'Jn'ted States Pe<, EPA/902/R-93-001f
Environmental Protection 902 .ary 1993
Agency
v>EPA Staten Island/New Jersey
Urban Air Toxics
Assessment Project
Report
Volume V
Risk Assessment and Statistical
Analyses
-------
ACKNOWLEDGEMENTS
This report is a collaborative effort of the staffs of the
Region II Office of the U.S. Environmental Protection Agency
(EPA), the New Jersey Department of Environmental Protection and
Energy, the New York State Department of Environmental
Conservation, the New York State Department of Health, the
University of Medicine and Dentistry of New Jersey and the
College of Staten Island. The project was undertaken at the
request of elected officials and other representatives of Staten
Island concerned that emissions from neighboring industrial
sources might be responsible for suspected excess cancer
incidences in the area.
Other EPA offices that provided assistance included the
Office of Air Quality Planning and Standards, which provided
contract support and advice; and particularly the Atmospheric
Research and Exposure Assessment Laboratory, which provided
contract support, quality assurance materials, and sampling and
analysis guidance, and participated in the quality assurance
testing that provided a common basis of comparison for the
volatile organic compound analyses. The Region II Office of
Policy and Management and its counterparts in the States of New
York and New Jersey processed the many grants and procurements,
and assisted in routing funding to the project where it was
needed.
The project was conceived and directed by Conrad Simon,
Director of the Air and Waste Management Division, who organized
and obtained the necessary federal funding.
Oversight of the overall project was provided by a
Management Steering Committee and oversight of specific
activities, by a Project Work Group. The members of these groups
are listed in Volume II of the report. The Project Coordinators
for EPA, Robert Kelly, Rudolph K. Kapichak, and Carol Bellizzi,
were responsible for the final preparation of this document and
for editing the materials provided by the project subcommittee
chairs. William Baker facilitated the coordinators' work.
Drs. Edward Ferrand and, later, Dr. Theo. J. Kneip, working
under contract for EPA, wrote several sections, coordinated
others, and provided a technical review of the work.
The project was made possible by the strong commitment it
received from its inception by Christopher Daggett as Regional
Administrator (RA) for EPA Region II, and by the continuing
support it received from William Muszynski as Acting RA and as
Deputy RA, and from Constantine Sidamon-Eristoff, the current RA.
The project has received considerable support from the other
-------
project organizations via the Management Steering Committee,
whose members are listed in Volume II.
11
-------
PREFACE - DESCRIPTION OF THE STATEN ISLAND/NEW JERSEY URBAN AIR
TOXICS ASSESSMENT PROJECT REPORT
This report describes a project undertaken by the States of
New York and New Jersey and the United States Environmental
Protection Agency with the assistance of the College of Staten
Island, the University of Medicine and Dentistry of New Jersey
and, as a contractor, the New Jersey Institute of Technology.
Volume I contains the historical basis for the project and a
summary of Volumes II, III, IV, and V of the project report.
Volume II of the report lists the objectives necessary for
achieving the overall purpose of the project, the organizational
structure of the project, and the tasks and responsibilities
assigned to the participants.
Volume III of the report presents the results and discussion
of each portion of the project for ambient air. It includes
monitoring data, the emission inventory, the results of the
source identification analyses, and comparisons of the monitoring
results with the results of other studies. Volume III is divided
into Part A for volatile organic compounds, and Part B for
metals, benzo[a]pyrene (BaP), and formaldehyde. Part B includes
the quality assurance (QA) reports for the metals, BaP, and
formaldehyde.
Volume IV presents the results and discussion for the indoor
air study performed in this project. It contains the QA reports
for the indoor air study, and a paper on the method for sampling
formaldehyde.
Volume V presents the results of the detailed statistical
analysis of the VOCs data, and the exposure and health risk
analyses for the project.
Volume VI, in two parts, consists of information on air
quality in the project area prior to the SI/NJ UATAP; quality
assurance (QA) reports that supplement the QA information in
Volume III, Parts A and B; the detailed workplans and QA plans of
each of the technical subcommittees; the QA reports prepared by
the organizations that analyzed the VOC samples; descriptions of
the sampling sites; assessment of the meteorological sites; and a
paper on emissions inventory development for publicly-owned
treatment works.
The AIRS database is the resource for recovery of the daily
data for the project. The quarterly summary reports from the
sampling organizations are available on a computer diskette from
the National Technical Information Service.
111
-------
STATEN ISLAND/NEW JERSEY
URBAN AIR TOXICS ASSESSMENT PROJECT
VOLUME V. RISK ASSESSMENT AND STATISTICAL ANALYSES
EPA/902/R-93-001f
TABLE OP CONTENTS
1. EXPOSURE AND HEALTH RISK ASSESSMENT 1-1
1.1 INTRODUCTION 1-1
1.2 AIR POLLUTANT CONCENTRATIONS IN THE SI/NJ UATAP
STUDY AREA 1-2
1.2.1 VOCs 1-3
1.2.2 Metals, Benzo[a]pyrene, and Formaldehyde . . 1-5
1.3 QUANTITATIVE RISK ASSESSMENT -
HAZARD IDENTIFICATION, REFERENCE CONCENTRATIONS,
AND INHALATION UNIT RISK FACTORS 1-5
1.3.1 Hazard Identification 1-5
1.3.2 Reference Concentrations and
Inhalation Unit Risk Factors 1-6
1.3.2.1 Reference concentrations for
noncancer toxicity 1-6
1.3.2.2 Unit risk factors for
carcinogenicity 1-8
1.4 LEVEL 1 RISK ASSESSMENTS 1-10
1.4.1 Exposure Assumptions 1-10
1.4.2 Results for the VOCs 1-12
1.4.3 Results for the Metals, Benzo[a]pyrene,
and Formaldehyde 1-14
1.4.4 Uncertainties and Limitations 1-17
1.4.5 Discussion 1-19
1.5 LEVEL 2 RISK ASSESSMENT (FOR VOCS ONLY) 1-20
1.5.1 Introduction 1-20
1.5.2 Concentration Data Used in the Level 2
Exposure Assessment 1-21
1.5.2.1 Residential site indoor and
outdoor data 1-21
1.5.2.2 Comparison of the outdoor air data
from the indoor air and the
ambient air portions of the
SI/NJ UATAP 1-22
IV
-------
1.5.3 Exposure Assessment 1-23
1.5.4 Risk Assessment Results and Discussion . . 1-25
1.5.5 uncertainties and Limitations 1-26
1.6 ADDITIVE RISK ASSESSMENT 1-26
1.6.1 Introduction 1-26
1.6.2 Noncancer Analysis 1-28
1.6.3 Analysis for the Carcinogens 1-29
1.6.4 Uncertainties in the Additive Risk
Assessment 1-30
1.7 GENERAL CONCLUSIONS 1-31
1.8 ACKNOWLEDGEMENT 1-32
1.9 REFERENCES 1-32
TABLES, FIGURES , AND MAPS 1-36
2. STATISTICAL ANALYSES 2-1
2.1 INTRODUCTION 2-1
2.2 ADJUSTING FOR METHOD BIAS 2-1
2.3 ANALYSIS OF VARIANCE RESULTS USING ADJUSTED DATA . 2-8
2.4 INTERPRETATION OF ANOVA RESULTS 2-8
2.5 ANALYZING DAILY CONCENTRATIONS AT
SITES 3,6 AND 8 2-14
2.5.1 Choice of Time Unit 2-14
2.5.2 Comparisons of Sites 3 and 6 ....... 2-19
2.6 CONCLUSIONS 2-22
ANOVA RESULTS 2-24
VOLUME V APPENDICES
APPENDIX A - COMPARISON OF THE SI/NJ UATAP AND THE
UATMP STUDIES A-l
APPENDIX B - DESCRIPTION OF THE 1988 UATMP SITES .... B-l
APPENDIX C - SUMMARY OF POPULATION ANALYSIS C-l
-------
1. EXPOSURE AND HEALTH RISK ASSESSMENT
1.1 INTRODUCTION
A primary objective of the Staten Island/New Jersey Urban
Air Toxics Assessment Project (SI/NJ UATAP) was to assess
exposure and health risk arising from inhalation of airborne
toxic pollutants in the area.1 The outdoor air sampling sites
and sampling frequencies were chosen so that spatial and temporal
concentration differences of air toxics within the Staten
Island/New Jersey region could be determined, and exposures and
associated health risks in the communities surrounding each site
could be estimated.
Volume II of the project report provides a detailed
description of the SI/NJ UATAP. For the reader's orientation,
Tables V-l-la through 2b listing the sampling sites and the
chemicals for which the samples were analyzed, and Map V-l-1
shows the locations of the monitoring sites. Staten Island is
bordered on the west by a complex of major industries including
pharmaceutical plants, oil refineries, and chemical storage
facilities. Other industrial sources of pollution include sewage
treatment plants and the 1400-acre Fresh Kills Landfill.
This volume provides the exposure and risk assessments for
the study chemicals for which toxicological information (e.g.,
inhalation reference concentrations and carcinogen unit risk
factors) and air concentration data from the study are available.
These conditions limited the scope of the quantitative risk
assessment to 22 of the 402 study chemicals—11 volatile organic
compounds (VOCs), 9 metals, benzo[a]pyrene (BaP), and
formaldehyde.3
Exposure to air pollutants in the project study area was
characterized qualitatively by comparing the measured pollutant
levels with levels of those pollutants in other urban areas of
the United States.
1 SI/NJ UATAP report, Volume II, Section 1.3.
2 In this tally, m- and jo-xylene are counted as one chemical.
3 Note that formaldehyde is a VOC, that its segregation from
other VOCs in this report is a consequence of its collection
by a different method from that used for the other VOCs.
l-l
-------
The health risks associated with the exposure were
characterized by conducting a quantitative health risk
assessment. In a quantitative health risk assessment, data on
pollutant effects determined from experimental exposure of
laboratory animals or, for some pollutants, from human exposure,
are used to estimate the risk (likelihood or probability) of such
health effects at the pollutant concentrations measured in the
project area.
These risks are expressed either by comparing the measured
concentrations with levels that are considered to be
substantially without appreciable risk (for noncancer effects) or
by estimating the increased risk of cancer from exposure to the
measured pollutant levels. Both types of risk estimates are
termed "increased" risk to indicate that they do not express the
total risk of these effects. Many other environmental, socio-
demographic, and genetic factors contribute to an individual's
total risk of cancer and other health conditions.
Two approaches to quantitative risk assessment for the
project data are presented. The Level 1 risk assessments assume
that an individual is exposed for an entire lifetime to the
annual average air pollutant concentration recorded at one of the
project monitors for the period from October 1, 1988, through
September 30, 1989. The Level 2 risk assessments include both
indoor and outdoor monitoring data, and assume that body weight
and inhalation rate vary over the lifetime of the individual.
VOCs were addressed separately from metals, BaP, and formaldehyde
in the Level 1 risk assessments, so that there are two Level 1
risk assessments in this report. The Level 2 risk assessment
addressed only the 13 VOCs quantitated in indoor air during the
period from July 10, 1990, through March 19, 1991.
The risk assessments presented in this report employed
methodologies outlined in EPA guidelines. (U.S. EPA, 1986a,
1986b, 1986C, 1992).
Note that these risk assessments for the SI/NJ UATAP are not
complete assessments of air pollution risk for Staten Island and
nearby New Jersey/ since (1) the study compounds do not represent
all the pollutants in ambient air, and (2) exposure via routes
other than direct inhalation (i.e., ingestion and dermal) from
ambient air are not addressed.
1.2 AIR POLLUTANT CONCENTRATIONS IN THE 8I/NJ UATAP STUDY AREA
Tables V-l-3 and V-l-4 summarize the annual average
concentrations of the air pollutants monitored in the SI/NJ
UATAP. In general, these air monitoring results show that air
1-2
-------
pollutant levels in the region are in the ranges for those
pollutants in other urban areas in the United States. This is
illustrated by the data from other urban areas summarized in
Figures V-l-1 through 10 (VOCs), and Figures V-l-ll through 24
(metals, BaP, and formaldehyde).
Data from the EPA 1988 and 1989 Urban Air Toxics Monitoring
Program (UATMP) studies (U.S. EPA, 1989b; and U.S. EPA, 1989c)
were selected for this comparison since they provided
concentration data for virtually all of the SI/NJ UATAP study
chemicals for a large number of urban locations nationwide. Some
of the 1988 UATMP sites are in highly industrialized locations,
while others are in residential sections of urban areas. The
SI/NJ UATAP sites are in residential neighborhoods: on rooftops
of fire stations, schools, a police department, a post office,
and a pumping station; and at ground level at a hospital, a park,
and a private home. However these sites are generally very close
to highly industrialized areas, as evidenced by the
microinventory of emissions sources within on kilometer of each
monitor.4 A description of the 1988 UATMP sites and a brief
comparison of the SI/NJ UATAP and the UATMP studies are in the
appendix of this volume. The site reports prepared for the SI/NJ
UATAP are in Volume VI of the six-volume SI/NJ UATAP report.
When comparing data from different studies, differences in
reported concentrations should not be construed as significant
without knowledge of limitations in data quality. Even within
the SI/NJ UATAP set of data, apparent differences in
concentration should not be assumed to be statistically
significant. In the cases of the SI/NJ UATAP concentration data
for a limited set of chemicals, the statistical analysis section
of this volume presents intersite concentration differences that
were found to be statistically significant.
1.2.1 VOCs
For ten VOCs, Figures V-l-1 through 10 compare the minimum,
median, and maximum of the SI/NJ UATAP site annual averages for
the period from October 1988 through September 1989 (excluding
Piscataway, the background site for the VOCs) and the annual
averages for Piscataway for the same period, to the UATMP VOC
annual average concentrations for 1988 (October 1987 through
September 1988) and 1989 (January 1989 through December 1990).
The.VOC comparison data for 1988 are from air monitoring sites in
19 cities. The 1989 data were collected in 12 cities. Six
cities provided data in both years. The 1989 data include two
This microinventory is found in Volume III, Part A, of the
SI/NJ UATAP report.
1-3
-------
monitoring stations each for two of the cities (Washington, D.C.,
and Wichita, Kansas). The UATMP sites are listed below.
1988 UATHP sites
Atlanta, GA
Burlington, VT
Dallas, TX
Hannond, IN
Lansing, KY
Midland, HI
Sauget, IL
Baton Rouge, LA
Cleveland, OK
Dearborn, HI
Houston, TX
Louisville, KY
Port Huron, HI
Birmingham, AL
Chicago, IL (Carver H.S. and Washington, H.S.)
Detroit, MI
Jacksonville, FL
Miami, FL
Portland, OR
1989 UATHP sites
Baton Rouge, LA
Dallas, TX
Miami, FL
St. Louis, MO
Wichita, KS #1
Camden, NJ
Ft. Lauderdale, FL
Pensacola, FL
Washington, DC #1
Wichita, KS #2
Chicago, IL (Carver H.S. and Washington, H.S.)
Houston, TX
Sauget, IL
Washington, DC #2
Figures V-l-1 through 10 are described below. Note that
these descriptions do not ascribe significance to the magnitudes
of the differences observed.
The annual average concentrations of some of the VOCs
differed widely between the two monitoring stations in Washington
and Wichita. Large differences were observed in the annual
average concentrations of some chemicals in the same cities in
different years, e.g., in the cases of benzene and
trichloromethane in Dallas. Some all-city median chemical
concentrations also varied widely between the two groups (1988
and 1989 groups) of cities. This demonstrates the variability
that may be found within an urban area in the same year, and
between sets of data for different years.
Annual average chemical concentrations at the other SI/NJ
UATAP monitoring sites were generally within the range of the
annual average concentrations for the same chemicals at the UATMP
sites. For the xylenes and trichloromethane, the SI/NJ UATAP
data are at the low end of the range of UATMP concentrations.
The dichloromethane levels at the SI/NJ UATAP sites were higher
than levels in many of the other cities. The tetrachloromethane
and benzene levels at the SI/NJ sites were higher than those at
the 1988 UATMP cities; but at the low end of the range of
concentrations of the 1989 UATMP cities.
In general, then, for the periods compared, exposure to
these ten VOC compounds in the SI/NJ UATAP study area are in the
range of exposures in other urban areas. Dichloromethane levels
were at the high end of the range of levels from the other
cities.
1-4
-------
1.2.2 Metals. BenzoTalpyrene. and Formaldehyde
For the metals, BaP, and formaldehyde, Figures V-l-11
through 24 compare the medians of the annual average
concentrations for the SI/NJ UATAP sites with the annual average
concentrations for the 1988 UATMP sites.
Concentrations of molybdenum, nickel, and vanadium at the
SI/NJ UATAP sites were higher than those reported for the UATMP
sites; and the concentration of chromium was higher than at most
of the UATMP sites. The risks associated with these
concentrations are discussed in section 1.4.5. The median annual
average concentration of nickel at the SI/NJ UATAP sites is two
to ten times higher than concentrations at the UATMP sites, with
the exception of Louisville, KY, where the annual average
concentration of nickel was about 1.5 times higher than the SI/NJ
UATAP median.
1.3 QUANTITATIVE RISK ASSESSMENT - HAZARD IDENTIFICATION,
REFERENCE CONCENTRATIONS, AND INHALATION UNIT RISK FACTORS
As defined by the National Academy of Sciences in 1983
(National Research Council, 1983), risk assessment involves one
or more of the following steps:
o hazard identification,
o dose-response assessment,
o exposure assessment, and
o risk characterization.
Risk assessment has been used extensively by regulatory
agencies to compare potential risks from different chemicals,
from exposure through different media (air, water, soil), and
from exposure in different geographic areas. It is important to
note that the risks presented are probabilities involving
assumptions that may lead to over- and underestimates of risk.
1.3.1 Hazard Identification
The 40 chemicals studied are known to cause, or are
suspected of causing adverse health effects. The health hazards
that may arise from exposure to the chemicals studied in the
1-5
-------
SI/NJ UATAP include non-cancer toxicity (e.g., effects on the
liver or central nervous system)5 and cancer.
1.3.2 Reference Concentrations and Inhalation Unit Risk Factors
Inhalation reference concentrations (RfCs)6 and/or
inhalation unit risk factors (lURFs) to assess noncancer toxicity
and for cancer, respectively, have been identified for 167 of the
study chemicals for which concentration data are available.
Tables V-l-6 and 7 list the available RfCs and lURFs.
1.3.2.1 Reference concentrations for noncancer toxicity
Where available, RfCs were used in the risk assessment for
noncancer health effects. Currently, only a small number of
inhalation RfCs is available on the Integrated Risk Information
System (IRIS) (U. S. EPA, 1990a), which contains the EPA's
consensus toxicological information on approximately 500
chemicals.8 Table V-l-6 identifies those study chemicals for
The potential target organs for the non-cancer health effects
of the study chemicals are listed in Table V-l-17 in
conjunction with the discussion of additive risk.
In this report, "reference concentration" and "RfC" are used
as generic terms, not specific to the reference concentrations
provided by EPA's Reference Concentration/Reference Dose
Workgroup. See discussion in Section 1.3.2.1.
The xylenes (for which RfCs were recently withdrawn from IRIS),
and lead and zinc (for which National Ambient Air Quality
Standards, and not RfCs are used) are excluded from this count.
The system is updated on a monthly basis to reflect currently
available toxicological and regulatory information. IRIS is
available to the public through the National Library of
Medicine's TOXNET System.
1-6
-------
which RfCs were available on IRIS or in the HEAST9 (U.S. EPA,
1992b), or were provided by the New York State Department of
Health (NYSDOH).
Inhalation RfCs appearing on IRIS represent EPA consensus;
they were developed by EPA's Reference Dose/Reference
Concentration Workgroup. This workgroup is comprised of senior
scientists from the different Program Offices and Regions within
EPA who have expertise in inhalation and oral toxicology, and
risk assessment.
An IRIS inhalation RfC considers toxic effects for both the
respiratory system (portal-of-entry) and for effects peripheral
to the respiratory system. The RfC is expressed in units of
milligrams/cubic meter (mg/m3) . In general, the inhalation RfC
is an estimate (with uncertainty spanning perhaps an order of
magnitude) of a daily inhalation exposure to the human population
(including sensitive subgroups) that is likely to be without an
appreciable risk of deleterious effects during a lifetime. RfCs
were derived according to the "Interim Methods for Development of
Inhalation Reference Doses" (U. S. EPA, 1990b) developed by EPA
scientists and peer-reviewed. The RfC methodology has been
reviewed by the Science Advisory Board, a panel that reviews
scientific documents for EPA. Note that RfCs can be derived for
noncarcinogenic health effects of carcinogenic compounds.
The development of the RfC involves an analysis of the
available toxicological data to identify the No Observed Adverse
Effect Level (NOAEL). The NOAEL is defined as an exposure level
at which there are no statistically or biologically significant
increases in the frequency or severity of adverse effects between
the exposed population and appropriate controls. While effects
may be produced at this level, they are not considered adverse
per se, or precursors to specific adverse effects. When research
results yield several NOAELs for different adverse effects from a
chemical, the regulatory focus is primarily on the lowest one.
To protect sensitive subpopulations (children, the elderly,
etc.) exposure should be limited to a fraction of the NOAEL by
introducing suitable factors from 10 to 100,000. The RfC is
calculated using the following equation:
9 The Health Effects Assessment Tables (HEAST) document is
developed by the Office of Solid Waste and Emergency Response
and the Office of Research and Development. The health
effects information in these tables is regarded as provisional
risk assessment information in that, except where the
information is referenced to IRIS, it does not represent an
EPA-wide consensus, although its inclusion does indicate
concurrence by individual EPA program offices.
1-7
-------
NOAEL
RfC = .
(Modifying Factor) X (Uncertainty Factor)
Uncertainty factors (factors of 10) are intended to account
for
(1) the variation in sensitivity among the
members of the human population;
(2) the uncertainty in extrapolating from animal
data to human exposures;
(3) the uncertainty in extrapolating from data
obtained in a study that is less than
lifetime exposure; and
(4) the uncertainty in using the Lowest Observed
Adverse Effects Level (LOAEL) when a NOAEL
has not been or cannot be determined.
Based on the Workgroup's assessment of the overall data base for
a chemical, a factor of 10 is assigned for each applicable source
of uncertainty, and the factors are multiplied to yield an
uncertainty factor. The maximum uncertainty factor is 3000; this
takes into account the compounding of error likely when
uncertainty from all four sources affects the derived RfC.
The modifying factor ranges from 1 to 10, depending on the
overall database (i.e., the number and quality of studies
available, quality of data, etc.) for the chemical; the default
value is 1.
Summaries of the bases for the RfCs developed by EPA and
NYSDOH are provided in Volume VI, Appendices. Also in Volume VI
are memoranda concerning the differences in development of the
former HEAST RfC for chromium and the current NYSDOH RfC for
chromium (Dollarhide, 1992), and the status of the RfC for xylene
(Poirier, 1992).
1.3.2.2 Unit risk factors for carcinogenicity
Table V-l-7 summarizes the unit risk factors available for
the study chemicals, their Chemical Abstract
Service Numbers, and the Carcinogenic Weight of Evidence
classifications. The Weight of Evidence classification was
developed by EPA's Carcinogen Risk Assessment Verification
Endeavor (CRAVE) Workgroup, composed of senior scientists from
EPA program offices selected for their expertise in assessing
carcinogens. CRAVEfs review process involves an extensive
analysis of the available toxicological, scientific and cancer
1-8
-------
information on the chemicals. Based on this review, CRAVE then
assigns a Weight of Evidence Classification to the chemicals as
outlined in EPA's Risk Assessment Guidelines for Carcinogens
(U.S. EPA, 1986a). The EPA classification is as follows:
Group A - Human carcinogen (based on human epidemiological evidence)
Group B - Probable human carcinogen
Bl - indicates limited evidence from human studies
B2 - indicates sufficient evidence from animal studies, but
inadequate evidence from human studies
Group C - Possible human carcinogen (based on animal evidence)
Group D - Not classifiable as to human carcinogenicity (based on lack
of evidence)
Group E - Evidence of non-carcinogenicity for humans
Chemicals ranked as Group D carcinogens lack adequate data for
the development of cancer dose-response information; they are
treated as non-carcinogens until additional research information
becomes available and the chemical can be reclassified. The
CRAVE Workgroup reviews data on an on-going basis; updates are
provided on IRIS.
Three of the metals and one of the VOCs in the SI/NJ UATAP
have been classified as, or associated with compounds classified
as, Group A carcinogens, one metal and formaldehyde are Group
Bl. Five VOCs, one metal, and BaP are Group B2. Three VOCs and
two metals are Group D. Of the remaining chemicals analyzed
during the project, but excluded from quantitative risk
assessments for carcinogenicity, two are Group D (1,1,1- and
1,1,2-trichloroethane); unit risk factors are available for two,
but no valid ambient air concentration data were reported
(chloromethane and beryllium); and 16 are unclassified.
When assessing cancer risks, EPA assumes that the
carcinogenic substances cause some level of risk at any exposure
level; that is, a zero-threshold for adverse effects is assumed.
