MEASURED '
6930
99
91
*
26
*
*
562
41
23
105
6
34
7560
26
15500
DATA BASES
MODEL
PREDICTED
3860
0.1
0
*
0.5
0.5
*
236
30
*
0.01
0.02
67
7120
6
11320
LITERATURE
#
#
#
20-400
#
#
2000
"; #
I
#
#
#
#
f
#
* No data available
# Basin-specific data not available
VI-11
-------
-------
CHAPTER VII
ASSUMPTIONS AND UNCERTAINTIES
Many assumptions and uncertainties are associated with the
quantification of cancer risk as a result of community exposure to
routinely released toxic air pollutants. In deriving a value for the
potency of a carcinogen (i.e., the unit risk factor) and applying that
factor to calculated cancer impacts, the following assumptions, which
inject a considerable degree of uncertainty into the analysis, are made:
o The response of humans to the substance is qualitatively and
quantitatively the same as in test animals;
o The effects- of the substance at a very high dose can accurately be
extrapolated to a very low dose by mathematical models containing
assumptions on the relation between dose and response;
The routes of exposure do not
quantitative results of the test;
affect the qualitative or
o All of the substance which is inhaled is absorbed into the body;
and
o An average person weighs 70 kg and breathes 20 m3 of air per day.
As a result of these assumptions, unit risk factors are considered
plausible, 95-percent, upper-bound estimates, i.e.; the risks are not
likely to be higher, but could be considerably lower. However, for
known human carcinogens,. CAG usually presents a most likely estimate,
not a 95-percent, upper-bound value. Because CAG has currently
characterized only 55 substances as to their carcinogenic potency, risk
estimates for most of the hundreds of chemicals present in urban ambient
air cannot be calculated.
In addition, there are several assumptions which relate to the
quantification of exposure and dose which can cause the risk analysis to
either overestimate or underestimate the cancer impact. Risk
assessments assume that people are exposed to the estimated
concentrations for 70 years, 24-hours a day. This is an overestimate of
the lifetime of most emission sources. In addition, most people change
homes and move around during each day. Population growth estimates are
also often not sufficient to quantify the 70-year exposed population for
calculations of the number of excess cancer cases. However, these
assumptions provide consistency in comparing relative risks between
different sources and can be used to ensure that an individual source
does not incur more than a standard amount of risk per unit of time.
At present, risk assessment methods for carcinogens in the ambient air
assume that indoor concentrations are the same as outdoor. If outdoor
concentrations do not penetrate completely indoors, then estimates of
VII-1
-------
risk have been overstated since more time is spent indoors.
Additionally, indoor sources of air pollutants are not addressed.
Certain pollutants may be present indoors at much higher concentrations
than outdoors and may make a significant contribution to the estimated
risk associated with exposure to air pollutants.
Another assumption made is that all risks are additive, even though
certain combinations of exposures may have synergistic (greater than
additive) effects, antagonistic (less than additive), or other types of
interactions.
There are risks that cannot yet be quantified using exposure models.
These risks are from exposure to compounds formed in the atmosphere
(e.g., formaldehyde). Literature data indicate that these risks may be
significant. Other chemicals may be transformed to less potent species
in the . air (e.g., reduction of hexavalent chromium to trivalent
chromium) and an overestimate of the risk would result.
While these assumptions and ensuing uncertainties must be considered in
evaluating results of this type of assessment, it is currently the best
available technique to estimate the magnitude of the risks and has been
employed by many agencies for regulatory decisions. The assumptions
used are intended to be health protective, yet have some bearing on
reality. ,
VII-2
-------
CHAPTER VIII
CONCLUSIONS AND RECOMMENDATIONS
An urban air toxic exposure and risk assessment model has been developed
and applied to the Basin. The technical approach for application of
this model can be used:
o To determine the magnitude of areawide risks and excess cancer
cases associated with toxic air pollutants emissions;
o To evaluate potential impacts of criteria pollutants control
strategies .on toxic air pollutants;
o To develop and prioritize a toxic air pollutants control program;
o To evaluate potential impacts of proposed new and modified sources
of toxic air pollutants emissions.
