Hazard Ranking System Issue Analysis:
Carcinogenic Risk Analysis of the
Air Pathway Target Distance Limit
MITRE
-------
Hazard Ranking System Issue Analysis:
Carcinogenic Risk Analysis of the
Air Pathway Target Distance Limit
Thomas F. Wolfinger
June 1987
MTR-86W140
SPONSOR:
U.S. Environmental Protection Agency
CONTRACT NO.:
EPA-68-01-7054
The MITRE Corporation
Metrek Division
7525 Colshire Drive
McLean, Virginia 22102-3481
-------
"-\
Department Approval:
MITRE Project Approval:
11
-------
ABSTRACT
This report presents an analysis of the target distance limit
employed in the EPA Hazard Ranking System (HRS) air pathway. The
target distance limit is defined as the maximum distance used in
determining the target population in the air pathway.
The report presents estimates of the general level of cancer
risk arising from air emissions from uncontrolled waste sites and
examines the implications of the analysis for the targets category
of the HRS air pathway. The principal conclusions reached in the
analysis are: (1) simple risk analysis techniques can be fruitfully
employed in HRS issue analyses, although their use in assessing
actual site risks based on the data developed during site
inspections is unwarranted, (2) the uncertainty associated with the
results is high and generalizations must be made with caution,
(3) cancer risks to individuals living beyond 4 miles from the
boundary of a site from long-term exposures to air emissions from
most uncontrolled waste sites are probably very low (as are their
risks from chronic exposures to non-carcinogenic contaminants with
safe exposure thresholds), (4) risks from these effects to
individuals living beyond 1/4 from the boundary of a site are also
probably low, (5) risks to individuals residing with 1/4 mile of
site boundaries are difficult to assess using techniques such as are
employed here and are probably higher than indicated in this study,
and (6) particulate emissions from large sites may pose a higher,
potentially unacceptable, cancer risk than is indicated by this
analysis.
The major implications for the HRS air pathway are as follows:
(1) the target distance limit cannot be definitively set based on
the results of this analysis, further analysis of sub-chronic risks
must be undertaken before a revision of the four-mile limit is made,
(2) cancer risk should not be emphasized in the HRS air pathway, and
(3) differential cancer risks between sites can best be reflected in
the toxicity and waste quantity components of the HRS waste
characteristics category.
Suggested Keywords: Superfund, Hazard ranking, Hazardous waste, Air
emissions.
111
-------
ACKNOWLEDGMENT
The author would like to acknowledge the participation of
Michael Dusetzina of the Halted States Environmental Protection
Agency, Office of Air Quality Planning and Standards. Mr. Dusetzina
exercised the Human Exposure Model and produced the data on which
the analysis is based. This study would not have been possible
without his participation and assistance.
IV
-------
TABLE OF CONTENTS
LIST OF ILLUSTRATIONS vii
LIST OF TABLES vii
1.0 INTRODUCTION 1
1.1 Background 1
1.2 Issue Description 3
1.3 Purpose and Organization of Beport 5
2.0 APPROACH 7
2.1 Focus of the Study 7
2.2 Simplifying Assumptions 11
2.3 Measures of Risk 24
2.4 Overview of the EPA Human Exposure Model 27
2.4.1 The Receptor Grid 28
2.4.2 The HEM Dispersion Module 28
2.4.3 The Population Allocation Module 32
2.4.4 The Exposure/Risk Module 34
2.4.5 HEM Output 35
2.5 Site Selection 36
2.6 Analysis Framework 37
3.0 CARCINOGENIC RISK ESTIMATES 41
3.1 MEI Risk 41
3.2 Risk to the Average Exposed Individual 45
3.3 Population Incidence 49
4.0 STUDY LIMITATIONS 53
4.1 limitations in the Emission Rate Assumptions 53
4.2 Limitations Associated with the Potency Factors 55
4.3 Limitations Arising from Simplifying Assumptions 58
4.4 Limitations of the HEM 62
4.5 Limitations in the Site Selection Process 63
4.6 Overall Implications for the Analysis 66
-------
TABLE OF CONTENTS (Concluded)
Page
5.0 CONCLUSIONS OF THE STUDY AND IMPLICATIONS FOR 71
THE HRS AIR PATHWAY TARGET DISTANCE CATEGORY
5.1 Risks from Air Contaminant Releases from Waste Sites 71
5.2 Implications for the HRS Air Pathway Targets Category 75
6.0 REFERENCES 79
APPENDIX A - SELECTED INFORMATION ON EMISSION RATES 83
vi
-------
LIST OF ILLUSTRATIONS
Figure Number
2-1
2-2
Basic Human Exposure Model (HEM) Structure
HEM Receptor Grid
Page
29
30
Table Number
2-1
2-2
3-1
3-2
3-3
3-4
4-1
4-2
LIST OF TABLES
Selected Information on Waste Site
Emission Rates
Carcinogenic Potency Factors
Summary Statistics of Maximally Exposed
Individual (MEI) Risk
Summary Statistics of Risk to Average
Exposed Individuals (AEI Risk) Living
Within Specified Distances
Summary Statistics of Risk to Average
Exposed Individuals (AEI Risk) Living
Within Specified Distance Ranges
Summary Statistics of Population Incidence
Comparison of Geographic Distribution
of Sites Selected for Analysis
Summary of Study Limitations
Page
14
17
43
46
48
50
65
67
Vll
-------
1.0 INTRODUCTION
1.1 Background
The Comprehensive Environmental Response, Compensation, and
Liability Act of 1980 (CERCLA) (PL 96-510) requires the President to
identify national priorities for remedial action among releases or
threatened releases of hazardous substances. These releases are to
be identified based on criteria promulgated in the National
Contingency Plan (NCP). On July 16, 1982, EPA promulgated the
Hazard Ranking System (HRS) as Appendix A to the NCP (40 CFR 300;
47 FR 31180). The HRS comprises the criteria required under CERCLA
and is used by EPA to estimate the relative potential hazard posed
by releases or threatened releases of hazardous substances.
The HRS is a means for applying uniform technical judgment
regarding the potential hazards presented by a release relative to
other releases. The HRS is used in identifying releases as national
priorities for further investigation and possible remedial action by
assigning numerical values (according to prescribed guidelines) to
factors that characterize the potential of any given release to
cause harm. The values are manipulated mathematically to yield a
single score that is designed to indicate the potential hazard posed
by each release relative to other releases. This score is one of
the criteria used by EPA in determining whether the release should
be placed on the National Priorities List (NPL).
-------
During the original NCP rulemaking process and the subsequent
application of the HRS to specific releases, a number of technical
issues have been raised regarding the HRS. These issues concern the
desire for modifications to the HRS to further improve its
capability to estimate the relative potential hazard of releases.
The Issues Include:
Review of other existing ranking systems suitable for
ranking hazardous waste sites for the NFL.
Feasibility of considering ground water flow direction and
distance, as well as defining "aquifer of concern," In
determining potentially affected targets.
Development of a human food chain exposure evaluation
methodology.
Development of a potential for air release factor category
In the HRS air pathway.
Review of the adequacy of the target distance specified in
the air pathway.
Feasibility of considering the accumulation of hazardous
substances in Indoor environments.
Feasibility of developing factors to account for
environmental attenuation of hazardous substances in ground
and surface water.
Feasibility of developing a more discriminating toxicity
factor.
Refinement of the definition of "significance" as it relates
to observed releases.
Suitability of the current HRS default value for an unknown
waste quantity.
Feasibility of determining and using hazardous substance
concentration data.
-------
Feasibility of evaluating waste quantity on a hazardous
constituent basis.
Review of the adequacy of the target distance specified in
the surface water pathway.
Development of a sensitive environment evaluation
methodology.
Feasibility of revising the containment factors to increase
discrimination among facilities.
Review of the potential for future changes in laboratory
detection limits to affect the types of sites considered for
the NPL.
Each technical issue is the subject of one or more separate but
related reports. These reports, although providing background,
analysis, conclusions and recommendations regarding the technical
issue, will not directly affect the HRS. Rather; these reports will
be used by an EPA working group that will assess and integrate the
results and prepare recommendations to EPA management regarding
future changes to the HRS. Any changes will then be proposed in
Federal notice and comment rulemaking as formal changes to the NCP.
The following section describes the specific issue that is the
subject of this report.
1.2 Issue Description
Several issues relevant to the HRS air pathway have been raised
by Congress and by public comments on the NPL and NPL rulemaking
actions. An analysis of these issues and options for revising the
HRS air pathway developed as a result of the analysis are presented
in a separate report (Wolfinger, 1986). That separate report
-------
focuses on revising the air pathway to reflect the potential of
sites to emit air contaminants in the absence of observed air
releases. An additional issue was identified in the course of that
analysis regarding whether the target distance limit in the air
pathway should be revised. The target distance limit is defined as
the maximum distance used in determining the target population for
the HRS air pathway. Currently, the air pathway target distance
limit is set at four miles.
This issue arises because current knowledge of atmospheric
residence times for common waste contaminants (generally 3 to 70
days; Cupitt, 1980) indicates that these contaminants may remain in
the atmosphere for relatively long periods of time and hence may be
transported over long distances.* Given these residence times,
transport distances would easily exceed the four miles currently
employed as the HRS air pathway target distance limit. As a result,
the number of people potentially at risk from site emissions could
be much greater than would be indicated using a four-mile limit.