In developing unit risk .factors, EPA typically uses a non-
threshold, linearized, multistage model to extrapolate from high-
dose data of animal tests to the low doses typically resulting
from human exposure to low concentrations in ambient air. Using
this model, a cancer slope factor (CSF)—proportion affected per
unit of dose—is developed by CRAVE for each chemical. The CSF,
expressed as (milligrams of substance per kilogram [kg] body
weight per day [d])'1, and weight of evidence can be used to
1-9
-------
For this risk analysis, the Inhalation Unit Risk Factor
(IURF) was used. The IURF is a quantitative estimate of the
increased probability of developing cancer from a 70-year
lifetime continuous exposure to a concentration of one microgram
of a given pollutant per cubic meter air. The lURF's assume that
the individual weighs 70 kg (154 Ibs.), and that the rate of
inhalation is 20 cubic meters/day over a 70-year period.
The CSFs and lURFs in Table V-l-7 that were found in IRIS
reflect the EPA consensus for these chemicals as of January 28,
1992. As new toxicological information becomes available, the
CRAVE Workgroup reviews it and, if warranted, changes the CSF and
IURF to reflect the new data.
The chromium IURF is based on chromium VI (valence +6), for
which carcinogenic data were available. The amounts of chromium
VI in the ambient air samples were not determined. In this risk
assessment, chromium VI is treated as the only carcinogenic
component of the total (reported) chromium concentration10, and
the total chromium is assumed to contain 1 or 10% chromium VI.
This range was chosen because of the absence of site-specific
data on sources of chromium VI versus chromium III.11 This may
lead to over- or underestimates of excess risk.
1.4 LEVEL 1 RISK ASSESSMENTS
Table V-l-8 summarizes the availability of RfCs, lURFs, and
ambient air concentration data for quantitative risk assessment.
1.4.1 Exposure Assumptions
The basic exposure variable used in the Level 1 risk
assessment is the annual average ambient air concentration for
the second year of the study (October 1988 through September
1989). The second-year data from each sampling organization were
selected for risk calculation since these data are regarded as
10 Chromium has also been found in the +3 valence state;
chromium III has not been determined to be carcinogenic.
11 Discussions with OAQPS and review of EPA's sludge regulations
indicated that chromium VI is not as stable as chromium III;
therefore, it is expected that the latter oxidation state
would be more prevalent in the ambient air.
1-10
-------
more self-consistent than the first-year data12, and since one
year is commonly used for assessing chronic risks. Annual
averages were calculated as follows:
Annual average = (Zn{x{) / (Znt) ,
where n( = number of samples in the i"1 quarter, and
x> = average concentration in the i* quarter.
For the three sites Bayley-Seton, Eltingville, and Dongan Hills,
these annual averages are biased towards the first and second
quarters of the year because of the greater sampling frequency
during those quarters.13
The Level l risk assessment assumes that an individual would
be exposed for 70 years to the concentrations identified in
Tables v-1-3 and V-i-4. in addition, the individual is assumed
to weigh 70 kgs, and inhale 23 cubic meters14 of ambient air
during 24-hour exposures every day of those 70 years.
These assumptions are used commonly for screening risk
assessments such as those presented here (e.g., U.S. EPA, 1989d;
U.S. EPA, 1990c).
Anecdotal information and the 1990 census data suggest that
many residents in the study area spend a good part of their time
either living in the community or working there. However, it was
not .possible to determine the percentage of the community that
would match the 70-year assumption used in this risk assessment.
(A summary of the population analysis is included in the appendix
of this volume.) For an unknown percentage of the community,
12 The_Quality Assurance Section supports this conclusion.
During the beginning of the first year of monitoring,
organizations were de-bugging their operations; sampling and
analysis by each organization was more constant through the
course of the second year. See Volume III, Part A, Section 2,
of the SI/NJ UATAP report.
13 The effect of the differences in sampling frequency is most
pronounced in the annual averages listed for Dongan Hills for
hexane, benzene, toluene, and m- and p-xylenes. If the
average of all of the sample concentrations is compared to
annual averages computed by averaging the quarterly averages,
the differences are that the latter set of averages is lower
than the former by 0.14 ppb (16%), 0.53 ppb (27%), 0.57 ppb
(14%), and 0.36 ppb (14%), respectively.
14 This value is based on 16 hours of light activity and 8 hours
of resting, assumptions that are part of the Reference .Man
scenario (ICRP, 1981).
1-11
-------
these calculations might lead to errors in risk estimates for the
following reasons:
Residents might not live in the community for an entire
lifetime (70 years). Current data in the Exposure
Factors Handbook (U. S. EPA, 1989a) indicate that the
average person (50th percentile) moves every 9 years
with an upper bound estimate (90th percentile) of 30
years. If an individual moves to an area with no
exposure to these chemicals after 30 years or 9 years,
the individual's exposure would be reduced.
Conversely, if relocation resulted in exposure to
higher concentrations than those in the study area, the
lifetime risks presented here would be underestimates.
- Residents might spend a portion of the day away from
the area.
Residents might also spend part of the year on
vacations or outside of the area.
The Level l approach to the exposure assessment, with its
use of outdoor air concentrations and default (standard)
assumptions, does not address the health impact of episodic high
exposures and other short-term (acute and less than 1 year)
exposure variations relative to the annual average
concentrations, or of activity patterns and indoor exposures.
The Level 2 risk assessment includes both indoor and outdoor VOC
exposure data.
One of the exposure (dose) assumptions in the Level 1
analyses is 23 m3/day as an inhalation rate. This inhalation
rate assumption (constant) is not the standard assumption used by
EPA. This modified inhalation rate has been incorporated into
the inhalation RfCs and lURFs, yielding the modified (adjusted)
RfCs and lURFs listed in Tables V-l-6 and 7.15
1.4.2 Results for the VOCs
For noncancer health impacts, the Table V-l-3 ambient air
concentrations were compared to the Table V-l-6 reference
15 The standard RfCs and lURFs are based on a standard
inhalation rate = 20 m3/day at a temperature of 25 °C,
lifetime = 70 years. The Level 1 risk assessment deviates
from this standard scenario both in inhalation rate and
temperature assumed (20 °C) .
1-12
-------
concentrations. The formalization used for this comparison was
the Hazard Quotient (HQ), where HQ = ambient concentration/
reference concentration.16'17
Table V-l-9 presents the HQs calculated for 9 VOCs. The HQs
are less than 0.1 for all compounds except benzene,
tetrachloromethane, and, for two sites, tetrachloroethene; only
benzene has an HQ greater than 1. According to the current
perspective of exceedance of an RfC, an HQ greater than 1
suggests a need for further analysis of the bases (e.g., RfC,
concentration data) of the calculated HQ. From an additive risk
perspective, a mixture of chemicals (here, ambient air) with HQs
less than one and a common target organ may result in a risk of
concern for that target organ (i.e., a hazard index greater than
1 for the target organ). See Section 1.6 for further detail.
For potential individual cancer risk estimates,
multiplication of the modified unit risk factors (modified to
reflect an inhalation rate of 23 m3/day rather than 20 m3/day) by
the Table V-l-3 concentrations of the pollutants generated the
Table V-l-10 estimates of potential excess lifetime cancer risk
per million for 6 VOCs. Since the estimates apply to a 70-year
period, the annual risks would be lower.
The resulting estimates of potential excess cancer risk
range from 0.4 to 61 per million over a lifetime (e.g. 70 years).
For benzene, tetrachloromethane, and trichloromethane— compounds
detected at all SI/NJ UATAP monitoring sites—estimated excess
lifetime cancer risks are higher than 10/million (10's) at some of
the sites. For benzene and tetrachloromethane, the estimated
excess cancer risks were found to be consistent across all
16 This deviates from the definition of hazard quotient in that
the definition specifies that the subject exposure be of
duration similar to that in the study from which the
reference concentration was derived. The HQs in this risk
assessment were calculated without regard for likely
differences between the subject chronic exposures and the
durations of the exposures leading to the reference
concentrations. However, most of the reference doses are
derived from chronic exposure studies.
17 The HQ does not define a dose-response relationship;
therefore, its numerical value should not be construed to be
a direct estimate of risk. (Adapted from the discussion of
Hazard Index in "Guideline for the Health Risk Assessment of
Chemical Mixtures," [U. S. EPA, 1986b]).
1-13
-------
sites.18 While the estimated potential excess cancer risks
calculated for trichloromethane at the Staten Island sites of
Susan Wagner, PS-26, Port Richmond, Pump Station, Great Kills,
and Tottenville were 1.7 to 8 times higher than at the New Jersey
and remaining Staten Island sites, this result must be viewed
with caution since the analytical methodologies were not
equivalent at all sites.19 (See Section 2 of Volume III, Part A,
for further detail.) Trichloromethane is very volatile;
breakthrough may have occurred for this compound at the New
Jersey sites and at the Bayley-Seton, Eltingville, and Dongan
Hills sites in New York, but, due to differences in effectiveness
of the sorbents used in sample collection, not at the other six
Staten Island sites. For further discussion of apparent site-to-
site differences, see the statistical analysis of the VOCs data
in Section 2 of this volume.
1.4.3 Results for the Metals. Benzoralpyrene. and Formaldehyde
Table V-l-3 summarizes the ambient air concentrations for
the metals, BaP, and formaldehyde. Tables V-l-11 and 12 present
the Level 1 risk calculations for the non-cancer and cancer
effects of the study chemicals for which reference concentrations
(RfCs) and/or unit risk factors (URFs) were available. In the
case of lead, concentrations were compared to the National
Ambient Air Quality Standard (NAAQS) for lead.2.0 For zinc,
concentrations were compared to the NAAQS for PM-1021.
18 While the statistical analysis (Section 2 of this volume)
found statistically significant intersite differences in the
concentrations of benzene and tetrachloromethane, the risk
assessment was not sensitive to these differences.
19 The sites listed were run by NYSDEC; the remaining sites were
run by NJDEP/NJIT and CSI.
20 The current NAAQS for lead is under review; it might be
revised downward to reflect current toxicological and
epidemiological evidence of the neurotoxic effects of lead.
Thus, this risk assessment for lead is not regarded as
conservative.
21 PM-10 is particulate matter with an aerodynamic diameter less
than or equal to 10 microns, the size range considered
respirable.
1-14
-------
The hazard quotients for the non-cancer effects were less
than 1 for cadmium, manganese, mercury, zinc, and formaldehyde22;
the concentrations reported for lead were 3 to 7% of the NAAQS
for lead. The HQ values calculated for nickel were close to 1,
suggesting a need for further information on the nickel species
present in the atmosphere, and further refinement of the RfC for
nickel. The hazard quotient for total chromium was calculated
using two different RfCs—one formerly from EPA23, the other from
NYSDOH. The former EPA RfC is regarded as more appropriate for
chromium VI than for total chromium since it is based on an
occupational study of workers exposed primarily to chromic acid
containing Cr VI. The justification for using 10% and 1% of the
total chromium concentration as the chromium VI concentration
follows under the discussion of cancer risk. The NYSDOH value
yielded HQs from 0.1 to 0.3; while the former EPA value yielded
HQs from 0.07 to 1.5, depending upon the site and the percent
chromium VI assumed. These results suggest the need for further
data on the chemical composition of the chromium in ambient air,
and for EPA's forwarding an RfC for chromium.24
As shown in Table V-l-12, the estimated increases in
probability of developing cancer were in the range from 0.24 to
37 in a million. Cancer risks were less than 1 in a million
(1Q-6) for BaP and formaldehyde; as mentioned in Section 1.1, this
risk for formaldehyde is likely to be an underestimate due to
ozone interference with the sampling method. For arsenic,
cadmium, chromium, and nickel, they were greater than 1 in a
million.
Since chromium in its hexavalent oxidation state is the only
chromium species believed to be carcinogenic, cancer risk for
chromium was calculated using the assumptions that 10% or 1% of
the total chromium concentrations was chromium VI. Chromium VI
22 The value used as an RfC for formaldehyde was developed as a
guideline concentration for short-term (1- to 4-hour)
exposures. In addition, ozone interference with the
formaldehyde sampling method tends to lead to formaldehyde
readings that are less than actual. If a chronic RfC were
lower than the short-term RfC and if ozone interference did
occur, then the HQ derived from the ratio of the low annual
average to the high reference concentration might tend to
underestimate the noncancer risk from formaldehyde.
23 The EPA RfC for chromium was withdrawn from the HEAST as of
the 1992 update. The RfC for chromium is under review at
EPA.
24 An ongoing EPA cancer and noncancer effort for chromium is
expected to provide an RfC for chromium in 1993.
1-15
-------
is found at percentages substantially higher than 10% only in
locations adjacent to major emission sources (e.g., chromium
electroplaters, cooling towers using chromate as a bactericide)
or in areas immediately adjacent to hazardous waste sites
containing chromium-laden slag (Lioy et al., 1992a). In an
effort to be conservative25, the 10% value was included in the
calculation. This assumes that a source with emissions similar in
composition to incinerator emissions was located just beyond one
kilometer from each site; no such source was within one kilometer
of any of the three New Jersey monitors for which valid data are
available.26
The carcinogenicity of nickel, like that of chromium, varies
with the compound in which the metal is presented. However, the
specific compounds in the ambient air samples were not
determined. Cancer potency factors have been derived for nickel
subsulfide and for nickel refinery dust. Nickel subsulfide is
associated primarily with a special process sometimes used in the
refining of nickel. Thus, use of the nickel subsulfide cancer
potency factor or RfC in a quantitative risk assessment for
nickel in ambient air was regarded as inappropriate. The cancer
potency factor and RfC for general refinery dusts, composed
mostly of nickel oxide and elemental nickel, were considered more
appropriate; they were used to calculate risks where nickel
subsulfide is not known to be present.
25 This use of the term "conservative" indicates an intent to err
in the direction of protecting human health by overestimating
risk, avoid underestimating.
26 In this part of New Jersey, stacks from chromium-emitting
sources are less than 1000 feet high, low enough that the
greatest impact from the source emissions would be within 1
km of the stack, and decrease rapidly beyond 1 km. Thus,
sources farther than 1 km from the monitor were regarded as
without significant impact on the monitor. Assuming that all
other sources contributing Cr VI to the total chromium
contain less than 10% Cr VI, 10% would be the highest
proportion of Cr VI in the sample. Based on this assessment
of dispersion characteristics of incinerator emissions, an
assumption of 10% chromium VI was considered conservative
from a risk assessment perspective. (Lioy, 1992b.)
If the general scenario above is inappropriate for specific
sources and monitor sites in this study, this estimate of 10%
as the hexavalent proportion of the total reported chromium
concentrations may not be conservative. In the absence of
better site-specific information, however, there is precedent
within EPA for use of the 10% assumption.
1-16
-------
1.4.4 Uncertainties and Limitations
A list of the uncertainties and limitations in approach
affecting the Level 1 risk assessments follows.
1. The non-cancer and cancer risk characterizations presented
are based on the assumption that the annual average
concentration derived from one year of monitoring data
reflects an individual's exposure to a given pollutant at a
site for a 70-year period. Since emissions and, hence, air
quality vary from year to year with such emission source
variables as automotive usage and fuel composition, and
industrial plant operations and production, and since the
amount and direction of variation is unknown, it is unclear
how much this assumption affects the calculated risks.
For example, auto usage or fuels may change, and few plants
in the area will operate or emit air pollutants at the same
levels for 70 years, though the area in which they are
located may remain industrial. Thus, future exposures could
be lower or higher than the SI/NJ UATAP monitoring data
indicate. The controls mandated by the Clean Air Act
Amendments of 1990 should continue to lower the
concentrations of many of the pollutants measured in this
study. Future pollutant exposures and risks will be lower
than risks based on 1988-1989 concentration data if the
control steps actually do lower the airborne concentrations,
and if these reductions are not offset by future growth.
In addition, these risk assessments do not address the
consequences of short-term peak exposures (e.g., as a result
of periodic releases from point sources) to concentrations
higher than the annual average concentrations. However the
risk assessment for such exposures would require the use of
health effect dose-response estimates tailored to the
particular exposure assessment, and concentration data
focusing on peak concentrations.
2. The calculated excess risks assume continuous outdoor
exposure, without addressing the potential exposures from
indoor environments in which many people in this country
spend much of their time. Indoor concentrations of certain
pollutants (e.g., formaldehyde and several VOCs) are
commonly several times higher than outdoor concentrations.
Thus, risk estimates based on outdoor air concentrations
alone may underestimate the contribution of such pollutants
to total risk, in contrast, for a pollutant with incomplete
penetration into the indoor environment from outdoor sources
and no indoor sources, risk estimates based on outdoor air
concentrations alone may lead to a higher estimate of the
contribution of such pollutants to total risJC.
1-17
-------
3. The analyses assume that people are continuously exposed to
air toxics at the average levels measured at the monitoring
station. This assumption does not consider such exposure
variables as a person's moving throughout the urban area,
spending time outside of the area (e.g., during vacations,
work), changing homes several times during a lifetime, or
living closer to a source than the location of any of the
monitors; or differences in lifestyle. (Section 1.4.1
contains further detail.) An overestimate of risk from
exposure to air in the study area may result; and an over-
or underestimate of risk from exposure to air in all
locations, depending on whether air pollutant concentrations
in the other locations are higher or lower than in the study
area.
4. The RfCs for the individual chemicals were derived using
different methods (available NOAEL, extrapolation from oral
RfD to inhalation RfC, etc.) that include different
uncertainty adjustments and modifying factors
5. In developing linearized unit risk factors, EPA uses a non-
threshold linearized multistage model, which is linear at
low doses, to extrapolate from high-dose experimental data
to the low doses typically caused by exposure to ambient air
pollutants. In other words, carcinogenic substances are
assumed to cause some risk at any exposure level. If the
true dose-response relationship at low doses is less than
linear (e.g., has a threshold), then the unit risk estimates
based on EPA lURFs would tend to be high, and therefore
overestimate the risk.
The unit risk factor is based on the upper bound of a 95%
confidence interval; if the true unit risk values are less
than that upper bound, then the calculated risks might be
overestimates.
6. The cancer Weight of Evidence, Inhalation Unit Risk Factors,
and reference concentrations reflect the current state of
toxicological knowledge for the specific chemicals. As more
scientific information is acquired, these values could
change significantly, as they have in the past, and thus the
magnitudes and relative contribution of particular
pollutants to estimated risk can change. The result is a
degree of uncertainty that cannot be assessed.
7. The risk estimates presented do not address the potential
for the ambient air mixture of pollutants to exhibit
biological activity that is synergistic, additive, or
antagonistic relative to their individual effects. An
additive risk assessment is presented in Section 1.6.
1-18
-------
8. The pollutants monitored do not include all pollutants
present in ambient air.
9. Uncertainty resulting from issues related to chromium and
nickel are discussed in Section 1.4.3.
10. Particles collected were <50 microns in aerodynamic
diameter, a size range that includes particles larger than
the 10-micron aerodynamic diameter considered the upper end
of the respirable-size range. Thus, risk estimates driven
by the concentration of respirable-size particles may be
overestimated.
11. Errors or limitations in the reported concentrations affect
the reliability of the risk estimates. The direction of the
impact on the risk estimate varies with the chemical.
Discussions of data quality are found in Volume III, Parts A
and B. Note the following:
Chemical analytical standards for accuracy were not run
for mercury at carteret, Elizabeth, and Highland Park;
and for arsenic at Susan Wagner High School (SW) and
PS-26.
The analytical recovery of BaP was poor for samples at
SW and PS-26 to the extent that the reported BaP
concentrations at these sites should be regarded as
minimum values. The recovery of nickel from samples at
these same sites was 75%.
An ozone interference with the formaldehyde collection
method used in the ambient air portion of this project
resulted in the reporting of formaldehyde
concentrations as less than actual; no correction based
on ozone concentration is available.
The availability of concentration data for a chemical
at some sites but not others indicates that the omitted
data did not meet sampling and/or analysis data quality
objectives for the project, or that quantitation of
that chemical was not performed by the analytical
lab(s) connected with the site(s).
1.4.5 Discussion
This Level l risk assessment focused on the chemicals for
which health information (IURFS and RFCs) or a NAAQS, and
quality-assured/quality-controlled data were available.
1-19
-------
Of the VOCs, only benzene was present at concentrations
exceeding its RfC. Benzene, tetrachloromethane, and
trichloromethane concentrations yielded estimated excess lifetime
cancer risks that exceeded ten per million. The risk estimates
for these chemicals were similar to those estimated for the UATMP
sites. See Tables V-l-13 and 14.
The risk estimates for chromium and nickel in the study area
were higher than the estimates for the 1988 UATMP sites. These
results suggest the need for further analysis of the chemical
composition of the chromium (i.e., valence state) and nickel
(dust, subsulfide, etc.) in ambient air, and of the information
supporting the RfCs for chromium. Again, apparent site-to-site
differences in estimated risks may not be statistically
significant. No effort has been made to characterize the
accuracy or precision of the UATMP annual averages and no
statistical tests were performed on the differences in
concentrations between the SI/NJ UATAP data and the UATMP data.
1.5 LEVEL 2 RISK ASSESSMENT (FOR VOCS ONLY)
1.5.1 Introduction
People are exposed to air contaminants outdoors; at work or
school; in cars, buses, and trains; and in their homes. The air
inside buildings and vehicles comes from outside and so generally
contains the same contaminants as the outdoor air. However,
there are many indoor sources that can increase the level of air
contaminants inside houses, offices, schools, and other
buildings. Smoking and other activities also increase
contaminant levels inside vehicles.
The Level 1 risk assessment bases the estimates of lifetime
average air contaminant exposure on outdoor contaminant levels
measured during this project. The Level 2 assessment supplements
the Level 1 risk characterization by indicating how exposure and
risk levels differ when considering both indoor and outdoor
contaminant levels for four residences in the study area.
Studies by the National Academy of Sciences (NAS, 1981) and
the National Institute of Occupational Safety and Health (NIOSH,
1989) have shown that as buildings become more air-tight (in
response to demands for energy conservation), and the exchange of
air between inside and outside decreases, there is a potential to
increase the concentrations of pollutants within the building.
Contaminants that have been found frequently indoors include
formaldehyde, combustion products (i.e., from stoves, heaters,
1-20
-------
and smoking), bacteriological contaminants, and some outdoor air
pollutants that penetrate indoors.
To assess the potential risks from indoor air pollutants,
NYSDOH carried out sampling for VOCs in four homes and two nearby
outdoor sites in the study area for a period of eight months.
These four homes are not to be regarded as representative of
indoor air in the study area.
The Level 2 risk assessment is provided for VOCs only.
Particulates data were not collected indoors, and results from
the indoor air formaldehyde samples were invalid due to sampler
malfunctions. Data from other studies show that indoor
formaldehyde concentrations are generally greater than outdoor
concentrations. For example, data on the National Ambient
Volatile Organic Compounds (VOC) Data Base (Shah and Heyerdahl,
1988) indicate a median ambient concentration of 4.1 ppb, and a
median indoor concentration of 42 ppb. This national median
indoor concentration, twice the RfC, corresponds to an estimated
increased lifetime cancer risk of over 100 per million.
Concentrations of the naturally-occurring radioactive gas
radon were also measured as part of the indoor air study. The
indoor levels ranged from 0.19 picoCuries per liter (pCi/1) to
1.86 pCi/l; the outdoor levels ranged from 0.30 to 1.37 pCi/1.
While radon exposure via indoor air is a significant contributor
to estimated excess cancer risk, it is not addressed in detail in
this report because of the non-anthropogenic nature of this air
pollutant. The results, including estimated potential excess
cancer risk, are discussed more thoroughly in Volume IV.
1.5.2 Concentration Data Used in the Level 2 Exposure Assessment
1.5.2.1 Residential site indoor and outdoor data
The Level 2 risk assessment is based on yoc concentration
data from four homes and two nearby outdoor sites that were
monitored as part of the indoor air portion of the SI/NJ UATAP.
Two homes were located in Staten Island, and two were located in
Carteret, New Jersey. While Volume IV of the project report
presents those data, certain information is repeated here to
establish the context of the indoor and outdoor data. The
indoor/outdoor VOC concentration ratios for the indoor air
portion of the project are summarized in Tables V-l-15 and 16.
Statistical tests, the results of which appear in Tables V-l-15
and 16, were performed to determine whether there were
significant differences between the mean indoor and outdoor VOC
concentrations for each house. See Volume IV, Indoor Air, for a
detailed discussion of'these results.
1-21
-------
1.5.2.2 Comparison of the outdoor air data from the indoor air
and the ambient air portions of the SI/NJ UATAP
Tables V-l-17 and 18 compare ambient data collected at PS-26
in Staten Island and Carteret High School in New Jersey by the
New York State Department of Health (NYSDOH) from July 1990
through March 1991, to data collected at PS-26 and another
location in Carteret from October 1988 through September 1989.