The assumptions built into the model limit its application in
interpreting the risk and excess cancer cases estimates. Some
assumptions lead to a potential underestimation of the risk to the
population, while others result in an upper-bound estimate of the cancer
risk. An understanding of these assumptions is needed in evaluating the
uncertainty associated with the estimated risks.
Even with the uncertainties in the modeling approach, the results can be
used to indicate the relative importance of. the individual carcinogenic
species and the relative contribution of individual source categories to
the total risk from a specific carcinogenic pollutant.
Results of this study show that of the carcinogenic pollutants
evaluated, both measured and model-predicted ambient concentrations of
benzene and hexavalent chromium have the greatest potential impact on
the Basin's population. Calculations of risk based on literature data
for ambient formaldehyde concentrations indicate a relatively
significant potential impact in the Basin from this pollutant. The
total number of lifetime excess cancer cases estimated from either
ambient concentration data base is approximately 20 to 30 percent of the
50,000 lifetime cancer cases expected in the Basin if approximately two
percent of all cancer cases are due to environmental pollution.
Recommendations for refining this methodology would include:
o Reduce the limitations of the model's application by enhancing the
ability to treat population mobility, different microenvironment
exposures, and multiple pathway exposures;
o Maintain and upgrade toxic emission inventory efforts on a routine
basis to characterize all sources of selected toxic air
VIII-1
-------
pollutants, including both permitted and non-permitted point
sources, and motor vehicle sources;
o Develop analytical techniques for the sampling and analysis of
selected ambient air toxics and for quantifying emissions from
existing sources; and
o Maintain District's ambient monitoring networks for the selected
gaseous organics and include ambient toxic metal compounds as
we! 1.
VIII-2
-------
REFERENCES
ARB, 1985. Letter from H. Wong-Woo, Deputy Executive Officer, State of
California Air Resources Board to J.A. Stuart, Executive Officer, South
Coast Air Quality Management District, August 22, 1985.
ARB and DOHS, 1986. Staff Report: Public Hearing to Consider the
Adoption of a Regulatory Amendment Identifying Chlorinated Dioxins and
Dibenzofurans as Toxic Air Contaminants. Release data: June 6, 1986
Adopted August 21, 1986.
Anderson, 6.E., C.S. Liu, H.Y. Holman and J.P. Killus, 1980. "Human
Exposure to Atmospheric Concentrations of Selected Chemicals." Systems
Applications, Inc., San Rafael, California, 1980.
Anderson, G.E. and G.W. Lunberg, 1983. "User's Manual for SHEAR: A
Computer Code for Modeling Human Exposure and Risk from Multiple
Hazardous Air Pollutants in Selected Regions." Report prepared for
EPA/OAQPS, 1982.
Czuczwa, J. and R.A. Hites, 1984. "Environmental Fate of Combustion
Generated Polychlorinated Dioxins and Furans," Environ. Sci. Techno!.
18(6); 444-50. ~ '.
Doll, R. and R. Peto, 1981. "The Causes of Cancer: Quantitative
Estimates of Avoidable Risks of Cancer in the United States Today,"
Journal of the National Cancer Institute^ June 1981.
EPA, 198S. The Air Toxics Problem in the United States: An Analysis of
Cancer Risks for Selected Pollutants. May 1985.
Federal Register, 1986. Guidelines for Carcinogen Risk Assessment, Vol.
51, No. 185 pp. 33992-34067, September 24, 1986.
Grosjean, D. 1982. "Formaldehyde and Other Carbonyls in Los Angeles
Ambient Air," Environmental Research and Technology 16mi 254-262.
Grosjean, D., R.D. Swanson, and C. Ellis. 1983. "Carbonyls in Los
Angeles Air: Contribution of Direct Emissions and Photochemistry." The
Science and the Total Environment 29: 65-85.
Liu, C.S., S. Van, G.E. Anderson, and G. Lunberg, 1986. "Urban Air
Toxics Exposure and Risk Assessment: A Modeling Approach." In-
Proceedings 79th APCA Annual Meeting. Minneapolis, Minnesota, June 22-
27, 1986.
Roberts, E., 1985. South Coast Air Quality Management District, El
Monte, CA. Personal communication, August 26, 1985.