However, the concentrations to which people living far from the site
would be exposed would, on average, be very low due to dilution.
Hence, the risk associated with exposures at long distances could be
low, possibly negligible. Nonetheless, if the contaminants released
*The estimates of atmospheric residence time provided reflect the
effects of atmospheric physio-chemical processes such as photo-
chemical oxidation and deposition on contaminant concentrations.
As such, residence time reflects the fraction of emitted contamin-
ant emissions remaining in the atmosphere, not the contaminant
concentration.
-------
from the site do not have safe exposure thresholds (for example,
carcinogens) even a low exposure would induce some non-zero risk of
adverse effects from the exposure.
Individual risks, such as the probability that an exposed
individual will develop cancer, will decline with distance as
exposure concentrations decline due to dilution. In contrast, other
risk measures, such as overall cancer incidence in the exposed
population, can increase with distance if population increases
faster with distance than average concentrations decline. Thus, the
counteracting effects of population increase with distance and
contaminant concentration decline with distance, in determining
risk, may or_ may not result in significant risks at distances
greater than four miles. Therefore, given a definition of an
acceptable risk,* dilution and population growth counteract to
determine the point at which the incremental risk from any specific
waste site becomes acceptable. The point at which risk generally
becomes acceptable could be used to determine the air pathway target
distance limit.
1.3 Purpose and Organization of Report
The principal purpose of this report is to investigate the
general level of carcinogenic risk associated with air releases from
*The establishment of an acceptable risk level for air emissions
from uncontrolled waste sites is beyond the scope of this paper.
Such a decision is most properly an EPA policy decision that takes
into account public perceptions of acceptability, as well as the
costs and benefits associated with the risk.
-------
uncontrolled waste sites and to determine whether a change is
warranted in the target distance limit in the air pathway. In
particular, the study is intended to determine whether a
modification to the current HRS target distance limit of four miles
is warranted, based on the potential for carcinogenic effects to
arise from air emissions from uncontrolled waste sites in exposed
populations residing beyond four miles from the site. The study
also examines the potential for carcinogenic effects in exposed
populations residing within four miles of the site.
Additionally, the study is intended to provide information that
can be used to address other questions concerning the HRS air
pathway. These questions include the definition of risk measures to
be employed in the air pathway, the determination of the relative
weight of carcinogenic effects in assessing toxicity in the air
pathway and the determination of the relative importance of air
pathway exposures in determining overall site risks in comparison to
ground water and surface water exposures.
The body of the report is organized in four chapters. Chapter 2
presents an overview of the approach used to analyze the target
distance limit. The results of the analysis are presented in
Chapter 3. Chapter 4 discusses the limitations in the approach and
their implications for the analysis. Chapter 5 presents the
conclusions of the analysis and discusses their implications for any
modification of the HKS air pathway.
-------
2.0 APPROACH
The approach employed in analyzing the air pathway target
distance limit was to determine the distances from uncontrolled
waste sites (i.e., CERCLA sites) at which incremental cancer risks
(i.e., risks above the "normal" background level) from air emissions
would be deemed acceptable. These distances might then be used in
determining the target distance limit, given a definition of
acceptable risk. This determination was made using incremental
lifetime (70-year) cancer risk estimates produced by the EPA Office
of Air Quality Planning and Standards Human Exposure Model (HEM)*
for 25 selected** final and proposed National Priorities List
sites. The availability and ease of use of this model, as well as
its constraints, determined to a large degree the type of analysis
conducted.
2.1 Focus of the Study
The focus of the target distance analysis is on the incremental
risk, at increasing distances, of carcinogenic effects arising from
long-term, sustained exposures (i.e., chronic) to atmospheric
contaminants arising from emissions from uncontrolled (and thus
generally inactive) waste sites. The analysis assumes that any
concentration, no matter how small, induces some risk of developing
cancer to individuals in the exposed population (i.e., there is no
*Further information on HEM can be found in Battye et al., 1985.
**Details of the selection process are provided in Section 2.5.
-------
threshold for carcinogenic effects). As Indicated by Haemisegger
et al. (1985), the no-threshold assumption for carcinogenic effects
has strong support in the scientific community. Further, as
indicated in Battye et al., 1985, EPA has reviewed the literature
regarding the carcinogen!city of volatile organic contaminants
potentially emitted from hazardous waste treatment, storage, and
disposal facilities and found no compelling scientific reason to
abandon the no-threshold assumption.
The primary reasons for this focus on carcinogenic effects was
the availability of both the HEM and supporting data on carcinogenic
potency* for several contaminants of interest and the importance of
the no-threshold assumption in analyzing risks at long distances
from the site, as discussed below. Although HEM could be modified
for use in other types of analysis (for example, non-carcinogenic
risk from chronic exposures), resource and data constraints
precluded modifying HEM or employing another model such as the
Industrial Source Complex Model (ISC).** The no-threshold
assumption is important because of its conservatism. The lack of a
threshold implies that any dose, no matter how small, will result in
some risk to an exposed individual. The alternate assumption, that
*For purposes of this analysis, carcinogenic potency is defined as
the probability that an exposed individual will develop cancer
arising from a sustained exposure to 1 ug/m^ of contaminant for
a 70-year period.
**For further information on the Industrial Source Complex
Short-term and Long-term models, the reader is referred to United
States Environmental Protection Agency, 1986a.
-------
a "no-effect" threshold exists implies that individuals exposed to
doses smaller than the threshold experience no risk. Dilution in
the atmosphere will probably reduce ambient concentrations below any
threshold as the distances increase from the emission point.
Nonetheless, population increases with distance will probably result
in significant increases in the number of people exposed at most
sites. Thus, the no-threshold assumption will generally result in
higher estimates of risk than the assumption of a "no-effect"
threshold. From this perspective, the no-threshold assumption is
desirable.
This focus on carcinogenic effects was deemed reasonable for
two reasons. First, EPA focused on cancer as the effect of interest
in its scoping study of the air toxics problem in the United States
(Haemisegger et al., 1985). The reasons for EPA's focus include the
importance of cancer as a cause of death in the United States, the
existence of several, ubiquitous, carcinogenic air pollutants (e.g.,
benzene, arsenic and vinyl chloride), the ability to quantify
carcinogenic risk, and the degree of concern, on the part of the
public, about the link between environmental pollution and cancer.
This emphasis on cancer does not imply that EPA is not concerned
with other types of human health or environment effects either
arising from long-term or short-term exposures.
Second, available data do not indicate that gaseous or
particulate contaminant concentrations sufficient to induce acute
-------
effects are generally achieved off-site from inactive facilities.*
Since, CERCLA is intended to address off-site risk from air releases
and not risk, to on-site workers [CERCLA Sec. 101(9) and 101(20)(A)],
the exclusion of risk of acute effects from the analysis is
warranted.
Finally, the assessment of sub-chronic effects (i.e., those
arising from exposures of 10 to 90 days duration) is severely
limited due to deficiencies in models relating site emissions to
levels of sub-chronic effects and deficiencies in the data necessary
to employ such models. Study limitations precluded the acquisition
and modification of a medium-term (e.g., monthly) dispersion model
(such as the ISC) or modification of HEM to create such as an
assessment model. Moreover, regardless of the availability of
applicable dispersion models, the assessment of sub-chronic effects
would be severely handicapped by the lack of data on sub-chronic
*However, a large body of evidence indicates that waste sites can
pose an acute risk to individuals working on the site as a result
of inhaling emitted contaminants. This evidence includes the case
histories of individuals who have died or become ill from breathing
vapors released from waste sites (e.g., hydrogen sulfide). Such
evidence is further supported by some of the available monitoring
data. For example, on-site ambient concentrations of trichloro-
ethylene have been detected at three times the threshold limit
value at the Kin-Buc Landfill (United States Environmental
Protection Agency, 1982). The potential for exposure to
concentrations that might result in acute effects is further
evident in the emphasis placed in EPA manuals on monitoring air
concentrations to ensure investigator safety during site and
remedial investigations (United States Environmental Protection
Agency, 1984a and 1985).
10
-------
effect thresholds for most contaminants of concern and the lack of
human exposure models relating concentrations to severity of effect
(e.g., number of people experiencing the effect). Similar data and
exposure model limitations limit the analysis of chronic,
non-carcinogenic effects as well.
2.2 Simplifying Assumptions
Given the limited focus of the study and the resource
constraints, several simplifying assumptions were made during the
study including:
Uniform, constant emission rate
Linear dose-response function with a uniform slope (potency
value)
Site-specific, but otherwise homogenous, unchanging
population
Simple climatological dispersion model
No transformation or deposition of contaminants
Use of Stability Array (STAR) meteorological data
Sustainability of annual average concentrations
This section describes these assumptions. The effect of these
assumptions on the study conclusions is discussed in the Chapter 4.
The first set of assumptions made in the study concern the
contaminant emission rates and characteristics. These assumptions
are among the most critical in terms of their affect on the validity
of the study conclusions. The study assumes a uniform, constant
emission rate of 100 kilograms per year with a carcinogenic potency
11
-------
-5 3
of 10 m /ug* for all sites regardless of the size or other
characteristics of the sites. The analysis did not attempt to
account for differences between the sites studied in terms of the
rates of contaminant releases (e.g., either daily variations or
variations in average annual emission rates) or the nature of the
contaminants that are released. Thus, the analysis tacitly assumes
that there are no significant effects of short-term deviations from
average doses on long-term cancer incidence.
