Assessing Multiple Pollutant Multiple Source Cancer Risks From Urban Air Toxic Summary of Approaches and Insights from Completed and Ongoing Urban Air Toxics Assessment Studies


                           EPA-450/2-89-010
                           April 1989
ASSESSING MULTIPLE POLLUTANT

MULTIPLE SOURCE CANCER RISKS

     FROM URBAN AIR TOXICS
     Summary of Approaches and Insights From
           Completed and Ongoing
       Urban Air Toxics Assessment Studies
   U. S. ENVIRONMENTAL PROTECTION AGENCY
           Office Of Air and Radiation
      Office Of Air Quality Planning And Standards
      Research Triangle Park, North Carolina 27711

-------
This report has been reviewed by the Office of Air Quality Planning
and Standards, U. S. Environmental Protection Agency, and has been
approved for publication.  Mention of trade names or commercial
products does  not constitute endorsement or recommendation for use.
                      EPA 450/2-89-010

-------
                         TABLE  OF CONTENTS
EXECUTIVE SUMMARY	i
       Background	i
       Purpose of Report	iv
       Report Organization	iv
       Report Emphasis	v
       The Changing Nature of Urban Assessments	vii
       References	viii
Chapter 1. An Introduction to Urban Air Toxics Studies	1
       1.1     Defining the Urban Air Toxics Problem	1
       1.2     Summary of the Completed and Ongoing Urban Air Toxics
              Assessments	2
                 The Six Months Study	3
                 The Integrated Environmental Management Project (IEMP)	..4
                 Clark County Study	5
                 South Coast—MATES	6
                 5 City Controllability Study	....8
                 Motor Vehicle Study	..9
                 Integrated Air Cancer Project (IACP)	11
                 TEAM (Total Exposure Assessment  Methodology) Studies	12
                 Staten Island/New Jersey Study	14
               -  Southeast Chicago Study	14
                 Urban Air"Toxics Monitoring Program....	'.	15 .
                 Bay Area Toxics Monitoring Study (BAAQMD, 1987)	16
       1.3     Other Urban Air Toxics Assessment Activities	17
       References	:	20
Chapter 2. Ambient Air Quality Monitoring	23
       2.1     Use of Ambient Air Monitoring Data in Multi-Pollutant, Multi-
              Source Urban Assessments	23
       2.2.    Decisions Affecting Monitoring  Plan Development	25
                 Studies  with Stated Goals for Air Quality Monitoring  Programs	27
       2.3.    Pollutant Coverage	33
                 Trends Toward Broader Pollutant Coverage	38
       2.4.    Site Selection	39
                 Formal [Quantitative] Approaches to Site Selection	39
                 Informal [Qualitative] Approaches to Site  Selection	45
                 Physical Siting Criteria Common to Ail Studies..	47
       2.5     Sampling Periods, Frequency and Duration	48
                 Sampling Period	49
                 Sampling Frequency	51
                 Sampling Duration	52
       2.6     Sampling and Analytical Techniques	54
                 Volatile Organics	54
                 Semivolatile Organics	58
                 Metals	58
                 Aldehydes/Ketones	,	59
       2.7.    Evolving Monitoring Technologies.	59

-------
                 TEAM 59
                 The Integrated Air Cancer Project	61
                 Special Monitoring Techniques	62
                    TAGA®	.62
                    ROSE  62
       2.8    Insights into the Use of Monitoring in Air Toxics Programs....	.62
                 Cost-Saving Measures	62
                 General Issues	66
Chapter 3. Emission Inventories	73
       3.1    Use of Emission Inventories in Multi-Pollutant, Multi-Source
             Urban Air Toxics Assessments	73
       3.2    Pollutant Coverage..	74
                 Treatment of Polycyclic Organic Matter (POM)	80
                 Treatment of Chromium	,	81
                 Treatment of Beryllium and Nickel	81
                 Secondary Pollutants	82
       3.3    Source Coverage	,	82
       3.4    Estimating Emissions	83
                 Emissions Estimation from Existing Data Bases	83
                 Source-Specific Survey Data	84
       3.5    Spatial and Temporal Resolution	86
                    Grid Size and Grid Cell Resolution	86
                    Point Source Resolution	87
                    Area Source  Allocation	88
                    Temporal Allocation	*....	89
       3.6    Emission Inventory Quality Assurance	I.........:....	89
                 Emission Inventory Review.	90
                 Data Verification	90
               ,  Monitoring vs. Modeling  Comparisons	91
       3.7    Insights into Compiling Inventories for Urban Air Toxics
             Assessment	91
       References	;	94
Chapter 4. Dispersion Modeling	95
       4.1    Use of Models in Estimating Exposure in Multi-Pollutant, Multi-
             Source Urban Assessments	95
       4.2    Decisions Affecting Modeling Protocols	97
       4.3    Model Selection	103
                 Model Characteristics	103
                 Rationale for Model Selection	106
       4.4    Release Specifications for Emissions	110
       4.5    Selection of Receptor Network Array	Ill
       4.6    Meteorological Data	119
                 Treatment of Meteorological Parameters	120
                 Treatment of Meteorological Variability	122
       4.7    Decay, Deposition, and Transformation	124
                 Decay 124
                 Deposition	124
       4.8    Model Execution	125
       4.9    Model Performance Evaluation	127
                 Evaluating Model Performance	127

-------
                 Improving Model Performance	129
       4.10   Insights into the Use of Modeling in Air Toxics Evaluations	137
                 Modeling Techniques	137
                 Cost-Saving Measures	142
       References	'.	144
Chapter 5. Exposure and Risk Assessment	146
       5.1    The Use of Exposure and Risk Assessment in Air Toxics Studies	146
       5.2    Methodological  Issues	149
                 Exposure Assessments	149
                 Risk-Assessment	155
       5.3    Comparison of Approaches to Interpreting Exposure and Risk	161
                 Ranking the Relative  Importance  of Pollutants	162
                 Ranking the Relative Importance of Sources or Source
                        Categories	162
                 Evaluating Exposure and Risk Patterns	163
       5.4    Insights into the Use of Exposure and Risk Assessment in Multiple
             Air Toxics Studies	169
       References	171
Chapter 6. Control Strategy Simulation  and Evaluation	173
       6.1    The Use of Control Strategy Simulation and Evaluation in Urban
             Air Toxics Assessments	~..-.	173
       6.2    .Comprehensive vs. Site-specific Strategy Analyses	174
                 Extent of Source Coverag'e.,.1	175
                 Multiple vs. Limited  Control Options	175
               -  Inclusion of Source Growth and Retirement	."	175
       6.3    Control Strategy"Simulation Procedures in the 5 City
             Controllability Study	i...	176
                 RIM Operation	176
       6.4    Control Strategies  Evaluated in  5 City Controllability Study	178
       6.5    Assumptions Inherent in 5 City Controllability Study Control
             Analysis	179
       6.6    Control Strategy Evaluation  in  the lEMPs	180
                 Philadelphia IEMP Control  Analysis	180
                 Baltimore IEMP Control Analysis	184
                 Santa Clara IEMP  Control Analysis	185
                 Analysis of Areawide vs. Most Exposed Individual (MEI) Risk
                        Reduction	186
       6.6.   Insights on Control Strategy Simulation and Evaluation	188
Chapter 7. Computerized Data Handling	191
       7.1    Data Handling Considerations in Urban Air Toxics Studies	191
                 Study Type	192
                 Computer Accessibility	192
       7.2    Data Handling Aspects of Emission Inventory/Dispersion Modeling
             Studies	193
                 PIPQUIC	193
                 Other Data Handling Systems	199
                 Cost Saving Techniques	205
       7.3    Data Handling Aspects of Ambient Air Monitoring Studies	205
       7.4    Insights on Computerized Data  Handling	206

-------
Chapter 8.  Emerging Methods: Receptor Modeling and Biological Testing.....	210
       8.1    The Use of Receptor Modeling in Urban Assessments	210
                 Pollutants and Source Categories Addressed	....212
                 Source Signature Testing and Tracer Analysis.....	.212
                 Measured Ambient Air Quality Data Sets	213
                 Statistical Techniques Used to Estimate Apportionment	214
                 Spatial  and Temporal Representativeness of Results	215
       8.2    Comments on the Use of Receptor Modeling for Urban Air Toxics
              Studies	216
       8.3 -    The Use of Biological Testing in Urban Air Toxics Studies	217
       8.4    Comments on the Use of Biological Testing in Air Toxics Studies	219
                 Relative Importance of Gas Phase Pollutants	219
                 Indoor/Outdoor Differences	220
                 Use of Biological Testing to Set Priorities	220
       References	221
Appendix A. Glossary of Terms	A-i
Appendix B. Evaluation of Optimum Size for Air Toxics Monitoring Network
       Based on Spatial Correlations of Concentration	B-i
       Introduction	B-i
       B.I    Method	B-i
       B.2    Example Application:  Philadelphia IEMP	B-iii
                 Siting Analysis Based on Modeled Data	B-iii
                 Siting Analysis Based oil Measured  Data	„	B-iv
                 Method Evaluation	B-vii

-------
EXECUTIVE SUMMARY
Background

Increasing attention is being given to the problem of air toxics exposures in urban
areas. This attention initially focused on maximum exposures of individuals living near
large point sources. Lately, however, some attention has shifted to the potential for
widespread risks in urban communities resulting from smaller, more dispersed emitters of
air toxics.