The data collected during the quarter beginning April 1989 are
not included so that data collected during the same seasons
(although different years) can be compared. In the discussion
that follows, the data from the ambient air (two-year) portion of
the SI/NJ UATAP are referred to as the UATAP data, and the indoor
air data are referred to as the NYSDOH data. Trichloromethane
(chloroform), tetrachloromethane (carbon tetrachloride),
trichloroethene (trichloroethylene), and tetrachloroethene
(tetrachloroethylene) were not detected in enough NYSDOH samples
to make valid comparisons. At PS-26, for six chemicals, the mean
concentrations reported by NYSDOH were higher than the mean
values reported by the UATAP; ratios of NYSDOH means to UATAP
means at PS-26 ranged from 1.3 to 3.5. For Carteret, this ratio
ranged from 0.8 to 4.4. Mean NYSDOH values for hexane and
benzene at Carteret High School were lower than those reported
for the UATAP Carteret site. Mean NYSDOH values for all other
chemicals at Carteret High School were higher than the mean
values reported for the UATAP Carteret site.
The results of this limited study conducted by NYSDOH are
generally in good agreement with the indoor sampling results of
the TEAM study (U.S. EPA, 1987b) and the VOCs data base, and the
two-year ambient air sampling portion of the SI/NJ UATAP.
In summary, the outdoor air data from the two portions of
the project are consistent considering that they were collected
in different years. Some differences between the two sets of
data were expected due to such factors as different
meteorological conditions, variations in source emissions,
interlaboratory differences in accuracy, and, in the case of
Carteret, different monitoring sites. In all cases, however, the
NYSDOH concentrations remained in the <10-ppb range, and the
outdoor levels for the residential indoor study did not differ
substantially from those reported for the ambient air portion of
the SI/NJ UATAP. The maximum differences for compounds available
from both data sets was 2.2 ppb (for toluene in Travis, Staten
Island), with the NYSDOH value being higher. Such differences
are regarded as without significant impact on the risk
assessment, such that the shorter-term, eight-month data set for
two outdoor sites is regarded as appropriate for use in the Level
2 risk assessment, where indoor and outdoor concentrations
measured simultaneously are preferable.
1-22
-------
1.5.3 Exposure Assessment
Increased health risks associated with exposure to the VOC
compounds which were measured in the ambient air and inside homes
are estimated under the assumption that an individual is exposed
to the average measured levels for his or her entire lifetime.
For this purpose, the lifetime is divided into three segments:
from birth to two years of age, from three to eighteen years of
age and from nineteen to seventy years of age. Average body
weights and inhalation rates are assigned to each of these age
intervals in calculating the individual's dose from inhalation of
pollutants in the air. The individual is assumed to spend a
portion of every day indoors and the remainder outdoors. This
division of time is assumed to vary with age according to a
specific scenario. The inhalation rate is also assumed to vary
between indoor activities and outdoor activities.
The general equation and the data used in making the Level 2
exposure estimates are presented Table V-l-19, along with a
sample calculation of lifetime average daily dose for a given
residence. The inhalation rates and body weights were derived
from data in the "Exposure Factors Handbook" (U.S. EPA, 1989a),
Tables 3-1 and 3A-2. From Table 3-1, resting, light exercise,
and moderate exercise exercise inhalation rates (cubic meter per
hour, m3/h) for a 10-year-old child were used for the age
interval three to 18; and the rates for an average adult were
used for the age interval 19 to 70. Data for the inhalation rate
of the infant (birth to two years old) were derived from Table
3A-2 in the handbook, which lists resting minute ventilation
rates (liters per minute, 1/m) for the infant. The range is from
0.25 to 2.09 1/m, with a mean value of 0.84 1/m. No data are
given for periods when the infant is exercising (for example,
crying, or, at a later age, crawling or walking). To account for
such periods of increased minute ventilation in the present
analysis, the upper end of the range of minute ventilation rates
for the resting infant was used to characterize the infant's
inhalation rate for the entire day. This ventilation rate of
2.09 1/m corresponds to an inhalation rate of 0.125 m3/n. Table
V-l-i9a summarizes the inhalation rates assumed in this exposure
assessment.
Table v-l-19b indicates the number of hours each day that
the individual is assumed to be indoors and outdoors during each
age interval. The further assumptions are made that half of the
time indoors is spent resting, and the other half is spent
performing light exercise; and that the time outdoors is spent at
the moderate exercise level of activity. The Table V-l-19a
inhalation rates, the Table V-l-19b scenarios for the numbers of
hours per day spent indoors and outdoors, and the aforementioned
activity assumptions were used to calculate the Table V-1-19C
1-23
-------
volumes of air inhaled indoors and outdoors, and the total volume
of air inhaled each day. To account for the possibility of very
different lifestyles leading to a wide variation in the amount of
air inhaled outdoors, an alternative scenario is presented for
the adult. This scenario, which appears in Tables V-l-19b and
19c as scenario 2 , assumes that the adult spends seven hours a
day outdoors, in contrast to one and one-half hours per day spent
outdoors in the other scenario for the adult age interval.
Lifetime average exposure and risk are estimated by
combining the assumed inhalation volumes and body weights with
indoor and outdoor pollutant concentration data and toxicological
data for each air pollutant.
The individual's daily exposure to a pollutant for each age
interval is calculated by multiplying the volume of inhaled
indoor air by the concentration of the pollutant in indoor air,
doing the same with the outdoor air volume and concentrations,
and adding the two products to arrive at the total quantity of
pollutant inhaled daily in units of micrograms per day (nq/d) .
This quantity is divided by the body weight corresponding to the
subject age interval to calculate an average daily pollutant dose
for that age interval in micrograms per kilogram of body weight
per day (/xg/kg/d) . The general equation for calculating the
average daily dose of an inhaled pollutant for a given age
interval is as follows:
Dose = [(Inhaled Indoor Air x CJ + (Inhaled Outdoor Air x
Weight
The units of dose are in M<3/kg-day.
Doses are calculated for an infant, a child, and an adult, using
the indoor and outdoor concentrations of pollutants in Tables V-
1-20 through 23. Lifetime average daily exposure doses are then
calculated as the age-weighted average of doses during infancy,
childhood, and adult life, as follows:
Lifetime average daily exposure =
[(2/70) x Infant Dose] + [(16/70 x Child Dose] +
[(52/70) x Adult Dose].
A sample calculation is provided in Table V-l-19.
The calculated results for each pollutant (expressed in /jg/kg-
day) are listed in Tables V-l-20 through 23.
1-24
-------
1.5.4 Risk Assessment Results and Discussion
Tables V-l-20 through 23 present the estimated increased
lifetime cancer risks and the HQs calculated using the lifetime
average daily dose estimates. Calculation of cancer risks
employed the cancer potency factor. Calculation of the HQs
entailed conversion of exposure from units of /xg/kg-d to the
units of concentration used for the RfCs, ^g/m3. The device for
this conversion was calculation of a composite air concentration
that would deliver the lifetime average daily exposure assuming
an inhalation rate of 20 m3/d for a 70-kg individual, where 100%
of the inhaled pollutant is transported across interfaces to
reach the organ where it causes/results in adverse effects.27
(See example in Table V-l-20.) The use of 100% as the portion
crossing interfaces to the target organ is a default assumption;
informtion on absorption, partition coefficients, or
pharmacokinetic factors was not pursued.
Comparing the calculated exposures for the two activity
scenarios for individual chemicals points out the relative
importance of indoor and outdoor contamination levels for the
individual's total exposure. Under scenario 2, the individual
inhales a greater total volume of air each day. When indoor
concentrations are considerably higher than outdoor
concentrations, as is the case of chloromethane in Tables V-l-22
and 23, scenario 2 yields a slightly higher lifetime average dose
than is estimated under scenario 1 because of the greater total
inhaled air volume. When the average outdoor concentration of a
pollutant is greater than the average indoor concentration, as is
the case of dichloromethane in Table V-l-20, the difference
between the two scenarios is even greater, reflecting the
combined effect of the greater amount of time the individual
spends outdoors under scenario 2 and the higher outdoor
concentration.
The risk results for the Level 2 risk assessment were
consistent with those for the Level 1 risk assessment for the
VOCs. The estimated lifetime risks of cancer were in the range
of one to 90 per million, with benzene, tetrachloromethane, and
trichloromethane each yielding risks of about 50 per million.
27 An alternate approach to the HQ is a time-weighted average of
the indoor and outdoor air concentrations, where the time
spent indoors and outdoors is an average across age intervals.
In such an approach, the RfC is regarded as protective of the
most sensitive population, so that attention to age intervals
and activity scenarios is unnecessary. This approach may be
legitimate for the RfCs from IRIS, but perhaps not for those
from other sources.
1-25
-------
For noncancer toxicity risks, benzene was the only measured
pollutant with an HQ greater than 1.
1.5.5 Uncertainties and Limitations
In addition to items 1, 4, 5, 6, and 7 in Section 1.4.4, the
uncertainties and limitations section of the Level 1 risk
assessment, there are a number of uncertainties and limitations
associated with the indoor air risk assessments. They include
the following:
l. The analysis assumes that current equipment (i.e., stoves,
cleaning products, insulation, etc.) will remain constant
over the next 70 years. However, with increased knowledge
of potential indoor air pollutants, it is anticipated that
these products might change and the level of indoor air
pollution levels might be reduced.
2. The pollutants monitored do not include all pollutants
present in ambient air or in indoor air.
3. The analyses assume that people are exposed to either the
average concentration at the indoor monitor or the average
concentration at the outdoor monitor for a lifetime.
Movement within the house or beyond the neighborhood of the
house is not considered. An over- or underestimate of risk
may result from this assumption.
4. The analysis assumes that the concentrations found in the
homes are representative of all indoor contaminants.
However, the level of contaminants in other indoor areas
(i.e., schools, malls, automobiles, or work environment)
might vary. The potential impacts on the risks are either
over- or underestimates which cannot be determined.
1.6 ADDITIVE RISK ASSESSMENT
1.6.1 Introduction
Risk estimates based on consideration of exposure to one
chemical at a time might significantly underestimate the risks
associated with simultaneous exposure to several substances.
The additive risk assessment considers simultaneous exposure to
chemicals in the ambient air. The methodology used in this
analysis was based on the EPA guidelines for risk assessment of
1-26
-------
chemical mixtures (U.S. EPA, 1986b); it assumes that dose-
specific information on the toxicity of the chemical mixtures
(i.e., ambient air and indoor air) was not available.
The additive risk assessment relies on the median risks
calculated for each chemical for all the sites in the Level 1
(ambient air) risk assessment, which used concentration data for
the year October 1988 through September 1989. This covers the
potential noncancer toxicity risks from 18 of the study
chemicals; and the cancer risks from 12 chemicals. The Level 1
risk estimates are presented in Tables V-l-9 through 12.
No similar assessment of additive risk has been done using
the indoor air (Level 2) results. Indoor air monitoring was
conducted at four homes to supplement the ambient air results by
providing examples of the difference between indoor and outdoor
air contaminant levels. Whereas the ambient monitoring stations
were selected to characterize air quality throughout the entire
study area, the indoor air sampling was much too limited to draw
general conclusions about indoor exposure throughout the study
area. Thus, it would be inappropriate to use these indoor air
data to characterize additive risk for the study area.
As shown in Table V-l-24, the number of sites for which data
were available varied greatly for the metals and to a lesser
extent for the VOCs. Considering the uneven availability of data
for the sites, use of the medians of the concentrations for all
sites was regarded as an equitable approach to characterizing
risk for the study area as a whole. The use of median
concentrations is based on the premise that no statistically-
significant intersite differences in risk were found in a
monitoring network considered representative of the study area.
The known errors in this premise are as follows:
Statistically significant intersite differences in
tetrachloroethene concentrations resulted in a difference in
HQs that may be significant in an additive risk
assessment.18 The largest intersite difference was between
the high at the Dongan Hills monitor and the low at the
Susan Wagner High School monitor; on the basis of the
unadjusted, reported concentrations, the HQs were 0.2 and
0.05, respectively.
The particulates/formaldehyde network of monitors was much
less extensive than the VOCs network; it is less likely to
represent the study area as a whole. Thus the use of the
median for all sites is less justifiable for the
particulates/formaldehyde risks than for the VOCs risks.
28 Refer to the statistical analysis of the VOCs data presented
in Section 2 of this volume.
1-27
-------
Chromium concentrations are known to be high in soils in
Hudson County, New Jersey, due to past
disposal/dissemination practices. If the soils are
resuspended and transported so that they reach project
monitors in New Jersey but not in Staten Island, then the
ambient air concentrations in New Jersey might be higher
than would be found in Staten Island.
As is the case for the Level 1 risk assessment from which
the medians are derived, the risks in Table V-l-24 are based on
the assumptions that a 70-kg individual inhales 23 m3 of ambient
air daily and continuously for 70 years, and that the pollutant
concentrations in the air are constant at the median annual
average concentrations reported for the monitors. These
estimates might potentially overestimate risks in the event that
the individual moves from the study area to an area with lower
concentrations of these contaminants.
1.6.2 Noncancer Analysis
Analysis of noncancer health effects was carried out using
the Hazard Index (HI) approach outlined in EPA's chemical
mixtures guidelines (U. S. EPA, 1986b). This approach assumes
that simultaneous exposures to several chemicals at
concentrations less than their respective reference
concentrations could result in an adverse health effect. It also
assumes that the risk of an adverse effect is proportional to the
sum of the ratios of the exposure concentrations to their
respective RfCs, i.e., to the sum of the Hazard Quotients. A
screening HI for chronic health effects was calculated as
follows: for each chemical, the median HQ was selected for the
sites in Tables V-l-9 and 11; then the sum of these median HQs
was calculated. This is considered a screening HI since it is
developed without regard for target organ of toxicity29; if this
total were less <1, then separation of the HQs by target organ
would not be pursued. Table V-l-25 lists the median HQs and the
screening HI. Since the screening HI was 4 or 5, depending on
which RfC was used for chromium, the chemicals were grouped by
target organ (system) of toxicity, and their HQs totaled for
those organs (systems).
29 The so-called target organ of toxicity is the organ or system
upon which the RfC is based; other organs or systems might be
adversely affected by exposure to the pollutant, but these
effects occur at exposures higher than those at which the
target organ is adversely affected.
1-28
-------
The resulting groups were for respiratory tract irritation,
liver effects, hematopoietic effects30, kidney effects, and
central nervous system effects. As summarized in Table V-l-25,
this second analysis indicated that the HI for respiratory tract
irritation was l or 2, depending on which RfC was used for
chromium; and the HI for hematopoietic effects was 2.
Review of the data indicates that three chemicals are
primarily responsible for the Hi's exceedence of unity. Nickel
and chromium each contributed as much as 1, depending on which
RfC was used for chromium, to the respiratory tract irritation
HI; while benzene alone accounted for the hematopoietic HI.
1.6.3 Analysis for the Carcinogens
The additivity model was also used to estimate the combined
cancer risk from exposure to VOCs, metals, and BaP. The combined
cancer risk was estimated by simply adding the estimated risks
for each of the pollutants according to the following equation:
Total Cancer Risk = E Risk;, where i is the i*
substance.
This approach to estimating the potential incremental individual
lifetime cancer risk for simultaneous exposure to several
carcinogens is based on EPA's guidelines for carcinogens (U. S.
EPA, 1986a). It is an approximation of an equation for combining
risks accounting for the joint probabilities of the same
individual's developing cancer as a consequence of exposure to
two or more carcinogens. This approach assumes independence of
action by the chemicals involved; or more specifically, it
assumes that no synergistic or antagonistic chemical interactions
occur, and that the carcinogens will produce cancer. If these
assumptions are incorrect, either an over- or underestimate of
risk could result.
Table V-l-24 provides a summary of the cancer analysis using
the median cancer risks for each chemical at the sites appearing
in Tables V-l-10 and 12. The additive cancer risk is calculated
using the alternative assumptions that 10% or 1% of the total
reported chromium concentration is chromium VI. When the median
estimated cancer risks for individual chemicals are added, the
total excess cancer risk per million is 123 or 96, depending on
whether 10% or 1% chromium VI, respectively, is assumed. Benzene
(40), chromium at 10% (30), arsenic (20), and tetrachloromethane
30 This refers to effects on the blood and blood-producing organs
and cells.
1-29
-------
(12) each contributed more than 10 per million to the increased
probability of developing cancer; and trichloromethane, nickel,
chromium at 1%, and cadmium each contributed about 5 per million
to the increased risk of cancer.
1.6.4 Uncertainties in the Additive Risk Assessment
Sources of uncertainty in the additive risk assessment are
those provided for the Level 1 risk assessment, in addition to
those listed below.
1. As noted in Section 1.6.1, the median concentrations used
for the particulates and formaldehyde in general, and
especially for chromium, may not equitably represent
concentrations in the entire study area due to the limited
monitoring network for those chemicals, and uneven
availability of data at the various sites.
For certain of the VOCs, statistically significant intersite
differences in concentration were found; however, with the
possible exception of tetrachloroethene, these VOCs provided
only small contributions to the additive risk assessment.
2. In evaluating the non-carcinogen HI, it is important to note
that the level of concern does not increase linearly as the
RfC is approached or exceeded because the RfCs are not of
equal accuracy or precision, and are not based on the same
severity of effect.
3. The RfCs for the individual chemicals were derived using
different methods (available NOAEL, extrapolation from oral
RfD to inhalation RfC, etc.) that include different
uncertainty adjustments and modifying factors, adding an
additional level of uncertainty associated with the HI.
However, two chemicals contributed most to the HI for
respiratory tract irritation; and only one contributed to
the HI for hematopoietic effects.
4. The IURF is based on a carcinogen potency factor that is an
upper 95th percentile estimate of potency, and, because
upper 9§th percentiles of probability distributions are not
strictly additive, the total additive cancer risk estimate
might become artificially more conservative as risks from a
number of different carcinogens are summed. However, it
should be noted that seven pollutants contributed most to
the total cancer total risk for the sites.
5. While the carcinogens that were analyzed had different
Weights of Evidence for human carcinogenicity, the cancer
1-30
-------
risk equation summed all carcinogens equally, giving as much
weight to Group B or C as to Group A carcinogens.
6. The actions of two different carcinogens might not be
independent; they might exhibit synergistic or antagonistic
effects. This toxicological effect could not be
accommodated in this analysis.
1.7 GENERAL CONCLUSIONS
Quantitative estimates of cancer and noncancer toxicity
risks were made for 21 of the 40 study chemicals—22 volatile
organic compounds (VOCs), 16 metals, benzo[a]pyrene, and
formaldehyde—for which adequate toxicological data and
concentration data were available. Only chronic risks based on
average exposures were considered. Estimated increased lifetime
cancer risks for individual pollutants were in the range of 0.4
to 61 per 1,000,000 for the Level l analysis (ambient air), and 1
to 80 per million for the Level 2 analysis (ambient and indoor
air, VOCs only). The Hazard Quotients for non-carcinogenic
effects were below one for all pollutants except benzene,
chromium, and nickel. The estimated risks for chromium and
nickel are believed to be conservative, i.e., err in the
direction of overestimating risk. However, since the specific
chemical species of chromium and nickel in the ambient air
samples were not measured, assumptions were made in selecting and
applying toxicological criteria.
The additive risk assessment for noncancer toxicity by
target organ, and for cancer for all pollutants combined, yielded
a maximum Hazard Index of 2 (hematopoietic effects and
respiratory tract irritation), and a cumulative cancer risk of 96
or 123 per million, depending on the reference concentrations and
chromium VI assumptions used in the estimates.
The estimated cancer and noncancer toxicity risks associated
with benzene were consistently higher than those estimated for
the other pollutants addressed in the Level 1 and Level 2
analyses. The next highest estimated risks in the Level 1
analysis were associated with nickel, chromium, and formaldehyde;
while the next highest estimated risks in the Level 2 analysis
(VOCs only) were from trichloromethane and tetrachloroethene.
Using the 1988 and 1989 Urban Air Toxics Monitoring Program
(UATMP) studies as the basis for comparison, the risk estimates
for nickel and chromium were higher for the study area than for
other urban areas nationwide. An assessment of the significance
of the magnitudes of the differences is not offered; some of the
differences might be attributable to differences in sampling and
1-31
-------
chemical analysis for the studies. This demonstrates the
importance of considering differences in data quality, especially
when comparing the results from different organizations.
The risk assessments provided here based on the SI/NJ UATAP
data do not cover the risk for all pollutants in the ambient air
and indoor air to which residents of the study area are exposed.
Additional uncertainties associated with the Level 1, Level 2,
and additive risk assessments have been listed with the
assessments. Beyond the uncertainties associated with the most
common risk assessment assumptions are those specific to (1) the
reference concentration for chromium, (2) the concentration data
for chromium and nickel, (3) recognition of a hazard quotient as
being of sufficient magnitude to warrant further refinement of a
risk assessment for noncancer toxicity (i.e., assessing a Hazard
Quotient or Hazard Index as significant from a risk assessment
perspective), and (4) the diversity of scientific interpretation
that underlies toxicological dose-response information from
various governmental agencies. Since the concentration data for
chromium and for nickel, and the reference concentration for
chromium have been critical to the magnitude of the cancer and
noncancer toxicity risk estimates for this study, further
refinement of these inputs to the risk assessment are
recommended.
1.8 ACKNOWLEDGEMENT
This section was prepared by Ms. Marian Olsen and Ms. Carol
Bellizzi of the Region II office of U.S. Environmental Protection
Agency; Dr. Paul Lioy of the University of Medicine and Dentistry
of New Jersey (UMDNJ), Robert Wood Johnson School of Medicine;
Dr. John Hawley of the New York State Department of Health; Ms.
Joann Held and Ms. Olga Boyko of the New Jersey Department of
Environmental Protection and Energy; and Mr. Robert Majewski and
Mr. Tom Gentile of the New York State Department of Environmental
Conservation. The final draft was reviewed within EPA and
modified to meet requirements for a risk assessment issued as an
EPA report.
1.9 REFERENCES
Dollarhide J.S. (1992) RfC for total chromium, memorandum of
August 26 to M. Olsen. Cincinnati, OH: U.S. Environmental
Protection Agency, Office of Research and Development,
Environmental Criteria and Assessment Office.
1-32
-------
ICRP. (1981) International Commission on Radiological
Protection report of the Task Group on Reference Man. New York:
Pergammon Press.
Ktsa V., Lioy P.J., Chow J.C., Watson J.G., Shupack S., Howoell
T., Sanders P. (1992) Particle size distribution of chromium:
total and hexavalent. Aerosol Research and Technology: in press.
Lahre T. (1990) Calculation of cancer risks from 1988 UATMP data,
memorandum of October 3. Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Air Quality Planning
and Standards.
Lioy P.J., Daisey, J.M. (1987) Toxic air pollution, a
comprehensive study of non-criteria air pollutants. Ann Arbor,
MI: Lewis Publishers, Inc.
Lioy P.J., Freeman N.C.G., Wainman T., Stern A.H., Boesch R.,
Howell T., Shupack S.I. (1992a) Microenvironmental analysis of
residential exposure to chromium laden wastes in and around New
Jersey homes. Risk Analysis: in press.
Lioy P.J. (1992b) Personal communication in August to C. Bellizzi
concerning proportion of total reported chromium concentration
that is hexavalent. Piscataway, NJ.
National Academy of Sciences. (1981) Indoor pollutants.
Washington, DC: National Academy Press.
Poirier K.A. (1992) Memorandum of September 3 to M. Olsen:
reference concentration (RfC) listings for xylene (mixed and o-,
m-, and p-isomers) in HEAST. Cincinnati, OH: U.S. Environmental
Protection Agency, Office of Research and Development,
Environmental Criteria and Assessment Office.
National Institute of Occupational Safety and Health. (1989)
Indoor air quality. Cincinnati, OH: Centers for Disease Control,
Occupational Safety and Health Administration.
National Research Council. (1983) Risk assessment in the
federal government: managing the process. Washington, DC:
National Academy Press.
Shah J.J., Heyerdahl E.K. (1988) National ambient volatile
organic compounds (VOC) data base update. Research Triangle
Park, NC: U.S. Environmental Protection Agency, Office of
Research and Development, Atmospheric Sciences Research
Laboratory, Atmospheric Chemistry and Physics Division; EPA
report no. EPA/600/3-38/010a. NTIS no. PB88-195631.
1-33
-------
U.S. Environmental Protection Agency. (1978) National primary and
secondary ambient air quality standards for lead. 43 FR 46258,
October 5, 1978.
U.S. Environmental Protection Agency. (1986a) Guidelines
for carcinogen risk assessment. 51 FR 33992, September 24, 1986.
U.S. Environmental Protection Agency (1986b). Guidelines
for the health risk assessment of chemical mixtures. 51 FR
34014, September 24, 1986.
U.S. Environmental Protection Agency. (1986c) Guidelines for
estimating exposures. 51 FR 34042-34054, September 24, 1986.
U.S. Environmental Protection Agency. (1987a) The total exposure
assessment methodology (TEAM) study: summary and analysis: volume
I. Research Triangle Park, NC: Office of Acid Deposition,
Environmental Monitoring and Quality Assurance.; EPA report no.
EPA/600/6-87/002a.
U.S. Environmental Protection Agency. (1987b) The total exposure
assessment methodology (TEAM) study: Elizabeth and Bayonne, NJ,
Devils Lake, ND and Greensboro, NC: volume II, parts 1 and 2.