R-l
-------
Salas, L.J., and H. Singh. 1986. "Measurements of Formaldehyde and
Acetaldehyde in the Urban Air," Atmospheric Environment 20f-6): 1301-
1304.
U.S. Bureau of the Census, (1981), "Census of Population and Housing,
1980: Summary Tape File 1, Technical Documentation." Prepared by Data
User Service Division, Washington, D.C.
U.S. Bureau of the Census (1980), "Geographic Base File/DIME (GBF/DIME),
1980: technical documentation." Prepared by Data User Service Division,
Washington, D.C.
Versar, Inc. 1984. Hazardous Air Pollutants -- Air Exposure and
Preliminary Risk Appraisal for 35 U.S. Counties. Prepared by Versar
Inc. for the U.S. Environmental Protection Agency, Office of Policy
Analysis, Washington, D.C., Contract #68-01-6715, September 1984.
Zwiacher, W.E., et al., 1983. "Emissions of Potentially Toxic/Hazardous
Air Contaminants in the South Coast Air Basin," South Coast Air Quality
Management District, September 1983.
Zwiacher, W.E., L.D. Yuhas, J.L. Whittaker, J.S. Fakhoury, W. Rogers, R.
Olivares, E. Sunico, and S.M. Weiss, 1985. Emissions of Potentially
Toxic/Hazardous Air Contaminants in . the South Coast Air Basin. 1984
Update. South Coast Air Quality Management District, September 20, 1985.
R-2
-------
APPENDIX A
SPATIAL DISTRIBUTION OF POINT SOURCE
AIR TOXICS EMISSIONS
-------
-------
k
18 2
55 S5
3NIH1HON Wlfl
IB S
O
CO
Di
0
S
o
» 1
CO
CO
-r
S o
P £ OC
»
0
1 1
o
2
Di
8-
CO
A-l
-------
§
a
__ _\
7
j:
UJ
UJ
M
GO
U_
O
CO
O
tt
CO
CO
3-3
«w i *"
O I i I
V) =5
< CJ
UJ >
o
co
o
Q_
O
O
O2
CO
Q_
CO
S5
m
ft
C«H) 3NIH1MON Win
A-2
-------
4i
3?
1
3NIH1MON Hin
II K K
CO
_Q
9
Q
3
9
UJ
CD
O
CO
O
»«
CO
CO
m en
5? IS
9
r co
o
a.
o
o
14
H-
CD
»i
ce
co
IH
a
o_
co
A-3
-------
I
3NIH1HON Win
A-4
-------
C«H) 3NIH1HON
A-5
-------
^"^n
X
i I 8
<»M) 3NIH1HON Hin
8
ss
Q
9
o
to
o
t «
CO
CO
/
in
*
co
o
Q_
U_
CO
m
CO
Q.
co
A-6
-------
-------
Q
3
t.
>S.
w
UJ
CD
5
O
a
UJ
UJ
UJ.
O
CO
O
a.
u.
o
CD
»i
CK
cn
t«
o
R
K
(«M) 3NIH1HON HJJ1
A-8
-------
C»M) 9NIH1HON H1H
A-9
-------
X
-I
L/
2
JtS 9) 9)
SNIH1HON
2
^
CO
_o
a:
LL!
o
en
C3
HI
CO
CO
o «t LLJ
~ iu °
is « o
o
Q_
O
O
»«
f
Z3
m
^<
o;
CO
i«
o
Q_
CO
A-10
-------
3NIH1MON HJJ1
A-ll
-------
§ S 6
(«M) 3NIH1HON Win
A-12
-------
(«M) 3NIH1HON Hin
A-13
-------
3NIH1HON NIP
A-14
-------
C«H) SNIHIHON Win
A-15
-------
3NIH1HON Hin
A-16
-------
(«M) CNIH1HON Win
A-17
-------
8
ft
3
g: £
3NIH1MON Win
in
£
A-18
-------
C«H) 3NIH1MON Wlf!
A-19
-------
-------
APPENDIX B
SPATIAL DISTRIBUTION OF MODEL-PREDICTED
AMBIENT CONCENTRATIONS
-------
-------
CO
UJ
c
CO
ce
UJ
CO
«c
en
CO
o
o
CO
cn
Lu-
en
LU
dl
Lu
CD
B-l
-------
ca
»«
i
in
S7.
in
!