A uniform annual emission rate was selected since it is not
possible to calculate site-specific annual emission rates within the
constraints of this project. The calculation of site-specific
annual emission rates would require the use of different emission
estimation equations for each site depending on the type of
emissions sources present on the site (e.g., surface impoundments).
There are no emission estimation procedures currently available that
are specifically applicable to uncontrolled waste sites and there
are questions about the applicability of the existing active site
emission rate equations to uncontrolled waste sites (Wolfinger,
1986).
Moreover, these emission rate equations usually address
instantaneous or short-term emission rates. Their extension to
*This potency is the probability of developing cancer resulting from
a continuous lifetime exposure to 1 ug/nr of contaminant. It is
equivalent to a conventional potency of 0.035 mg/kg/d under the
assumptions used in this study. These assumptions are discussed
later in this section.
12
-------
annual emissions would involve integrating emissions estimates for
numerous different site-specific meteorological conditions to form a
long-term average rate. Such a procedure would be further
complicated by the fact that many sites have numerous emission
sources of differing types. Finally, data are lacking for many of
the factors included in these emissions estimation equations (e.g.,
age of the site and waste constituent concentrations). Further
information on available methods for estimating site-specific
emission rates can be found in Breton et al., 1983 and 1984.
The emission rate of 100 kilograms per year was chosen for
several reasons. First, it is not an unreasonable value given the
range of estimates provided in the literature. An overview of the
available information on emission rates is presented in Table 2-1.
These estimates reflect numerous different assumptions about site
conditions. More complete information is presented in Appendix A.
The table also provides the 100 kilogram per year equivalent to the
emission rate expressed in the units used by the respective authors.
2
For example, assuming a site of area 10,000 m , 100 kilograms per
2
year is equivalent to 0.32 ug/m /s. This value can be compared
2
with the values of 0.4 to 18.4 ug/m /s reported in Hwang, 1985.
Inferring long-term emission rates from the values in this table
should be viewed with caution since the authors cited do not indicate
whether they believe the rates presented can be sustained for a
period of one year. In fact, it is unlikely that many of the cited
13
-------
TABLE 2-1
SELECTED INFORMATION ON WASTES SITE EMISSION RATES
Emission Rate
Source (units)
Contaminant
Surface Contaminated
Landfill Impoundment Landtreatment Soil
100 Kg/yr
Equivalent
Hwang, 1985 (ug/m2/s)
Toluene 1.48
1,1,1-Trichloroethane
Methylene Chloride 0.4
Tetrachloroethylene 0.76
Benzene
Chlorobenzene
Baker, 1985 (g/s)
Vinyl Chloride
Shen, 1981, (ug/s)
Aroclor 1242
Shen, 1982b (ug/s)
Aroclor 1242
Springer, Thibodeaux
and Chatrathi, 1983
(mg/m2/d)
Aroclor 1248
as liquid
in sludge
0.02-0.28
8-184
35-5680
0.08-2.65
0.30-0.91
18.4
15.3
4.7
1.1
2.4
0.32
0.32
0.32
0.32
0.32
0.32
0.003
3171
3171
27.4
27.4
-------
TABLE 2-1 (Concluded)
Source (units)
Contaminant
Emission Rate
Surface Contaminated
Landfill Impoundment Landtreatment Soil
100 Kg/yr
Equivalent
Thibodeaux, 1981
(g/m2/d)
Benzene
Chloroform
Vinyl Chloride
Aroclor 1248
Shen, 1982a (g/s)
Benzene
Thibodeaux et al.,
1982 (kg/d)
Benzene
Toluene
Total Hydrocarbons
1,1-Dichloroethane
Total Chlorinated
Hydrocarbons
Caravanos and Shen,
1984 (g/min)
Benzene
Carbon Tetrachloride
Trichloroethylene
89-563
340-1820
826-3650
0.001 - 0.01
5.5
0.1-0.38
1.2-0.48
2.6-5.2
9.0-0.11
50.0-1.1
0.7- 4.5
1.4-11.1
0.8- 7.9
0.027
0.027
0.027
0.027
0.003
0.274
0.274
0.274
0.274
0.274
0.190
0.190
0.190
-------
rates could be sustained for such a period of time, much less the
70-year lifetime assessed in this study. As indicated by this
table, the variation in reported emissions estimates is high and is
a major source of uncertainty in the results of the analysis.
In the opinion of the author, the uncertainties inherent in
carcinogenic risk assessments indicate that only order of magnitude
(or greater) differences in risk estimates are important in
analyzing the air pathway target distance limit (see Sections 4.1
and 4.2). Thus, given the values cited in Table 2-1, the set of
reasonable values for a sustained rate would be limited to the values
0.1, 1.0, 10, 100 and 1000 kilograms per year. The value of 1000
kilograms per year, equivalent to 70 metric tons emitted over the
period of the analysis, was deemed to be too high. Available
information on the quantities of contaminants in NPL sites indicate
that the values of 0.1 and 1.0 kilograms per year are probably too
low. This leaves 10 and 100 kilograms per year as choices for the
analysis.
The study also assumes that a linear, no-threshold dose-response
function is appropriate for the contaminants in question at the
expected concentrations. The use of such a model for low-dose risk
characterization is generally accepted, given that there is no
conclusive evidence indicating that a different model is
preferable. Table 2-2 presents a fairly comprehensive list of
potency factors for airborne carcinogens for use in such a model.
16
-------
TABLE 2-2
CARCINOGENIC POTENCY FACTORS
(ug/m3)-l
Chemical
Acrylamide
Acr yl onl tr He
Allyl Chloride
Asbestos
Arsenic
Benzene
Benzidene
Benzo(a)pyrene
Benzyl Chloride
Beryllium
Bis ( chl or o methyl) ether
1,3 Butadiene
Cadmium
Carbon Tetrachloride
Chloroform
Chromium
1,1 Dichloroethylene
1,2 Di chl or oe thane
Factor
1.7 x 10~5
6.9 x 10~5
6.8 x 1(T5
5.5 x 10~8
1.0 x 10~5
1.4 x ID"2*
7.4 x 10~6
6.9 x 10~6
6.6 x 10-2
1.7 x 10~3
3.3 x 10~3
1.2 x ID"5
1.4 x 10~3
4.0 x 10~4
0.93**
4.6 x 10~7
1.7 x 10~3
2.3 x 10~3
1.5 x 10~5
1.0 x 10~5
1.2 x 10~2
3.3 x 10~4
1.0 x 10~5
2.6 x 10~5
7.0 x 10~6
Reference
3
1
3
3
3
1
1
2,3
1
1
3
3
1
3
1
3
1
3
2,3
3
1,3
1
1
3
2
*Potency factor is in units of cc/fiber.
**This value corresponds to the probability (p) that an individual
exposed to 1 ug/m3 of bis(chloromethyl) ether continuously for
70 years will develop cancer, assuming p = 1 - exp(-pf x d), where
pf is the potency factor and d is the dose arising from a
continuous exposure of 1 ug/m3. The linear model assumed in
this study does not apply to this contaminant at this level of
exposure. As of NPL Update 5 (51 FR 21099, 10 June 1986), this
contaminant has not been identified at any NPL site during a site
inspection.
17
-------
TABLE 2-2 (Continued)
Chemical
Di chlorome thane
(Methylene Chloride)
Die thanolamine
Dimethylnitrosamine
Dioctyl Phthalate
Ethyl Acetate
Ethyl ene
Epi chl or ohydr in
Ethylene Dibromide
Ethyl ene Oxide
Formaldehyde
Melamine
Methyl Chloride
4,4 Methylene Dianiline
Nickel
Nitrobenzene
Ni tr os o mor ph ol in e
P en ta chl or o ph en ol
Perchloroethylene
Polychlorinated Biphenyls
Polynuclear Aromatic
Hydrocarbons
Propylene Dichloride
4, 4, 150 Propyliene Di phenol
Propylene Oxide
Styrene
Terephthalic Acid
Titanium Oxide
Trichloroethylene
Vinyl Chloride
Vinylidene Chloride
Factor
4.1 x 10 ~6
1.8 x 10~7
1.1 x 10~7
5.4 x 10-3
1.3 x 10-7
5.0 x ID"7
2.7 x 10~6
2.2 x 10-7
5.1 x 10~4
1.0 x 10~4
3.6 x 10~4
6.1 x 10~6
4.1 x ID"7
1.4 x ID"7
2.1 x 10~5
3.4 x 10~4
3.3 x 10~4
1.2 x 10-7
2.5 x 10~5
3.9 x 10~7
4.9 x ID"7
1.7 x 10~6
1.2 x 10~3
1.7 x 10~3
7.2 x ID'7
1.4 x 10~6
1.2 x 10~4
2.9 x ID"7
1.8 x 10~8
5.6 x 10-7
1.3 x 10~6
4.1 x 10~6
7.1 x 10"6
2.6 x 10~6
4.2 x 10~5
Reference
i
J_
2,3
3
3
3
3
3
3
3
1
2,3
2,3
3
3
3
1
3
3
3
3
1
2,3
3
1
3
3
3
3
3
3
1
2,3
1
3
3
18
-------
TABLE 2-2 (Concluded)
Reference
1 - Adapted from United States Environmental Protection Agency,
Superfund Public Health Evaluation Manual, (EPA
540/1-86/060, OSWER Directive 9285.4-1), United States
Environmental Protection Agency, Washington, DC, October
1986b.