"Recent Environmental Protection Agency (EPA) studies1 suggest that multiple
pollutant, multiple source exposures to air toxics may be causing 1,000 to 2,000 excess
cancer cases annually in the United States and average lifetime individual cancer risks
ranging from 1 in 10,000 to 1 in 1,000 (10~4 to 10'3) in urban areas. These studies further
suggest that air emissions from small point, area, and highway vehicle sources may be the
largest contributors to urban air toxics cancer incidence.

Under EPA's National Air Toxics Strategy (EPA, 1985), a major activity has been
the assessment of high risk urban problems. Numerous assessment efforts have been
initiated in the past five years or are under way in many U.S. cities to evaluate both the
nature and magnitude of the air toxics problem in urban environments. Some of these
studies are examining the potential for risk mitigation through various, control
alternatives. EPA, in carrying out the National Air Toxics Strategy, has been encouraging
urban areas to engage in these kinds of assessment activities. Moreover, EPA is promoting
1 Haemisegger, et al., 1985; Manale, et al., 1987; EPA, 1986; Hinman, et al., 1986; EPA,
1987a; Shikiya, et al., 1987; Shikiya, et al., 1988; EPA, 1987b; EPA, 1988.

-------
the coordination of State and local air toxics control programs with ozone (63) and
particulate matter (PMio) control programs to assure that, whenever possible, future
State Implementation Plans (SIPs) incorporate measures that reflect co-control of air
toxics. (EPA, 1987c) .

Current Assessment Approaches—Most urban air toxics assessments to date have
used either of two basic approaches: ambient air monitoring and emissions modeling.
Many, in fact, have used a combination of both approaches to minimize the weaknesses of
each.

The ambient air monitoring approach is conceptually simple. Measured ambient
levels of air toxics in an urban area are multiplied by cancer unit risk factors to calculate.
individual risks. (Unit risk factors relate specific probabilities of contracting cancer to
lifetime exposures to 1 ug/m3 of a specific pollutant.) Individual risks are then multiplied
by population data to calculate aggregate cancer incidence.. Because of data limitations,
such analyses are usually fairly crude. For example, some studies may assume that one or
several measurements of each, .pollutant represent long-term, .areawide population
exposures. Ideally, ambient assessments of urban air toxics risks .would be more
sophisticated, involving many more ambient samples at more representative times and
locations.

The second major approach involves dispersion modeling of emissions. In this
approach, an emission inventory is compiled for the air toxics of concern and is modeled
(using various long-term dispersion models) to predict ambient air concentrations over the
urban area. These modeled ambient air concentrations are then used to estimate individual
risks and aggregate incidence as if they were measured data. The dispersion modeling
approach requires a comprehensive emission inventory of point, area, and highway vehicle
sources.

There are advantages and disadvantages to each of these two assessment
approaches. The use of reliable and representative ambient air monitoring data avoids the
errors inherent in emission inventories and dispersion models, which can be considerable.
In addition, the ambient air monitoring approach allows one to handle secondarily formed
u

-------
pollutants that are not emitted directly (e.g., photochemically formed formaldehyde) or
pollutants  that exist  in  the atmosphere resulting  from gradual  global buildup of
background levels  (e.g.,  carbon tetrachloride)  or transport  from other urban areas.
(Ambient monitoring data can also be used to help verify modeling results.)

       Important advantages of the dispersion modeling approach are that it allows one
to predict risk reductions as a function of anticipated emission changes  and it allows
ambient air levels to be projected at many more receptors than may be possible in most air
sampling networks.  It also  allows one to handle pollutants (e.g., hexavalent chromium)
for which ambient sampling methods are not yet available or  cost-effective. Dispersion
models also allow one to estimate the impact of specific sources on particular receptors,
which is not possible with ambient monitoring data.

       Future Assessment Approaches—Alternative approaches  are being  developed for
assessing  the..urban air toxics problem,  the most notable being  EPA's Integrated Air
Cancer Project (IACP) and Total Exposure Assessment Monitoring (TEAM) studies.  The
IACP is being conducted by.the  EPA  research labs  in Research. Triangle Park, North
Carolina.   It  is a  long-term  interdisciplinary research  program  aimed at developing
scientific methods and data bases to identify the major sources of carcinogenic chemicals
emitted into  the air  or  arising from  atmospheric  transformations.   This project is
performing extensive  ambient  air  monitoring, receptor modeling, bioassays,  and
atmospheric transformation  studies to identify the principal airborne carcinogens and, to
the extent feasible, quantify the human cancer risk they may pose.  Thus far, the project
has concentrated on the  effects  of emissions from wood burning and mobile  sources.
 Subsequent phases of this project will add other residential combustion sources  (e.g., oil
 combustion) and industrial sources. (Lewtas, 1987)

        The TEAM approach (Ott, 1986) assesses exposures by  measuring pollutant levels
 in an individual's immediate physical environment throughout the day. In contrast to a
 fixed monitoring station  that collects pollutants as they  are carried by the ambient (or
 indoor) air to the location of the devices, TEAM monitoring devices are located on  vests
 worn by individuals during their daily routine.   In addition, TEAM studies measure the
 concentration of chemicals in an individual's  drinking water and exhaled breath to
                                         iii

-------
determine the actual amount of the pollutant that has been taken up by the individual.
Personal sampling directly addresses the variability of exposures within a population.
TEAM may best contribute to future urban assessments by improving the exposure
assumptions that are typically made in current assessments.

Purpose of Report

The primary purpose of this report is to assist State, local, and other agency
personnel by describing methods that have been used in assessing multiple source,
multiple pollutant risks from air toxics exposures in urban areas.

This report is a synthesis of assessment methods that have been made to date to
quantify multi-pollutant, multi-source urban air toxics exposures and risks. It does not
constitute formal EPA requirements for conducting an urban risk assessment of air toxics,
nor does it recommend a single approach. Instead, it identifies techniques that others have
elected to employ, and it offers insights tha't may assist the reader in selecting a particular
set of techniques for use in a given locale. In some cases, this report discusses problems
encountered in previous studies so that prospective study managers might learn some
lessons. A State or local agency will have to decide from the procedures which (if any) to
adapt in its particular urban assessment based on program goals, timing, and resources. -

Report Organization

Chapter 1 of this report presents a brief overview of the major studies completed
in the past several years that have assessed the multi-pollutant, multi-source urban air
toxics problem. This overview is provided because these studies are cited throughout the
body of this report. To a limited extent, the major findings of each completed study are
summarized to give the reader a sense of the available evidence that suggests the existence
of an urban air toxics cancer problem.

Chapters 2 through 7 describe specific activities that are common to urban air
assessments. These major activities are highlighted below:
• Ambient air monitoring;
IV

-------
• Emission inventorying;
• Dispersion modeling;
• Exposure and risk assessment;
• Control alternative simulation and evaluation; and
• Data processing.

For each of these activities, a synopsis is given in the respective chapter of the
procedures employed in the various urban assessments, and insights are offered on the
suitability of these procedures in particular applications. It should be noted that not all
of the above activities would necessarily be involved in a given urban assessment.

Chapter 8 summarizes various assessment methods that are currently under
development and may be employed in future urban assessments. These methods, which
include bioassays, receptor modeling, and personal monitoring, are being developed by
EPA research laboratories. Although not the focus of this report, it is useful to be aware
of these evolving techniques.

A glossary of terms relating to urban risk assessment is included in Appendix A
for those not conversant with the terminology used herein. Appendix B discusses a
method for representative monitoring site location relating to material discussed in
Chapter 2.

Report Emphasis

Cancer Risk—Cancer has been the major focus of urban air toxics studies to date.
As such, this report focuses primarily on techniques employed by various study managers
to assess cancer risks. These techniques generally involve the evaluation of long-term
(e.g., annual average) exposures to multiple pollutants. Techniques for evaluating specific
non-cancer risks associated with short-term (acute and subchronic) exposures are only
discussed briefly, reflecting the limited work done in this area. Note that this report does
not deal with the development of dose-response relationships and cancer unit risk factors.
It is assumed that potency data for different pollutants are available from other sources.

-------
Measures of Cancer Risk—Two common measures of excess cancer are incidence
and individual risk. Excess cancer incidence is a measure of the number of excess cancer
cases associated with air toxics exposures within an urban area, whereas individual risk is
the probability that an individual will contract cancer from a particular exposure.

There are many variations of each of these measures. For example, incidence can
be additive, covering multiple pollutants. Incidence can reflect areawide cancer cases or
can be specific to particular "hotspots," such as neighborhoods or industrial parks.
Incidence can reflect the number of cases expected in a single year or within a 70-year
"lifetime." Incidence is sometimes population-normalized, for example, adjusted per
million population. A commonly used measure in multi-pollutant, multi-source urban air
toxics assessments is annual, areawide excess cancer incidence.

La the same way, individual risk can refer to the cancer probability associated with
multiple or single pollutant exposures, the probability associated with the average
exposure within a broad area or with "the exposure at a particular receptor, or the
probability of contracting cancer in a single year or during a 70-year lifetime.

Screening Assessments—Most of the urban assessments completed to date are
best described as screening (or scoping) studies, performed to yield an order-of-magnitude
estimate of the relative nature of the urban cancer problem, rather than to provide
absolute predictions of incidence and individual risks. Screening studies are often
considered more credible for prioritizing subsequent assessment efforts than for defining
direct regulatory measures or for predicting risks to specific individuals by specific
sources and pollutants. If, for example, the screening analyses described herein would
point to a single point source or a cluster of sources as posing particularly high risks,
more detailed assessment techniques might be needed, especially in situations (e.g., permit
review or standard setting) where specific regulatory actions were contemplated.