CM
CM
I
I
i
in
C9
CO
O
esj
ca
' UJ
C£.
CJ
14
u.
Q
LU
LU
UJ
CD
LU
CD
OH
LU
co
03
co
o
o
en
Q
LU
I
O
LU
or
Q_
UJ
Q
O
Win
B-Z
-------
E
J£
O7
UJ.
ro
i
CD
UJ
CD
H^
U.
UJ
i« CO
i
-c o
US. CO
C3
UJ
ca
UJ
O
ONIHiHON NiO
B-3
-------
&
co
or.
CQ
tn LU
"C Q£
LU j
5'
<
CD
CD CO
«c -=c
cc: CD
LLJ C_5
CD
CO
CD
LU
ce
LU
a
o
3NIH1MON HIP.
B-4
-------
LU
03
UJ
S cc
o <
< CQ
I i !
CD CO
CD <
CC CD
CD
LU
CO
UJ
CD
LU
a
Q_
LU
O
) 3NIH1MON Win
-------
CO
CD
UJ.
o
CD CO
U_ -<
CD CD
CD CC
_I <
LU CO
< CD
ce o
i O
CO
: LU
O
t 1
Q
LU
ce
Q-
O
CmM) 3NIH1HON N10
B-6
-------
Ctt
CD
CO
CD
O
LLJ
("00 3NIKLHON Win
B-7
-------
CO
I
ea
LU
CO
CD
LU
CD
LU
CD "Z.
CC i<
CO CO
i< -C
Q CD
uj ce
CO
<
CD
3 ce CD
S LU CD
U. > CO
LU
o
ti
a
LU
a
-o
£
8
fe § i
CmM) ONIHiHON
b-8
-------
CO
-<
CCS
cc
CC-
B-
S3
co co co co
3NIH1MON NiH
B-9
-------
co
-=:
ct:
LU
i -c:
CQ ct:
co-
cci
cc
ti
CO
CD
CJ
CO
cc:
o_
LU
CD
CD
3NIH1HON Win
B-10
-------
CO
CD
!
CD
C3
>- CO
rD m
CJ
o- fv*
cn
CC CD
LlJ C_J'
cc
Q
LU
C3
UJ
cc:
Q_
LU
Q
CD
HiH
B-ll
-------
CO
G
LU
G
O
2: cc
CD -C
Ce CD
CD
ce
IE
LU CO
s: <
G
I iI c_>
of ^
LU ZD
> G
-<: co
i LU
O.
LU
LU
CC
Q_
LU
) 3NIH1HON Win
B-12
-------
CO
ce:
CO UJ
CD
or.
CO
O'
cc
cn
UJ
cc
o..
CD
CD
) ONIHiHON Win
B-13
-------
3
I
in
in
!
CM
CM.
I
*
tn
CO
o
I
ce:
i
z.
LU
O
LU
LU
CO
1
£
g
t t
J_
tn
uu
g
ID
<*
rf
1
CQ
UJ
ce.
'IS
l«4
u.
LU
CD
rv*
C3
I
o
OS
I. t
ULJ
Q_
LU
ce
LU
<
CQ
ce
1 -«
-<
CO
«c
o
o
x:
1
CD
CO
LU
LU
ce
Q_
o
3NIH1HON Win
B-14
-------
3
?
IS
i
m
>
CM
I
CD
3
in
is
1 :::
en
CD
LU
=D
CD
LU
CD
<
cc
UJ
CD
CO
-c-
o
CD
co
LU
C3
LLJ
ce
Q_
LU
CD
CD
3NIH1HON WiP.
B-15
-------
I
O3
IT5
CM
CM
1
!
in
o
CL:
O
CJ
LLJ
LU 2:
^_^ ^" '
az co
C5 <
__ ' CD
in
c_> ce
i i K i
CE: <
i
! I
T-. CO
» -<
^"1 t *
» <_>
V 1
ce: co
LU
LU
en
LU
Ccl
o_
LU
o
o
CmH) 3NIH1HON Win
B-16
-------
8
co
UJ
CJ
rn co
i -c
UJ -CD
o
cc cc
i i- en
o: <
i" o
CJ
UJ
ce:
UJ
CO
-< t
z. -z.