2 - Battye, William et al., Preliminary Source Assessment for
Hazardous Waste Air Emissions from Treatment, Storage and
Disposal Facilities (TSDFs), (Draft Final Report), GCA
Corporation, Bedford, MA, February 1985.
3 - Haemisegger, Elaine et al., The Air Toxics Problem in the
United States; An Analysis of Cancer for Selected
Pollutants, (EPA-450/1-85-001), United States Environmental
Protection Agency, Washington, DC, May 1985.
19
-------
Most, but not all, of these substances have been identified at
uncontrolled waste sites. Potency values in the references cited
o
were given either in units of m /ug (Battye et al., 1985 and
Haemisegger et al., 1985) or were adapted from values expressed in
units of (mg/kg/d) (United States Environmental Protection
Agency; 1986b). In the latter case, the values were converted
assuming a continuous 70-year exposure, an inhalation rate of
3
20 m per day, and a body weight of 70 kilograms. These
assumptions are those recommended in the referenced report. In all
cases, these potency values reflect inhalation potency according to
the references.
5 3
As can be seen from this list, the value chosen (10 m /ug)
is a reasonable "mid-range" value, particularly given that some
uncontrolled waste sites do not contain carcinogens.* A uniform
value was applied since data are lacking on the carcinogenic potency
factors for many contaminants commonly found in uncontrolled waste
sites (even suspected carcinogens such as tetrachloroethane) and
because of difficulties in identifying the contaminants available
for migration into the air on a site-specific basis.
The values of 100 kilograms per year and 10 m /ug are
less critical than they might seem because of the structure and
assumptions of a linear, no-threshold dose response model such as is
used in HEM. Viewed as a set of equations, the model is linear in
*In terms of this study, contaminants that are not carcinogens can
be viewed as having a carcinogenic potency of zero.
20
-------
both the emission rate and the potency. This implies that an order
of magnitude increase in either one (or in the product of the two)
will induce a proportional increase in the estimates of risk. Thus,
the results from the model when an emissions rate of 10 kilograms
_c 2
per year and a potency of 10 m /ug are assumed are one-tenth
the results when an emission rate of 100 kilograms per year and a
5 3
potency of 10 m /ug are assumed.
The second set of assumptions concern the type and demographic
characteristics of the populations assessed in the study. The study
employed actual population data for 25 uncontrolled waste sites.
The data used are taken from the 1980 Census and manipulated in the
HEM population allocation module discussed below. However, the
study addresses only individuals residing on or near the sites
(i.e., within 60 miles) and assumes that the individuals are exposed
to average annual contaminant concentrations based on their place of
residence (as determined by their Census enumeration district and
the HEM population allocation module). The study does not address
non-resident individuals, such as workers, who might be exposed to
site emissions either routinely or on a one-time basis.
The study also assumes that the exposed populations are
demographically homogenous and identical. As a result, the analysis
does not address the additional risk posed to sensitive populations.
This assumption permits a simple conversion of potency in terms of
1 3
(mg/kg/d) to potency in terms of m /ug as discussed
21
-------
previously. The study does not address differences between
individuals (such as age, sex, or smoking habits) that might affect
their exposure levels, their contaminant intake and retention rates,
or the characteristics of their response to contaminant doses. The
study also does not account for any differences in the population
characteristics that might be associated with different sites.
These assumptions permit the use of a single potency factor for all
individuals exposed to emissions from a site. The assumption,
coupled with the assumption that contaminant emission rates and
characteristics are identical for all sites, permits the use of a
single potency factor throughout the study. These assumptions are
necessitated both by considerations of simplicity and by gaps in the
available data on dose-response relationships.
Further, the study assumes that the population does not change
over time. Population growth and other demographic changes over
time are not considered in the analysis.
The third assumption concerns the applicability of the
dispersion model used to estimate exposure concentrations. The
study ignores certain site-specific characteristics that might
affect exposure concentrations, such as terrain. This assumption
permits the use of the simple Gaussian plume model employed in HEM.
The study also assumes tacitly in using such a dispersion model that
wind speed, direction, and stability vary only over time, not over
distance. These parameters are assumed to be spatially invariant
22
-------
within the transport distances studied. The temporal, frequency
distribution of these parameters varies, however, between sites.
The fourth simplifying assumption is that the contaminants
are neither transformed nor deposited during transport within the
distances studied. This assumption is reasonable for gaseous
contaminants given the data available on atmospheric residence times.
The assumption is less reasonable for particulates. However, it is
necessary since HEM does not account for deposition.
The fifth assumption made concerns the meteorological, data
applicable to the sites. Since no site-specific meteorological data
of sufficient duration and quality for use in the analysis were
available for uncontrolled waste sites, the study assumed that the
meteorological data from the closest weather station that collected
Stability Array (STAR) data were applicable to the sites. The
impact of this assumption on site selection is discussed later.
Moreover; the study assumes that these data are applicable to the
70-year period of the study.
The sixth assumption made is that the average annual
concentrations estimated by HEM are sustained for the 70-year period
of the study. This assumption derives from the assumptions that the
emission rate is constant and the meteorological characteristics are
fixed for the 70-year study period.
Finally, the study assumes that the emissions arise from the
center of the sites.
23
-------
2.3 Measures of Risk
Three generic measures of risk, defined below, were analyzed:
MEI risk, AEI risk and population incidence. Although these measures
are most commonly used in assessing cancer risks, their use is not
restricted to cancer analyses. For the purposes of assessing cancer
risks, MEI risk is defined as the lifetime probability that a
maximally exposed individual (MEI) will develop cancer. MEI risk is
dimensionless, since it is strictly a probability. However, MEI
risk is frequently expressed in terms of the expected number of
cancer cases per unit of maximally exposed population (e.g., 1 case
per 100 people).
Maximally exposed individuals are those individuals, in the
exposed population, that are expected to be exposed to the highest
ambient concentration (and thus receive the highest dose) of the
contaminant in question from the source in question. Contrary to
intuition, the maximally exposed individual is not necessarily the
individual living closest to the site. Local atmospheric patterns
and site-specific emission patterns may cause an individual who
lives farther from the site to be exposed to a higher long-term
average concentration, receive a higher dose, and be at greater risk
(all other factors being equal).
The average exposed individual risk (AEI risk) of developing
cancer is defined here as the average lifetime probability that an
individual in the exposed population will develop cancer. As in the
24
-------
case of MEI risk, AEI risk is dimensionless. AEI risk is also
frequently expressed as a cancer rate (i.e., expected number of
cancer cases per unit of exposed population). For example, the AEI
risk from a site might be expressed as 1 case per 100,000 people,
corresponding to a probability of 10 . In terms of the estimates
produced by HEM, AEI risk is the sum, over all applicable receptor
locations,* of the product of the population at the receptor
location, the concentration at the receptor location, and the potency
of the contaminant, all divided by the total applicable population
(i.e., the product of the potency and the population-weighted average
of the ambient concentrations).
Population incidence is defined as the number of incremental
cancer cases expected in the exposed population over a 70-year period.
Population incidence is frequently divided by 70 to yield an "annual"
incidence for convenience of comparison. Thus, unlike MEI risk and
AEI risk, population incidence is not a rate or probability. It is
expressed as the expected increase in the number of cases in the
exposed population. For example, assume a site has an exposed
population of 10,000 and an AEI risk of 0.001 (1 in 1000). Also
assume a second site has an exposed population of 1000 and an AEI risk
of 0.01 (1 in 100). The second site has an AEI risk 10 times higher
than the first site, yet the expected incremental cancer incidence in
*A receptor location in HEM is a point in space at which individuals
(receptors) are exposed to ambient contaminant concentrations.
25
-------
the exposed populations for both sites are the same, 10 cases.
Similarly-, two sites may have the same AEI risk, but have different
population incidences due to differences in the size of the exposed
population. Thus, although population incidence and AEI risk are
related, they present different perspectives on the risk of cancer
from contaminant releases from a site.
MEI risk and population incidence are used in numerous EPA
analyses including Clean Air Act Section 112 analyses of National
Emission Standards for Hazardous Air Pollutants, Clean Air Act
Section 108 and 109 analyses for National Ambient Air Quality
Standards, and Toxic Substances Control Act Section 4(f) regulatory
analyses. EPA has also suggested the use of MEI risk and population
incidence as primary risk measures in the recent proposal of a
regulatory program for land disposal prohibitions under the
authorities in the Resource Conservation and Recovery Act (51 PR 1635,
14 January 1986). AEI risk is currently not considered directly in
EPA analys es.
EPA has not set an acceptable risk level to be used uniformly.
-4
Generally, MEI risks above 10 have been considered unacceptable,
_Q
while risks below 10 have been considered acceptable (Thomas,
1984). The acceptability of MEI risks between 10~^ and 10 is
assessed on a case-by-case basis incorporating, for example,
considerations of the size of the exposed population. The Superfund
Public Health Evaluation Manual also indicates that remedial actions
26
-------
under CERCLA should be designed to reduce risk levels to within this
range (United States Environmental Protection Agency, 1986b).