Risk Assessment vs. Risk Management—Quantitative risk assessment is a process
by which a factual base of information is used to estimate the probability of incurring
some risk (e.g., cancer) due to exposure to a specific chemical or chemical mixture. In
contrast, risk management is the decision-making process by which some action is taken or
VI

-------
some policy is formed concerning a potential risk to the environment and/or to human
health. Risk management differs from risk assessment in that management of risk usually
considers political, economic, and social issues in the decision-making process. This
document mainly deals with the methods to assess urban air toxics risks, including the
projection of risk reductions associated with control alternatives.

The Changing Nature of Urban Assessments

There are myriad techniques, assumptions, and data involved in urban
assessments, many of which may change in the next few years. Cancer unit risk factors,
for example, are subject to significant, often order-of-magnitude, changes as new dose-
response data become available. Ambient air monitoring and source sampling techniques
for air toxics are evolving, allowing certain compounds to be measured accurately at
levels not previously possible. Techniques are also becoming available to evaluate
mutagenicity of specific urban air compounds-as they age and transform photochemically.

It is important to be aware of the changing nature of urban assessments for two
reasons. First, it means that-no-, two completed assessments have likely used the same
procedures, have made the same assumptions, or have used the same data in estimating
urban risk. If one examines the emission factors, unit risk factors, modeling assumptions,
pollutant coverage, etc., in the various studies, significant divergence is often apparent.
Hence, care needs to be taken when "relating results from one study to another to avoid
"apples and oranges" comparisons.

Second, the information presented in the following chapters will undoubtedly
change as the perception of the urban problem matures and as more assessment options
become available. Moreover, the underlying emission inventory and ambient air quality
data bases upon which such assessments are based, should improve with time. Thus,
those responsible for urban risk assessments should keep abreast of these changes and
reflect them as much as possible.
vu

-------
                              EXECUTIVE  SUMMARY
References
EPA,  1985.   "A  Strategy to Reduce Risks  to  Public  Health from Air Toxics," U.S.
Environmental Protection Agency, Washington, D.C.

EPA, 1986. Integrated Environmental Management Project: Philadelphia Phase n Report.
Regulatory Integration Division, Washington, D.C.

EPA, 1987a.  Kanawha Valley West Virginia Toxics Screening Study Report.
Environmental Services Division (Region HI) and the Regulatory Integration Division,
Philadelphia, Pennsylvania.

EPA,  1987b.  National Air Toxics Information  Clearinghouse Report;  "Qualitative and
Quantitative  Carcinogenic  Risk  Assessment,"  EPA 450/5-87-003, U.S. Environmental
Protection Agency and STAPPA/ALAPCO, Washington, D.C., pp. 11-1 through 11-6.

EPA,  1987c.  "State Implementation Plans;  Approval of Post-1987 Ozone and Carbon
Monoxide Plan Revisions for Areas Not Attaining the National Air Quality Standards;
Notice,"  Federal  Register,  Vol.  52, No. 22fi, Part II, Tuesday, November 24,  1987.
p. 45072.

EPA,  1988.  Preliminary results from 5-City Controllability Study being conducted  by
E.H. Pechan  and  Associates under  contract  to  U.S. Environmental Protection Agency,
Research Triangle  Park, N.C. (Tom Lahre, EPA project officer).

Haemisegger, E.,  et al.  1985.   "The Air Toxics Problem in the United States: An
Assessment  of Cancer Risks for Selected Pollutants," U.S. Environmental Protection
Agency, Washington, D.C.

Hinman, K., et al.,  19861   "Santa Clara Integrated Environmental Management Project.
Stage Two Report," U.S., Environmental Protection Agency, Washington, D.C.

Lewtas, J., and L.  Cupitt, 1987.   "Overview  of the Integrated Air  Cancer Project,"
Proceedings  of the 1987 EPA/APCA Symposium on Measurement of Toxics and Related
Air Pollutants, Research Triangle Park, N.C., pp.  555-561.

Manale, A, et al.,  1987.  "Baltimore Integrated Environmental Management Project.  Phase
H Report:  Air Toxics," (Draft) U.S. Environmental Protection Agency, Washington, D.C.

Ott, W., et al., 1986. "EPA's Research Program on Total Human Exposure," Environmental
International, Vol. 12, pp. 475-494.

Shikiya, M.,  C. Liu, E. Nelson, and R. Rapoport. 1987.  The Magnitude of Ambient Air
Toxics Impacts from Existing Sources in the South Coast Air Basin: 1987 Air  Quality
Management Plan Revision Working Paper Number 3. South Coast Air Quality
Management District, June 1987, El Monte, California.
                                       vm

-------
Shikiya, M., W.  Barcikowski, and M. I. Kahn.  1988.  Analysis of Ambient Data from
Potential Toxics "Hotspots" in the South Coast Air Basin.  South Coast Air Quality
Management District,  October 1988.
                                        IX

-------

-------
                                    CHAPTER i

                 AN INTRODUCTION TO THE URBAN AIR TOXICS
                                     STUDIES
1.1
The following subjects are covered in this chapter:

•      Defining the urban air toxics problem

•      Summary of completed and ongoing urban air toxics assessments

•      Other urban air-toxics assessment activities


Defining the Urban Air Toxics Problem
       The perception has been evolving over the last few years that an air toxics "soup"

exists over cities that is imposing an uncertain, but possibly significant, public health risk

of increased cancer. This concern has been described by the following terms:

       •      Urban hot spot problem

       •      Multiple source, multiple pollutant problem

       •      High-risk urban problem

       •      Urban soup

Whatever term one uses to describe the phenomenon, the concerns are basically the same:

       •      That traditional air toxics programs, which typically emphasize major
              point sources and single pollutant analyses, may not be addressing  the
              most significant cancer risks in urban areas;

       •      That criteria pollutant control programs,  which are known to effect some
              air  toxics reductions through  indirect  means,  may  not be sufficiently
              addressing certain air toxics problems;

-------
       •      That small point and area sources and road vehicles may be causing more
              of a problem than has been recognized heretofor in air toxics programs;
       •      That simultaneous  exposures to urban air  toxics  mixtures  from many
              sources can lead to significant aggregate risks in urban areas, risks that are
              not being realistically assessed  by current pollutant-by-pollutant risk
              assessments; and
       •      That byproducts of irradiated and aged urban air mixtures may be much
              more potent than "fresh" (i.e., point of release) emissions traditionally
              analyzed in some risk assessments.

       Thus, whether one calls the problem "urban soup" or something else,  the point of
conducting an urban air toxics assessment  is the same—to  shed light  on these concerns
and determine whether additional air toxics  controls are needed.

1.2    Summary of the Completed and Ongoing Urban Air Toxics Assessments

       This report is a compilation of procedures that are being employed for conducting
multi-pollutant, multi-source urban air toxics assessments.  The procedures described in
this report are derived from afhumber of urban studies either" completed in the last several
years or currently under way.

       The purpose of this  chapter is to introduce the  reader to the  studies alluded to
throughout the remainder of this report.  Hence, what follows is a brief discussion of the
purpose and approach of each study, along with an overview of its major findings.  As a
rule, a short descriptive title is assigned to each study (e.g., the "Six Months Study") so
that hereafter, for ease of reference, the full  title of each assessment need not be repeated.

       The reader is cautioned that all of the risk and incidence values  cited herein, while
appearing very precise and certain in some cases, are acknowledged to be very preliminary
and tentative by the respective study authors and by the authors of this report. The many
uncertainties, assumptions, and limitations inherent in the procediires used in  these
studies are documented in the remainder of this report, and should be kept in  mind before
using any of the following data and conclusions.

-------
The Six Months Study

This study is formally entitled The Air Toxics Problem in the United States: An
Analysis of Cancer Risks for Selected Pollutants. (Haemisegger, 1985) It represents EPA's
first comprehensive analysis of the air toxics problem and provides a basis for the
Agency's National Air Toxics Strategy. (EPA, 1985) The Six Months Study is regarded as
a "scoping" study only, useful in a relative sense to yield rough approximations of total
incidence and individual risks.

Three major analyses were undertaken as part of the Six Months Study to estimate
cancer incidence and individual lifetime risks. An "Ambient Air Quality Study" used
ambient data for 5 metals, 10 volatile organic compounds, and benzo(a)pyrene [B(a)P] to
assess risks. Two other analyses—a "NESHAP Study" and a "35-County Study"—used
emission estimates and exposure models to estimate incidence and maximum individual
risks associated with the pollutants selected. In these analyses, a "B(a)P surrogate"
approach was used to estimate cancer incidence associated with products of incomplete
combustion (PIC). To do this", a dose-response coefficient relating lung cancer incidence
and PIC was generated from epidemiological studies, and cancer incidence associated with
PIC exposure was estimated by applying this dose-response coefficient to B(a)P levels.
Finally, quantitative risk assessments available from other EPA activities for asbestos,
radionuclides, and gasoline marketing were incorporated into the study.

The Six Months Study shows that both point and area sources contribute
significantly to the air toxics problem. Large point sources are associated with high
maximum individual lifetime risks (commonly 10"4 to 10"2), whereas the additive lifetime
individual risk due to simultaneous exposure to 10 to 15 pollutants range from
10'4 to 10'3. The latter risks do not appear related to specific point sources, but instead,
are associated with the complex mixtures typical of urban ambient air. For the pollutants
examined, additive cancer incidence averaged about 6 excess cases per million people per
year.

This study very tentatively suggests that EPA's criteria pollutant programs have
significantly reduced air toxics levels. An analysis of 16 pollutants was completed using

-------
both monitoring and emission data in order to evaluate progress made on air toxics
between 1970 and 1980. The estimated cancer incidence rate for these air pollutants in
1980 was less than half that for 1970.