2: t
-ar
C3
UJ
CJ
LU
a:
o_
UJ
a
3NIHM3N Win
B-17
-------
CO
cc.
CD
(_>
LU
ce:
o
CO
cri
cc:
oo
>- t
2: co
a
CJ
ce
LL.
CD
CO
UJ
en
LiJ
Q
UJ
Q_
LU
C3
B-18
-------
CO
o
LU
O
O
O
CO ^
UJ
_J
X
LU
r* cc
' LU
CD --»
CO
CQ
CO
o
3:
CD
CO
LU
LU
1
O
LU
ct:
a.
LU
CD
C3.
Will
b-19
-------
-------
APPENDIX C
SPATIAL DISTRIBUTION OF INDIVIDUAL CANCER RISKS
AND NUMBER OF EXCESS CANCER IN THE SOUTH
COAST AIR BASIN
-------
-------
CO
-=r
CD
ct:
en
UJ
£
r-* ^-» ^» ^»
CO CO CO CO
ONIHiaON Win
< CO
I « <
C_J CD
CO
co
co co
LU
ce.
CS
a
.CO
ct:
.
CD LU
Q± CD
Q_
Q_
LU LU
i ce:
CJ ^5
i t CO
a o
LU D_
cc: x
D_ LU
C-l
-------
ca
LU
LU
LU
CO
ad
ce
<
i
-< CO
>< -e:
O CD
CD f_>
co-
co z:
-<: i
is: CD
co co
ce LU
i i
<
zn :z
a i<
ii
> LU
a aa
o 3:
CD LU
cc oa
LU 2:
o_ 1 CO
Q O
LU Q_
CC X
Q_ LU
CD
3NIH1HON Win
c-2
-------
LU
g
I
CB
-a
[
s
J-
8
! CO
I CO
O
C_> C-J
CD
co a:
CO t
o
is= CO
CO
11 UJ
ce a:
t2 oo _j
cs
. ' ct:
t*t 11 *
^ OQ
C3 I
CD ~Z.
\ UJ
f>^ tt
UJ CO
o_ s:
Q_ -<
C3 CO
UJ CD
CC Q_
Q_ X
UJ
SB
ID
9°
1
in
S
3NIH1HON Win
C-3
-------
LJJ
CO
H- CQ
f<
:se o:
ti
f^*\ «^^
LU
I I
-C CO
It -- m
LU rs:
Q_ -<:
Q_
a
LU LU
i cr:
CJ =3
ii CO
C2 C5
LU Q-
ce x
Q_ LU
LU
a
C-4
-------
2
s
s;
i
±{
Q o
03 2 =i
CD
CD
cc:
LU
Q_
Q_
cc:
CD
CD
cJ
CD
LU
Q_
CD
CD
eg
CO
D_
X
) ONIHiHON Win
C-5
-------
U_ CO
1< <
_i CQ
a: cc:
a en
CD I
co rs
CO CD
-C CO
:s: LU
co 3:
ii i
cc
CD
i«
U_
G
CQ
LU LU
Q_ <.
Q_ CO
cn
LU O
CD Qi
LU Z3
CC CO
a_ CD
o_
i X
LU LU
a
CD
C-6
-------
UJ
s:
»t
i
UJ "Z.
U_ ii
ii CO
; -c
-T- C°
P1 QC
UJ CO
_
O
CO
CO
CO
CO
11
a:
Ci
= 5
o
cp
'ce
UJ
o_
Q_
LU
co
G
o_
X
UJ
a
CD
("M) 3NIH1HON
C-7
-------
H-
£5
i?
!
or.
rf a
=r: d
C5
ri
ffl
.2
X, ^
or-
LU-
ca
Qr-
CD I
u: co
i -<
-c: c;
CO-
cc
CO
ce:
^i
-5
CD
CO
i
«>
§5
CD
CQ
CO
LU
Q
ii
ce:'
LU
s:
LU
rv
UJ
Q_
Q_
UJ
CD
CD
UJ
CJ- U.J
CD
LU
ce:
o_
CO
CD
Q_
X
CD
C-8
-------
CD
UJ
u_
CD
LLJ
CO
co-
cri
co
<:
CD
UJ
on
co CD
i< CO
CE:
a
cn 2:
CD "Z.