Further, as indicated in Travis et al., 1987, EPA has suggested a de
-5 -4
minimis MEI risk level of 10 to 10 for small populations and
10 to 10 for large populations. These authors also indicate
that in the context of occupational decisions, the Federal government
_3
has always acted to reduce MEI risks when it exceeds about 4 x 10
in small populations or about 3 x 10 in large populations (e.g.,
the entire population of the United States). Further, according to
the authors' analysis, the Federal government has never acted when
-4 -6
MEI risks were below 10 for small populations, or 10 for
large populations. EPA has not set acceptable levels for AEI risk.
Outputs from HEM can be used to estimate all three of these risk
measures, as discussed below.
2.4 Overview of the EPA Human Exposure Model
The EPA Human Exposure.Model (HEM) is a straightforward exposure
model that relates emissions of contaminants from a source to the
exposure and risk at receptor locations at varying distances from the
site. These estimates are developed in three basic steps. First,
annual average contaminant concentrations are estimated at each
receptor location using a long-term dispersion model. Second, the
population allocation component of the model estimates the number of
people at each receptor location. The concentration and population
estimates are then used in the third step to calculate exposure/risk
27
-------
estimates. Figure 2-1 illustrates the sequence of calculations in
the model. The following sections describe the model in greater
detail.
2.4.1 The Receptor Grid
HEM uses a circular receptor grid in estimating concentrations
and exposure. The model assumes a virtual point source (i.e., an
emission source of essentially zero radius) located at the center of
a collection of concentric rings. The radius of each ring is
specified by the user. The current analysis employed rings of the
following radii:
1/4 miles 3 miles 25 miles
1/2 miles 4 miles 50 miles
1 mile 10 miles 60 miles
2 miles 15 miles
The receptor grid is formed by the intersection of these rings with
16 wind direction radial lines drawn from the point source. These
radials are 22.5 degrees apart. An example of an HEM receptor grid
is illustrated in Figure 2-2.
2.4.2 The HEM Dispersion Module
The dispersion model employed within the HEM dispersion module
is a simplified climatological, steady-state Gaussian plume model*
designed to estimate long-term arithmetic average pollutant
concentrations at user-specified receptor locations. The HEM
*Further information on Gaussian plume models can be found in Hanna,
Briggs, and Hosker, 1982.
28
-------
Emissions
Data
(Identity of Contaminant)
Potency
Factors
Location
Data
I
Dispersion
Model
Concentration
Data
I
Exposure/
Risk Model
I
Exposure
Risk
Estimates
Meteorological
Data
Population
Data
I
Population
Allocation
Model
I
1980
Census
Data
FIGURE 2-1
BASIC HUMAN EXPOSURE MODEL (HEM) STRUCTURE
-------
to
en
c
c
CD
O
C
o
O
o
CD
CL
C/D
U)
o
60 Miles
50 Miles
25 Miles
10 Miles
Wind Direction
Radials
(16 in HEM)
Receptor Points
FIGURE 2-2
HEM RECEPTOR GRID
-------
dispersion model estimates the annual average ambient concentrations
of a pollutant at each receptor location arising from constant
emissions at the center of the receptor grid. The model can account
for chemical transformation but not for deposition of the pollutants.
However, as discussed previously, the analysis assumes that
contaminants neither transform nor deposit within the area of
concern. Alternately, the model can be viewed as assuming that all
deposited materials are immediately resuspended.
An annual total emissions rate (kilograms per year) is required
as input to the dispersion model. The model calculates the
equivalent emissions rate in units of kilograms per second. The
model assumes the emissions rate is constant throughout the year
regardless of site-specific conditions, including meteorological
conditions. This assumption is possibly a weakness in the model
when analyzing waste site emissions. The rate of emissions from
waste sites is sensitive, for example, to changes in temperature,
pressure and other meteorological variables. Thus, the assumption
that emission rates are constant may result in an overestimate (or
underestimate) of the concentrations associated with any set of
meteorological conditions. The potentially non-linear relationship
between emission rates and meteorological conditions indicates that
these errors may not average out.
The dispersion model employs the Stability Array (STAR) data
from the closest weather station (usually an airport) as
31
-------
meteorological input. These data are available from the National
Climatic Center in Ashville, North Carolina. The STAR data include
wind speed, wind direction and atmospheric stability class
information. The model manipulates these data to create a joint
wind speed/velocity/stability class frequency distribution. This
distribution, generally reflecting five years of data, is assumed to
be the same in all years of the analysis.
The airborne contaminant concentration at each receptor
location is calculated using the Gaussian dispersion model under the
meteorological conditions that define each class in the frequency
distribution. The annual average concentration at each receptor
location is calculated as the average of the calculated receptor
location concentrations, weighted by the frequency.
2.4.3 The Population Allocation Module
The number of people allocated to each receptor location is
determined in the population allocation module. As discussed
previously, no attempt was made to supplement the estimates produced
by the population allocation module with additional, site specific
information such as the locations of residences near the sites.
This module is an adaptation of the RADII5 computer program
developed by the Bureau of the Census (Moon, 1985 and Dusetzina,
1985). RADII5 employs the 1980 Census data to determine the number
of people living within a given distance of a user-specified
location. The location is given in terms of latitude/longitude
32
-------
coordinates. This location data requirement affected the choice of
sites selected for the target distance analysis, as discussed below.
The module uses two different allocation procedures to allocate
people to receptor locations, depending on the distance from the
site. One procedure is used for receptor locations within about 2.2
miles (3.5 kilometers) of the site, while a different procedure is
used for receptor locations lying beyond about 2.2 miles. Different
procedures are required due to the size of the grid cells within
about 2.2 miles, which in many cases are smaller than the Census
enumeration districts used to determine population in the RADII5
program. An allocation procedure is employed to distribute a
portion of the population in each enumeration district to each
receptor location within the district. Due to this limitation in
the allocation procedure, the risk estimates produced by HEM under
the study assumptions are suspect for distance below about 2.2
miles.
A different problem, however, pertains for longer distances.
This problems arises whenever the enumeration districts do not
wholly fall within the rings specified in the analysis. In such
cases, people residing within one ring may be allocated to the next,
more distant ring due to the assumptions in the allocation
procedure. This problem declines in importance as distance
increases because differences in ambient contaminant concentrations
between rings level off.
33
-------
2.4.4 The Exposure/Risk Module
The preceding modules produce estimates of (1) the annual
average contaminant concentration at each receptor location and (2)
the population at each receptor location. The exposure/risk module
calculates the incremental lifetime risk of developing cancer for
the population located at each receptor location, as a result of
continuous exposure to the calculated contaminant concentrations.
These calculations use the above concentration estimates and a
cancer potency factor in a linear model. The model assumes that the
probability of an individual developing cancer is the product of the
potency factor and the concentration to which the individual is
exposed. The expected number of cases in the exposed population is
simply the sum of the probabilities that each individual in the
exposed population will develop cancer.
The contaminant-specific potency factor is defined as the
lifetime probability that an individual would develop cancer from
2
inhaling air containing an average concentration of 1 ug/m of the
contaminant continuously for his lifetime. These potency factors
*
are usually calculated from q.. potency values* assuming that
each individual weighs seventy kilograms and breathes 20 cubic
meters of air per day, as discussed previously. Generally the
*
q.. potency factors are determined as the upper limit of the 95
percent confidence interval on the linear term in a multi-stage
*Further In forma tl"n on q* values can be found in Office of
Technology Assessment, 1981.
34
-------
probability model* of cancer effects versus dose. The data used to
develop these models is generally based on studies of cancer in
animals. As stated previously, the study assumes that the potency
factor applies to all exposed individuals regardless of age, sex, or
other demographic or health characteristics.
2.4.5 HIM Output
HEM produces estimates for several risk-related variables that
can be used in calculating risk measures. The principal HEM
variables used in the target distance analysis are:
Maximum contaminant concentration to which someone is exposed.
Number of people exposed to this maximum concentration.
Minimum contaminant concentration to which at least one
individual is exposed.
Number of people exposed to at least this minimum
concentration.
Population incidence (i.e. the sum, over all receptor
locations, of the product of the average annual concentration
at the location, the population at the location, and the
potency factor).
Values for additional variables are also available from HEM but are
not relevant to this analysis.
HEM produces estimates of each of these variables, for each
waste site, as a function of the distances specified. Within HEM,
differences in site estimates reflect differences in assumptions of
emission rates, emission potency factors, atmospheric conditions and
*Further information on such effects models can be found in Office
of Technology Assessment, 1981.
35
-------
demographic characteristics. The model is linear in the first two
factors, but not in the latter two factors. For example, a 10
percent decrease in the emission rate will result in a 10 percent
decrease in all concentration and exposure variables but will not
affect the population variable. However, this analysis assumes
fixed values for the emission rate and potency factor. Thus, in
this study, differences in risk estimates between sites reflect
differences in site-specific meteorology and population
characteristics only.
2.5 Site Selection
An analysis of several sites is needed to ensure that the
results reasonably reflect the range of site meteorological and
demographic conditions encountered at uncontrolled waste sites.