Research Triangle Park, NC: Office of Acid Deposition,
Environmental Monitoring and Quality Assurance. Available from
NTIS, Springfield, VA: PB88-100078.
U.S. Environmental Protection Agency. (1989a) Exposure
factors handbook. Washington, DC: Office of Health and
Environmental Assessment, Office of Research and Development; EPA
publication no. EPA/600/8-89/043.
U. S. Environmental Protection Agency. (1989b) Nonmethane
organic compound monitoring program, final report, volume II:
urban air toxics monitoring program. Research Triangle Park, NC:
Office of Air Quality Planning and Standards; EPA publication no.
EPA-450/4-89-005.
U. S. Environmental Protection Agency. (1989c) Urban air
toxics monitoring program, draft final report. Research Triangle
Park, NC: Office of Air Quality Planning and Standards; EPA no.
68D80014.
U. S. Environmental Protection Agency. (1989d) Analysis of air
toxics emissions, exposures, cancer risks and controllability in
five urban areas, volume I, base year analysis and results.
Research Triangle Park, NC: Office of Air Quality Planning and
Standards; EPA report no. EPA-450/2-89-012a.
1-34
-------
U. S. Environmental Protection Agency. (1990a) Integrated
Risk Information System (IRIS). Washington, DC: Office of Health
and Environmental Assessment, Office of Research and Development.
Available from: TOXNET, a data base accessible through the
National Library of Medicine.
U. S. Environmental Protection Agency. (1990b) Interim
methods for development of inhalation reference concentrations.
Washington, DC: Office of Research and Development; EPA
publication no. EPA/600-8-90/066A.
U.S. Environmental Protection Agency. (1990c) Cancer risk from
outdoor exposure to air toxics, volume 1. Research Triangle
Park, NC: Office of Air Quality Planning and Standards; EPA
report no. EPA-450/l-90-004a.
U.S. Environmental Protection Agency. (1991) Health effects
assessment summary table. Washington, DC: Office of Solid Waste
and Emergency Response and Office of Research and Development;
EPA report no. OERR 9200.6-303 (91-1). NTIS No. PB91-921199
U. S. Environmental Protection Agency. (1992a) Guidelines for
exposure assessment. 57 FR No. 104, 22888-22937, May 29, 1992.
U.S. Environmental Protection Agency. (1992b) Health effects
assessment summary tables, annual update. Cincinnati, OH: Office
of Emergency and Remedial Response; and Environmental Criteria
and Assessment Office, Office of Health and Environmental
Assessment, Office of Research and Development; EPA report no.
OHEA ECAO-CIN-821.
U.S. Environmental Protection Agency. (1992c) Memorandum to M.
Olsen concerning carcinogenic characterizations for
perchloroethylene and trichloroethylene. Cincinnati, OH: Office
of Research and Development, Environmental Assessment and
Criteria Office.
Wallace, L.A. (1987) Project summary, the total exposure
assessment methodology (TEAM) study. Washington, DC: U.S>
Environmental Protection Agency, Office of Acid Deposition,
Environmental Monitoring and Quality Assurance; EPA report no.
EPA/600/S6-87/002.
1-35
-------
Table V-1-1a; Sampling Sites - Staten Island
Site
Code
Community
Site Name
1 Westerleigh
Susan Wagner H.S.
2 Travis
P.S. 26
Sampling Type Frequency
NYSDEC Sorbent every sixth day
Tenax "
Canister "
Hi Volume Filter "
Formaldehyde "
Meteorology continuous
NYSDEC Sorbent every sixth day
Tenax "
Canister "
Hi Volume Filter "
3 Annadale. Eltingville Tenax
Fire Station 167 Canister
every day'
every eighteenth day2
4 Great Kills
Fire Station 162
5 Port Richmond
Post Office
6 Oongan Hills
Fire Station 159
ing Statio
Landfill,
Mon
Near Landfill, near
Staten Island Mall
NYSDEC Sorbent every sixth day
Canister every eighteenth day
NYSDEC Sorbent every sixth day
Tenax "
Canister "
Formaldehyde "
Tenax every day'
Canister every eighteenth day2
NYSDEC Sorbent every sixth day
Canister every eighteenth day
Meteorology continuous
8 Clifton
Tenax
Bayley Seton Hospital Canister
every day1
every eighteenth day2
9 Tottenville
Fire Station 151
NYSDEC Sorbent every sixth day
Tenax "
Canister every eighteenth day
Meteorology continuous
10 Arthur Kill. Rossville Hi Volume Filter every sixth day
New York Telephone
1 Through 3/89. Every sixth day to 9/89.
1 Rotated among sites 3, 6, and 8 on a monthly basis.
* Data did not meet OA requirements.
" Some data did not meet QA requirements.
' Samples taken only occasionally, not enough data to report.
Operating dates
10/87- 9/89
10/87-12/87
4/88- 9/89
10/87- 9/89
7/88- 9/89
4/88- 9/89
10/87- 9/89
5/88-10/88*
8/88- 9/89
10/87- 9/89
10/87- 9/89
12/87- 9/89
9/88- 9/89
9/88- 9/89
10/87- 9/89
6/88- 1/89"
6/88- 9/89
7/88- 9/89
10/87- 9/89
1/88- 9/89
10/88- 9/89
10/88- 9/89
10/88- 9/89
10/87- 9/89
2/88- 9/89
10/87- 9/89
6/88- 1/89"
7/88- 9/89
4/88- 9/89
3/88-9/89"
1-36
-------
Table V-l-lb; Sampling Sites - New Jersey
Sampling Type Frequency
Site
Code
Community
Site Name
A Elizabeth
Mattano Park
B Carteret
Police Station
C Sewaren
Tenax
Canister
High Volume Filter
Formaldehyde
Tenax
Canister
High Volume Filter
Formaldehyde
Tenax
every sixth day
every sixth day
Glen Cove School Canister
D Piscataway
Pvt. Residence
E Highland Park
Fire Station
—Newark Airport
—Elizabeth
NJDEP Trailer
Tenax
Canister
Formaldehyde
every sixth day
every eighteenth day1
every sixth day
every eighteenth day1
every sixth day
High Volume Filter every sixth day
Operating
dates
5/88- 9/89
5/88- 9/89
5/88- 9/89
5/88- 9/89
10/87- 9/89
10/87- 9/89
10/87- 9/89
11/87- 9/89
11/88- 9/89
12/87- 9/89
11/88- 9/89
11/87- 9/89
11/88- 4/892
5/89- 9/893
11/88- 9/89
Meteorological Data
Meteorological Data
hourly
continuous
routine NWSb data
1 Rotated between sites C and D on a monthly basis.
2 Analysis by EPA contractor lab.
3 Analysis by NJIT.
• National Weather Service
b Data were not used due to equipment problems.
1-37
-------
Table V-l-lc: Abbreviations used for site names
SW Susan Wagner High School
PS-26 PS-26 is a public school in Travis
ELTVL Eltingville, Annadale
GT KLLS Great Kills
PRT RCH Port Richmond
DONGAN Dongan Hills
PUMP Pumping Station near Staten Island Mall, near
Freshkills Landfill
B-STN Bayley Seton Hospital in Clifton
TOTT Tottenville
ELIZ Elizabeth
CART Cartaret
SEW Sewaren
PSCAT Piscataway
HIPRK Highland Park
1-38
-------
Table V-l-2a; VOCs Analyzed During Project
Chloromethane1
Dichloromethane (Methylene Chloride)
Trichloromethane (Chloroform)
Trichloroethane, 1,1,1-
Trichloroethane, 1,1,2-
Tetrachloromethane (Carbon Tetrachloride)
Trichloroethylene
Tetrachloroethylene (Tetracholoroethene, perchloroethylene)
1,1-
1,2- (Ethylene Dichloride)
Dichloroethane,
Dichloroethane,
Tribromomethane (Bromoform)
Benzene
Toluene
Hexane
Xylene, o-
Xylene, m- and
Ethlybenzene
Chlorobenzene
Styrene
Dichlorobenzene, o- (1,2 - dichlorobenzene)
Dichlorobenzene, m- (1,3 - dichlorobenzene)
Dichlorobenzene, p/- (1,4 - dichlorobenzene)
~ 2
1 No valid data obtained from ambient air monitoring.
2 Not separated by analytical method.
1-39
-------
Table V-l-2b; Particulate Species Analyzed During Project
Arsenic
Barium
Benzo[a]pyrene
Beryllium'
Cadmium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Mercury
Molybdenum
Nickel
Selenium2
Vanadium
Zinc
1 Never detected.
2 No valid data were obtained.
1-40
-------
Map V-l-1
(Vigure XI -1)
SI/NJ UATAP
Monitoring Locations
SITES:
1 •
2. Travis
3. Aun.nlale
4. Cnrat KKI9
5. Pert Richmond
6. Dongon Hills
7L Pumping Station"
a. cimon
9. ToUenvflle
A. Elizabeth
B. Cartorat
C. S^woron
0. Piscatavay
E. Highland Pork
V.F
KEY:
V - VoMIU Orgonk Compound*
P - PoHlctHole»-Troc« Metals A BaP
W - Meteorology
F •• rormaldehyd*
-------
TABLE V-1-3; AMBIENT AIR CONCENTRATIONS FOR THE VOLATILE ORGANIC CQMPOUMDS-
AHNUAL AVERAGES FOR OCTOBER 1988 THROUGH SEPTEMBER 1989*
Concentrations, ppb
NEU JERSEY
CHEMICAL
DICHLOROMETHANE
TRICHLOROMETHANE
TETRACHLORONETHANE
TRICHLOROETHENE
TETRACHLOROETHENE
HEXAME, n-
BENZENE
TOLUENE
XYLEME, o-
XTLENES, p- and •-
ETHVL8ENZENE
CART
--
0.02
0.12
0.05
0.17
1.08
1.48
3.80
0.40
1.16
.
EJUZ
--
0.02
0.13
0.04
0.21
0.91
1.45
3.62
0.40
1.06
.
SEW
--
0.02
0.15
0.04
0.21
O.B4
1.16
2.88
0.31
0.85
.
P5CAT K1PRK
..
0.02
0.11
0.05
0.13
0.50
0.97
2.11
0.24
0.52
.
sw
0.47
0.07
0.09
0.10
0.18
-
0.77
2.42
0.30
0.92
-
fS-26
0.93
0.10
0.10
0.12
0.19
-
1,27
3.88
0.44
1.42
.
PRT RCH
0.85
0.08
0.09
0.13
0.24
-
1.34
4.25
0.55
1.76
.
PUMP
0.76
0.15
0.10
0.27
1.09
-
1.16
3.87
0.46
1.50
-
GT ELLS
0.50
0.06
0.10
0.07
0.20
-
0.92
2.89
0.37
1.19
-
NEU YORK
TOTT
0.60
0.07
0.16
0.08
0.20
-
O.B6
2.81
0.32
1.04
-
SITES
B-STN
--
0.03
0.11
0.08
0.27
0.77
1.40
3.19
0.37
1.83
0.50
EITVL
--
0.04
O.U
0.06
0.21
0.89
1.50
3.45
0.38
1.89
0.53
DOHCAK
--
0.03
0.12
0.07
0.68
0.86
1.96
4.10
0.42
2.48
0.66
* Annual avg, * , where t\ = number of sanples in the ith quarter,
In, . x, * avg. cone, in the ith quarter. Includes annual averages based on > 38 sanples.
- Samples not collected at this site.
-- Submitted data not good; inappropriate collection method.
Note: In the absence of acccuracy and precision characterization of the reported annual averages, differences
should not be assumed to be statistically significant.
1-42
-------
TABLE V-1-4; AHBIEHT AIR CONCENTRATIONS* FOR METALS. BEN20[a]PYRENE, AND FORMALDEHYDE-
ANNUAL AVERAGES FOR OCTOBER 1988 THROUGH SEPTEMBER 1989
Chemical
Concentrations, ng/m3
u>
ARSENIC
CADMIUM
CHROMIUM, total
LEAD2
MANGANESE
MERCURY
NICKEL
VANDIUM
ZINC
BaP
FORMALDEHYDE3
--- Indicates that
1 m
CART EL1Z
4.2 1.6
27 16
43.1 38.6
21.6 14.8
0.5 0.5
28.2 23.6
116.1 113.9
0.20 0.19
no data are available.
=",*.
H1PRK
2.1
12
91.1
13.3
0.5
22.4
97.8
0.14
camnl oc in fh»
SU PS -26 PRT RICH
.H C.J
!!).£ 10. D
IV. 1 <:U.c
113.2 96.5
0.15 0.21
2524 ---- 2137
(2.02 ppb) (1.71 ppb)
ith nuarttfr
Sn, KI = avg. cone, in the ith quarter. Includes annual averages based on >_ 38 samples.
2 Highest quarter concentrations, not annual average concentrations.
3 Formaldehyde concentrations Mere converted from ppb to ng/m3 with the following formula (MU, molecular weight, of formaldehyde is 30.03):
ng/m3 = (ppb)(MU)(1000)/24.04 I, where 24.04 I is volume of ideal gas at 20°C.
Note: Apparent site-to-site differences in concentration may not be statistically significant.
-------
Table V-l-5: Key to Figures V-l-1 through 10, bar charts
comparing annual average concentrations for
the SI/NJ UATAP sites to those for the UATMP sites
Min. Ann. Avg., Ned. Ann. Avg., Hax. Ann. Avg. are, respectively, the minimum,
median, and maximum annual averages for the specified group of site(a)a—SI
being the SI/NJ UATAP sites, and U being the UATMP sites. Ann. Avg. is the
annual average concentration for the specified site.
1988
SI SI/NJ UATAP sites (Oct. 1987 - Sept. 1988)
PI Piscataway (from SI/NJ UATAP)
U 1988 UATMP sites (Oct. 1987 - Sept. 1988)
A Atlanta, GA
B Baton Rouge, LA
C Birmingham, AL
D Burlington, VT
E Chicago, IL (Carver H.S. and Washington, H.S.)
F Cleveland, OH
G Dallas, TX
H Dearborn, MI
I Detroit, MI
J Hammond, IN
K Houston, TX
L Jacksonville, FL
M Lansing, KY
N Louisville, KY
0 Miami, FL
P Midland, MI
Q Port Huron, MI
R Portland, OR
S Sauget, IL
1989
SI SI/NJ UATAP sites
PI Piscataway (from SI/NJ UATAP)
U 1989 UATMP sites (Jan. 1989 - Jan. 1990)
a Baton Rouge, LA
b Camden, NJ
c Chicago, IL (Carver H.S. and Washington, H.S.)
d Dallas, TX
e Ft. Lauderdale, FL
f Houston, TX
g Miami, FL
h Pensacola, FL
i Sauget, IL
j St. Louis, MO
k Washington, DC #1
1 Washington, DC #2
m Wichita, KS #1
n Wichita, KS #2
1-44
-------
FIGURE V-l-1
Dichloromethane -
SI/NJ UATAP Compared to UATMP
I 98Q
O V
I
'
1
ni
V
| 1 Mln. Ann. Avg .
CITtE&
Mod. Ann. Avg. [ D Max. Ann. Avg. L\1 Ann. Avg.
b. 1989
c_>
O 2
r I Mln. Ann. Avg.
•». -o o «> «, V^>ck>^>»»>*->>-^5>-C^
CITIES
Hi Mod. Ann. Avg. I! I Max. Ann. Avg. KN Ann. Avg.
1-45
-------
FIGURE V-l-2
Trichloromethane •
SI/NJ UATAP Compared to UATMP
I OSS
Ann. Avg. k\l Ann
Mod. Ann. Avg. I1 II Ma
I I FV/lln. Ann. Avg.
Mad. Ann. Avg. d] Max. Ann. Avg. K] Ann. Avg
U Mln. Ann. Avg
1-46
-------
FIGURE V-l-3
Tetrachloromethane -
SI/NJ UATAP Compared to UATMP
1
.18 —
.ia -
.14 -
-12 -
O.I -
-08 -
.OO -
.04 -
.02 -
T
•
•
*
X
'
-
•1
1 m
K
-^d^i^msi^^sEL-^i^^a^ L^-^-,
. o c. Ann. AVQ. C^Ann. Av«
b. 1989
0.2
g> O.I S
o
8 °'1
o.os
•
CITIES
CUlWIIn. Ann. Avg. • Mad. Ann. Avg- DH Max. Ann. Avg. CXI Ann. Avg.
1-47
-------
FIGURE V-l-4
Trichloroethylene -
SI/NJ UATAP Compared to UATMP
1988
Mod. Ann. Avg. I ! I Max. Ann. Avg. KN Ann. Avg
I I Min. Ann. Avg
b. 1989
1 .0
0.8
go.e
O O.4
O
0.2 -
0.0
S1
1
H Mln. Ann. Avg.
CITIES
Mad. Ann. Avg. E 3 Max. Ann. Avg. C3Ann. Avg.
1-A8
-------
FIGURE V-l-5
Tetrachloroethylene -
SI/NJ UATAP Compared to UATMP
1988
S -I
A -
3 -
1 -
Or/lin. Ann. Avg.
CITIES
Mad. Ann. Avg. GUI Max. Ann. Avg. C^Ann. Avg.
1989
.r,
5
1.2 -
1 -
o.a
o.e -
CITIES
. Ann. Avg. • M«d. Ann. Avg. O Max. Ann. Avg. [SJAnn. Avg.
1-49
-------
FIGURE V-l-6
Benzene -
SI/NJ UATAP Compared to UATMP
a. 1 988
CH Min. Ann. Avg. •is/led. Ann. Avg- B Mmx. Ann. Avg. Q Ann. Avg
b. 1
s.o -<
4.0
O 2.O
u
1.O -
0.0
Mln. Ann. Avfl.
CITIES
Mad. Ann. Avg. Q Max. Ann. Avg. QAnn. Avg.
1-50
-------
FIGURE V-l-7
Toluene -
SI/NJ UATAP Compared to UATMP
1 988
IB -i
18
11
12
1O —
IV/lln. Aon. Avg.
fr- >• ^^-v^-^-^O^ O
-------
FIGURE V-l-8
Xylene (m- and p-) -
SI/NJ UATAP Compared to UATMP
a. 1 988
10.0
8.0 -
£ 6.0-
O 4.O
u
2.0-
O.O
»—
,
,
•
.
I
r
.
•*
*
,
'
1
f-
'
'"
.
f
1
'
f
.'
-
-
I N.
1
«
"
\
k
.-V
••
to
-
•
s
\
\
\
s
\.,^
CITIES
. Ann. Avg. • Mod. Ann. Avg. O Max. Ann. Avg. S3 Ann. Avg.
8.O -
7.0-*
6.0
ai 5.0 -
a.
a.
§3.0-
2.0-
1.0
0.0
CITIES
UMin. Ann. Avg. • Mod. Ann. Avg. E3 Max. Ann. Avg. El Ann. Avg.
1-52
-------
FIGURE V-l-9
Xylene (o) -
SI/NJ UATAP Compared to UATMP
1 988
2 —
1 -
rci
CITIES
J Min. Ann. Avg- Hr^ad. Ann. Avg-
t>.
IV/l
• x. Ann. AVQ. KM Ann. Avg.
Ann. Avg- M»d. Ann. Avg. Max. Ann. Avg. El Ann. Avg
1-53
-------
FIGURE V-l-10
Ethylbenzene -
SI/NJ UATAP Compared to UATMP
•
-»
1
o
V •«» O -O
I I Min. Ann. Avg.
CITIES
Mod. Ann. Avg. I 3 A/lax. Ann. Avg.
Ann. Avg.
b
•
LJ Min. Ann. Avg. Hi Mad. Ann. Avg. I J Max. Ann. Avg. S3 Ann.
1-54
-------
10
8
FIGURE V-l-11
Comparison of SI/NJ UATAP Data (10/88-10/89)
with 1988 UATMP Data (10/87-10/88)
ARSENIC
E
o
c
I
o
4
13
I
I I
I I
I \^ \^ I
I
w^g>vv
Location
Medoi for Susan Wagner H.S.
and PS-26
1-55
-------
FIGURE V-l-12
14
o 10
8
Comparison of SI/NJ UATAP Data (10/88-10/89)
with 1988 UATMP Data (10/87-10/88)
CADMIUM
o
o
^>^>V^>>"
Location
1-56
-------
, 30
25
FIGURE V-l-13
Comparison of SI/NJ UATAP Data (10/88-10/89)
with 1988 UATMP Data (10/87-10/88)
CHROMIUM
E
o
I 15
o
o
0>
co
* 5
<
0
X '//.
; /// ///
\ l^ I
1 1^ 1^ T
^
p>S>Vi
>^ rt>^ A^
^o<°° sX
Location
* Hic^teid Park NJ (SI/NJ UATAP background site)
** Ktocfan for Cartoret and Efzatwth
1-57
-------
FIGURE V-l-14
Comparison of SI/NJ UATAP Data (10/88-10/89)
with 1988 UATMP Data (10/87-10/88)
COBALT
§2.5
,0
7a>
2
g
I
I 1.5
c
o
o
-------
FIGURE V-l-15
Comparison of SI/NJ UATAP Data (10/88-10/89)
With 1988 UATMP Data (10/87-10/88)
COPPER
* Highland Park NJ (SI/NJ UATAP background she)
** Median far Carteret, Bizabefi, PS-26, and
Susan Wagner H.S.
1-59
-------
FIGURE V-l-16
E 10,000
I
8,000
if
.g
"Jo
£ 6,000
8
2 4,000
V^W"
Location
* H&isrrt Park NJ (SI/NJ UATAP backyound site)
*• Mednn for Cateret, Biz^eti. PS-26. and
Susai Warier H.S.
1-60
-------
FIGURE V-l-17
Comparison of SI/NJ UATAP Data (10/88-10/89)
with 1988 UATMP Data (10/87-10/88)
LEAD - Highest Quarterly Average
•^ soo
CD OUU
c
o
•t; Af\n
(\J *HJU
t:
8
£ or\n
O o*-MJ
o
ly average
5 100
S3
Q-
n\ /"I
o>
-
I
,»
'
X" x
/V
^
I
'//,
'//,
I
^/
I
~t>^
P71
l
. u
//^
^\.
I
rt
.
I
v\.-
' '
//
I
-------
FIGURE V-l-18
Comparison of SI/NJ UATAP Data (10/88-10/89)
with 1988 UATMP Data (10/87-10/88)
MANGANESE
600
.0
psoo
c
o
la 400
O)
g 300
o
o
|200
Q)
C
<
100
0
VTA I7771 I/%H ^
V/A
[7771 [7771
[7771 [7771 [7771
Location
* H&isnd Pak NJ (SI/NJ UATAP background site)
** Median for Garten*. Elzabeti, PS-26, and
Susan Wagner H.S.
1-62
-------
10
^ 8
c
o
2 fi
c 6
CD
o
CO
2
a
« 2
(0
c
0
FIGURE V-l-19
Comparison of SI/NJ UATAP Data (10/8&-10/89)
with 1988 UATMP Data (10/87-10/88)
MOLYBDENUM
Location
* Median for Susan Wagner H.S. and PS-26
1-63
-------
40
FIGURE V-l-20
Comparison of SI/NJ UATAP Data (10/88-10/89)
with 1988 UATMP Data (10/87-10/88)
NICKEL
3
O
c 30
o
I
"E
o
>
o
o
0
D)
CO
g> 10
co
To
n
I o
z / / /
\ I
\ \
I I I I
Locatbn
* Hi^il£nd Park NJ (SI/NJ UATAP back^ound site)
** Mecfan for Carter*, Bizabeth, PS-26,and
Susan Wagner H.S.
I I
1-64
-------
20
1
c" 15
o
2
c
2 10
FIGURE V-l-21
Comparison of SI/NJ UATAP Data (10/88-10/89)
with 1988 UATMP Data (10/87-10/88)
VANADIUM
0
'
I n^
,*^>>>*>^^v>^<^
Location
Median for Sus»i Wagner H.S. and PS-26
1-65
-------
1,
FIGURE V-l-22
Comparison of SI/NJ UATAP Data (1Q/88-10/89)
with 1988 UATMP Data (10/87-10/88)
ZINC
=J
t
cf
g
I
"E
8
c:
O
O
0>
CO
CO
1,
VTA
VTA
T^ \
i r i^ \
T^ I
4&***!S*!!!^^
^\R* V^^^ ^y^ v/^ i'^' y*
90*
Location
* HigNcnd Park NJ (SI/NJ UATAP background site)
** Medfem for Carteret, Bizabeti, P&-26. end
Susan Wacpier H.S.