CD UJ
I '*
ft" m
UJ 2:
Q_ «C
Q_
< CO
f^~\ f"^
UJ Q_
CE: x
Q_ UJ
UJ
a
CD
3NIH1MON NIP
C-9
-------
LU
i CO
n= co
CO
O CJ
CO
co or
CO f
CO
hi
OS
CO
LU
=3
u.
a
CD
t
cc
LU
Q_
Q_
LU
O
ce:
o
ce:
LU
Q_
CD
LU
LU
ce
a_
LU
a
o
LU
ce
CO
C3
D_
X
LU
ffi
c*^- r^ r*» c**
3NIHiyON Win
C-10
-------
UJ
s:
I i
LU
nr CQ
i
i CC
en
LU i
i co
-< -<
CD
co 3:
CO i
< :=}
CD
^Z CO
CO
ce re
i i LU
^- 2:
1 I 1 I '
en _i
i- ir:
CQ
o_
Q_
a i
UJ
i LU
o ce
i =>
Q CO
LU c;
QC Q_
Q_ X
LU
a
C«M> ONIHiHON NIP.
C-ll
-------
Q; 14
.^ =
I
83
\
7
JL; jr>
wj OO
!"*» !*
CO CO
en to
CWii)
ro en
win
^
g
g
Lu'
re co
i -c:
ii CO
cc:
CD <
LU <
1
<; i
>i co
(_; -<
CD CD
CO O
CO
CO CD
T< CO
CM
-J 3=
C f
07
B'
C3
1-71
i i
cn
CD
CC _i
Q_
Q_
CD
LU UJ
i ce
c;- =D
i < CO
CD. CD
LU Q_
CC X
Q_ LU
CD
CD
$
m 05
C-12
-------
CD
LU
a
CO
CO
CO
CD
ce:
H1
<
I
CO
CO
LU
CO-
CO
cc
c;
CO
LU
CJ I
LU
UJ C->
0£ X
ZD LU
O
CQ
f
ce:
LU
Q_
Q_
LU
LU
LU
LU
CD
CO
CD
Q_
X
it LU
cn s:
I I ! I 1
ce i
Q_ LU
LU
cn
CD
3NIH1HON NiH
C-13
-------
F= o
fr-4
p* rv*
UJ Q
I <
co z
s
"2
«-M
ca S
a
-z.
CD >-i
a
r
^^.
--of
^
§
cn
i
I
S
CO
Q -C
LU 02
O -C
O
CO f
CO CO
CO O
LU
O =5
CE: co
O LU
jg ^
s
2 LU
S CO
o y
X
UJ LU
=« .^> LU
U. O OQ
'. O
C£. \
Q_ LU
CO
^^j fT^
LU Q_
I X
O LU
C3
LU
LU
LU
_J U_
LU i
Win
cn
m tn
cn cn cn
C-14
-------
r~i CO
LU -<
1 CD
<
i ce
CJi ii
CO
CO t
< co-
co o
Ic! CJ
CO
< 3:
<_> I
ce o
LU CO
un z
75
°co
CO
UJ LU
OS CJ
=> X
CD LU
U_ Q
CD
OD
t
t±
LU
Q_
=D
CD
LU
CJs
LU
Q_
_]
LU
CD
CD
LU
ce
CO
CD
Q_
X
LU
LU
3NIH1HON WiP.
C-15
-------
-« -J.
2
cd
I -C
CO 35
a
7 §
S 5
V
£
in
in
in in
S K
en on
S <£
CO
OS
cn
co-
co
CO
*2
' M . _
l_- CD
co or
UJ C_i
o <
x ct:
LU (
CD
CQ
CC
C5
cn
ce:
Q_ CD
a
LU
Cj
IH
CD
LU
a
CO
CD
a..
x
LU
LL.
CmM) SNIH1HON NIP.
C-16
-------
8 £
8
C""* f^*» c*^
3NIH1HON W1P.