The Integrated Environmental Management Project (TEMPI

Under the Integrated Environmental Management Project (IEMP), EPA's Office of
Policy, Planning, and Evaluation (OPPE) conducted a number of urban (or "geographic")
studies to evaluate the multi-media aspects of various environmental issues, with general
emphasis on toxics exposures from drinking water, hazardous wastes, and air. The
geographic areas covered by the IEMP studies to date are Philadelphia, Pennsylvania
(EPA, 1986); Baltimore, Maryland (Manale, 1987); Santa Clara ("Silicon Valley"),
California (Hinman, 1986); and Kanawha Valley, West Virginia (EPA, 1987a). Another
IEMP is under, way in Denver, Colorado, at this writing.

The concept of integrated environmental management developed out of EPA's
recognition that there are potential drawbacks to the traditional approaches for developing
environmental regulations. Agencies have traditionally focused on individual industries,
pollutants, and media—thereby potentially limiting their ability to determine where,
among the various media, their resources could best be employed to optimize health
protection. Moreover, the traditional approach may not ensure that pollution controls are
not merely shifting risk from one medium to another. Thus, comparing the different media
risks to help set priorities allows environmental managers to focus limited resources in a
manner that will achieve the greatest public benefit.

The cancer assessment approaches used in each of the four IEMP studies
completed to date are fundamentally similar. First, each study modeled emissions data to
project ambient air concentrations. These projected concentrations were then distributed
over a population grid to estimate exposures, which, in turn, were multiplied by unit risk
factors to estimate individual cancer risks and incidence. Some ambient air measurements
were made as part of these studies to complement the modeling results and to help check
for errors and omissions in the inventory.

-------
1
The results of the lEMPs generally support the conclusions of the Six Months
Study. Excess.cancer incidence from exposures to 10 to 20 toxic air pollutants ranges
from about 1 to 18 cases per year per million persons. The relative contributions of point
and area sources vary, with area sources generally predominant (except in Kanawha Valley
with its numerous, large chemical complexes). A combination of metals (e.g., chromium),
polycyclic organic matter (e.g., B(a)P), and toxic volatiles (e.g., benzene) account for most
of the excess urban cancer incidence. Additive individual lifetime cancer risks to the most
exposed individuals around point sources vary considerably, ranging from 10'4 to as high
as 10"2. Average individual lifetime risks of 10"4 are common in many areas not impacted
by particular point sources.

Several of the lEMPs attempted to evaluate noncancer risks. The only pollutant
present at levels associated with noncancer_health effects was benzene, both around a
large steel manufacturing complex and near major traffic intersections. One IEMP study
also suggested that ambient xylene concentrations near large wastewater treatment plants
may have the potential for causing noncancer health effects.

Clark County Study

In early 1985, after reviewing EPA's Six Month Study, the Clark County Health
District attempted to address the magnitude and nature of the air toxics problem in
Las Vegas, Nevada, using a simple ambient monitoring approach. (EPA, 1987b) The
District initially estimated average, annual Valley-wide levels for 11 chemical substances,
based on short-term data, and multiplied these levels by unit risk factors to arrive at an
additive cancer incidence. The District had some knowledge of the emissions and/or air
quality levels of the 11 substances in the Las Vegas Valley. Most air quality
measurements were obtained from an established station in east central Las Vegas (which
has a history of high CO and TSP levels) and from a station in the southeast Las Vegas
Valley (which has a history of high ozone levels and complaints of chlorine odor and eye
irritation). The toxics data were derived from the District's efforts to characterize urban
haze and to develop specific hydrocarbon profiles to support ozone modeling exercises.
The urban haze research provided short-term data for various metals and carbonaceous
material.

-------
Results of the initial screening compilation show that products of incomplete
combustion (PIC), asbestos, benzene, perchloroethylene, and chromium account for most
of the cancer incidence from air toxics exposures in the Valley. Following the initial
screening, the District then refined its estimates of annual averages for particular
substances of interest and recalculated at the annual Valley-wide cancer incidence.
Totaling the chemicals, the District calculated that the annual incidence is 4.3 to 9.4
cancers per million people per year. PIC is the overwhelming contributor to total
incidence from air toxics generated almost equally by three source categories: (1) cars
using leaded gasoline, (2) diesel trucks and buses, and (3) wood burning fireplaces. Other
pollutants of lesser importance include arsenic, benzene, cadmium, and chromium.

South Coast—MATES

The South Coast Air Quality Management District (SCAQMD) has conducted one
of the most comprehensive studies to date of the urban air toxics problem. This study,
termed the Multiple-Air Toxics Exposure Study (or "MATES"), is a multiple, year
evaluation of community exposure to air, toxics in the Los Angeles metropolitan area.
MATES has been the focal point for urban air toxics study by SCAQMD since 1985;
however, various other independent activities that have been ongoing have provided data
for MATES. These other activities include certain monitoring and emission inventory
activities, an "in-vehicle" characterization of exposures to commuters inside their vehicles,
and the development of a dispersion/exposure/risk model named "SCREAM" (South Coast
Risk and Exposure Assessment Model). SCREAM is an enhanced version of EPA's Human
Exposure Model (HEM). (Barcikowski, 1988)

The MATES project is documented in a series of working papers, each reflecting
major components of activity. This series includes, in order:
• A quality assurance plan (SCAQMD, 1985)
• A monitoring site selection plan (SCAQMD, 1987a)
• . An emissions data summary (SCAQMD, 1987b) -
• A modeling results summary (SCAQMD, 1987c)

-------
• A monitoring results summary (SCAQMD, 1988).

The MATES project utilized data from two different ambient sampling efforts.
One set of data came from ongoing sampling activities at existing SCAQMD sampling
sites. The second effort was carried out by SCAQMD as part of MATES and involved
setting up ten new/temporary sampling sites in areas suspected of having elevated
concentrations, based on dispersion modeling of emissions data and various on-site
criteria.

Risk estimations at the sampling locations indicate that there are areas within the
South Coast Basin where the combined impact of multiple air toxics are significantly
higher than the Basin average. The combined individual risks from simultaneous exposure
to the 20 pollutants under study (14 organic gases and 6 metals) ranged from about 1 in
10,000 to 1 in 1,000, predominantly due to benzene and chromium. Higher concentrations
of most of the pollutants were measured in" the winter than in the summer because of such
meteorological conditions during the winter months as low ambient temperatures and
longer hours of surface inversions leading to reduced vertical mixing. (Barcikowski, 1988)

The modeling effort in MATES used, as input, emission inventories of point, area,
and highway vehicle emissions compiled in 1982 and updated to 1984. The estimation of
population exposures to one or more air toxics was conducted by first using dispersion
modeling of emissions data to calculate the long-term concentrations at centroids of
census areas and then multiplying the calculated concentrations with the population that
each centroid represents. The areawide risks, in terms of incremental cancer cases, were
then calculated by multiplying the population exposure with the chemical-specific unit
risk factors developed by the California Department of Health Services. A linear
relationship was assumed and the exposure/risks associated with multiple sources and
with species of air toxics are assumed additive. SCREAM was used to apportion the
number of excess cancer cases by source category and by pollutant, and to identify high-
risk chemical substance and source categories. SCREAM can also be used to identify high-
risk locations and to estimate control measure effectiveness in reducing exposure, cancer
risk, and number of cases.

-------
Of the 20 air toxics studied in the emission inventory/dispersion modeling portion
of MATES, benzene and hexavalent chromium 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 in 10,000 or higher.
As an upper-bound estimate, this portion of the study found that about 160 excess cancer
cases would result annually in the Basin due to the combined exposure to all 20
pollutants, or about 15 excess annual cases per million population. (Barcikowski, 1988;
Shikiya, 1987c)

Subsequent activities include the enhancement of SCREAM and. the development
of various risk mitigation strategies. The SCREAM computer model is being enhanced and
will include options for population mobility, indoor-versus-outdoor exposures, and
noninhalation routes of exposure. SCREAM will also be enhanced to provide better
estimates of individual risk and community, .cancer burden in regions and subregions of the
Basin.

The two major toxic air contaminants of concern, benzene and hexavalent
chromium, will be reduced through control efforts at the local District level and by
programs at the State level. SCAQMD has recently adopted a measure to control
emissions from chrome platers and is moving forward with rulemaking to control
hexavalent chromium emissions from cooling towers. Options for reducing benzene
contents in motor vehicle fuels are being evaluated. More stringent requirements on
motor vehicles that are aimed at reducing hydrocarbon emissions will also reduce
emissions of benzene from motor vehicles. (Barcikowski, 1988)

5 Citv Controllability Study

This study (EPA, 1988a), currently under way, is modeling ambient exposures to
25 air toxic compounds in five major urban areas. It also estimates how exposure levels
are likely to change in the future as a result of alternative control scenarios. The urban
areas were chosen to represent a cross-section of potential problem types and to ensure
geographical diversity. Air toxics exposures in the five study areas were estimated via
EPA's Human Exposure Model (HEM). A 1980 baseline emissions inventory was

-------
established to quantify emissions data for input into the model. Emissions data were
developed by various techniques, building on the source data in EPA's National Emissions
Data System (NEDS) and special State and local inventories. Various 1995 projection
inventories, reflecting expected and alternative control scenarios, will be simulated to
evaluate the effectiveness of various mitigation options. Risks from secondary (i.e.,
photo chemically formed) formaldehyde were estimated by superimposing measured
concentrations over the modeling domain. Polycyclic Organic Matter (POM) risks were
estimated using "comparative potency factors" for individual source categories rather than
the "B(a)P surrogate approach" used in the Six Months Study. Comparative potency
factors are being developed in EPA's Integrated Air Cancer Program (IACP) and relate the
potency of various organic mixtures to known carcinogens by comparing their respective
mutagenicities. (Lewtas, 1987)

The primary measure of risk in this study is aggregate cancer incidence.
Preliminary results suggest that cancer incidence in the five cities ranges from about 2 to
10 excess cases annually per- million persons, in general agreement with other urban
assessments. Urbanwide individual lifetime risks range from about 1.5 x 10~4 to 7 x 10"4.
The major pollutants of concern are polycyclic organic matter (POM), hexavalent
chromium, 1,3-butadiene, formaldehyde, and benzene. Small point and area sources
contribute most to total incidence, with road vehicles, cooling towers, chromium platers,
solvent use, and wood combustion being predominant.