1-66
-------
FIGURE V-l-23
Comparison of SI/NJ UATAP Data (10/88-10/89)
with1988 UATMP Data (10/87-10/88)
BENZO(A)PYRENE
o
^
c
o
^ 4
c
o
o
CO
13
C
< 0
Z&
\ / / /
r/ />
/ ,
'/
>££VV'
Location
* Highland Park NJ (SI/NJ UATAP background site)
** Median for Carteret, Stzabeti, PS-26, and
Susan Wagner H.S
1-67
-------
FIGURE V-l-24
Comparison of SI/NJ UATAP Data (10/88-10/89)
with 1989 UATMP Data for
FORMALDEHYDE
n
CL
CL
g a
nj
•*—
8
02
o
^
^<^>*v^<>^>^<
fV KkO A\V ^r-A ^f-5- -c^V .c^
^* *tf* ^^ ^^ v*.^ **5>*
sv
^X" s*>^rf^
^V^>°"^ ^
!*»*
Location
• Median far Susan Wagnef HS and Port Richmond
1-68
-------
TABLE V-1-6; NONCANCER INHALATION REFERENCE CONCENTRATIONS (RFC'S)
CHEMICALS
CHLOROMETHANE
DICHLOROHETHANE
TRICHLOROMETHANE
TETRACHLOROHETHANE
TRICHLOROETHENE
TETRACHLOROETHENE
HEXANE, n-
BENZENE
TOLUENE
XYLENE
ETHYLBENZENE
FORMALDEHYDE
LEAD
CHROMIUM
NICKEL, refinery dust
BENZO(a)PYRENE
ARSENIC
CADMIUM
MERCURY
MANGANESE
VANADIUM
ZINC
CAS NUMBER
74-87-3
75-09-2
67-66-3
56-23-5
79-01-6
127-18-4
110-54-3
71-43-2
108-88-3
1330-20-7
100-41-4
50-00-0
7439-92-1
7440-47-3
00-02-0
50-32-8
7440-38-2
7440-43-9
7439-97-6
7439-96-5
7440-66-6
ug/ffl3
per ppb
at T=20 °C *
2.10
3.53
4.96
6.39
5.46
6.89
3.58
3.26
3.83
4.41
4.35
1.33
-
-
-
-
-
-
-
INHALATION
RFC
(mg/ro3)
0.831
3.02
-
-
-
-
0.201
-
0.4002
_3
1.02
0.0301 **
0.00154
0.00011S, 0.00000239
0.000020'
-
-
0.00002'
0.000302
0.000402
0.00025'
0.05"
ADJUSTED
RFC
(for 23 nS/d)
342 ppb
739 ppb
6.05' ppb
0.329'ppb
4.21' ppb
4.35' ppb
48.6 ppb
0.644'ppb
90.8 ppb
-
200 ppb
19.6 ppb, 26000 ng/m3
1300 ng/m3
87.0' 8. 1.738 ng/m3
17.4 ng/m3
,
17.4 ng/m
261 ng/m3
3
348 ng/m3
217 ng/m3
1
43500 ng/m3
See next page for footnotes.
1-69
-------
TABLE V-1-6
FOOTNOTES
* For conversion from units of RfC, weight/volume, to units of reported concentrations, volume/volume) and, thus, the adjusted RfC. Based on
density of ideal gas at 20 °C. Sample conversion of inhalation reference concentration to ppb for benzene:
2.1E-3 mg/m3 x (20/23) x (1 ppb/(3.26 ug/m3]) x (1000 ug/mg) = 5.6E-1 ppb
**This is a short-term guideline concentration for 1- to 4-hour exposures; a chronic RfC was not available.
- Information not available or not provided.
1 MTSOOH
January 1992 IRIS.
RfC appeared in HEAST of 1991. but not in 1992 update of HEAST (Poirier, 1992).
NAAQS for lead, not an RfC (U.S. EPA, 1978).
For total chromium.
Used for hexavalent
-------
TABLE V-1-7; CANCER INHALATION UNIT RISK FACTORS
CHEMICALS
CHLOROMETHANE
D I CHLOROMETHANE
TRICHLOROHETHANE
TETRACHLOROMETHANE
TRICHLOROETHENE
TETRACHLOROETHENE
NEXANE, n-
BENZENE
TOLUENE
XYLENE
ETHYLBENZENE
FORMALDEHYDE
LEAD
CHROMIUM, hexavalent
NICKEL, refinery dust
BENZOtalPYRENE
ARSENIC
CADMIUM
MERCURY
MANGANESE
VANADIUM
ZINC
CAS NUMBER
74-87-3
75-09-2
67-66-3
56-23-5
79-01-6
127-18-4
110-54-3
71-43-2
108-88-3
1330-20-7
100-41-4
50-00-0
7439-92-1
7440-47-3
00-02-0
50-32-8
7440-38-2
7440-43-9
7439-97-6
7439-96-5
HEIGHT OF
EVIDENCE
C
B2
B2
B2
B23
C3
.
A
D
D
D
B1
B2
A
A
B2
A
B1
D
0
-
-
ug/m3
per ppb*
2.10
3.53
4.96
6.39
5.46
6.89
3.58
3.26
3.83
4.41
4.35
1.33
.
.
;
.
.
-
-
-
-
CANCER
SLOPE FACTOR
(mg/kg/d)1
6.3 E-03*
7.5 E-03
8.1 E-02
1.3 E-01
6.0 E-033
2.0 E-033
-
2.9 E-02
-
-
-
4.5 E-02
-
4.1 E+01
8.4 E-01
7.3 E+00
-
-
-
-
-
INHALATION UNIT
RISK FACTOR**1
(per ug/m3)
1.8E-04*
4.7E-07
2.3E-05
1.5E-05
1.7E-06*
5.8E-072
-
8.3E-06
-
-
-
1.3E-05
-
1.2E-02
2.4E-04
1.7E-032
4.3E-03
1.8E-03
-
*
"
Inhalation unit risk
(actor at 23 m3/day
4.3E-06 per ppb
1.9E-06 per ppb
1.3E-04 per ppb
1.1E-04 per ppb
1.1E-05 per ppb
4.6E-06 per ppb
-
3. IE-OS per ppb
-
-
—
2.0E-05 per ppb,
1.5E-08 per ng/m3
-
1.4E-05 per ng/m3
2.8E-07 per ng/ni3
a.OE-06 per ng/m3
4.9E-06 per ng/m3
2.1E-06 per ng/m3
"
"
See FOOTNOTES on next page.
1-71
-------
TABLE V-1-7. continued; CAHCER INHALATION UNIT RISK FACTORS
*For conversion of Inhalation Unit Risk Factor from units of per M9/m3 to units of per ppb; assumes an ideaV gas at 1=20 °C, P=1 atw. Sec sample
calculations on next page.)
••Assunes 70-kg person inhaling 20 tS air of ambient concentration composition for 2« h each day of a 70-year lifetime.
'From January 1992 IRIS unless noted as otherwise.
AHEAST 1992.
21991 H5AST
3In 1992 HEAST. but not IRIS; under review at EPA for reclassifteat ion as C-B2 continuum (U.S. EPA. 1992c).
Sample conversion of inhalation unit risk from cancer$/(ug/m3) to cancers
per ppb for benzene:
[B.3E-6 cancers/
-------
TABLE V-1-8; AVAILABILITY OF INHALATION UNIT RISK FACTORS, INHALATION REFERENCE
AMBIENT CONCENTRATIONS (OCTOBER 1988-SEPTEMBER 1989) FOR THE LEVEL
chemical
chloromethane
dichloromethane
trichloromethane
tetrachloromethane
dichloroethane, 1,1-
dichloroethane, 1,2-
trichloroethene
trichloroethane, 1,1,1-
trichloroethane, 1,1,2-
tetrachloroethene
tribromomethane
hexane
benzene
toluene
xylene, o-
xylene, m- and p-
styrene
ethylbenzene
chlorobenzene
dichlorobenzene, 1,2-
dichlorobenzene, 1,3-
dichlorobenzene, 1,4-
formaldehyde
arsenic
bar inn
beryllium
cadmium
chromium
cobalt
copper
iron
lead
CONCENTRATIONS, AND
1 RISK ASSESSMENT
Inhalation
Unit Risk
Factor*
yes
yes
yes
yes
no
no
yes
no (D)
no (0)
yes
no
no
yes
no (D)
no (0)
no (D)
no
no (D)
no
no
no
no
yes
yes
no
yes
yes
yes
no
no
no
no
Reference
Concentration
no
yes
yes
yes
no
no
yes
no
no
yes
no
yes
yes
yes
no
no
no
yes
no
no
no
no
no
no
no
no
yes
yes
no
no
no
no'4
Ambient
cone.
no
yes
yes
yes
no'
no2
yes
yes
no3
yes
no4
yes
yes
yes
yes
yes
yes*
yes
no6
no7
no8
no'
yes10-"
yes'2
yes12
no*
yes'0
yes13
yes12
yes'0
yes10
yes10
1-73
-------
TABLE V-1-8. continued
Chemical
manganese
mercury
molybdenum
nickel
selenium
vanadium
zinc
benzo[a]pyrene
Inhalation
Unit Risk
Factor*
no (0)
no (D)
no
yes
no
no
no
yes
Reference
Concentration
yes
yes
no
yes
no
yes
no15
no
yes"
yes'
yes'
yes"
no
yes'
yes'
yes'
Footnotes
(D) Indicates that the chemical is a Group D carcinogen,
a classification for which unit risk factors are not developed.
1 Low X>ndl for NYSOEC. high for CSI (No NJIT data)
2 Low X>mdl (NYSOEC and CSI; no NJIT data)
3 Almost never detected (NYSOEC and CSI; no NJIT data)
4 Almost never detected (CSI; no NYSOEC or NJIT data)
5 Almost always detected, only CSI (No NYSOC or NJIT data)
6 Low/medium X>mdl for NYSOEC, low for CSI (No NJIT data)
7 Medium/high X>mdl for NYSOEC, low for CSI (No NJIT data)
8 Lou X>mdl (NYSOEC and CSI; no NJIT data)
9 Low X>mdl (NYSOEC and CSI; no NJIT data)
10 Data for NYSOEC and NJIT sites; CSI did not conduct monitoring for particulates.
11 Low number of samples for NJIT sites. Ozone interference with sampling method for all sites
resulting in negative bias in reported concentrations.
12 Data for two NYSOEC sites only.
* Beryllium was never detected in ambient air.
13 Data for three NJIT sites only, including background site.
U National Ambient Air Quality Standard (NAAOS) for lead, not a reference concentration (RfC).
15 NAAQS for PM-10, not an RfC.
1-74
-------
TABLE V-1-9; MONCANCER RISK ESTIMATES, HAZARD QUOTIENTS* FOR VOLATILE ORGANIC COMPOUNDS
BASED ON ANNUAL AVERAGE
CHEMICALS
DICHLORONETHAHE
TRICHLOROHETHANE
TETkACHLOROMETHANE
TRICHLOROETHENE
rET«ACHLORO£THEHE
HEXANE, n-
BENZEHE
TOLUENE
XVLENE. o-**
XYIENES, p- and ro-**
ETHYLBENZENE
CART
--
0.003
0.4
0.01
0.04
0.02
3
0.04
...
...
-
CONCENTRATIONS FROM OCT. 88 THROUGH SEPT. 89
HAZARD OUOTIENTS
HEW JERSEY SITES
ELIZ SEW PSCAT HIPRK
..
0.003 0.003 0.003
0.4 0.4 0.3
0.01 0.01 0.01
0.05 0.05 0.03
0.02 0.02 0.01
322-
0.04 0.03 0.02
...
...
-
NEW YORK SITES
SW PS-26 PRT RCH PUMP
0.0006 0.001 0.001 0.001
0.01 0.02 0.01 0.03
0.3 0.3 0.3 0.3
0.02 0.03 0.03 0.06
0.04 0.04 0.06 0.3
.
1222
0.03 0.04 0.05 0.04
.
CT KLLS IpJT
0.0006 0.0008
0.01 0.01
0.3 0.5
0.02 0.02
0.05 0.05
-
2 2
0.03 0.03
~.
---
-
B-STN
--
0.005
0.3
0.02
0.06
0.02
3
0.04
...
---
0.002
ELTVL
--
0.007
0.4
0.01
0.05
0.02
3
0.04
...
---
0.003
PONGAM
--
0.005
0.4
0.02
0.2
0.02
4
0.04
-•-
---
0.003
* Hazard Quotient = Anfcient cone./Reference cone.
-- Submitted data were invalid.
Data not collected at this site.
** Uhite an RfC for xytenes uas available from the 1991 HE AST, it uas absent from the 1992 update (March 1992).
--- HEAST RfC withdrawn.
Note: Apparent site-to-site differences in calculated Hazard Quotients may not be statistically significant.
1-75
-------
TABLE V-1-10: CANCER RISK ESTIMATES, EXCESS LIFETIME (70 YEARS) CANCERS FOR VOLATILE ORGANIC COMPOUNDS-
BASED ON ANNUAL AVERAGE CONCENTRATIONS FROM OCT. 88 THROUGH SEPT. 89
CANCER RISK X 10*
CHEMICALS
OICHLOROMETHANE
TRICHLOROMETHANE
TETRACHLOROMETHANE
TRICHLOROETHENE
TETRACHLOROETHENE
HEXANE, n-
BENZENE
TOLUENE
XYLENE. o-
XYLENES, p- and m-
ETHYLBENZENE
NEW JERSEY SITES
CART
2.3
13
0.5
0.8
ELIZ
2.6
U
0.4
1.0
SEW
2.8
17
0.5
1.0
PSCAT HIPRK
2.7
12
0.5
0.6
su
0.9
8.8
10
1.1
0.8
PS -26
1.8
12
11
1.3
0.9
NEW YORK SITES
PRT RCH
1.6
10
9.5
1.4
1.1
PUMP
1.4
19
11
2.9
5.0
GT KLLS
1.0
7.8
11
0.8
0.9
TOTT
1.1
8.9
18
0.9
0.9
B-STN
4.5
12
0.8
1.2
ELTVL
4.7
16
0.7
1.0
DONGAN
4.6
14
0.7
3.1
46
45
36
30
24
39
42
36
29
27
44
47
61
Unit risk factor is not available.
-- Submitted data were invalid.
— Data not collected at this site.
Note: Apparent site-to-site differences in calculated risk may not be statistically significant.
1-76
-------
TABLE V-1-1V.
NOMCANCER RISK ESTIMATES, HAZARD QUOTIENTS* FOR METALS, BENZOtcriPYRENE, AND FORMALDEHYDE-
BASEO OH ANNUAL AVERAGE CONCENTRATIONS FOR OCT. 88 THROUGH SEPT. 89
Chemical
Hazard Quotient
ARSENIC
CADMIUM
CHROMIUM
using former HEAST
assuring 10X Cr
assuming 1X Cr
using NYSDOH RfC2
LEAD3
MANGANESE
MERCURY
NICKEL
VANADIUM
ZINC4
BaP
FORMALDEHYDE
CART
0.2
RfC1
VI 1.5
VI 0.15
0.3
0.03
0.06
0.002
2
0.003
EL1Z
0.1
0.9
0.09
0.2
0.03
0.04
0.002
1
0.003
H1PRK
0.1
0.7
0.07
0.1
0.07
0.04
0.002
1 1
0.002
SW PS- 26
0.1 0.1
0.03 0.04
0.04 O.OS
1
0.1 0.1
0.003 0.002
0.1
PORT RICH
0.09
* Hazard Quotient = Ambient air concentration/Reference Concentration
1 This RfC, appearing in the 1991 HEAST but absent from the 1992 update, is 2 ng/m3. For the purpose of the risk assessments for this project, this
RfC is treated as appropriate for hexavatent chromium (Cr VI), but not for Cr III, which is treated as noncarcinogenic. Total chromium, and not
Cr VI or Cr III, Mas quantitated; 10X or IX of the total chromium concentration is assumed to be Cr VI. See text for further detail.
2 The RfC from NYSOOH is 100 ng/m3.
3 The NAAQS for lead was used in the absence of an EPA-approved RfC. Since the NAAOS averaging period is a calendar year, the highest quarterly
average concentration for each site was used instead of the annual average concentration.
4 Based on the NAAQS for PM-10.
Note: Apparent site-to-site differences in calculated hazard quotients may not be statistically significant.
1-77
-------
TABLE V-1-12; CANCER RISK, ESTIMATED EXCESS INDIVIDUAL LIFETIME (70 YEARS) CANCER RISK
BASED ON ANNUAL AVERAGE CONCENTRATIONS FOR OCT. 88 THROUGH SEPT. 89
Chemical
Excess cancer risk x 10*
CART
EUZ
HIPRK
SU
PS-26
21
5.6
0.41
PORT RICH
ARSENIC 18
CADMIUM 8.7 3.7 4.3 5.0 5.2
CHROMIUM (10X)1 37 22 17 ----
CHROMIUM
-------
TABLE V-1-13; COMPARISON OF NONCANCER RISK ESTIMATES FOR AMBIENT AIR - HAZARD QUOTIENTS FOR SI/NJ UATAP AND UATMP STUDIES
CHEMICAL
DICKLOROMETHANE
HEXANE
TR1CHLOROMETHANE
TETRACHLOROMETHANE
BENZENE
TRICHLOROETHENE
TOLUENE
TETRACNLOROETHENE
ETHYLBENZENE
XYLENE. m- and p-'
XYLENE. o-4
FORMALDEHYDE
HAZARD QUOTIENT*
SI/NJ UATAP'
Min.
0.0006
0.0089
0.012S
0.3103
1.3750
0.0108
0.02
0.0362
0.0083
Max.
0.0011
0.0193
0.0938
0.5517
3.5000
0.0730
0.041
0.2868
0.0110
Med.
0.0008
0.0154
0.0250
0.3793
2.2679
0.0189
0.033
0.0553
0.0088
1988 UATMP2 (19 Cities)
Min.
0.0001
-
0.0063
0.0138
0.5714
0.0027
0.016
0.0105
0.0027
Max.
0.0009
.
2.0313
0.1724
6.0893
0.4243
0.15
1.0526
0,0962
Med.
0.0005
.
0.0625
0.034S
1.5536
0.0162
0.026
0.0184
0.0065
1989 UATMP1 (12 Cities)
Min.
0.0002
.
0.0025
0.4483
1.0714
0.0027
0.012
0.0184
0.0020
Max.
0.0048
-
1.6312
0.8276
7.0893
0.2405
0.15
0.1026
0.0228
Med.
0.0002
0.0250
0.6552
3.0714
0.0432
0.034
0.0447
0.0073
0.0872
0.103
0.0949
0.0671
0.1814
0.0971
ARSENIC
CADMIUM
CHROMIUM
using former HEAST RfC4S
assuring 10X Cr VI
assuming IX Cr VI
using NYSDOH RfC*
COBALT
COPPER
IRON
LEAD7
MANGANESE
MERCURY
MOLYBDENUM
NICKEL
VANADIUM
ZINC9
fiENZOfeJPYRENE
0.0093
0.1059
0.92
0.092
0.18
.
-
-
0.0111
0.0423
0.0019
-
1.1235
0.0894
0.0022
""
0.0108
0.2471
1.6
0.16
0.31
.
-
.
0.0351
0.0617
0.0019
-
1.6588
0.0994
0.0027
"
0.0100
0.1235
1.2
0.12
0.24
_
-
.
0.0267
0.0429
0.0019
-
1.2882
0.0941
0.0026
~
0.0070
0.0294
0.088
0.0088
0.017
.
-
-
0.0077
0.0583
0.0000
-
0.1647
0.0288
0.0006
•"
0.0210
0.7824
1.5
0.15
0.29
.
-
-
0.3385
1.4049
0.0000
-
2.0000
0.0841
0.0252
~
0.0083
0,0529
0.33
0.033
0.064
. -
.
-
0.0308
0.0823
0.0000
-
O.Z235
0.0306
0.0021
"
.
-
-
-
-
_
-
-
-
-
-
-
-
-
-
"
Note: Apparent site-to-site differences in calculated hazard quotients may not be statistically significant.
SEE FOOTNOTES ON NEXT PAGE.
1~79
-------
TABLE V-1-13. CONTINUED
FOOTNOTES
* Calculated using the minimum, maximum, and median annual average concentration data for sites in the study indicated, and adjusted reference concentrations
(RfCs) from Table V-1-6.
1 Based on concentration data for the period 10/88 through 9/89; does not include the background site (Piscatauay).
2 Based on concentration data for the period 9/24/87 through 10/6/88.
3 Based on concentration data for the period 1/22/89 through 1/17/90.
4 The RfC in the 1991 HEAST was withdrawn and, thus, absent from the 1992 update; no RfC is available currently.
- Either RfC or concentration data is/are not available.
5 This RfC, appearing in the 1991 HEAST but absent from the 1992 update, is 2 ng/m3; the RfC was withdrawn pending a public meeting scheduled for discussion of
the RfC. For the purpose of the risk assessments for this project, this RfC is treated as appropriate for hexavalent chromium (Cr VI), but not for Cr III,
which is treated as noncarcinogenic. Total chromium, and not Cr VI or Cr III, was quantitated; 10X or 1X of the total chromiun concentration is assumed to be
Cr VI. See text for further detail.
6 The RfC from NYSDOH is 100 ng/m3.
7 The NAAOS for lead was used in the absence of an EPA-approved RfC. Since the NAAOS averaging period is a calendar year, the highest quarter concentration for
each site was used instead of annual average concentration.
8 Based on the NAAQS for PM-10.
Note: Apparent site-to-site differences in calculated hazard quotients may not be statistically significant.
1-80
-------
TABLE V-1-14; COMPARISON OF CANCER RISK ESTIMATES FOR AMBIENT AIR - ESTIMATED EXCESS INDIVIDUAL LIFETIME CANCER RISK FOR SI/NJ UATAP AND UATMP STUDIES
CHEMICAL
INCREASED LIFETIME CANCER RISK*
SI/NJ UATAP1
DICHLOROMETHANE
HEXANE
TRICHLOROMETHANE
TETRACHLOROHETHANE
BENZENE
TRICHLOROETHENE
TOLUENE
TETRACHLOROETHENE
ETHYLBENZENE
XYLENE, m- and p-
XYLENE. o-
FORMALDEHYDE
ARSENIC
CADMIUM
CHROMIUM4
assuning 10X Cr VI
assuning 1X Cr VI
COBALT
COPPER
IRON
LEAD*
MANGANESE
MERCURY
MOLYBDENUM
NICKEL
VANADIUM
ZINC
B(A)P
Min.
8.9E-07
-
2.6E-06
9.9E-06
2.4E-05
4.4E-07
-
6.0E-07
-
-
-
3.5E-05
1.8E-05
3.8E-06
.
2.2E-06
2.2E-07
-
-
-
-
-
-
-
5.3E-06
-
-
3.0E-07
Max.
1.8E-06
-
1.9E-05
1.8E-05
6.1E-05
3.0E-06
-
5.0E-06
-
-
-
4.1E-05
2. IE-OS
8.8E-06
3.8E-05
3.8E-06
-
-
-
-
-
-
-
7.9E-06
-
-
4.2E-07
Med.
1.3E-06
-
5.26-06
1.2E-05
3.9E-05
7.7E-07
-
9.7E-07
-
-
-
3.8E-05
2.0E-05
4.4E-06
3.0E-05
3.0E-06
-
-
-
-
-
-
-
6.1E-06
-
-
3.4E-07
1988
Min.
1.1E-07
-
1.3E-06
4.4E-07
9.9E-06
1.1E-07
-
1.8E-07
-
-
-
-
1.4E-05
1.0E-06
2.1E-06
2.1E-07
-
-
-
-
-
-
•
7.8E-07
-
-
6.4E-08
UATMP' (19
Max.
1.4E-06
•
4.2E-04
5.5E-06
1.1E-04
1.7E-05
-
1.8E-05
-
-
-
-
4. IE-OS
2.8E-05
3.5E-05
3.5E-06
-
-
-
-
-
-
-
9.5E-06
-
-
1.0E-05
Cities)
Med.
8.4E-07
-
1.3E-05
1.1E-06
2.7E-05
6.6E-07
-
3.2E-07
-
-
-
-
1.6E-05
1.9E-06
7.8E-06
7.8E-07
-
-
-
-
-
-
-
1.1E-06
-
-
3.7E-07
1989 UATMP3 (12 Cities)
Min.
2.5E-07
-
5.2E-07
1.4E-05
1.9E-05
1.1E-07
-
3.2E-07
-
-
-
2.8E-05
.
-
-
-
-
-
-
-
-
-
-
-
-
-
Max.
7.8E-06
-
3.4E-04
2.6E-05
1.2E-04
9.8E-06
-
1.8E-06
-
-
-
7.6E-05
.
-
-
-
*
-
-
-
-
~
-
-
*•
*
Med.
4.0E-07
-
5.2E-06
2.1E-05
5.3E-05
1.8E-06
-
7.8E-07
-
-
-
4.1E-05
.
-
•
•
•
-
-
-
-
•
*
•
*
*
Note: Apparent site-to-site differences in risks may not be statistically significant.
SEE FOOTNOTES ON NEXT PAGE.
1-81
-------
TABLE V-1-U. CONTINUED
FOOTNOTES
* Calculated using the mini nun, maximum, and median annual average concentration data for sites in the study indicated, and
adjusted inhalation unit risk factors (lURFs) from Table V-1-7.