8 S s
ii CO
3: -C
CD
CTr
LU CC
CJ
G
co
CO
CO IE
UJ I
co ID
-C G
C_j CO
CC LJ_
LU a;
CJ I
1*^
o
OS
=5
C3
CO
CO
X
LU
cn
^
ID
ci
CD
!
CC
LU
Q_
Q_
LU
LU CO
1 CD
Ct_
LU
LU
CC
o_
LU
C-17
-------
LU
CD
CO
CO
CO
<
GO
ce
CO
o
CO
UJ
CO
o
CD
CO
LU
E
-Si
LU 2:
;3 <. UJ
o
%> LU
CO
CO
a-
CD
GQ
i
ce
UJ
o_
Q_
LU
C3
LU
CC
Q_
O
o:
OQ
LU
CD
UJ
CO
o
Q_
X
LU
LU
O LU
21 U_
^
OT5
-------
§
e g
o
i i
a
-i a:
7
\
^^
s
CO
CD
Q£
CO
CJ 0
CD
CD
CO CO
LU
CO UJ
-C 31
<_> I
LU
CD O*
5*1' »*H
LU
O
m^ *
S
O O
X i
a LU
z 2:
=3 LU
C3 I
QD >-
LU LU
Q_
Q_ O
r*^ t i f
LU ce
I ^D
O CO
ii CU
a o_
UJ X
Q± LU
Q_
LU
LU i«
a i
en
8
en
s s
HiH
C-19
-------
en
CO
cd
CO
CO
U.J
CO
CO.
o
C_>
O
o^
-------
CD
CO
CO
CO
LU
CO
CO
CD
ce:.
I !
-<
CO
<
CD
CD
CO
LU
ce:
UJ
t-1 -C
CM 0
I
O CO
CO
1. 1 LU
£ o
=3 X
CS 'UJ
CD
CO
cc
LU
'o_
LU
o
ce
Q_
LU
ca
CD
LU
LU
O
ce:
CD
ce
LU
Q_
CD
LU
on
CO
CD
Ct_
X
LU
LU
C-21
-------
a: co
i -c
ii CD
3=
CC
CD i«
CO
0£
o
CD
CO
CO
CO CO
O LU
CO LU
CO _!
LU >-
O IT
X 1
LU LU
CD
en ct:
z. o
CD 3:
CD O
cb cc
d LU
LU CC
CJ CO
a o_
LU x
ce LU
Q_
O LU
rs: LJ_
ob gi co S
CWH) 3NIH1HON WiH
C-22
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-450/4-88-013
2.
3. RECIPIENTS ACCESSION NO.
TITLE AND SUBTITLE
South Coast Air Quality Management District
Multiple Air Toxics Exposure Study, Work Paper
No. 3 '. r;
5. REPORT DATE
November 1988
6. PERFORMING ORGANIZATION CODE
AUTHOR(S)
South Coast Air Quality Management District
8. PERFORMING ORGANIZATION REPORT NO.
PERFORMING ORGANIZATION NAME AND ADDRESS
South Coast Air Quality Management Division
9150 Flair Drive
El Monte, CA 91731
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
2. SPONSORING AGENCY NAME AND ADDRESS
Air Quality Management Division
OAR, OAQPS, AQMD, PCS (MD-15)
Noncriteria Pollutant Programs Branch (MD-15)
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
S. SUPPLEMENTARY NOTES
EPA Project Officer; James H. Southerland
6. ABSTRACT
The South Coast Air Quality Management District of California has completed a
Multiple Air Toxics Exposure Study (MATES) which examines the additive risk from a
number of air toxics on an urban area. This project, though partieally funded by
EPA, is an example of how a State or local agency may approach assessing their
local air toxics risks as is encouraged by EPA's Urban Air Toxics Program which
results from EPA's Air Toxic Strategy. -This report is a summary of the methods
used by this California agency. Though not intended as an endorsement of the
entire contents of the report, EPA is reproducing their report (working paper
number 3) to benefit and encourage other agencies which may be contemplating such
an assessment.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (TllisReport)
21. NO. OF PAGES
132
20. SECURITY CLASS (Tllispage)
22. PRICE
EPA Form 2220-1 (R«v. 4-77) PREVIOUS EDITION is OBSOLETE
-------
------- |