An important finding of this study is that considerable air toxics control will be
achieved by 1995 from measures reasonably anticipated to be in effect because of State
Implementation Plans (SIPs), New Source Review (NSR), the Federal Motor Vehicle
Control Program (FMVCP), and National Emission Standards for Hazardous Air
Pollutants (NESHAPs). As suggested by the Six Months Study, some of this reduction is
indirectly due to criteria pollutant measures rather than to direct controls on air toxics.

Motor Vehicle Study

This study, entitled Air Toxics Emissions from Motor Vehicles. (Carey, 1987),
focused primarily on cancer risks posed by road vehicle emissions in the United States. It

-------
is not an urban assessment per se, but the results are useful in an urban soup context as
most vehicular travel is in urban areas. The report considered all vehicular air emissions
for which EPA has cancer unit risk estimates. Specific pollutants and pollutant categories
included are diesel particulate, formaldehyde, benzene, gasoline vapors, gas phase
organics, organics associated with non-diesel particulate, dioxins, asbestos, vehicular
interior emissions, and metals.

Several unique aspects to this study are worth highlighting. First, an activity
pattern model was used to reflect changing exposures to sets of population groups as they
change location from one microenvironment to another during their day-to-day activities.
These microenvironments included street canyons, parking garages, and indoor air, as well
as various urban and suburban settings. This approach for estimating exposures differs
markedly from all other urban assessments to date. Second, this study included an
approach to evaluate risk from polycyclic organic matter (POM) that used a comparative
potency method. In this method, the potency of organic extracts of diesel particulate and
gasoline particle organics was- compared with the potencies of extracts from sources for
which epidemiological data were available (Albert, 1983). This contrasts with the so-
called "B(a)P surrogate approach" used in EPA's Six Months Study wherein B(a)P was
used as a surrogate for all POM.

This study associated 385 to 2,286 annual excess cancer cases nationwide with air
toxics emissions from road vehicles. From 2 to 11 excess cancer cases are predicted per
year per million people in urban settings. This incidence drops roughly 40 percent by
1995 because of more stringent diesel particulate standards for both light and heavy duty
vehicles and because of the increasing use of three-way catalyst equipped vehicles coupled
with the phaseout of non-catalyst equipped vehicles. The pollutants contributing most to
total cancer incidence are, in rough order of importance, diesel particulate, 1,3-butadiene,
benzene, gasoline particulate-associated organics, formaldehyde, and asbestos. The bulk
of the formaldehyde risk is due to secondary (i.e., photochemically produced)
formaldehyde.
10

-------
Integrated Air Cancer Project flACPl

The Integrated Air Cancer Project (IACP) is being conducted by the EPA research
labs in Research Triangle Park, North Carolina. It is an interdisciplinary research program
(Lewtas, 1987) aimed at developing scientific methods and data bases to identify the major
sources of carcinogenic chemicals emitted into the air or arising from atmospheric
transformations. This project is performing extensive ambient air monitoring, receptor
modeling, bioassaying, and atmospheric transformation studies. Thus far, the study has
concentrated on the effects of emissions from wood burning and mobile sources.
Subsequent phases of this project will add other residential combustion sources (e.g., oil
combustion) and industrial sources.

The major long-term goals of the project are as follows:
1.
2.

3.
To identify the principal airborne carcinogens.
To determine which emissions sources are the major contributors of
carcinogens to ambient air. Major carcinogenic emissions sources will be
determined by-field studies with simultaneous emission characterizations
and ambient monitoring, followed by source apportionment calculations.
To improve estimates of comparative human cancer risk from specific air
pollution emission sources. The study will develop a comparative
methodology to evaluate and apply short-term mutagenesis and animal
carcinogenesis data on emission sources. Improved human exposure
estimates will be developed for complex emission products and individual
carcinogens, including transformation products.
Field studies have been conducted to date in Raleigh, North Carolina,
Albuquerque, New Mexico, and Boise, Idaho, with plans to add Roanoke, Virginia. Most
of the work that has been reported has been in Raleigh and Albuquerque, and has focused
mainly on methods development. The Boise results are not available as of this writing.

An important result has been the observation of elevated mutagenicities of
irradiated and aged gas phase mixtures. The mutagenicities of wood smoke, auto exhaust,
and several common industrial organics (toluene, propylene, and acetaldehyde) all
increase by an order-of-magnitude or more after the mixtures are irradiated and allowed to
react for several hours. These results suggest that the transformations of complex
11

-------
mixtures may contribute significantly to the total burden of mutagens or carcinogens in
the environment and should be considered in assessing risk. (Shepson, 1987)

TEAM (Total Exposure Assessment Methodology! Studies

Current urban assessments traditionally assume that the individual's entire
exposure to air toxics is a result of continuous exposure to outdoor air. The simplistic
assumptions and procedures involved in most current assessments, therefore, do not
realistically approximate actual personal risks from other routes of exposure, such as
from indoor air, workplace exposures, automobile interiors, etc.

In contrast, EPA's TEAM studies (Ott, 1986; Ott, 1988) have been designed to test
statistical and chemical methodologies for estimating total human exposure across
different microenvironments and various media. This research program has sought to
establish, for each of 20 or so organic chemicals, the relative importance of certain routes
of exposure (air, -water) and whether predictable correlations exist between exposure and
1°. * - — — —
body burden (as measured through analysis of breath).

The TEAM approach assesses exposures by measuring the amounts of pollutants
in an individual's immediate physical environment throughout the day. In contrast to a
fixed monitoring station that collects pollutants as they are c.arried by the ambient (or
indoor) air to the location of the devices, TEAM monitoring devices are located on vests
worn by individuals throughout their daily routine. In addition, TEAM measures the
concentration of chemicals in an individual's drinking water and exhaled breath in order to
determine the actual amount of the pollutant that has been taken up by the individual.
Personal sampling directly addresses the variability of exposures within a population and
should help to improve the exposure assumptions that are typically made in urban
assessments.

Since 1979, the TEAM studies have measured the personal exposures of hundreds
of people across the United States to 20 to 26 volatile organic compounds (VOCs) in air
and drinking water. The chemicals were selected on the basis of their toxicity,
carcinogenicity, mutagenicity, production volume, presence in the air or drinking water at
12

-------
the field sites, existence of National Bureau of Standards (NBS) permeation standards, and
amenability to collection on the sorbent Tenax.  The following cities are among those that
have been or are being evaluated:  Bayonne-Elizabeth,  New Jersey; Greensboro, North
Carolina; Devil's Lake, North Dakota; the California cities of Los Angeles and Antioch;
Pittsburgh, Pennsylvania; and Baltimore, Maryland. The latter application was coupled
directly with the Baltimore TEMP in order to evaluate the appropriateness of the use of
ambient monitoring and modeling data as proxies  for total individual exposures in risk
assessments.  The interpretation of the Baltimore TEAM results is still in progress at this
writing.  Results from the TEAM  studies "indicate  that even  in  major  chemical
manufacturing and petroleum refining areas such as Los Angeles and northern New Jersey,
personal  exposures  to  many  toxic  and  carcinogenic  VOCs  exceeded  outdoor
concentrations by  100 to 400 percent.  Breath measurements of the 550 residents who
took part in the study were significantly correlated with the preceding air exposures.  It
was concluded that personal  activities' such as smoking,  using  room air deodorizers,
wearing dry-cleaned-clothes, and even using hot water were responsible for a major
"portion of most people's exposure to benzene, para-dichlorobenzene, tetrachloroethylene,
and chloroform, respectively."   (Wallace, 1988)  Prior TEAM studies  (Ott,  1985)  have
shown that levels of  11 important organic compounds,  some of which are regarded as
potential carcinogens, are found to be significantly higher indoors than outdoors.  These
chemicals included chloroform, 1,1,1-trichloroethane,  benzene, carbon tetrachloride,
trichloroethylene, perchloroethylene,  styrene,  meta- and  para-dichlorobenzene,
ethylbenzene, and jrylene isomers. The sources appear to be inside the home,  probably in
furniture,  paint,  solvents, drapes, carpets, spray  cans, clothing, and construction
materials.

        In addition to  the above described method, which directly measures daily exposure
profiles for a representative cross-section of population within an area, TEAM  is also
constructing  estimates of personal exposure profiles by combining information on human
activities  with microenvironmental  field data.  In  essence,  concentration is measured in
selected microenvironments in  which a person will be exposed throughout a  day.  Then,
an integrated  exposure is computed as the product of (1) the  concentration of each
pollutant  encountered in each microenvironment and (2) the time the person spends there,
                                        13

-------
divided by the total time of exposure. For VOCs, typical microenvironments may include
gas stations, dry-cleaning stores, freshly painted rooms, and households where solvents
are stored indoors. The construction of time-activity exposure profiles allows the
extrapolation of the results to larger populations and other locales, using models that
incorporate human activity patterns. (Ott, 1985).