1 Based on concentration data for the period 10/88 through 9/89; does not include the background site (Piscatauay).
2 Based on concentration data for the period 9/24/87 through 10/6/88.
3 Based on concentration data for the period 1/22/89 through 1/17/90.
4 The available IURF is for hexavalent chromium (Cr VI). The risk assessments for this project assume that Cr VI is the only
carcinogenic chromium component of the reported total chromium concentrations, and that 10X or 1X of the total reported
concentrations is Cr VI. See text for further detail.
5 While lead is classified as a Group B2 carcinogen, an IURF for lead is not available.
1-82
-------
v-i-15; Indoor/Outdoor Ratios and Correlation Coefficients
between Indoor Air and Corresponding Outdoor Air
Concentrations
compound
Carteret, New Jersey
Site 0030-B1
Site 0030-B2
I/O
I/O
chloromethane
dichloromethane
hexane
chloroform
1,1,1-
trichloroethane
benzene
trichloro-
ethylene
toluene
tetrachloro-
ethylene
ethylbenzene
n^p-xylene
0-xylene
l.l
0.4*
0.9
1.8*
0.9
0.9
5.1*
1.6*
1.9*
1.4*
1.2
1.1
0.34
0.33
0.47
0.50
0.11
0.65
0.33
0.12
0.09
0.54
0.47
0.37
0.51
0.53
0.63
0.51
0.67
0.65
0.37
0.00
0.37
0.45
0.49
0.45
1.2
0.5*
1.9*
2.9*
0.5*
1.6*
2.3*
2.0*
1.7*
2.1*
2.0*
2.0*
0.05
0.30
0.54
0.33
0.07
0.36
0.38
0.04
0.61
0.22
0.36
0.05
0.43
0.45
0.57
0.43
0.07
0.60
0.46
0.30
0.63
0.50
0.74
0.53
* = p<0.05
j/O = mean indoor air concentration divided by the corresponding
mean outdoor air concentration
•p = Pearson correlation coefficient
e = Spearman correlation coefficient
1-83
-------
Table V-l-16: Indoor/Outdoor Ratios and Correlation Coefficients
between Indoor Air and Corresponding Outdoor Air
Concentrations
Compound
Staten Island
Site 7097-2A
chloromethane
dichloromethane
hexane
chloroform
1,1,1-
trichloroethane
benzene
trichloro-
ethylene
toluene
tetrachloro-
ethylene
ethylbenzene
m,p-xylene
o-xylene
* - p<0.05
I/O = mean indoor
I/O
2.2*
0.8*
2.1*
1.7*
0.85
1.7*
0.56*
2.0*
0.83
1.7*
2.0
1.5
P
0.26
0.77
0.76
0.19
0.73
0.67
0.28
0.09
0.88
0.64
0.51
0.37
air concentration
S
0.18
0.73
0.73
0.34
0.64
0.66
0.48
0.21
0.86
0.68
0.51
0.40
divided
Site
I/O
2.5*
3.0*
1.7*
3.4*
1.0
1.4*
0.74
1.7*
1.1*
1.1*
1.1*
0.95
by the
7097-2B
P
0.28
0.20
0.48
0.22
0.43
0.58
0.43
0.20
0.86
0.74
0.76
0.55
S
0.23
0.11
0.63
0.28
0.11
0.35
0.55
0.43
0.78
0.51
0.84
0.83
corresponding
mean outdoor air concentration
P « Pearson correlation coefficient
S » Spearman correlation coefficient
1-84
-------
V-l-17; Comparison of Ambient Air Data for PS-26 (Travis,
Staten Island)—NYSDOH* (7/90-3/91) vs.
UATAP** (10/88-3/89 and 7/89-9/89)
DATAP**
chloromethane
dichloromethane
hexane
chloroform
n
NA
41
NA
41
mean
(PPb)
NA
0.93
NA
0.11
n
36
44
36
44
NYSDOH
mean ratio* difference15
(PPb)
0.6
1.2 1.3 +0.24
1.2
c
1,1,1-
trichloroethane
carbon
41
0.49
44
0.7
1.4
+0.20
tetrachloride
benzene
trichloroethylene
toluene
tetrachloro-
ethylene
ethylbenzene
jn/p-xylene
0-xylene
41
41
26
41
41
NA
41
41
0.11
1.29
0.08
4.04
0.18
NA
1.47
0.45
44
44
36
44
44
36
36
44
C
1.7 1.4
c
6.1 1.5
c
0.9
3.1 2.2
1.4 3.5
-
+0.47
-
+2.19
-
-
+1.72
+1.13
* NYSDOH refers to the indoor air portion of the SI/NJ UATAP.
** UATAP refers to the ambient air portion of the SI/NJ UATAP.
jjA -not available
a -ratio of NYSDOH/UATAP
b -difference equals NYSDOH minus UATAP
c -low frequency of detection prevents calculation of a valid mean
1-85
-------
Table V-l-18; Comparison of Ambient Data for Carteret, New Jersey—
NYSDOH* (7/90-3/91) VS.
UATAP** (10/88-3/89 and 7/89-9/89)
UATAP**
chloromethane
dichlorome thane
hexane
chloroform
1,1,1-
trichloroethane
carbon
tetrachloride
benzene
trichloroethylene
toluene
ethylbenzene
m/p-xylene
o-xylene
n
NA
NA
25
40
40
40
40
40
40
NA
40
40
mean
(ppb)
NA
NA
1.09
0.01
0.58
0.11
1.54
0.04
4.11
NA
1.29
0.43
NYSDOH
n
36
42
36
42
42
42
42
36
42
36
36
44
mean
(ppb)
0.7
2.2
0.8
C
2.6
C
1.4
C
6.0
0.9
3.1
1.4
ratio* difference1
-
- -
0.8 -0.24
- -
4.4 +1.98
- -
0.9 -0.14
-
1.4 +1.76
- -
2.2 +1.72
3.5 +1.13
* NYSDOH refers to the indoor air portion of the SI/NJ UATAP.
** UATAP refers to the ambient air portion of the SI/NJ UATAP.
NA -not available
a -ratio of NYSDOH/UATAP
b - difference equals NYSDOH minus UATAP
c - low frequency of detection prevents calculation of a valid mean.
1-86
-------
Table V-l-19; Level 2 Exposure Assessment Scenarios and
Sample Calculation
General equations and explanation
The general equation used to calculate lifetime average daily
dose is as follows:
D=Z[(V70)DAi],
where DAi = average daily dose for a given age
interval, and
tj = number of years covered by that
age interval.
The general equation used to calculate average daily dose for
each age interval is as follows, where C, and C0 are variables;
and V, and V0 are selected constants that vary with age and level
of activity, and B is a constant that varies with age:
DA= (VAIC, + VAOC0))/WA,
where DA = average daily dose, jug/ (kg-d) , for a given
age interval;
VAI = volume of indoor air inhaled per day, m3/d;
VAO = volume of outdoor air inhaled per day,m3/d;
CAI = average concentration of pollutant in
indoor air, /ig/ni3;
CAO = average concentration of pollutant in
outdoor air, /xg/m3; and
WA = average body weight, kg, for given age
interval.
Of the numerous options for the constants V,, V0/ and B, several
sets were chosen and presented as exposure scenarios. Inhalation
rate varies with age and activity level (e.g., resting, light
exercise, or moderate exercise). The number of hours spent
indoors and outdoors varies with age and occupation/lifestyle.
The tables below present the body weights and information used to
derive the inhaled volumes of air.
1-87
-------
Table V-1-19. continued
Table V-1-19a: Body Weight and Inhalation Rate for Three Age Intervals and Three Activity Levels
Bodv
Inhalation Rate. m3/d
Age, years
0-2
3-18
19-70
10
40
70
Resting
01?5
0.4
0.5
Light exercise
1
0.6
Moderate exercise
0155
3.2
2.1
Table V-1-19b; Hours Per Day Spent Indoors and Outdoors by Age Interval
Hours per day
Age, years
0-2
3-18
19-70
Scenario 1
1 ndoors Outdoors
23.5 0.5
20.4 3.6
22.5 1.5
Scenario 2
Indoors C
23.5
20.4
17
lutdoors
0.5
3.6
7
Table V-1-19c: Volume of Inhaled Air by Age Interval (based on information in Tables V-1-19a and 19b)
Age, years
Volume of air inhaled. m3/d
Scenario 1
0-2
3-18
19-70
Indoors
2.9375
14.28
12.375
Outdoors
0.0625
11.52
3.15
Daily
Total
3.00
25.80
15.52
Indoors
2.9375
14.28
9.35
Scenario 2
Outdoors
0.0625
11.52
14.7
Daily
Total
3.00
25.80
24.05
Sample Calculation
Sample Level 2 exposure calculation of lifetime average daily dose for tetrachloroethene:
Scenario 1 for Home 0030-B1 (Table V-1-20)
Vlf m3/d
C,, rag/m3
v0, mVd
C0, tng/m3
U, kg
0 thru 2 vrs
2.9375
3.5
0.0625
1.8
10
3 thru 18 vrs
14.28
3.5
11.52
1.8
40
19 thru 70 yrs
12.375
3.5
3.15
1.8
70
D- (2/70)[<2.9375)C3.5>+(0.0625)(1.8)3/10 +
(16/70) [(14.28X3.5)+(11.52X1.8)1/40 *
(52/70) [(12.375X3.5)+(3.15X1.8)1/70
- 0.02970 + 0.4041 + 0.5198
0- 0.9536
1-88
-------
TABLE V-1-20: ESTIMATED EXCESS INDIVIDUAL LIFETIME CANCER RISKS AMD HAZARD QUOTIENTS,
BASED ON MEASURED AIR QUALITY JULY 1990 - MARCH 1991 for
Home 0030-81, Ambient monitor 0030-B3
POLLUTAHT
chloromethane
di chIoromethane
hexane
t r i choloromethane
tet rachloromethane*
benzene
trichloroethene
toluene
tetrachloroethene
ethylbenzene
m,p-xylene
o-xylene
Conc.Qig/m3)1
in out
Life.Avg.Expos.'
(ug/kg- day)
Cancer
Slope factor
Sccn.1 Seen.2 (»g/kg-d)-
1.5
3.1
2.7
1.7
1.3
4.2
5.5
35.2
3.5
3.9
12.6
5.4
1.4
7.9
3.0
1.0
1.3
4.5
1.1
22.2
1.8
2.8
10.6
5.0
0.47
1.47
0.89
0.47
0.43
1.38
1.33
10.01
0.95
1.15
3.85
1.69
0.59
2.34
1.17
0.54
0.54
1.79
1.28
11.59
1.06
1.36
4.75
2.14
6.30E-06
1.60E-06
B.10E-05
1.30E-04
2.90E-05
1.70E-05
1.80E-06
Est.Lifetiroe
Cancer Risk3
Seen. 1
3.0E-06
2.4E-06
-
3.8E-05
5.5E-05*
4.0E-05
2.3E-05
-
1.7E-06
-
-
-
Seen. 2
3.7E-06
3.7E-06
-
4.3E-05
7.0E-05*
5.2E-05
2.2E-05
-
1.9E-06
-
-
-
Ref.
Cone.4
(ug/m3)
826
3000
200
30.4
2.10
2.10
23.00
400
30.0
1000
300
700
Composite Air
Concentration5
(ug/nt3)
Seen. 1 Seen. 2 Seen. 1 Seen. 2
Hazard Quotient
6
1.65
5.15
3.12
1.66
1.49
4.63
4.64
35.03
3.32
4.01
13.48
5.92
2.06
8.20
4.10
1.88
1.89
6.27
4.48
40.55
3.72
4.77
16.62
7.47
0.002
0.002
0.016
0.055
0.71*
2.3
0.20
0.088
0.11
0.002
0.003
0.020
0.062
0.90*
.0
.195
.10
.12
0.004 0.005
1
Indoor and outdoor concentrations
Lifetime Average Exposure.
"Scen.1" and "Seen.2" (scenarios 1 and 2).
3 Estimated Excess Individual Lifetime Cancer Risk
Estimated Excess Individual Lifetime Cancer Risk
See Table V-1-19 for sample calculation of lifetime average daily dose, and for meaning of
. Sample calculation for benzene, scenario 1:
= (Life.Avg. Expos.) x (Cancer Slope Factor)
= 1.38 jig/kg-d) x [2.9E-05/ (0g/kg-d}]
= 4.0 x 10-5
* Reference Concentration
5 Sample calculation for benzene, scenario 1: Composite Air Concentration = (Life.Avg. Expos. )x (1/volume inhaled daily) x (Weight of Individual).
= {1.38 /ig/kg-d) x (1/20m3/d) x (70 kg)
= 4.83 fig/nS
6 For this risk assessment. Hazard Quotient = Composite Air Concentration/Reference Concentration.
* Tet rach loromethane was never detected in the indoor air portion of the SI/NJ UATAP. 1.3 ng/m3 is half the weighted average mdl. The detection
limit was 5.4 fig/m3 from 7/10/90 to 10/2/90, and 1.2
from 10/14/90 to 3/19/91.
1-89
-------
TABLE V-1-21: ESTIMATED EXCESS INDIVIDUAL LIFETIME CANCER RISKS AND HAZARD QUOTIENTS,
BASED ON MEASURED AIR QUALITY JULY 1990 - MARCH 1991 for
Home 0030-B2, Ambient monitor 0030-B3
POLLUTANT
chloromethane
dichloromethane
hexane
trichloromethane
tetrachtoromethane
benzene
trichloroethene
toluene
tetrachloroethene
ethylbenzene
m,p-xylene
o-xylene
Cone.
in
out
1.6
3.6
5.7
2.8
1.4*
7.0
2.5
45.0
3.1
5.8
21.4
10.3
1.4
7.9
3.0
1.0
1.3*
4.5
1.1
22.2
1.8
2.8
10.6
5.0
Life.Avg.Expos.
(ug/kg- day)
Scen.1
Cancer
Slope Factor
Seen.2 (ug/kg-d)-'
0.48
1.58
1.56
0.72
0.45
2.00
0.65
12.19
0.87
1.57
5.80
2.78
0.60
2.43
1.74
0.75
0.56
2.33
0.70
13.44
0.99
1.72
6.41
3.07
6.30E-06
1.60E-06
8.10E-05
1.30E-04
2.90E-05
1.70E-05
1.80E-06
Est. Lifetime
Cancer Risk 3
Seen. 1
3.0E-06
2.5E-06
-
5.9E-05
5.8E-05*
5.8E-05
1.1E-05
-
1.6E-06
-
-
-
Seen. 2
3.8E-06
3.9E-06
-
6. IE-OS
7.3E-05*
6.7E-05
1.2E-05
-
1.8E-06
-
-
-
Ref.
Cone.4
Composite Air
Concentration5
(ug/m3)
(ug/m3) Seen. 1
826
3000
200
30.4
2.10
2.10
23.0
400
30.0
1000
300
700
1.69
5.52
5.45
2.53
1.56
7.02
2.29
42.65
3.03
5.48
20.29
9.73
Seen. 2
2.10
8.51
6.09
2.62
1.96
8.14
2.46
47.05
3.47
6.03
22.44
10.73
Hazard Quotient'
Seen. 1 Seen. 2
0.002
0.002
0.027
0.083
0.748
3.3
0.099
0.11
0.10
0.005
0.068
0.014
0.003
0.003
0.030
0.086
0.93*
3.9
0.11
0.12
0.12
0.006
0.075
0.015
Footnotes; See Table V-l-20.
1-90
-------
TABLE V-1-22: ESTIMATED EXCESS INDIVIDUAL LIFETIME CANCER RISKS AND HAZARD QUOTIENTS,
BASED ON MEASURED AIR QUALITY JULY 1990 - MARCH 1991 for
Home 7097-2A, Ambient monitor 7097-2C
Life.Avg.Expos.2
Cone.
POLLUTANT in
chtoromethane 2.8
dichloromethane 3.3
hexane 8.6
trichloromethane 1.7
tetrachloromethane* 1.3*
benzene 9.8
trichloroethene 1.0
toluene 46.8
tetrachloroethene 2.1
ethylbenzene 7.1
«,p-xylene 28.1
o-xylene 10.4
-------
TABLE V-1-23: ESTIMATED EXCESS INDIVIDUAL LIFETIME CANCER RISKS AND HAZARD QUOTIENTS,
BASED ON MEASURED AIR QUALITY JULY 1990 - MARCH 1991 for
Home 7097-2B, Ambient monitor 7097-2C
POLLUTANT
chloromethane
dichloromethane
hexane
trichloromethane
tetrachloromethane*
benzene
t r i chIoroethene
toluene
tetrachIoroethene
ethylbenzene
m,p-xylene
o-xylene
Conc.Oig/m3)1
in
3.1
12.4
7.3
3.6
1.3*
8.0
1.3
38.8
2.9
4.5
14.9
6.5
out
1.3
4.1
4.2
1.1
1.4*
5.6
1.7
23.5
2.6
4.1
13.9
6.8
Li fe.Avg. Expos.2 Cancer
(ug/kg- day) Slope Factor
Scen.1
0.82
3.16
2.03
0.90
0.42
2.32
0.46
10.94
0.89
1.40
4.67
2.13
Seen. 2 (uq/kg-d)-1
0.87
3.26
2.31
0.91
0.55
2.75
0.63
12.56
1.12
1.76
5.90
2.76
6.30E-06
1.60E-06
-
8.10E-05
1.30E-04
2.90E-05
1.70E-05
-
1.80E-06
-
-
-
Est. Lifetime
Cancer Risk3
Seen. 1
5.1E-06
5.1E-06
-
7.3E-05
Seen. 2
5.5E-06
5.2E-06
-
7.4E-05
5.5E-05*7.1E-05*
6.7E-05
7.8E-06
-
1.6E-06
-
-
8.0E-05
1.1E-05
-
2.0E-06
Ref.
Cone/
(ug/m3)
826
3000
200
30.4
2.10
2.10
23.0
400
30
1000
300
700
Composite Air
Concentration
(ug/m3)
Seen. 1
2.85
11.06
7.10
3.15
1.48
8.13
1.60
38.29
3.12
4.88
16.36
7.46
Seen. 2
3.04
11.41
8.07
3.20
1.92
9.64
2.19
43.97
3.90
6.14
20.64
9.65
Hazard Quot i ent6
Seen. 1
0.0035
0.0037
0.036
0.10
0.71*
3.9
0.069
0.096
0.10
0.0049
0.055
0.11
Seen. 2
0.0037
0.0038
0.040
0.11
0.91*
4.6
0.095
0.011
0.13
0.0061
0.069
O.OU
Footnotes: See Table V-l-20.
1-92
-------
V-l-24:
Arsenic
Benzene
Benzo(a)pyrene
pi
chloromethane
Formaldehyde
n-
summary of Risk (Cancer and Noncancer) from
Median Annual Average Concentrations of
Pollutants Addressed in the Level 1 Risk Assessment
Median
Cancer Risk
(per million)*
20
40
0.38
5.1
30d, 3e
1.2
0.26
Manganese
Mercury
/jetrachloroethene
rpetrach lor ome thane
,jirichloroethene
^,richloromethane
Vanadium
o-
p- and m-
Zinc
TOTAL
6.0
1.0
12
0.8
6.2
Median
Hazard Quotient*
_c
2
-
0.1
1.2f, 0.12", 0.2h
0.0008
0.03
0.1
0.2
0.031
0.04
0.002
1
0.05
0.4
0.04
0.02
0.009
0.1
-
—
0.003
Number
of Sitesb
2
12
4
4
2
6
3
2
6
4
4
2
4
12
12
12
12
12
2
12
12
4
123d, 96e
5.3f, 4.2g, 4.3h
background sites were excluded when determining the median annual
average concentrations.
b Dumber of sites contributing annual average concentrations to the
medians.
c Dash indicates that the Inhalation Unit Risk Factor or the Reference
Concentration was not available so that the cancer risk or Hazard
Quotient could not be calculated.
footnotes continued on next page.)
1-93
-------
Table V-l-24. continued
d Assuming 10% chromium VI. The cancer risks associated with chromium
exposure were calculated assuming that 10% or 1% of the total reported
chromium concentration is in the hexavalent oxidation state (Cr VI).
In the ambient environment, chromium is found in the Cr VI and Cr III
forms. Only Cr VI has been shown to be carcinogenic; the cancer unit
risk factor is based on exposure to Cr VI. See text for further
detail.
e Assuming 1% Cr VI. See footnote d above.
f Using the former HEAST RfC and assuming that 10% of the total reported
chromium concentration is Cr VI. See text for further detail.
8 Using the former HEAST RfC and assuming that 1% of the total reported
chromium concentration is Cr VI. See text for further detail.
h Using the NYSDOH RfC for total chromium. The NYSDOH RfC is based on
different toxicological studies from those used to develop the former
HEAST RfC. See text for further detail.
' In the absence of either an Inhalation Unit Risk Factor or a Reference
Concentration, the current National Ambient Air Quality Standard
(NAAQS) for lead was used in place of an RfC in the Hazard Quotient
calculation. EPA, currently reevaluating the standard, plans to
publish a notice in the Federal Register concerning a proposed new
NAAQS for lead.
1-94
-------
Table V-l-25; Noncancer Additive Risk Analysis by Target Organ
Effect Associated with Chemical Exposure
Respiratory Tract (Irritation)
Chromium
Formaldehyde
Manganese*
Nickel
Vanadium
Zinc
TOTAL HI, Respiratory
Liver Effects
Ethylbenzene**
Trichloromethane
Tetrachloromethane
Trichloroethene
Tetrachloroethene
Xylenes (m-)
TOTAL HI, Liver
Hematopoietic System
Benzene
TOTAL HI, Hematopoietic
Kidney
Cadmium
TOTAL HI, Kidney
Central Nervous System
Lead
Mercury***
Dichloromethane
Hexane n-
Toluene
Xylene, o-
TOTAL HI, CNS
Hazard Quotient
from Table V-l-24
1.2', 0.12", 0.2C
0.1
0.04
1.0
0.1
0.003
2', 1", lc
0.03
0.009
0.4
0.02
0.05
_d
0.5
2.0
2
0.1
0.1
0.03
0.002
0.0008
0.02
0.04
0.1
Footnotes
* Developmental toxicity and effects on the kidney are also associated with exposure to manganese.
*« Effects on the central nervous system have been associated with exposure to ethylbenzene.
*** Effects on the kidney have been associated with exposure to mercury.
a Assumes the former HEAST RfC and 10X Cr VI in the total reported chromium concentration.
b Assumes the former HEAST RfC and 1X Cr VI in the total reported chromium concentration.
c Assumes the NYSDOH RfC for total chromium.
d The RfC in the 1991 HEAST does not appear in 1992 update; no RfC is available currently.
1-95
-------
2. STATISTICAL ANALYSES
2.1 INTRODUCTION
The Staten Island/New Jersey Urban Air Toxics Assessment
Project included ambient air monitoring for a set of target
organic compounds at 13 sites in the area. The compounds
examined in this report include: benzene (BENZ), toluene (TOLU),
o-xylene (OXYL) , hexane (HEXA) , 1,1,1-trichloroethane (THE),
carbon tetrachloride (GARB), trichloroethene (TRIG),
tetrachloroethene (TECH), styrene (S), and m,p-xylene (MPXY).
The physical monitoring covered a two-year period beginning in
October 1987 and was conducted by three independent
organizations: the College of Staten Island (CSI), the New
Jersey Institute of Technology (NJIT) and the New York State
Department of Environmental Conservation (NYSDEC). Each
organization had complete responsibility for its sampling sites
and employed different sampling and analytical methodologies.
The purpose of this section is to present the results of a
statistical analysis of the project's volatile organic compound
(VOC) ambient monitoring data, to assess the site-to-site
differences for each compound, and to determine the concentration
levels that should be included in the risk assessment for the
project. This statistical assessment was performed by Research
Triangle Institute, under contract to EPA. It followed similar
work performed by the University of Medicine and Dentistry of New
Jersey (UMDNJ), and utilized the data files prepared by UMDNJ.
Sections 2 and 3 focus on the assessment of site-to-site
differences using "adjusted" data derived from collocated
measuring instruments which were made by PEI using a canister
technique. Section 4 examines the potential transformation of
the CSI data set by utilizing the logarithm of the concentrations
to assess differences among the sites it operated. Unadjusted
data are used since all measurements were made by a single
organization.
2.2 ADJUSTING FOR METHOD BIAS
The initial step in assessing organization bias was to plot
the data from collocated instruments by compound and
organization. For sites operated by NYSDEC, the data were
plotted by year to correspond to the two methods employed—
sorbent and ATD. After examining the plots, some data points
were deleted as outliers. These are listed in Table V-2-1. In
2-1
-------
addition, an assessment of bias was considered inappropriate for
TECH in NJIT sites because of the small number of collocated
samples and for HEXA in CSI sites in year 1 because of apparent
problems with the PEI canister.
After excluding the data points identified above, three
models were fitted to the collocated data. This was done by
compound for CSI and NJIT and by compound and method for NYSDEC.