However, the data requirements of the HEM, together with project
resource constraints, restricted the choice of sites suitable for
analysis. HEM requires that the latitude/longitude of the site be
provided as input within an accuracy of one second. Further, the
limited availability of site-specific meteorological data requires
that the site can be associated with a weather station collecting
STAR data. As a result of these considerations, 25 sites were
randomly selected for analysis from the December 1985 list of final
and proposed NPL sites subject to these data requirements. Because
of these restrictions, the geographic distribution of sites selected
may not be truly representative of the universe of uncontrolled waste
36
-------
sites. The distribution is somewhat biased towards the industrial
Midwest, an area of generally higher-than-average population density,
The implications of this bias for the analysis are discussed in
Section 4.5.
2.6 Analysis Framework
The HEM was used to provide-estimates of contaminant
concentration, exposure, and population at various distances from
the center of the 25 selected sites. These outputs were converted,
as necessary, to yield estimates of MEI risk, AEI risk and
population incidence for each site. Since MEI risk is defined as
the maximum risk to any individual in the exposed population, a
single estimate of MEI risk is made for each site. MEI risk varies
among sites, as does the distance from the site to the closest
receptor location associated with a maximally exposed individual.
This distance is a potentially important determinant of the MEI risk
and is a potentially significant determinant of the air pathway
target distance limit. The target distance analysis includes both
an analysis of MEI risk and an analysis of the distance to the
closest maximally exposed individual.
Two analyses of AEI risk are presented in this report. The
first addresses the probability that an average-exposed individual
living within a circle of specified radius about the center of the
site will develop cancer as a result of emission from the site. For
example, estimates are presented on the average risk to individuals
37
-------
living within 25 miles of the site. The second incremental analysis
examines the AEI risk to populations living between specified radii,
e.g., the average risk to an individual living between 25 and 50
miles from the center of the site. This analysis is presented for
six such areas:
0-4 miles 15-25 miles
4-10 miles 25-50 miles
10-15 miles 50-60 miles
An incremental AEI risk analysis for areas within 4 miles
(e.g., 1-2 miles) was not presented due to uncertainties in the
population allocation procedures employed within about 2.2 miles of
the sites (see Sections 2.4.3 and 4.4).
This dual approach was taken to examine the possibility of
setting the air pathway target distance limit based on either the
distance at which AEI risk would be acceptable or, alternately; the
smallest distance from the site beyond which the average risk to an
exposed individual living in this area would be acceptable.
The expected number of incremental cancer cases arising from
exposure to site emissions in the exposed population (i.e.,
population incidence) is examined for six concentric zones centered
at the center of the site. This analysis was performed to determine
first, if a significant number of cases could be expected in the
exposed population. If this proved to be so, the analysis would
examine the feasibility of setting the air pathway target distance
38
-------
limit based on the distance at which most (or all) of the expected
number of incremental cases would be accounted for.
Four descriptive statistics are presented for each of the risk
variables discussed above: average among sites, variation among
sites, minimum of all sites, and maximum of all sites. The
following chapter discusses the risk estimates produced during the
study.
39
-------
3.0 CARCINOGENIC RISK ESTIMATES
This chapter presents the estimates of carcinogenic risk from
releases of airborne contaminants from uncontrolled waste sites.
Results are presented for three risk variables: risk to the
maximally exposed Individual (MEI risk), average risk to exposed
Individuals (AEI risk), and expected lifetime incidence of increased
cancer In the exposed population (population incidence). In
addition, results of an analysis to determine the distance to the
maximally exposed Individual are also presented.
3.1 MEI Risk
In HEM, the risk to the maximally exposed individual at a site
is equal to the product of the maximum exposure concentration and
3
the potency factor (expressed in units of m /ug equivalents). The
maximum exposure concentration is the highest estimated concentration
at a receptor location with at least one individual. Thus, the MEI
risk at a site is determined by both population distribution and
meteorology.
The maximum exposure concentration is equal to the highest
estimated concentration only if someone is living at the associated
receptor location. Otherwise, the maximum exposure concentration is
less than the highest estimated concentration. Similarly, the
maximum exposure concentration does not necessarily occur at the
closest receptor location because of meteorological considerations.
Nor does It necessarily occur in the first ring around the site.
41
-------
Furthermore, the maximum exposure concentration does not necessarily
occur at the closest receptor location that has an individual.
Meteorological conditions (e.g., prevailing winds) may result in a
higher average concentration occurring at a location further from
the site along a different radial.
Estimated MEI risks for the 25 sites examined ranged from
9 6
4 x 10 to 1 x 10 (see Table 3-1). The average was estimated
at 5 x 10 . This indicates that the chance that the maximally
exposed individual will develop cancer during his lifetime as a
result of continuous exposure to emissions from the site is about 1
in 2 million on average. Inter-site variation, as reflected in the
standard deviation of the site-specific estimates about the average,
was about one-half the average, indicating only a small degree of
site-to-site variation in MEI risk. Actual site-to-site variation
in MEI risk, due to actual variations in emission rates and waste
composition, may be higher than indicated. No analysis was
performed to determine the relative roles of meteorology and
population distribution in determining these differences.
Table 3-1 also presents information on the distance to the
maximally exposed individual and on the risk to that individual.
The table indicates that at no site examined did the MEI live within
1/4 mile of the estimated center of the site. The result is
potentially misleading since some of the sites are larger than 1/4
mile in radius, while distance in this analysis is measured from the
42
-------
TABLE 3-1
SUMMARY STATISTICS OF MAXIMALLY EXPOSED INDIVIDUAL (MEI) RISK
Distance Ring
(miles)
0-1/4
1/4-1/2
1/2-1
1-2
2-3
greater than 3
Number of Sites
0
8
8
8
1
0
MEI Risks
Average
NA
5
7
3
0.04
NA
(10-7)
Range
NA
3-7
4-11
1-10
0.04
NA
NA - Not applicable.
43
-------
center of the site. Further, as stated previously, this study does
not address workers at the site, only individuals residing on or
near the site. Thus, at some of the sites, no one lives within 1/4
mile of the site center since they would then also be living on the
site. However, at three sites, individuals were identified as
living within 1/4 mile of the site but, in each case, the risk to
these individuals was lower than the MEI risks. This illustrates
that distance alone does not determine the MEI risk. Meteorology
and geographic population distribution are important determinants as
well.*
At eight of the sites examined, the MEI lived between 1/4 and
1/2 mile of the site. The risks to these maximally exposed
individuals ranged between 3 x 10 and 7 x 10 with an average
of 5 x 10 . At eight other sites, the MEI lived between 1/2 and
1 mile of the site. Contrary to intuition, the range of risks for
these individuals ranged from 4 x 10~ to 11 x 10 and averaged
about 7 x 10 , a slight increase over the 1/4 to 1/2 mile ring.
Both the range of the risks and the average risks declined for the
eight sites at which the MEI lived between 1 and 2 miles of the
site. At only one site did the MEI live beyond 2 miles from the
site. The MEI risk at this latter site was 4 x 10 (the minimum
estimated).
*An additional important determinant of MEI risk, actual location
of residence, is not considered in this study since people are
assigned to receptor locations, not actual residence locations.
44
-------
The table indicates that a target distance limit of three miles
in the air pathway would account for the MEI at most sites. The
analysis indicates that the probability that the MEI at a final or
proposed NPL site would be beyond three miles is small. The bias in
the site selection procedure discussed previously, however;
indicates that the distance to the maximally exposed individual at
the selected sites may be an underestimate if compared to all final
and proposed NPL sites. This possibility arises since, because of
the location bias, individuals may live closer to the selected
sites, on average, than to NPL sites in general (a corollary of the
possible higher-than-average population density associated with the
selected sites). However, a distance of three miles may not account
for the maximally exposed individual at non-NPL sites again due to
possible differences in population density (non-NPL sites may have
lower average population densities than NPL sites, a possibility
arising from the importance of population in determining air pathway
scores in the current HRS).
3.2 Risk to the Average Exposed Individual
Table 3-2 presents the estimates of AEI risk to the population
living within specified distances from the center of the site.
Under the assumptions of the analysis, the risk is small at all
distances examined, ranging, on average, from 1 chance in
100 million within 4 miles to about 1 chance in 1 billion within
60 miles. The risks, on average, to the population living closer
45
-------
TABLE 3-2
SUMMARY STATISTICS OF RISK TO AVERAGE EXPOSED INDIVIDUALS
(AEI RISK) LIVING WITHIN SPECIFIED DISTANCES
Parameter
Average
Inter-site
Variation**
Maximum
Minimum
4
1 E-8
1 E-8
6 E-8
2 E-9
10
5 E-9
4 E-9
2 E-8
1 E-9
Distance
15
3 E-9
3 E-9
1 E-8
6 E-10
(miles)*
25
2 E-9
2 E-9
7 E-9
3 E-10
50
1 E-9
1 E-9
7 E-9
2 E-10
60
9 E-10
1 E-9
7 E-9
1 E-10
*Measured from center of site.
**Defined as the standard deviation of the site-specific estimates
about the average.
46
-------
than 4 miles will be higher, although the average will never exceed
the MEI risk. Thus, the AEI risk to the population living closer
than 4 miles from the site will not exceed 5 x 10 , on average.