Staten Island/New Tersey Study

The Staten Island/New Jersey study (EPA, 1986b), begun in 1986, was initiated to
provide data on the scope and magnitude of the urban air toxics problems in the Staten
Island/New Jersey (Middlesex and Union Counties) areas. This study will rely primarily
on ambient monitoring to characterize air quality in the study area for selected toxic
pollutants. The measured data will be used to estimate exposures and to understand
better the interrelationship between indoor and ambient exposures. Plans are for emission
inventory data to be compiled to formulate hypotheses linking major contaminants to
potential sources and "to evaluate, general abatement strategies.

The Staten Island area was selected for intensive study based on the high density
of industry in the area, a long-standing history of odor complaints leading to a perception
that an air toxics problem exists, and the high quality of technical expertise available from
State and local organizations to study the situation. As of this writing, no results have
yet been reported.
,/
Southeast Chicago Study

The Southeast Chicago study (Summerhays, 1987) was initiated by EPA as a result
of citizen concerns regarding the environmental safety of the heavily industrialized area in
the vicinity of Lake Calumet, located in southeast Chicago, Illinois. A broad range of
environmental concerns was raised relating to air, water, and land pollution issues. As a
response, an emission modeling study was undertaken by EPA's Region V office to
ascertain whether air toxics exposures may be causing significant cancer risks.. The initial
step was the development of an air toxics emission inventory over a broad area of
Chicago, encompassing the receptor area of concern. The inventory included 47 suspected
14

-------
carcinogens—22 nonhalogenated compounds, 17 halogenated compounds, and 8 inorganic
species (mostly metals). As part of the inventory process, questionnaires were sent to 29
of the 88 point sources in the area.

       Dispersion modeling using ISC (for point sources) and CDM (for area sources)  was
performed to estimate ambient concentrations of air toxics in the  receptor grid.   The
resulting concentrations were applied to population data for the same receptor grid  and
multiplied by cancer unit risk factors to estimate individual risk and aggregate incidence.

       No results are  available  from this modeling study as of this writing.

       Urban Air Toxics Monitoring Program

       In 1987,  EPA initiated a screening  program  entitled the  Urban  Air  Toxics
Monitoring Program (EPA, 1987c). The primary purpose of the program is to collect air
quality data to  support State and local agency efforts to assess the nature and magnitude
of the urban air toxics problem.in their respective areas.  In .1987, 19 State and local
agencies  agreed to participate in this program. Depending on future funding allocation and
interest shown by State and local agencies, this program may be continued for  several
years.

       EPA's objectives in promoting and supporting this monitoring program are:
       •      To provide estimates of annual concentrations of selected air toxics;
       •      To provide information for prioritizing  and planning future work  and
              sampling on a more in-depth and pollutant-specific basis in local areas;
       •      To provide a means to identify prevailing pollutants  and possible source
              types that may need further assessment; and
       •      To identify a means to evaluate and prioritize future air toxics mitigation
              programs.

       The  program  calls for  State  and local agency personnel to collect ambient air
samples  for subsequent analysis either by EPA or an EPA contractor for  specific toxic
compounds.  Under EPA direction, a central contractor will bring the necessary sampling
train to each site, assemble the apparatus, instruct the sampling technician  on equipment
                                         15

-------
operation and maintenance, and provide detailed sample shipping information.
Participating personnel will schedule a meeting date and location for the contractor to set
up and demonstrate the sampling trains.

Three different types of ambient air samples are collected. The first set of air
toxics samples is collected in stainless steel canisters for 24-hour periods every 12 days
for one year. After sample collection, the canisters are air shipped to a central laboratory
for analysis. The samples are analyzed for about 30 volatile organic compounds by gas
chromatography equipped with multi-detector capability. The second set of samples is
collected in cartridges for determination of formaldehyde and other aldehydes. The third
set is total suspended particulate (TSP) matter collected from a high-volume air sampler
for determination of about 14 metals and B(a)P.

No results are available as of this writing.

Bay..Area Toxics Monitoring ....Study CBAAQMD. 1987)

The Bay Area Air Quality Management District (BAAQMD), in conjunction with
the California Air Resources Board, conducts ambient sampling for 11 volatile air toxics at
a network of 15 monitoring sites in the Bay Area, comprising one of the largest urban air
toxics networks operating in the United States. The network is oriented to population
exposure, with a background site north of San Francisco near the ocean. The pollutants
analyzed are believed by the Bay Area AQMD to represent a large part of the risk of
volatile toxic substances to the general population. Twenty-four hour samples are taken
in Tedlar bags and are analyzed by gas chromatography. Sampling frequency is twice per
month per site.

Measurements were started in 1986 and are ongoing. Results to date suggest the
benzene levels constitute the majority of risk associated with the measiired exposures and
are largely derived from vehicular activity.
16

-------
1.3    Other Urban Air Toxics Assessment Activities

EPA has been encouraging State and local air pollution control agencies to assess their
urban air toxics problems.  Numerous assessment activities have been undertaken in the
past several years in the areas of ambient air monitoring, emissions inventorying, exposure
and risk characterization, and mitigation analysis.  Figure 1-1 shows those U.S. cities
where some type of assessment activities have been initiated involving one or more of
these activities. Not all of these activities will lead to multi-pollutant, multi-source
cancer risk assessments. Table 1-1 lists the specific activities undertaken in 30 urban
areas having populations exceeding one million persons.
                                          17

-------
Chapter 1
References
Albert, R., et al., 1983. "Comparative Potency Method for Cancer Risk Assessment:
Application to Diesel Particulate Emissions," Risk Analysis, Vol. 3, No,, 2, pp. 101-117.

BAAQMD, 1987. Letter and Attachment from D.A. Levaggi to M. Stenburg, Bay Area Air
Quality Management District. 939 Ellis Street, San Francisco, California 94109.
September 10, 1987.

Barcikowski, W., 1988. South Coast Air Quality Management District. Letter with
attachments to Tom Lahre, U.S. Environmental Protection Agency, Research Triangle
Park, North Carolina. October 24, 1988.

Carey, P., 1987. Air Toxics Emissions from Motor Vehicles, EPA-AA-TSS-PA-86-5, U.S.
Environmental Protection Agency, Ann Arbor, Michigan,.

EPA, 1985. A Strategy to Reduce Risks_ to Public Health from Air Toxics, U.S.
Environmental Protection Agency, Washington," D.C.

EPA, 1986, Final Report of the Philadelphia Integrated Environmental Management
Project. U.S. Environmental Protection Agency, Washington, D.C.

EPA, 1987a. Kanawha Valley. Toxics Screening Study Final Report, U.S. Environmental
Protection Agency, Washington, D.C.

EPA, 1987b. National Air Toxics Information Clearinghouse report entitled "Qualitative
and Quantitative Carcinogenic Risk Assessment," EPA 450/5-87-003, U.S. Environmental
Protection Agency and STAPPA/ALAPCO, pp. 11-1 through 11-6.

EPA, 1987c. Urban Air Toxics Monitoring Program, EPA-450/4-87-022, U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina.

EPA 1988a. Preliminary results from 5 City Controllability Study being conducted by
E. H. Pechan and Assoc. under contract to U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina. (Tom Lahre, EPA project officer).

EPA, 1988b. Staten Island/New Jersey Urban Air Toxics Assessment Project. Revised
Status Report. U.S. EPA Region in.

Haemisegger, E., et al., 1985. The Air Toxics Problem in the United States: An
Assessment of Cancer Risks for Selected Pollutants, U.S. Environmental Protection
Agency, Washington, D.C.

Hinman, K., et al., 1986. Santa Clara Integrated Environmental Management Project.
Stage Two Report. U.S. Environmental Protection Agency, Washington, D.C.
20

-------
Lamason,  1988.  Personal communication.   U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina. August 1988.

Lewtas, J.,and  L.  Cupitt, 1987.  "Overview of the Integrated  Air Cancer Project,"
Proceedings of the 1987 EPA/APCA Symposium  on Measurement of Toxics and Related
Air Pollutants,  U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina, pp. 555-561.

Machlin, Paula R., 1986.   "Denver Integrated Environmental  Management  Project:
Approach to Ambient Air Toxics Monitoring," EPA Region 8, Denver Colorado, November
1986.

Manale, A., et  al., 1987.  "Baltimore Integrated  Environmental Management  Project.
Phase II Report:  Air Toxics," (Draft), U.S. Environmental Protection Agency, Washington,
D.C.

Ott, W., 1985.  "Total Human Exposure," Environmental Science Technology, Vol.  19,
No. 10, 1985.

Ott,  W.,  et al.,  1986.   "EPA's  Research  Program  on Total  Human Exposure,"
Environmental International, Vol. 12, pp. 475:494.

Ott, W., 1988.  "Human Exposure to Environmental Pollutants," Presented at the 81st
Annual Air Pollution Control Association Meeting, June 20-24, 1988, Dallas, Texas.

SCAQMD (1985). MATES Working Paper #1, Quality Assurance Plan, June 1985, South
Coast Air Quality Management District.

SCAQMD  (1987a).  MATES Working Paper #2, Air Toxics Monitoring Site Selection,
September  1987, South Coast Air Quality Management District.

SCAQMD (1987b).  MATES Working Paper #3, Toxic Emissions Data for the South Coast
Air Basin, April 1987,  South Coast Air Quality Management District.

SCAQMD  (1987c).  MATES Working Paper  #4,  Urban Air Toxics  Exposure Model:
Development and  Application, October 1987,  South  Coast  Air Quality Management
District.