The models were
Model A: E(\/PEI) = (a + BXr )
1/2
Model B: E(ln(PEI)) = In (a + BXr)
Model C: E(PEI) = a + 6Xr
where
E(VpEI)
E(ln(PEI))
expected value of the square root of the PEI
measurement corresponding to a given value of Xr,
expected value of the logarithm of the PEI
measurement corresponding to a given value of Xr,
2-2
-------
Table V-2-1
VALUES DELETED IN ASSESSING BIAS
Compound
MPXY
OXYL
T11E
TECH
TOLU
BENZ
Organization
NYSDEC
CSI
CSI
NYSDEC
NYSDEC
CSI
NYSDEC
cst
CSI
NYSDEC
Year
2
1
1
2
2
2
2
1
2
2
(6.6,
(5.8,
(1A
(8.6,
(0.3,
(2.7,
(6.7,
(6.4,
(13.4,
(0.3,
Values*
0.9), (0.9, 2.9)
0.9)
0.1)
0.8)
2.3)
0.2)
0.3)
1.6)
3.6)
2.6)
* The first value within the parenthesis is the PEI measurement.
2-3
-------
E(PEI) = expected value of the PEI measurement
corresponding to a given value of Xr,
X;. = reported measurement, and
a,B = parameters to be estimated.
Note that the parameters a and 6 have the same interpretation for
all three models: a = 0 implies an additive bias (of one method
relative to PEI) and 6=1 implies a multiplicative bias relative
to PEI. The models differ in the error structure—in Model A
the error is additive on the square root scale, in Model B the
error is additive on the logarithmic scale and in Model C the
error is additive on the original measurement scale. In each
case the variances of the errors are assumed stable on their
respective scales.
Estimates of a and 6 (i.e., a and b, respectively) for the
three models are given in Table 2 by compound, organization and
method. In all cases, the intercept, a, is positive and in most
cases it is statistically different from zero. In general, the
slope parameter estimate, b, is statistically less than one (some
exceptions are NYSDEC measurements of OXYL, THE and TECH). For
a given compound/organization/method combination, the
relationship between PEI measurements and an organization's
reported measurements is basically the same for all three models.
For instance, the first line of Table V-2-2 shows the following
relationships between PEI and CSI in the measurement of BENZ
Model A: PEI = 0.61 + 0.51 (CSI)
Model B: PEI = 0.57 + 0.52 (CSI)
Model C: PEI - 0.64 + 0.51 (CSI)
2-4
-------
Le V-2-2
|TER ESTIMATES FOR MODELS A, B, C
D TO COLLOCATED DATA
Model A
Compound
BENZ
HEXA
MPXY
OXYL
T11E
TECH
TOLU
Org.
CSI
NYSDEC
NYSDEC
NJIT
CSI
NJIT
CSI
NYSDEC
NYSDEC
NJIT
CSI
NYSDEC
NYSDEC
NJIT
CSI
NYSDEC
NJIT
CSI
NYSDEC
CSI
NYSDEC
NYSDEC
NJIT
a
Metprcept
- (-08)*
Af(.04)
S0(-10)
5 (.09)
6 (.07)
•1 (.09)
13 (.08)
/21 (.08)
c27{.14)
53(.12)
21 (.04)
.17 (.08)
.25 (.10)
.25 (.06)
.31 (.06)
.22 (.04)
.33 (.07)
.12 (.06)
.08 (.05)
.57 (.18)
.40 (.13)
.39 (.23)
1.1 6 (.27)
b
Slope
.51 (.07)
.76 (.04)
.39 (.10)
.50 (.08)
.54 (.09)
.51 (.10)
.49 (.05)
.81 (.06)
.81 (.13)
.58 (.13)
.97(.11)
1.06(.18)
.92 (,26}
.51 (.15)
.87 (.15)
.95 (.09)
.66 (.14)
.73 (.09)
1.29 (.13)
.66 (.05)
.70 (.04)
.77(.10)
.49 (.09)
Model
a
Intercept
,57(.Q8)
.29 (.04)
.47 (.09)
.39 (.08)
.37 (.06)
.33 (.07)
.43 (.08)
.14(.06)
.23(.13)
.41 (.08)
.21 (.04)
.14 (.05) 1
.23 (.08)
.22 (.05)
.33 (.06)
.22 (.04)
.30 (.07)
.17 (.06)
.10(.03) 1
.54(.14)
.40 (.10)
.44 (.20)
.82 (.20)
B
b
Slope
.52 (.08)
.72 (.05)
.39 (.11)
.54 (.10)
.51 (.09)
.59 (.12)
.47 (.06)
.82 (.07)
.82 (.14)
.72 (.15)
.92 (.12)
.02 (.14)
.89 (.25)
.53(.16)
.75(.16)
.89 (.10)
.69 (.17)
.61 (.13)
.17(.13)
.66 (.06)
.68 (.04)
.70(.11)
.60 (.10)
Mode!
a
Intercept
.64 (.09)
.25 (.05)
.52(.10)
.54(.10)
.36 (.09)
.52 (.11)
.44 (.09)
.36 (.10)
.35 (.17)
.70 (.15)
.21 (.04)
.33 (.12)
.29 (.11)
.29 (.07)
.27 (.08) 1
.24 (.05)
.37 (.08)
.09 (.06)
.16(.11) 1
.62 (.23)
.34 (.18)
.39 (.28)
1.50 (.33)
C
b
Slope
.51 (.06)
.80 (.04)
.40{.10)
.46 (.06)
.56 (.09)
.42(.10)
.50 (.05)
.74(.06)
.79(.13)
.46 (.13)
.98 (.09)
.89 (.21)
.90 (.27)
.48{.1S)
.01 (.14)
.96 (.08)
.64(.12)
.80 (.05)
.21 (.12)
.66 (.06)
.74 (.04)
.80 (.10)
.42 (.08)
Standard error of estiK60 in parenthesis.
2-5
-------
Other examples can be formed by the estimates given on any line
in Table V-2-2.
The choice of which model to use to make adjustments to data
reported by an organization was based, in part, on correlations
between the observed and predicted values of PEI measurements.
For a given compound/organization/method combination, these
correlations correspond to
, 1/2
Model A: corr(vpEI, (a + bXr ) )
Model B: corr(In(PEI), ln(a + bXr ))
Model C: corr(PEI, a + bXr )
The sets of correlations calculated for the three models are
shown in Table V-2-3. Ideally, the model with the highest
correlations is the best selection; however, there was not a
particular model that exhibited this characteristic across all
compounds and organizations. We selected Model B for two
reasons. First, in cases in which all three correlations were
relatively low, this model tended to have the highest
correlations (e.g., OXYL measurements by NYSDEC's ATD method).
Second, and more importantly, it was felt that standard
deviations of measurement errors are likely to increase
proportionately with concentration level. Having selected
Model B, adjustments were then made for all reported measurements
of compounds for which the organization's correlation shown in
Table V-2-3 was >0.60*. Adjustments based on correlations below
0.60 were not considered reliable. For a given
compound/organization/method combination, an adjusted measurement,
adjusted Xr, was calculated for each reported measurement, Xr, using
the following relationship:
Adjustments based on correlations below 0.60 would account for
less than 36% of the variation in the data. While this is an
arbitrary cut-off, it is logical approach for dealing with a
difficult situation.
2-6
-------
Table V-2-3
CORRELATION BETWEEN THE OBSERVED AND PREDICTED
VALUES OF PEI BY MODEL, COMPOUND AND ORGANIZATION
Compound
BENZ
HEXA
MPXY
OXYL
T11E
TECH
TOLU
Organization
CS1
NYSDEC
NYSDEC
NJIT
CSI
NJIT
CSI
NYSDEC
NYSDEC
NJIT
CSI
NYSDEC
NYSDEC
NJIT
CSI
NYSDEC
NJIT
CSI
NYSDEC
CSI
NYSDEC
NYSDEC
NJIT
Method
.
ATD
SOR
-
_
-
_
ATD
SOR
-
_
ATD
SOR
-
_
ATD
-
—
ATD
_
ATD
SOR
-
Model A
0.72
0.87
0.52
0.78
0.80
0.73
0.81
0.78
0.71
0.61
0.82
0.52
0.50
0.52
0.71
0.74
0.67
0.90
0.88
0.84
0.85
0.75
0.72
Model B
0.70
0.84
0.51
0.77
0.79
0.78
0.77
0.78
0.71
0.69
0.77
0.62
0.51
0.55
0.67
0.70
0.64
0.82
0.89
0.85
0.84
0.71
0.76
Model C
0.74
0.88
0.54
0.78
0.80
0.66
0.82
0.76
0.68
0.53
0.85
0.38
0.47
0.48
0.75
0.75
0.68
0.95
0.82
0.83
0.85
0.76
0.68
2-7
-------
adjusted Xr = a + bXr
where a and b are the values shown in Table V-2-2 for Model B.
Table V-2-4 shows the number of observations contained in the
adjusted database.
2.3 ANALYSIS OF VARIANCE RESULTS USING ADJUSTED DATA
Differences among sites operated by all three organizations
were assessed through the use of the two-way analysis of variance
(ANOVA) procedure applied to logarithms of adjusted measurements.
For a given compound, this procedure reguires adjusted
measurement data from s sites on each of d days. The sites are
coded as follows:
Susan Wagner
Travis (PS-26)
Eltingville
Great Kills
Port Richmond
Dongan Hills
1
2
3
4
5
6
Pumping Station
Bayley Seton Hospital
Tottenville
Elizabeth
Carteret
Sewaren
Piscataway
7
8
9
A
B
C
D
The sites included in a particular analysis of a compound are shown
in Table V-2-5. These sites were determined after examining the
availability of adjusted data (as reflected in Table V-2-4) and the
data collection periods for the various sites. For example,
Sites 2, 4, 7, C and D began operations in January 1989 and
Sites 3, 6 and 8 ceased operations in April 1989. Because of
this short overlapping time period, the number of days common to
all 13 study sites with usable data is very small. Therefore,
Sites 2, 4, 7, C and D were never included with Sites 3, 6 and 8
in any of the ANOVAs (see Table V-2-5). For some compounds (e.g.,
BENZ), multiple ANOVAs were made by dropping selected sites in
order to increase the number of days with usable data.
The ANOVA results (specifically the Student-Newman-Keuls
(SNK) test results) are given in Appendix A and are summarized in
Table V-2-6.
2.4 INTERPRETATION OF ANOVA RESULTS
The ANOVA results presented in Table V-2-6 and in Appendix A
depict the relative differences in the mean concentrations measured
at the various sites for each contaminant. For each SNK Analysis,
the sites are arranged in descending order of mean concentration,
with each underlined group of sites having mean concentrations
2-8
-------
measured at the various sites for each contaminant. For each SNK
Analysis, the sites are arranged in descending order of mean
concentration, with each underlined group of sites having mean
concentrations statistically indistinguishable from each other
(at the 0.05 level). Thus, for Analysis 7 toluene (TOLU), for
example, the first seven sites are not statistically different
from each other, but site 1 is different from all of the rest.
In this case, the concentration reported for the seven
indistinguishable sites would be the average of seven means.
More than one SNK Analysis was performed for every parameter
except o-xylene (OXYL). This was done as an effort to include as
many sites as possible. As can be seen for benzene (BENZ), the
largest possible data set, with 69 sets of simultaneous samples,
only included four sites. In order to add a fifth site to the
analysis, only 32 samples sets were available, while only 15 sets
could be used to analyze all eight sites. There are obvious
advantages and disadvantages to each approach.
It is also clear for Table V-2-6 that Analysis 7 is the
exception rather than the rule, presenting a single, obvious
partitioning of the data into mutually exclusive groupings.
Analysis 9 for m-, p-xylene (MPXY) is more typical, with three
overlapping groupings.
When several analyses for the same compound are made the
results should be used as follows: The comparison using the
largest number of sites should be used as a first cut, in order
to determine the breakdown of differences. The second
comparison, with the greater sample size and fewer sites, should
be used to make further comparisons, but only among those groups
which were compared. For example, analysis 1 indicates that site
6 was significantly different from sites 5 and 9 and site 1.
However, site 6 was indistinguishable from sites B, 8, and 3.
Using analysis 2 shows that site 6 was indeed different from
sites B, 8 and 3, but was indistinguishable from site A.
Analysis 3 shows that site B is different from sites 3 and 8.
2-9
-------
Table V-2-4
NUMBER OF OBSERVATIONS IN ADJUSTED DATABASE
Compound
Organization
NJIT
CSI
NYSDEC
Site
A
B
C
D
3
6
8
1
1
2
2
4
4
5
5
7
7
9
9
Method
.
-
-
-
_
-
-
SOR
ATD
SOR
ATD
SOR
ATD
SOR
ATD
SOR
ATD
SOR
ATD
BENZ
76
110
36
42
441
443
501
50
41
39
53
41
54
HEXA
74
110
36
42
445
447
504
MPXY
74
108
36
42
49
49
41
39
41
53
41
39
53
OXYL
434
438
488
49
41
39
53
41
53
T11E
76
108
37
42
439
434
485
50
41
39
53
41
54
TECH
432
432
484
49
41
39
53
41
53
TOLU
76
109
36
42
440
443
496
50
49
41
39
42
53
41
40
53
2-10
-------
Table V-2-5
SUBGROUPS OF ADJUSTED DATA ANALYZED
BY ANALYSIS OF VARIANCE
Analysis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Compound
BENZ
BENZ
BENZ
T11E
T11E
T11E
TOLU
TOLU
MPXY
MPXY
OXYL
HEXA
HEXA
TECH
Sites
3,
3,
3,
3,
3,
3,
3,
3,
A,
3,
3,
3,
3,
6,
6,
6,
6,
6,
6,
6,
6,
B,
B,
6,
6,
6,
6,
8,
8,
8,
8,
8,
8,
8,
8,
c,
8,
8,
8,
8,
A,
A,
A,
A,
A,
D,
1,
A,
1,
B,
B
B
B,
B
B
B,
B,
1,
1,
5,
B
B
5,
1, 5, 9
1, 5, 9
1, 5, 9
1, 5, 9
5, 9, 2, 4, 7
5, 9
9
9
Organizations
CSI, NJIT, NYSDEC
CSI, NJIT
CSI,
CSI,
CSI,
CSI,
CSI,
CSI,
NJIT
NJIT
NJIT, NYSDEC
NJIT
NJIT
NJIT, NYSDEC
NJIT, NYSDEC
, NYSDEC
NJIT, NYSDEC
CSI,
CSI,
CSI,
CSI,
NYSDEC
NJIT
NJIT
NYSDEC
2-11
-------
Table V-2-6
SUMMARY OF STUDENT-NEWMAN-KEULS TEST RESULTS
FOR FOURTEEN ANALYSES INVOLVING SELECTED SITES
FROM TWO OR MORE ORGANIZATIONS
Analysis Sample
No. Compound Size Site*
6 A B 3 8 5 9 1
1 BENZ 15
6 A B 8 3
BENZ 32
6 3 8 B
BENZ 69 —
A 8 B 3 5 6 9 1
T11E 13
A 8 B 3 6
5 T11E 31
8 B 3 6
6 T11E 63 . —
A B 6 5 3 8 9 1
7 TOLU 21 -
3 6 5 8 B 9 1
8 TOLU 46 -
2-12
-------
Table V-2-6 Cont'd
Analysis Sample
No. Compound Size Site*
5B7CA9D421
9 MPXY 24
B 5 9 1
10 MPXY 73
653891
11 OXYL 17
A B 3 6 8
12 HEXA 33
B 3 6 8
13 HEXA 70
685391
14 TECH 17 —
* Sites are arranged from left to right in descending order based on mean concentration (site
means on the log scale are given in Appendix A). Sites underlined are not significantly
different at the 0.05 level.
2-13
-------
2.5 ANALYZING DAILY CONCENTRATIONS AT SITES 3, 6 AND 8
2.5.1 Choice of Time Unit
Let Xy denote the concentration of a given compound at
site i (i=l,2,...,1) on day j (j=l,2, . .. , J) . Let Y;j = In(Xjj) .
The underlying model structure for the analysis, using a day as
the time unit of analysis, is:
Yy = M + fi + *j +
-------
out in the same manner, and results in three sums of squares and
associated mean squares, as shown below:
Source of Degrees of Sums of Mean
Variation Freedom Squares Squares
Days J-l SSD MSD=SSD/(J-1)
Sites 1-1 SSS MSS=SSS/ (1-1)
Error (J-l) (1-1) SSE MSE=SSE/ (J-l) (1-1)
Total IJ-l
Expected values of the pertinent mean squares under the two
scenarios are different, however, and are as follows:
Expected Value of Mean Squares for:
Source Scenario 1 (days random) Scenario 2 (days fixed)
Sites V. + V<4 + J0(*) Ve + J
Error Ve + v|4 Ve + Qt6)
In the above, the V, and VJ4 represent the error and interaction
variance components, respectively.
The test for site differences is formed by computing
F = MSS/MSE; as can be seen by the above table, this will be a
valid test under scenario 1 since the numerator and denominator
mean squares have the same expectation when the null hypothesis
is true. If days are considered fixed, however, the test will be
valid only if there are no site-by-day interaction effects (i.e.,
j(Jd) =0). If such interactions are present but are ignored in
the analysis, then the denominator mean square will be an
overestimate of the V€ and hence real differences will be less
likely to be detected than they should be.
In contrast with the other two organizations, the three CSI
sites furnished concentration data essentially on a daily basis.
While some argument can be made that the days for the other
organizations are representative of some larger set of days, the
use of all days for the CSI sites implies that days should be
considered a fixed component. Thus it is important to consider
whether it is reasonable to assume that there are no interaction
effects, and if not, whether there is another time unit (e.g., a
week) where the assumption of no site-by-time interaction might
be more tenable. The model indicates, in the absence of
measurement errors and interaction effects, that one site's set
of LN( concentrations) differs from those of another site by a
2-15
-------
constant—i.e., that one site has daily concentrations that
consistently differ from those of another by a constant
percentage. Assuming that measurement errors of different days'
samples are independent, then, in the absence of interactions,
the differences in LN(concentrations) for two sites should not
exhibit autocorrelations. Conversely, if autocorrelations in
such differences do exist, then interactions are indicated. (It
should be pointed out that lack of autocorrelations does not
preclude the existence of interactions, since the interaction
effects may not occur in a systematic temporal fashion.)
Autocorrelations of the differences in daily
LN(concentrations) for each pair of CSI sites are shown in
Table V-2-7. The correlations are given for lags 1 through 9,
where the lag-k correlation is defined as the correlation between
Dj and D:+k, where Oj denotes the difference in LN(concentrations)
for a given pair of sites on day j. Since sample sizes are large
for these series, even relatively small correlations are
statistically significant. Correlations significant at the 0.01
level (generally any correlation above about 0.13) are shown in
boldface type in Table V-2-7.
With the exception of carbon tetrachloride, all compounds
showed significant lag-1 correlations for at least one pair of
sites. The presence of a weekly cycle was also indicated for a
number of cases as evidenced by the significant correlations at
lags 6, 7, and 8. Toluene exhibited persistently high lag
correlations for differences involving site 8; this indicates
that these series of differences are nonstationary. A similar
pattern occurred for 1,1,1-trichloroethane for site 3 vs. site 8.
The results suggest that there are some compounds for which
the assumed lack of interaction is not a reasonable assumption.
Weekly concentrations were thus computed by averaging
concentrations over days; at least 5 days within a week were
required to form the weekly value. Autocorrelations among the
weekly data (log scale) were then calculated. These are given in
Table V-2-8. Many of the weekly correlations are larger than those
for the daily data, though because of smaller sample sizes
substantially fewer are statistically significant at the 0.01
level. Very few large autocorrelations were found after four
lags. The first four lag correlations for the site 6 vs. 8
difference were found significant for benzene, hexane, o-xylene,
and toluene.
2-16
-------
Table V-2-7
AUTOCORRELATIONS OF SITE DIFFERENCES
IN LN(DAILY CONCENTRATIONS)
Autocorrelations x 1 00, for
Compound
1,1,1-trichloroethane
Benzene
Carbon Tetrachloride
Hexane
o-Xylene
Styrene
Toluene
Tetrachloroethene
Trichloroethene
Sites
3,6
3,8
6,8
3,6
3,8
6,8
3,6
3,8
6,8
3,6
3,8
6,8
3,6
3,8
6,8
3,6
3,8
6,8
3,6
3,8
6,8
3,6
3,8
6,8
3,6
3,8
6,8
1
03
19
09
23
22
21
-03
09
10
15
15
20
19
19
32
17
21
25
22
28
32
23
14
19
15
20
09
2
06
21
-11
16
27
14
10
08
-07
03
13
09
10
14
16
00
11
15
11
20
23
10
02
04
12
11
09
3
-05
24
-02
10
10
13
-09
11
03
04
09
13
17
17
05
10
10
08
12
20
15
11
02
11
15
14
02
4
03
11
-04
02
10
13
-04
06
11
04
09
16
08
10
16
04
08
13
10
20
19
02
08
04
03
10
07
5
07
20
08
-02
13
14
-01
10
09
-05
06
11
00
08
14
10
09
07
03
23
23
05
08
11
-06
06
-09
6
-03
11
09
03
14
20
12
16
02
-03
08
16
09
22
28
18
05
05
09
23
29
10
08
15
07
09
02
Lag:
7
09
14
05
-09
14
14
07
20
08
-06
03
15
01
11
27
04
02
22
03
19
30
00
20
02
-06
14
-01
8
09
15
10
-05
10
16
07
10
06
02
16
11
03
05
14
06
09
10
09
23
21
05
00
-01
07
04
-02
9
03
18
10
02
04
02
08
03
13
00
12
06
13
19
10
-05
11
06
06
17
20
08
01
04
-02
02
-07
Autocorrelations shown in bold are statistically significant (0.01 level).
2-17
-------
Table V-2-8
AUTOCORRELATIONS OF SITE DIFFERENCES
IN LN(WEEKLY CONCENTRATIONS)
Autocorrelations x 100, for Lag:
Compound
1,1,1-trichloroethane
Benzene
Carbon Tetrachloride
Hexane
o-Xylene
Styrene
Toluene
Tetrachloroethene
Trichloroethene
Sites
3,6
3,8
6,8
3,6
3,8
6,8
3,6
3,8
6,8
3,6
3,8
6,8
3,6
3,8
6,8
3,6
3,8
6,8
3,6
3,8
6,8
3,6
3,8
6,8
3,6
3,8
6,8
1
23
24
-12
08
35
47
22
25
44
06
14
58
28
39
45
06
19
21
20
36
56
06
30
12
34
49
09
2
14
36
11
24
43
44
26
29
35
-02
29
48
16
30
53
-08
-19
27
26
51
63
32
28
04
35
49
39
3
09
17
-01
15
47
52
-12
-07
41
-04
18
43
11
10
42
03
-40
22
16
32
54
28
17
12
16
38
23
4
-01
40
09
34
38
40
14
04
31
31
26
50
24
28
49
23
-11
37
15
44
61
04
00
03
12
21
18
5
-08
43
10
-21
05
-06
21
33
22
-04
03
34
-03
18
23
04
16
17
-05
32
38
41
17
05
24
20
15
6
-03
40
04
08
21
27
17
30
26
08
10
45
17
12
28
06
06
07
-05
32
45
-14
-05
-18
-11
08
-02
7
03
34
-15
24
42
22
13
24
33
05
07
33
15
10
10
30
04
30
26
40
29
28
18
15
07
03
35
8
17
22
21
-04
10
-12
-20
-23
-14
-01
-05
24
04
26
09
01
03
24
01
22
28
-04
23
-13
20
12
-13
9
14
43
14
-10
-02
16
-22
-05
05
-07
-31
21
12
03
22
-04
16
32
13
22
23
27
24
17
03
34
00
Autocorrelations shown in bold are statistically significant (0.01 level).
2-18
-------
These results, together with those for the daily data, suggest
that a still longer averaging time (e.g., 4-wk average) might be
appropriate for some compounds and sites. Nonstationarity for
two of the toluene series was again indicated. This is
illustrated in Figure V-2-1, which gives a plot of the weekly
toluene
data for the site 6 minus site 8 differences. Similar, though
less pronounced, patterns were evident for several of the other
compounds. One implication of such nonstationarity is that
general conclusions regarding site differences are not possible,
since the differences depend on the particular time frame that is
examined (e.g., one site's levels may be higher in one season and
lower in another relative to another site). Also, if long-term
site means (or differences between them) are to be reported, then
the preferred time frame consists of whole years (i.e., 52 or 104
weeks rather than 75 weeks, say).
2.5.2 Comparisons of sites 3 and 6
The daily results for sites 3 and 6 (Table V-2-7) generally
showed significant autocorrelations only for the first few lags.
The weekly results for these same sites (Table V-2-8) showed no
significant lag correlations at the 0.01 level of significance.
For this pair of sites, the use of a weekly unit of analysis
would thus appear to be preferable . to the daily unit.
Consequently, for these two sites, paired t-tests were calculated
using the weekly unit. In the case of only two sites, this
procedure is equivalent to the two-way ANOVA and negates the need
for using the SNK procedure for comparing specific sites. The
results are summarized in Table V-2-9. Benzene, o-xylene, styrene,
toluene, and tetrachloroethene exhibited significantly higher
levels at site 6 than at site 3, while carbon tetrachloride
levels were significantly smaller for site 6 relative to site 3.