The site-to-site variation in AEI risk is moderate. Inter-site
-8 -9
variation was about 10 at 4 miles, declining steadily to 10
at 60 miles. However, the inter-site variation was approximately
equal to the average AEI risk at all distances. The maximum AEI
_Q _Q
risks range from about & x 10 at 4 miles to about 7 x 10 at
-9
25 to 60 miles. The minimum AEI risks range from about 2 x 10
at 4 miles to about 1 x 10 at 60 miles.
Table 3-3 presents the estimates of the AEI risk to the
population living beyond 4 miles of the site. As stated above, AEI
risk is small even within a distance of 4 miles and declines with
further distance. This is highlighted by the decline in average
risks to individuals living further from the site. The chance that
an average exposed individual would develop cancer as a result of
the emissions from a site average about 1 chance in 1 billion for
the population living between 4 and 10 miles from the site. AEI
risk declines to about 1 chance in 10 billion for the population
living between 50 and 60 miles of a site.
There is small site-to-site variation in these estimates as
well. The inter-site variation was 6 x 10 between 4 and 10
miles and declined steadily to 6 x 10 between 50 and 60 miles.
The relative variation was fairly constant at 50 percent. Estimated
47
-------
TABLE 3-3
SUMMARY STATISTICS OF RISK TO AVERAGE EXPOSED INDIVIDUALS
(AEI RISK) LIVING WITHIN SPECIFIED DISTANCE RANGES
Distance (miles)
Parameter
Average
Inter-site
Variation*
Maximum
Minimum
Number of
Sites**
4-10
1 E-9
6 E-10
4 E-9
7 E-10
24
10-15
7 E-10
3 E-10
1 E-9
3 E-10
24
15-25
4 E-10
2 E-10
8 E-10
2 E-10
24
25-50
2 E-10
9 E-ll
5 E-10
8 E-ll
24
50-60
1 E-10
6 E-ll
3 E-10
5 E-ll
24
*Defined as the standard deviation of the site-specific estimates
about the average.
**Model estimates indicated that no one lives beyond four miles of
one of the sites. Hence, this site was excluded from this
analysis.
48
-------
AEI risks ranged from 2 x 10 to 6 x 10 within 4 miles of the
site. In the area from 50 to 60 miles surrounding the site,
estimated AEI risk ranged from 5 x 10 to 3 x 10 .
3.3 Population Incidence
Population incidence is defined as the number of incremental
cancer cases that can be expected in the exposed population due to
the emissions from the site in question over a period of 70 years.
At each receptor location, the product of the population estimate,
the concentration and the potency factor is calculated. The sum of
the values for all receptors within a specified distance equals the
population incidence within that distance.
Table 3-4 presents the summary statistics of population
incidence for the sites examined in the analysis. As seen in the
table, population incidence is very small at all distances. The
-4
model indicates that, on average, one can expect 7 x 10
additional cancer cases in the exposed population over a 70-year
period within 4 miles of the site. The estimated population
_3
incidence increases to only about 2 x 10 cases in the exposed
population, on average, within 60 miles of the site. As would be
expected, population incidence continually increases with distance,
although at a decreasing rate. This illustrates that, given an
assumption that no safe exposure threshold exists, population
incidence will increase as long as population increases, despite
decreasing incremental exposure concentrations.
49
-------
TABLE 3-4
SUMMARY STATISTICS OF POPULATION INCIDENCE*
Distance (miles)
Parameter
Average
Inter-site
Variation**
Maximum
Minimum
4
7 E-4
8 E-4
4 E-3
3 E-5
10
1 E-3
8 E-4
5 E-3
8 E-5
15
1 E-3
1 E-3
6 E-3
1 E-4
25
2 E-3
2 E-3
6 E-3
2 E-4
50
2 E-3
2 E-3
1 E-2
2 E-4
60
2 E-3
3 E-3
1 E-2
2 E-4
*Expected number of incremental cancer cases in the exposed
population.
**Defined as the standard deviation of the site-specific estimates
about the average.
50
-------
There is moderate site-to-site variation in the estimates of
population incidence arising from differences in meteorology and
population distribution at each distance. Inter-site variation
-4 -3
increased from 8 x 10 at 4 miles to about 3 x 10 at 60 miles
probably reflecting increasing site-to-site variation in
population. Nonetheless, as in the case of AEI risk, the relative
site-to-site variation was approximately equal to the population
incidence at all distances. The maximum population incidence ranged
-3 -2
from about 4 x 10 cases within 4 miles of the site to 1 x 10
cases within 60 miles of the site. The minimum incidence ranged
from 3 x 10 within 4 miles to 2 x 10 within 60 miles.
These results indicate that incremental population incidence
does not provide a reasonable basis for establishing a target
distance limit in the HRS air pathway. Thus, an extension of the
target distance limit beyond 4 miles, to account for increases in
number of expected cancer cases in the general exposed population as
distance increases, is not warranted. Further, the results together
with the results of the analysis of AEI risk, indicate that cancer
incidence in the general exposed population probably should not be
reflected in the air pathway target category at all, since the
absolute magnitude of the estimates are very small even in the worst
case.
51
-------
52
-------
4.0 STUDY LIMITATIONS
Several important limitations affect the conclusions that can
be drawn from the analysis. Most of these limitations are common to
all assessments of the risk from releases of airborne carcinogens.
These limitations can be classified into five groups: emissions
rates, potency factors, study assumptions, HEM model structure, and
site selection. The following sections discuss these groups of
limitations and their implications.
4.1 Limitations in the Bnission Rate Assumptions
This section discusses the limitations arising from the choice
of the emission rate of 100 kilograms per year, as well as the
limitations arising from the assumptions of constancy and
uniformity. Some of these assumptions are common to all airborne
carcinogen risk assessments (e.g., sustainability of emission rates,
independence of meteorology and emission rate, and simplifications
arising from difficulties in estimating emission rates). Other
limitations arise from unique aspects of this analysis (e.g.,
uniformity in emission rates).
The analysis assumes that a constant emission rate of 100
kilograms per year applies uniformly to all of the sites studied.
The value of 100 kilograms per year was chosen after a review of the
limited body of information available on waste site emission rates.
Much of this information applies to active waste disposal sites or
is based on idealized models that do not reflect conditions common
53
-------
to uncontrolled waste sites. Few data are available on emission
rates for uncontrolled waste sites. However, even given the limited
coverage of the available information, it is clear that the
site-to-site variation in short-term emission rates is large (see
Appendix A). For example, estimated instantaneous emission rates of
2
benzene from landfills range between 0.047 ng/cm /sec and 2106
2
ng/cm /sec in different studies. (One hundred kilograms per year
2 2
is equivalent to about 0.032 ng/cm /sec for a 10,000 m
landfill.) Further, available models indicate particulate emission
rates could exceed 100 kilograms per year in locations that are both
windy and dry. Given some of the contaminant concentrations found
in soils of uncontrolled waste sites (on the order of 1 part in
100), emission rates for particulate carcinogenic contaminants might
reasonably exceed 100 kg/yr as well.
As stated previously, the assumptions that the average
concentration is sustained at each receptor location is derived from
the assumption that the constant emission rate is sustained. It is
doubtful that any but the largest sites could sustain a 100 kilogram
per year emission rate for 70 years. However, little information is
available on long-term site emission rates.
The study also assumes that the constant emission rate applies
to all sites at all times. This assumption is not common in
airborne carcinogenic risks assessments but rather arises from
limitations in resources and scope of this study and the resulting
54
-------
need for simplicity. This assumption ignores the wide variation
between sites in emission rate, quantity, and composition due to
differences in site characteristics such as waste characteristics,
containment, soil type and meteorology. The constant emission rate
also ignores the temporal variations in emission rates within a year
and between years. Generally, as a site ages its emissions rate
will decline as the quantity of contaminants remaining on the site
declines.
Overall, the emission rate assumptions are probably
conservative. The uniform emission rate of 100 kilograms per year
for 70 years may be an overestimate of the "average" of all sites,
particularly for sites containing volatile organics. It is less
likely to be an underestimate. Long-term sustained rates for most
sites will probably be lower, although short-term emissions at much
higher equivalent rates are likely. The lack of data to confirm or
dispute the emission rate assumptions implies a high degree of
uncertainty in the risk estimates produced in the analysis.
However; the net result is probably that the risk values will be
somewhat elevated and their uncertainty high.
4.2 Limitations Associated with the Potency Factors
Several of the limitations associated with the potency factors
are similar to the limitations in the emission rates. The
limitations associated with basic potency factors are common to all
risk assessments. The additional limitations associated with
55
-------
translating potency factors in dose units (e.g., mg/kg/d) to potency
3
factors in ambient units (e.g., m /ug) are unavoidable but,
nonetheless, common to many airborne carcinogenic risk assessments.
The first important limitation arises from the use of a uniform
5 3
potency factor of 10 m /ug (see Section 2.2). This limitation
is not common but, as with the emission rate, arises from the need
for simplicity in the analysis. An examination of the available
data in Table 2-2 on potency factors (expressed in these units)
indicates that this value is a reasonable mid-range value. It is,
however, low when compared with potency factors for airborne,
_2
carcinogenic inorganics (e.g., 10 for arsenic). The overall
3
range of "known" potency values (expressed in terms of m /ug) is
8
about 10 to 1.0. However, many of the substances found at
uncontrolled waste sites are not carcinogenic, e.g., iron. Thus,
the carcinogenic potency of emissions from some uncontrolled waste
sites may be as low as zero.