SCAQMD (1988).  MATES Working Paper #5, Analysis of Ambient Data  from Potential
Toxics "Hot Spots" in the South Coast Air Basin, September 1988, South Coast Air
Quality Management District.

Shepson,  P., T. Kleindienst, and E. Edney,  1987.  The Production  of Mutagenic
Compounds as a Result of Urban Photochemistry, EPA 600/3-87-020, U.S.  Environmental
Protection Agency, Research Triangle Park, North Carolina.

Summerhays, 1987.  "Air Toxics Emission Inventory for the Southeast Chicago Area."
U.S. EPA Region V.
                                       21

-------
Wallace,  L., et al., 1988.   "Preliminary Results from the  Baltimore  TEAM Study."
Presented  at the  81st  Air  Pollution   Control Association  Annual  Meeting,
June 19-24, 1988, Dallas, Texas.
                                        22

-------
                                     CHAPTER 2
                      AMBIENT AIR QUALITY MONITORING
2.1
The following subjects are covered in this chapter:
•      Use of ambient air monitoring data in urban assessments
•      -Pollutant coverage       .......
•      Site .selection
•      Sampling periods, frequencies, and duration
•      Sampling and analytical techniques
•      Evolving monitoring techniques
•      Insights into the use of monitoring in air toxics programs

Use of Ambient Air Monitoring Data in Multi-Pollutant. Multi-Source Urban
Assessments
       Multi-pollutant, multi-source urban air toxics assessments commonly use ambient
air monitoring data to help characterize exposures and risks related to toxic air pollutants.
Ambient air data are most  often  used in conjunction  with dispersion modeling of
emissions data to minimize the limitations inherent in using either type of data alone.

       Estimating  exposures using data from ambient air monitoring programs  in the
studies under review has been conceptually straightforward.  Ambient levels of air toxics
have been measured over urban areas  to yield estimates of population exposures.  These
                                        23

-------
exposure levels were then multiplied by the respective potencies (i.e., cancer unit risk
factors) of each  pollutant to estimate individual cancer  risks,  which, in turn, are
superimposed over  population  distributions to  estimate excess cancer  cases.  Such
analyses are often fairly crude because only a comparatively few measurements of each
pollutant are  available  to  represent long-term, area-wide population  exposures.
Fortunately, the increasing emphasis on ambient air toxics monitoring by States and local
agencies is providing more complete data for exposure and risk assessments.

       There are important advantages in using ambient air monitoring data as part of
urban assessments, the principal one  being  the  ability  to  directly  characterize
concentrations at key receptor sites of interest.   In this regard, the  use of reliable and
representative ambient air monitoring data, instead of modeled concentrations based on
emissions data, circumvents the errors inherent in emission inventories and dispersion
models, which can  be considerable.   In .fact, some pollutants that  may be covered by
ambient air quality monitoring programs may not be included in emissions inventories. In
addition, ambient air quality-monitoring allows one  to evaluate pollutants that are not
emitted directly, such as photochemically formed formaldehyde or peroxyacetyl nitrate
 (PAN)  or other  transformation products,  or to measure compounds that exist in the
 atmosphere because of gradual build-up of background levels (e.g., carbon tetrachloride) or
 transport from other urban areas.  Finally, ambient  air measurements provide a direct
 reference from which to assess dispersion model performance and emission inventory
 accuracy.

        Ambient air monitoring data were used in many of these studies for purposes in
 addition to multi-source, multi-pollutant cancer risk assessments.  For example, ambient
 air measurements were used in the Philadelphia and Kanawha Valley studies to help
 identify maximum individual exposures in the vicinity of major point sources. This might
 also be done to help verify that the  conditions  in an operating  permit  are being met.
 Ambient air measurements can also be  used  to assess the adequacy of an emission
 inventory and/or dispersion model are used to predict ambient air concentrations.  Poor
 agreement of measured and predicted ambient concentrations  may reflect  inventory error
                                          24

-------
(e.g., a major overlooked source) and/or poor model performance. Finally, ambient
monitoring data are now being used for mutagenicity bioassays and source
apportionments. Regulatory agencies may want to consider multiple program uses of any
measurements that are taken as part of an urban soup assessment to optimize the use of
its ambient monitoring resources.

The major disadvantages of ambient air monitoring are the high costs and
uncertainties involved with measuring low concentrations of specific air toxics. Obtaining
a sufficient number of samples to show variations in concentration in time and space can
be prohibitively expensive. Also, high detection limits, sample contamination (in the
field, in transit, and in the laboratory), uncertain identification of target compounds, and
other sources of error have hampered many data collection efforts; these should be
anticipated and carefully evaluated during Jhe design stages of monitoring programs.
Finally, ambient sampling and/or analytical procedures are not yet proven for all
pollutants (e.g., hexavalent chromium, 1,3 butadiene, acrylonitrile, and ethylene oxide).
All of. these disadvantages suggest that EPA and State and local agencies must optimize
the number of sampling sites and measurements in any assessment, and must coordinate
with other applied programs and the research community before initiating a major
sampling effort.

In the Evolving Monitoring Technologies section of this chapter, several promising
approaches are described to reduce the cost of monitoring programs for toxic air
pollutants. Techniques such as sample compositing, sampling over longer time periods,
and periodic expansion to a larger sampling network may dramatically increase the
information provided per monitoring dollar spent.

2.2. Decisions Affecting Monitoring Plan Development

Design of an effective monitoring plan is the most important step in the
development of a monitoring program. The two basic elements of concern are (1) clearly
specified data quality objectives (definition of the type, accuracy, and coverage of data to
be gathered) and (2) an adequate quality assurance plan against which achievement of data
25

-------
quality objectives can be reviewed.  Not only are these steps essential in ensuring the

cost-effective deployment of resources, their scope and design should closely support the

goals and objectives of the overall study.  Inadequacies in monitoring program design can

seriously compromise the ability of a study to define the exposures and potential risks it

seeks to investigate.


       As the studies reviewed here illustrate, cost constraints inevitably lead to design

compromises or trade-offs among multiple competing objectives.  Any ambient air quality

monitoring program is inherently expensive, but air toxics studies tend to be relatively

more expensive  than conventional  programs.   Sampling  techniques  are often not

standardized; per-sample laboratory costs are comparatively high,  and the number  of

compounds of potential interest is large.  Furthermore, in the  absence of specific  ambient

air quality standards for toxic air pollutants_ (which would  define specific statistics  of

regulatory concern), study objectives tend  to  include  multiple  aspects of air toxics

exposures: emission rates, fenceline concentrations, source-receptor relationships, average

population exposures, or maximum individual exposures. It is therefore highly important

for designers to reach firm agreement on a monitoring program's objectives, especially in

terms of how these objectives relate to the broad technical questions  of the study as a

whole.


       Examples  of technical  trade-offs that  should  be  addressed in  designing  a

monitoring plan include:

       «       Sample accuracy:  The most accurate sampling and laboratory procedures
               are not necessarily the most desirable for every study.  Achieving highly
               accurate results for individual samples may compromise  the number of
               sites  sampled, the number of compounds evaluated,  the frequency  of
               sampling, or the length of the  sampling period.  Undue emphasis  on
               individual sample accuracy may  actually reduce  rather than enhance the
               effectiveness of the measured data in supporting program goals.  This is
               especially true in cancer assessments because of  the inherent uncertainty
               that exists in cancer unit risk factors.

        •      Site selection!  Many factors influence site selection, including the gradient
               of concentrations expected in each area  (locations  with steep gradients
               require the  most  careful location of sites),  laboratory detection limits
               (location of sites in areas where  concentrations often fall  below detection
                                         26

-------
limits may waste resources), or air quality statistics of concern (average
exposures versus maximum exposures).

The studies under review here illustrate these points in various ways. Some of the
reports produced by these studies listed specific, formal monitoring objectives; for others,
we have had to infer monitoring goals from the stated overall study objectives and
technical approaches taken.

The advantage of presenting detailed, specific objectives prior to initiating a
monitoring program is that it provides a focal point for consensus between program
managers and technical analysts. In addition to helping these groups collaborate on a
detailed and cost-effective QA/QC plan, setting formal objectives provides a benchmark to
assess how well a program meets its goals.

Table 2-1 Summarizes the stated or implied monitoring goals of the eight studies
with monitoring programs. Before examining individual technical issues in more detail, it
is useful to discuss the stated goals of these eight monitoring programs and the extent to
which they were achieved.

Studies with Stated Goals for Air Quality Monitoring Programs

The IEMP programs in Kanawha Valley, Philadelphia, Baltimore, and Denver; the
Clark County Study; and the Urban Air Toxics Monitoring Program and the South Coast
MATES project all documented their monitoring objectives.