2-19
-------
PLOT OF WEEKLY DIFFERENCES (SITE6-SITE8) AS A FUNCTION OF TIME - TOLUENE
0.3
0.2
0.1
'c
o
7^
«J 0.0
V-i
u
C
0)
u
o -0.1
u
r-l
£ -0.2
to 3-
•* ,5
0 H
C -0.3
u
C
0)
M — o 4
0)
•H
O
-0.5
-0.6
-0.7
A
A
A A
A
A
A A A
A A A
A A A
A
A A A A
A A A A A TJ
A «'
A A C
A A A A n>
A <
A ^
A A 1
A A A
A A A AA
* A
AA A AA
A A A
A A
H A A A
A
h
A
h
0 10 20 30 40 50 60 70 80
WEEK
-------
Table V-2-9
TEST OF DIFFERENCES JN AVERAGE
LN(WEEKLY CONCENTRATIONS) FOR TWO CSI SITES
Compound
1,1,1-trichloroethane
Benzene
Carbon Tetrachloride
Hexane
o-Xylene
Styrene
Toluene
Tetrachloroethene
Trichloroethene
No. of
Weeks
56
54
56
57
56
44
56
55
57
Difference
(Site 6-Site 3)
-0.0528
0.1984**
-0.1567"
-0.0614
0.2152"
0.2031"
0.1070"
0.9044"
-0.0001
Differences are statistically significant (0.01 level).
2-21
-------
2.6 CONCLUSIONS
The statistical analysis presented here is necessarily quite
complex. It is an attempt to determine, after the fact, whether
site-to-site differences in VOC concentrations measured by
different organizations are statistically significant; the design
of the original study was not entirely suitable for that
analysis. The PEI reference samples which are used as the basis
for the site-to-site comparisons, were intended to be used for a
different purpose. They were to be used as a quality assurance
check on the assumption that the two-tube Tenax sampling scheme
would prove satisfactory and that the various organizations could
produce valid data. The reference samples did satisfy this
purpose and show that the two-tube adsorbent methods would be
effective and that the individual organizations could and did
produce valid data. However, neither the PEI reference samples
nor the specific collocated sampling events (so-called shootouts)
provided enough data to confirm, in every case, whether the
concentrations of VOCs produced by the different organizations
are directly comparable. This was because of the great
variability between organizations discovered during the
shootouts. It was not feasible to perform additional shootouts,
making it necessary to use the PEI data for this purpose as well.
The results of the analysis do indicate that, for certain
VOCs, the differences between certain sites were significant
during the times that the reference samples from PEI were
available. However, since those data sets were limited when
compared to the overall data set, all possible comparisons
between sites could not be performed.
When the analysis showed no statistically significant
differences, the results should be considered the same. Any risk
assessments or other uses for the data should use the actual
measured annual means but should consider the effects from those
VOCs to be indistinguishable across those monitoring sites.
When the analysis shows a significant difference between
sites for the limited data set for which the reference samples
are available, the overall results may be considered different.
Any risk assessment or other data use may utilize the actual
annual mean concentrations.
An additional benefit of this statistical analysis is the
confirmation that annual averaging is the most appropriate means
for utilizing the VOC monitoring data from this project. As
shown in the analysis of the daily PEI results, the
autocorrelations between consecutive samples could strongly
affect any short-term averaging of the data. Thus, the presence
of more frequent data, while serving to improve the precision of
the calculated means does not necessarily enhance the ability to
2-22
-------
look at shorter averaging times. The conclusion, then, is that
the use of annual averages for all of the VOC results, for all
organizations, is appropriate.
2-23
-------
ANOVA Results
2-24
-------
SAS
COMP=BENZ
Analysis of Variance Procedure
Student-Newman-Keuls test for variable: LCONC
NOTE: This test controls the type I experimentwise error rate under the
complete null hypothesis but not under partial null hypotheses.
Alpha= 0.05 df= 98 MSE= 0.038236
Number of Means 2345
Critical Range 0.1416934 0.1699247 0.1866214 0.1984399
Number of Means 678
Critical Range 0.207552 0.2149506 0.2211598
Means with the same letter are not significantly different.
SNK Grouping Mean N SITE
A
A
A
A
A
A
A
A
A
B
B
B
C
0
0
0
0
0
0
-0
-0
.3416
.2568
.2328
.2138
.1777
.0054
.0077
.3412
15 6
15 A
15 B
15 3
15 8
15 5
15 9
15 1
-------
SAS
COMP=BENZ
Analysis of Variance Procedure
Student-Newman-Keuls test for variable: LCONC
NOTE: This test controls the type I experimentwise error rate under the
complete null hypothesis but not under partial null hypotheses.
Alpha= 0.05 df= 124 MSE= 0.032802
Number of Means 2345
Critical Range 0.0896179 0.1074091 0.117917 0.1253449
Means with the same letter are not significantly different.
SNK Grouping Mean N SITE
B
B
B
B
B
B
B
A
A
A
0.
0.
0.
0.
0.
2981
2172
1698
1552
1366
32 6
32 A
32 B
32 8
32 3
2-26
-------
SAS
COMP=BENZ
Analysis of Variance Procedure
Student-Newman-Keuls test for variable: LCONC
NOTE: This test controls the type I experimentwise error rate under the
complete null hypothesis but not under partial null hypotheses.
Alpha= 0.05 df= 204 MSE= 0.043343
Number of Means 234
Critical Range 0.0698846 0.0836842 0.0918135
Means with the same letter are not significantly different.
SNK Grouping Mean N SITE
A 0.2562 69 6
B 0.1678 69 3
B
B 0.1372 69 8
C 0.0070 69 B
2-27
-------
SAS
COMP=T11E
Analysis of Variance Procedure
Student-Newman-Keuls test for variable: LCONC
NOTE: This test controls the type I experimentwise error rate under the
complete null hypothesis but not under partial null hypotheses.
Alpha= 0.05 df= 84 MSE= 0.056285
Number of Means 2345
Critical Range 0.1850504 0.2220264 0.2439191 0.2594301
Number of Means 678
Critical Range 0.2713976 0.2811157 0.2892762
Means with the same letter are not significantly different.
SNK Grouping Mean N SITE
A -0.2370 13 A
A
B A -0.3231 13 8
B A
B AC -0.4319 13 B
B C
B C -0.5217 13 3
B C
B C -0.5253 13 5
B C
B C -0.5619 13 6
B C
B C -0.5711 13 9
C
C -0.6915 13 1
2-28
-------
SAS
COMP=T11E
Analysis of Variance Procedure
Student-Newman-Keuls test for variable: LCONC
NOTE: This test controls the type I experimentwise error rate under the
complete null hypothesis but not under partial null hypotheses.
Alpha= 0.05 df= 120 MSE= 0.054839
Number of Means 2345
Critical Range 0.117768 0.1411583 0.1549755 0.1647444
Means with the same letter are not significantly different.
SNK Grouping Mean N SITE
A -0.2651 31 A
A
B A -0.3002 31 8
B
B -0.4130 31 B
C -0.5326 31 3
C
C -0.5718 31 6
2-29
-------
SAS
COMP=T11E -
Analysis of Variance Procedure
Student-Newman-Keuls test for variable: LCONC
NOTE: This test controls the type I experimentwise error rate under the
complete null hypothesis but not under partial null hypotheses.
Alpha= 0.05 df= 186 MSE= 0.050702
Number of Means 234
Critical Range 0.079148 0.0947893 0.1040068
Means with the same letter are not significantly different.
SNK Grouping Mean N SITE
A -0.3908 63 8
B -0.4746 63 B
B
C B -0.4989 63 3
C
C -0.5724 63 6
2-30
-------
SAS
- COMP=TOLU
Analysis of Variance Procedure
Student-Newman-Keuls test for variable: LCONC
NOTE: This test controls the type I experimentwise- error rate under the
complete null hypothesis but not under partial null hypotheses.
Alpha= 0.05 df= 140 MSE= 0.079053
Number of Means 2345
Critical Range 0.1715467 0.2055495 0.2256215 0.2398003
Number of Means 678
Critical Range 0.250727 0.2595813 0.2670143
Means with the same letter are not significantly different.
SNK Grouping Mean N SITE
A 1.1374 21 A
A
A 1.1238 21 B
A
A 1.0061 21 6
A
A 0.9678 21 5
A
A 0.9279 21 3
A
A 0.9192 21 8
A
A 0.9126 21 9
B 0.4798 21 1
2-31
-------
SAS
COMP=TOLU
Analysis of Variance Procedure
Student-Newman-Keuls test for variable: LCONC
NOTE: This test controls the type I experimentwise error rate under the
complete null hypothesis but not under partial null hypotheses.
Alpha= 0.05 df= 270 MSE= 0.092197
Number of Means 2 3 4 5 67
Critical Range 0.1246497 0.1492132 0.163671 0.1738793 0.1817284 0.1880855
Means with the same letter are not significantly different.
SNK Grouping Mean N SITE
A 0.9941 46 3
A
A 0.9870 46 6
A
A 0.9504 46 5
A
A 0.9470 46 8
A
A 0.8499 46 B
A
A 0.8173 46 9
B 0.5059 46 1
2-32
-------
SAS
COMP=MPXY
Analysis of Variance Procedure
Student-Newman-Keuls test for variable: LCONC
NOTE: This test controls the type I experimentwise error rate under the
complete null hypothesis but not under partial null hypotheses.
Alpha= 0.05 df» 207 MSE= 0.140176
Number of Means 23456
Critical Range 0.2130788 0.2551486 0.2799308 0.2974369 0.3109054
Number of Means 7 8 9 10
Critical Range 0.321817 0.3309595 0.3388223 0.3457033
Means with the same letter are not significantly different.
SNK Grouping
' Mean
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
A
A
A
A
A
A
A
A
A
A
A
C
C
C
C
C
C
C
0
0
0
0
-0
-0
-0
-0
-0
-0
N SITE
.157
.021
.012
.001
.012
.051
.200
.200
.281
.392
24
24
24
24
24
24
24
24
24
24
5
B
7
C
A
9
0
4
2
1
2-33
-------
SAS
- COMP=MPXY
Analysis of Variance Procedure
Student-Newman-Keuls test for variable: LCONC
NOTE: This test controls the type I experimentwise error rate under the
complete null hypothesis but not under partial null hypotheses.
Alpha= 0.05 df= 216 MSE= 0.17216
Number of Means 234
Critical Range 0.1353651 0.1620823 0.1778182
Means with the same letter are not significantly different.
SNK Grouping Mean N SITE
A 0.0122 73 B
A
A -0.0992 73 5
B -0.3072 73 9
C -0.5752 73 1
2-34
-------
SAS
COMP=OXYL --
Analysis of Variance Procedure
Student-Newman-Keuls test for variable: LCONC
NOTE: This test controls the type I experimentwise error rate under the
complete null hypothesis but not under partial null hypotheses.
Alpha= 0.05 df= 80 MSE= 0.042746
Number of Means 23456
Critical Range 0.1411255 0.1693531 0.1860725 0.1979139 0.2070663
Means with the same letter are not significantly different.
SNK Grouping Mean N SITE
A -0.4792 17 6
A
B A -0.6180 17 5
B
B -0.6770 17 3
B
B -0.6880 17 8
B
B -0.7686 17 9
C -1.0965 17 1
2-35
-------
SAS
COMP=HEXA
Analysis of Variance Procedure
Student-Newman-Keuls test for variable: LCONC
NOTE: This test controls the type I experimentwise error rate under the
complete null hypothesis but not under partial null hypotheses.
Alpha= 0.05 df= 128 MSE= 0.072304
Number of Means 2345
Critical Range 0.1309819 0.1569737 0.1723229 0.1831709
Means with the same letter are not significantly different.
SNK Grouping Mean N SITE
A 0.0047 33 A
A
A -0.0953 33 B
B -0.2730 33 3
B
B -0.3039 33 6
B
B -0.3317 33 8
2-36
-------
SAS
- - COMP=HEXA
Analysis of Variance Procedure
Student-Newman-Keuls test for variable: LCONC
NOTE: This test controls the type I experimentwise error rate under the
complete null hypothesis but not under partial null hypotheses.
Alpha= 0.05 df= 207 MSE= 0.060907
Number of Means 234
Critical Range 0.0822421 0.0984798 0.108045
Means with the same letter are not significantly different.
SNK Grouping Mean N SITE
A -0.2541 70 B
A
A -0.2892 70 3
A
A -0.3245 70 6
A
A -0.3319 70 8
2-37
-------
SAS
COMP=TECH
Analysis of Variance Procedure
Student-Newman-Keuls test for variable: LCONC
NOTE: This test controls the type I experimentwise error rate under the
complete null hypothesis but not under partial null hypotheses.
Alpha= 0.05 df= 80 MSE= 0.075218
Number of Means 23456
Critical Range 0.1872045 0.2246486 0.246827 0.2625349 0.2746756
Means with the same letter are not significantly different.
SNK Grouping Mean N SITE
A -0.6319 17 6
B -1.2228 17 8
B
C B -1.3039 17 5
C B
C B -1.3209 17 3
C B
C B -1.3882 17 9
C
C -1.5179 17 1
2-38
-------
APPENDIX A
BRIEF COMPARISON OF THE SI/NJ UATAP AND THE
1988 AND 1989 DATMP STUDIES
(from Section 3.5.1 of Volume III, Part A,
of the SI/NJ UATAP report).
1. SI/NJ UATAP (Staten island/New Jersey Urban Air
Toxics Assessment Project)
Sites; 13 ambient air sites for VOCs, 5 sites for metals and
B[a]P, and 2 sites for formaldehyde for the outdoor air portion
of the study. (Piscataway served as an upwind site for VOCs and
formaldehyde; and Highland Park, as an upwind site for metals and
B[a]P. Upwind refers to the W to SW wind direction, the
predominant wind direction for the project area.) Four indoor
and two outdoor sites were used for the 8-month indoor air
portion of the study.
Chemicals; 41 chemicals—23 VOCs31, 16 metals,
benzo[a]pyrene, and formaldehyde.
Sampling frequency: 24-h samples every 6th day; October '87
through September '89. (Only data for the period 10/88
through 9/89 have been included in this comparison). CSI
sampled daily during certain quarters. Annual averages are
arithmetic averages of all samples; for CSI sites, many more
samples were collected during some quarters than others.
Collection; For VOCs, NJIT and CSI used Tenax as the sole
adsorbent, and NYSDEC used a series of Tenax/Amersorb/carbon
in a single tube as the adsorbent. A combination of
canisters and periodic simultaneous sampling by the three
organizations was used as an indication of accuracy of the
sorbent methods, and as a basis for integration of the
inter-organizational set of data. For particulate matter
and B[a])P, high-volume samplers were used. For
formaldehyde, 2,4-dinitrophenylhydrazine-coated (DNPH-
coated) silica cartridges were used.
Analysis: For VOCs, NYSDEC and CSI used GC/MS; NJIT used -
GC/FID/ECD with confirmation by GC/MS. For metals, all
organizations used atomic absorption spectrophotometry.
In this tally, m- and g-xylene were counted separately,
although they were not distinguished by the analytical methods
used.
1-A-l
-------
Organization; Sites were run by three organizations—6 by
NYSDEC, 5 by NJDEP, 3 by CSI; each organization had a
different lab analyze samples collected at its sites.
Site selection; Residential neighborhood complaints,
availability, accessibility, security, absence of known
point sources nearby, geographic distribution, proximity to
breathing zone, in general conforming to the USEPA air
monitoring siting requirements.
Objectives; Characterization of ambient air quality, risk
assessment, and source identification. Further detail may
be found in the nine objectives stated in Volume II of the
SI/NJ UATAP project report.
2. 1988 and 1989 UATMP (Urban Air Toxics Monitoring
Program, USEPA)
Sites; 19 sites operated in the 1988 study (9/24/87 through
10/6/88). 14 sites operated in 1989 study (1/22/89 through
1/17/90)—6 sites were the same as in the 1988 study, and 8
sites were new. See tables for site locations.
Chemicals; 38 gaseous organic compounds; and metals, B[a]P,
and carbonyl compounds.
Sampling frequency; 24-h samples every 12 days.
Annual arithmetic averages are listed in the tables.
Collection; For VOCs, stainless steel canisters. For
particulate matter, high-volume filters. For carbonyl
compounds, DNPH-coated silica cartridges.
Analysis; GC/MD, with 3 detectors for identification (BCD,
PID, FID)32; FID for most quantitation, ECD for most
halogenated compounds. GC/MS for identification
confirmation of GC/MD results.
Organization; A contractor set up the sites. State or local
personnel collected samples and sent them to the EPA
contractor. The contractor analyzed all samples.
Site selection; OAQPS (Office of Air Quality Planning and
Standards) guidelines, with site selection by Regional EPA
offices and State offices together.
32 GC = gas chromatograph. MD = multi-detector. ECD = electron
capture detector. PID = photoionization detector.
FID = flame ionization detector. MS = mass spectrometer.
l-A-2
-------
Objective; Screening to help state and local agencies
determine if an air toxics problem existed, assess air
quality, provide focus for follow-up studies and risk
reduction activities.
l-A-3
-------
APPENDIX B
DESCRIPTION OF THE 1988 UATMP SITES
From "Calculation of Cancer Risks from 1988 UATMP Data, " (Lahre,
1990).
Atlanta, Georgia (AT GA)
-- site close to downtown
-- 1/2 mile or so from two freeways
-- adjacent to junk yard that occasionally burns insulated wire and auto
engines, etc.
-- adjacent to slightly used parking lot
-- mixed commercial and residential E and N of site, commercial in other
directions
Birmingham, Alabama (BH AL)
-- residential site location, across from police and fire department
-- closest major point source is U. S. Steel, within 1/2 mile
-- Koppers coking facility about 4 or 5 miles distant from site
Baton Rouge, Louisiana (BR LA)
-- located near State Capital
-- also site of Louisiana hourly monitoring station
-- several miles south of major petrochemical complex
... 2 synthetic organic chemical manufacturers
... 1 petroleum refinery
... 1 power plant
-- smaller petroleum refinery to NW
-- petroleum product tank farm to west of site
-- major highway within 1/2 mile
Burlington, Vermont (BR VT)
-- site generally commercial in nature
-- site within Burlington city limits
-- site located in municipal parking lot
-- 2 service stations across street
-- several parking decks within one block
Chicago Illinois (CH IL)
-- in Southeast Chicago
-- mixed residential/industrial area with many traffic arterials
-- major steel mill/coke oven within 1/2 mile
-- hazardous waste Incinerator nearby
-- garbage landfill within 1/2 mile
-- site location changed during '88 from Carver to Washington High School,
but still in same general area of Chicago
l-B-l
-------
Cleveland, Ohio (CL OH)
-- site located within industrial/warehouse area
-- major steel mill/coke oven within 1 mile
-- major traffic arterials nearby
-- some residential areas nearby
Dearborn, Michigan (DB MI)
-- Dearborn is a western suburb of Detroit
-- site located at school
-- primary land use is industrial, changing within 2 miles to residential
to the north, east and south
-- Ford Motor assembly, steel and glass plant located 1/2 mile W of
monitor
-- the Asphalt Products Co., a slag processing company and Detroit Lime
are within 1 mile SW
-- significant railroad activity adjacent to site
Dallas, Texas (DL TX)
-- commercial downtown area, not residential or industrial
-- no major point sources within 1 mile
-- mostly parking lots, minor arterials, commercial buildings immediately
adjacent to site
Detroit, Michigan (DT MI)
-- site located in mixed residential, industrial and commercial area
-- dominant Influence is area sources
-- site located immediately adjacent to freeway
-- General Motors plant located 3/4 mile N
-- Detroit incinerator located 1/2 mile SW
-- freeway immediately adjacent to site
Houston, Texas (HI TX)
-- site located at a school
-- near heavily industrialized East Houston
-- several miles from Houston Ship Channel
-- heavily automotive traffic In area
Hammond, Indiana (HA IN)
— site located In Industrial park
-- major petroleum tank farm within 1/2 nlle to U
-- little immediate traffic, no residences nearby
-- major freeway 1/2 to 3/4 miles away
l-B-2
-------
Jacksonville, Florida (JA FL)
-- site immediately surrounded by residences and city park
-- nearest major traffic artery about 500 feet away
-- Kraft pulp mill within 1 mile to NE
-- oil storage tanks and phosphate storage/loading facilities within 1 mile
to N
Louisville, Kentucky (L2 KY)
-- downtown site
-- site at 3-way street intersection below elevated freeway
-- service station 1 block W
-- multi-'story parking garage 1 block SW
Lansing, Michigan (LA NI)
-- site located at school
-- primary land use is commercial
-• land use to N and E changes in 2 miles to residential
-- land use in SE, SW and NW is a mixture of residential, commercial and
industrial
-- dominant immediate influence at site is area sources
-- within 1/2 mile of site is a heat treating company, 3 GM plants, and
a GMC assembly/powerhouse plant
--a wastewater treatment plant and a steel company are within 1 mile
Midland, Michigan (MD HI)
-- primary immediate land use is commercial
-- adjacent major traffic arterial
-- land use changes to residential within 2 miles to N, NE and NW
-- Dow Chemical is largest source nearby, occupying all of land to SW, S
and SE of plant within 1 mile
-- about 1 mile E of site is a metal coating company and a municipal sewage
treatment plant
Miami,Florida (HI FL)
-- residential/commercial site
-- site located on rooftop of several story commercial building
-- site across street from baseball stadium and parking lot
-- residential to NW, N and W
— mixture of warehouses and small shops to NE and S, Including bakery,
auto repair shops
-- jail for criminally Insane to E
-- no major point sources in proximity
l-B-3
-------
Portland, Oregon (P2 OR)
•- site located in Portland's NW industrial area
-- immediate proximity of extensive gasoline storage and transfer
facilities
-- nearby railroad switching yard with diesel engines
-- various other sources in general proximity (involving plastics, paints,
roofing materials, and "tall oil" combustion)
Port Huron, Michigan (PH HI)
-- site in parking lot of National Guard armory
-- primary immediate land use is residential
-- dominant local source influencing site 1s area sources
--land use changes within 1-2 miles to the E, W and NE to industrial
-- small industrial park with 30 small sources about 1/2 mile to E with
at least one formaldehyde emitter
-- hospital is 1 mile E of site
-- Canadian refineries and chemical companies 1.5 miles E of site
Sauget, Illinois (SA IL)
-- heavily industrialized area, in East St. Louis
-- Monsanto chemical plant within 500 meters
-- also in proximity are metal processors and a rubber recycler
l-B-4
-------
APPENDIX C
SUMMARY OF POPULATION ANALYSIS
The following is a summary of the population analysis
prepared by Marian Olsen for the SI/NJ UATAP.
Summary of the Population Analysis for the SI/NJ UATAP
Anecdotal data suggested that many residents had been so-
called Islanders for their entire lifetimes. While it is
difficult to verify this, analysis of the 1980 census data
indicated the following:
With few exceptions, the population in the communities
studied showed a steady increase from 1930 through
projections for 2010. (Union County showed a slight
decrease in 1980 and 1990, with projected increased in 2000
and 2010.)
In 1980, the data showed that owner- and renter-occupied
units occupied 20 or more years by the same person(s) in
Richmond, Union, and Middlesex Counties were, respectively,
17.2%, 21.6%, and 24.2%.
The percentage of employees living in the same county as
they work ranged from 42% for Richmond County to 72% for
Highland Park.
The majority of New Jersey workers in Union and Middlesex
Counties spends 21.1 to 23.4 minutes traveling from home to
work. In Richmond County, the average time spent commuting
one way is 42.9 minutes.
These preliminary data qualitatively suggest that many
residents in the study area spend a good part of their time
either living in the community or working there. However, it was
not possible to determine the percentage of the community that
would match the assumptions used in this report.
References
Nelson A. Rockefeller Institute of Government. (1991) 1991
New York State statistical yearbook, 16th edition, revised and
1-C-l
-------
expanded. Albany, NY: The Nelson A. Rockefeller Institute of
Government.
New Jersey State Data Center. (1991) NJSDC 1990 census
publication, New Jersey population trends, 1790 to 1990.
Trenton, NJ: New Jersey State Data Center, Division of Labor
Market and Demographic Research.
Staten Island Chamber of Commerce. (1991) Statistical guide
Staten Island, New York, 1991 - 1992. Staten Island, NY: Staten
Island Chamber of Commerce.
U. S. Department of Commerce. (1980a) Detailed housing
characteristics, New York, 1980 census of housing, volume 1,
chapter b: detailed housing characteristics, part 34, HC80-1-B34.
Washington, DC: U. S. Deptartment of Commerce, Bureau of Census.
U. S. Department of Commerce. (1980b) Detailed housing
characteristics, New York, 1980 census of housing, volume 1,
chapter b: detailed housing characteristics, part 34, HC80-1-B34.
Washington, DC: U. S. Department of Commerce, Bureau of Census.
U. S. Environmental Protection Agency. (1989) Exposure factors
handbook. Washington, DC: Office of Health and Environmental
Assessment, Office of Research and Development: EPA publication
no. EPA/600/8-89/043.
l-C-2
------- |