Kanawha Valley—Designers of the Kanawha Valley program listed five general
monitoring program objectives that clearly supported overall program goals for this
unique industrial area:1
These are quoted verbatim from the study report.
27-

-------
              Table 2-1.  Summary of Objectives for Air  Toxics
                Monitoring Programs Reviewed In This Study

Study
South Coast— MATES
Clark County
lEMPs
Philadelphia
Baltimore
Santa Clara
Kanawha
Denver
Urban Air Toxics
Program -
USES
Estimate
Average
Exposures
X
X
X
X
Evaluate
Model
Performance
X

Support
Emissions
Verification
X

X X
X X
-Goals Unspecified 	
X X
~~~

SCALES ADDRESSED
Define Define
Population Maximum
Exposures Concentrations
X X
X
X X
X X
	 Goals Unspecified 	
X
X
x
   1.     To 'qualitatively and quantitatively confirm  the  presence of targeted
          compounds in the  ambient air.   EPA Region Ill's minimum objective was
          to confirm the presence of compounds for which emissions  data were
          available  from the West  Virginia Air Pollution Control Commission's
          inventories.  There was  concern that some emissions data may have been
          outdated  or  otherwise inadequate and  that  modeled predictions  of
          concentrations  of  certain  pollutants could be in error, both because of
          technical difficulties in modeling the valley (see Chapter 4) and because of
          possible chemical reactions during transport.

          This  goal was largely achieved by  the program.2   Most predicted
          compounds were documented, though ambiguities remained with regard to
          ethylene oxide.
Although the overall objectives of this study were met fairly well, Phase I of the two-
phase monitoring plan developed for the study did not meet its study objectives. The
strategy of Phase I was to conduct a limited three- or four-day screening study using
bag sampling with GC  analysis  (and  some GS/MS analysis for confirmation) to rank
potential  sites  and select pollutants  for the  larger-scale  effort  in  Phase  II.
Unfortunately, the GC and GC/MS results  were not in agreement and no clear pattern
emerged to distinguish upwind and downwind sites near major sources.
                                    28

-------
2. To determine whether a more in-depth emissions verification of facilities
and processes is needed. To the extent that models could predict ambient
concentrations, the study sought to determine which compounds in the
inventories were least well characterized and required additional follow-
up.

This goal was also achieved. The study did find that certain measured
concentrations did appear to be significantly higher or lower than
predicted. As of this writing, however, the study has not reconciled
specific differences by following up to determine whether certain
emissions were overestimated or underestimated.

3. To assist in model selection for potential future study. Technical
difficulties in modeling the Valley (see Chapter 4) are substantial. To
fulfill this monitoring goal, the study data were to be interpreted to the
extent feasible to determine whether model modifications should be
attempted to better define transport and dispersion for this area.3

This goal has not yet been achieved because the necessary analysis has not
yet been done.

4. ..To help determine whether a more detailed monitoring program is needed
at a later date. If significant most exposed individual MEI concentrations
of study pollutants were detected in populated neighborhoods, and if
monitoring confirmed that modeled results reasonably predicted the
presence and locations of these MEI concentrations, additional monitoring
might be appropriate to confirm model predictions in other locations or for
additional compounds.

This goal was achieved; predicted high concentrations of certain
compounds were tentatively confirmed in target neighborhoods, indicating
the appropriateness of further monitoring.

5. To evaluate a newly designed 24-hour canister sampling system. The
Kanawha project used a relatively new evacuated canister system for
sample collection. The final goal of the program was to evaluate the
effectiveness of this method for the compounds of interest.

The program confirmed the general effectiveness of the canister system.4

Overall, the Kanawha program clearly stated its monitoring program goals and

documented its progress in meeting these goals in its final report. It was the only study.
3 Specific issues included whether increased use of turbulent intensity data and a
specific treatment for Valley wall reflection would be possible and desirable.
4 Difficulties were encountered, however, with the formaldehyde sampling techniques
used in the study, invalidating results for this pollutant.
29

-------
reviewed that tracked achievement of its stated goals. Further discussions in this chapter

and elsewhere will deal with how these program goals were reflected in details of the

monitoring plan itself.

Denver—The Denver IEMP project also listed general objectives for its monitoring

effort that directly supported overall study goals. They were, however, markedly

different in many respects than the Kanawha effort.

1. Collect ambient concentration data on selected toxic air pollutants that are
likely to he present in metropolitan Denver using the best available
sampling and analysis techniques. Denver's air pollution problem, is
considered to be regional: there are few industrial sources, and much of the
air pollution is attributable to complex meteorological conditions and
possibly transformation of pollutants from automobiles and other
dispersed sources. Study designers determined that in this case it was
necessary and appropriate to use the most current sampling methods'
available. The generalized, nature of the pollution implied that
concentrations would not vary extensively from site to site, permitting use
"of a small number of stations. The main issue of concern was to quantify
multiple pollutant species in depth, using the most current methods
available from EPA's Office of Research and Development.

This objective is expected to be achieved. Data analysis is still in progress.

2. Collect ambient concentration data during high episodes for the routinely
collected pollutants, and conduct broad-scan analyses** for those not
routinely measured and which may warrant further study. This program
objective called for increased 12-hour coverage of VOCs on days with peak
CO and TSP concentrations.

This objective was not fully achieved despite the use of independently
analyzed 12-hour samples taken during predicted peak air pollution days.
The goal of collecting additional samples on peak days was limited to one
extra day of sampling.

3. Produce ambient measurements that are spatially and temporally
representative of annual and acute [short-term6] exposures experienced by
the general population. This program objective overlaps somewhat with
objective 2. It called for distribution of sampling stations over the general
5 "Broad-scan" refers to a detailed GC/MS library search for each peak to attempt to
estimate the concentration of as many pollutants as possible. More typically,
analysis is limited to a preselected number of routinely quantified peaks.
6 Use of the word "acute" implies short-term concentrations, not concentrations high
enough to produce acute toxic responses in humans.
30

-------
metropolitan area, sited both to reflect annual average concentrations and
to capture high short-term concentrations.
Spatial representativeness was achieved, but, as noted, definition of short-
term peak concentrations does not appear to be fully achieved.
4. Help identify and prioritize future research needs. The program's final goal
was to set priorities for future studies in terms of compounds of interest
and spatial and temporal resolution. It was expected that some
compounds for which risk values are not available may be identified at
relatively high levels, and that final evaluations of risk would be
postponed until suitable potency estimates are forthcoming.
It is not possible to assess the attainment of this goal until data
interpretation is complete.

Philadelphia and Baltimore—These projects were among the earliest, if not the
first, projects to attempt urban scale monitoring for air toxics. They had one basic
monitoring goal—to use a source-oriented siting'approach to help evaluate and verify local
emission inventories. They were both partly successful in reaching this goal, though
perhaps more elaborate and comprehensively developed goal statements could have
improved the program results.

Clark County—The objective of this study was to determine annual average
concentrations of most volatile organic compounds, selected metals, and asbestos. These
compounds were expected to represent toxics of concern within the Las Vegas study area,
which is heavily influenced by mobile sources. The study appears to have largely
achieved its stated monitoring objectives.

Urban Air Toxics Monitoring Program—This is a continuing, multi-city program in
which each monitoring study reflects the general objectives of the overall program. These
objectives include the following:
1. To provide estimates of annual concentrations of selected air toxics.
2. To provide information for prioritizing and planning future work and
sampling on a more in-depth and pollutant-specific basis in local areas.
3. To provide a means to identify prevailing pollutants and possible source
types that may need further assessment.
31

-------
       4.      To identify a means to evaluate and rank  future air toxics mitigation
              programs.

       South  Coast—MATES—The  South  Coast Air Quality  Management  District

(SCAQMD) developed a very detailed monitoring protocol to cover all important aspects

of ambient sampling, as itemized below:

       •      Monitoring objectives

       •      Chemical species measured

       «      Sampling methods

       •      Analytical methods

       •      Identification of potential sites

       •      Sampling strategy

       •      "Data quality objectives    	

       •      Quality assurance^ considerations


       Specific monitoring objectives were:

       1.     Measure multiple toxic air contaminant concentrations in areas where the
              probability of the occurrence of elevated risk  is greatest.

       2.     Collect data representative of the potentially toxic chemical species that
              are expected to be significant in the South Coast Air Basin.

       3.     Measure both organic  compounds in the gaseous phase  and metals  on
              suspended particulate matter.

       4.     Collect samples that are representative  of a winter/spring" season and a
              summer/fall season, and from which meaningful estimates of annual
              average concentrations may be determined.

       5.     Collect data that can be  compared with information from other ongoing
              toxics programs in the SCAB.


       Details of the MATES monitoring program are given in (Shikiya, 1988).
                                        32

-------
2.3. Pollutant Coverage

Table 2-2 presents the pollutants covered in each, monitoring program. As is clear
from this table, studies to date have emphasized volatile organics and, to a considerably
lesser extent, metals. Coverage of semivolatile organics and aldehydes/ketones has been
very limited.

The focus on volatile organics was apparently a result of emphasis placed on
volatile organics in available emission inventories, which often preceded these studies.
Even where these inventories were developed expressly to evaluate toxic air pollution (as
in Philadelphia), they tended not to look comprehensively at all categories of potential
toxics, but to emphasize the high-volume, often carcinogenic, organic solvents. The
implicit assumption was that these high-volume carcinogens were likely to account for the
major fraction of the risk from toxic air pollutants. This bias toward volatile organics is
clearly evident in the IEMP studies in Philadelphia, Kanawha Valley, and Santa Clara,
whose goals were, explicitly linked to verification or review of available emission
inventories. These studies" included coverage of pollutants released from industrial
sources, as well as those related to the fueling, combustion, or evaporative emissions
from mobile sources. Pollutants in this latter group include benzene, toluene, ethyl
benzene, and the isomers of xylene.

Once the emphasis on volatile organics was established, most studies covered
pollutants that could be measured by readily available methods. For the majority of
studies conducted during the period of 1980 through 1985, Tenax®/GC or GC/MS
methods7 generally dictated pollutant coverage, producing a heavy emphasis on common
solvents such as perchloroethylene and trichloroethylene (solvents such as methylene
chloride and 1,1,1-trichloroethane tended to pose contamination problems). In addition,
7 The only other generally available method for organics sampling was to use charcoal
as the sorbent, but charcoal was generally considered substantially inferior to Tenax.
33