As in the case of the emission rate, the assumption that a
uniform potency factor applies to all sites ignores site-to-site
variation in waste composition. It is evident that the overall
carcinogenic potency of the contaminants on one site may be orders
of magnitude higher than those on another site due to differences in
waste composition among sites. Further, the choice of a single
potency value ignores synergistic or antagonistic interactions
between the various contaminants present on a particular site and
56
-------
between site contaminants and other contaminants in the environment
in determining the overall emissions potency.
Similarly, the analysis does not account for the interaction of
the exposure with other lifestyle-related risk factors, such as
smoking, in determining the probability that an individual will
develop cancer. Nor does the analysis account for any other
synergistic or antagonistic interactions between the exposure to the
contaminant and any other factor. The study assumes that the
potency factor applies to each individual uniformly irrespective of
the individual's age, sex or any other characteristic and thus,
risks to sensitive populations are not specifically addressed as
well. Many of these characteristics, particularly age, affect the
probability that an individual will develop cancer from a given
exposure during his lifetime.
There are further uncertainties associated with the basic
*
q potency factor values that limit the usefulness of the model
results. These uncertainties arise from limitations in the basic
toxicology data underlying the potency factors, the validity of the
use of models based on from animal data to in assessing human health
risks, and the validity of the models used to infer low dose
response relationships based on high dose response data. Battye
et al. (1985) present a more detailed discussion of potency factors
and their use in HEM. A more complete discussion of the overall
limitations of potency factors can be found in Office of Technology
Assessment, 1981 and National Research Council, 1983.
57
-------
Finally, as discussed previously, the potency values used in
*
this analysis were derived from q1 values using assumptions on
the weight and daily respiratory rate of an exposed individual
(assuming a homogenous population). These assumptions ignore the
potential for some portion of some contaminants to be exhaled and
thus not to be available to cause a harmful effect to the body
system.
The overall effect of these considerations on the validity of
the results of the target distance analysis is problematic.
Overall, they indicate that a very high degree of uncertainty is
introduced into those portions of the analysis sensitive to the
actual risk values (e.g., MEI risk). Many of these uncertainties,
however are also common to any cancer risk assessment.
4.3 Limitations Arising from Simplifying Assumptions
There are several limitations that arise from the simplifying
assumptions made during the study in additional to those related to
emission rates and potency. Again, many of these limitations are
common to airborne carcinogenic risk assessments and arise from gaps
in fundamental knowledge of atmospheric processes in general and
site-specific conditions in particular.
First, the necessary simplifying assumptions required to use a
the dispersion module of HEM are strictly true for many of the sites
examined. These assumptions include spatial invariance of wind
speed, wind direction, and atmospheric stability. Nonetheless,
58
-------
these simplifying assumptions are commonly made in many air quality
modeling analyses, even in employing approved air quality models in
analyses supporting air emission permit processes. The question is
whether the extent of the violation of assumptions in comparison to
reality is sufficient to invalidate the use of the model. These
assumptions have been accepted in other HEM analyses (e.g., Battye
et al., 1985). However, it is imperative that the truth of the
assumptions be considered in making judgments based on the modeling
results.
HEM has not been reviewed by EPA as part of its formal model
review process. It is thus difficult to delineate the limitations
of the model overall. However, HEM's limitations include all of
those generally applicable to any climatological dispersion model
such as the Climatological Dispersion Model (CDM; Busse and
Zimmerman, 1973 and Brubaker, Brown and Cirillo, 1977). According
to the recommendations in the EPA Guideline on Air Quality Models
(1986a), CDM is appropriate for use in the following situations:
Point and area sources
Urban areas
Flat terrain
Transport distances less than 50 kilometers
Long-term average concentrations
Many of the sites analyzed in this study, however, are neither in
urban areas nor are they surrounded by flat terrain. Further, the
analysis was conducted for distances up to 100 kilometers. Despite
59
-------
these recommendations, HEM is used by EPA in analyzing airborne
carcinogenic risk at distances up to 100 kilometers (Dusetzina,
1986).
The target distance analysis implicitly assumes that the terrain
around the site is flat for 60 miles; an assumption of questionable
validity in many parts of the United States. The apparent conflict
between these assumptions and the EPA recommendations does not
necessarily invalidate the use of the model and the results of the
analysis. It does indicate that the results should be viewed with
caution.
Second, the assumption that the contaminant is neither
transformed or deposited within the grid region may not be valid.
As stated previously, it may be valid for many gaseous contaminants
but it is unlikely to be valid for particulates. This assumption
may result in an overestimation of risks in the cases of chemicals
that transform into other, less potent chemicals or that deposit and
do not resuspend. Alternately, underestimates of risks may result
for chemicals that transform into other, more potent chemicals.
Further, lack of consideration of deposition ignores the potential
increase in unit risk associated with ingesting deposited material
as compared with inhaling suspended material. Overall, it is
difficult to determine whether HEM underestimates or overestimates
risks due to these assumptions.
60
-------
Third, the study assumptions that the distribution of
meteorological conditions, developed from the past five years of
applicable STAR data from the closest applicable weather station,
remains the same for all years of the analysis is open to question.
Given that five years of data determine the distribution, however;
this assumption is probably reasonable. The related assumption that
the meteorological conditions from the nearest weather station are
sufficiently representative of site meteorology for use in this
study is of greater, but otherwise indeterminate, concern.
Fourth, the study assumption that the concentrations are
maintained at a constant level over the 70-year period is open to
question. This is equivalent to assuming that the site sustains the
user-specified emissions rate for the 70-year study period and that
the distribution of meteorological conditions does not change from
year to year. This is a common, worst-case assumption arising from
the need to forecast long-term concentrations from shorter-term
(e.g., annual) information. This assumption is conservative,
however, as it is unlikely that emissions from a waste site will
remain constant over time (and would probably decrease).
A fifth limitation arises from the study assumption that
populations do not change over time. The study essentially assumes
that the number of people living in the affected area does not
change over time (i.e., it does not account for births, deaths or
migration into or out of the region), nor does it account for
61
-------
migration within the region. Essentially, the study assumes that
the same people remain at their assigned grid locations indefinitely.
However, given the limitations of local, long-term (e.g., 70 years)
forecasting models and uncertainties in the effects of short-term
exposures on long-term risk, these limitations may both be
unavoidable. Further, they may be acceptable since, on balance,
they would probably result in an overestimation of the risks from an
particular site.
A sixth limitation comes from the assumption that the emissions
arise from the center of the site. This assumption may result in
underestimation of the risks posed to surrounding populations due to
emissions arising from sources located near the site boundary. This
limitation increases in importance as the size and number of sources
on the site increase.
Finally, the study assumption that the dose-response
relationship can be adequately reflected using a linear, no-threshold
dose-response model is open to question. Numerous other models can
be employed (e.g., Weibull and probit models). However, in general,
the linear, no-threshold model produces higher risk estimates at the
low concentrations indicated by the dispersion model and is thus
considered to be suitably conservative.
4.4 Limitations of the HEM
The principal limitations that HEM places on the conclusions,
independent of the study assumptions, arise from the dispersion
model employed in HEM and the population allocation procedure.
62
-------
As stated previously, the dispersion model used in HEM is a
steady-state Gaussian plume model. These models in general have
many strengths, as indicated by their widespread use in regulatory
applications. However, models of this type have a number of
weaknesses that arise from their simplifying assumptions. Smith
(1980) presents a review of the strengths and weaknesses of these
models. The most important weaknesses, from the perspective of this
analysis, are the assumptions that the wind speed, direction and
stability are uniform over the entire grid region and that the
terrain in the grid region is flat. Additionally, concerns have
been raised about the validity of the use of Gaussian plume models
in uneven terrain at distances beyond 10 kilometers (American
Meteorological Society, 1980).
The population allocation procedure used in HEM was discussed
in Section 2.4.3. As noted in that section, the procedure has
difficulties at distances less than about 2.2 miles. In site-speclfc
analyses, EPA supplements the population allocations produced by HEM
with actual location information derived from USGS topographic
maps. This procedure was not followed during this study. Thus, the
study relies on the population allocation module results produced by
HEM for short distance risk analyses.
4.5 Limitations in the Site Selection Process
As discussed previously, 25 sites were selected from a list of
final and proposed NPL sites that met HEM data requirements.
63
-------
Table 4-1 presents a comparison between the regional distribution of
the selected sites and the distribution of final and proposed sites
through NPL Update 5. Overall the comparison is good. The
comparison shows complete agreement for three of the regions (III,
VIII, and X). The comparison indicates a slight bias towards
Regions I and VI and a larger bias towards Region V. The
comparisons also indicate a bias away from Regions II, IV, VII,
and IX. Overall, this comparison indicates that the distribution is
moderately biased towards areas of generally higher population
density with a higher degree of urbanization. This would indicate
that the population density surrounding the selected sites will be
greater, on average, than that of the NPL sites as a whole. The
potential effects of this geographic bias can be determined. Due to
the bias toward areas of higher urbanization and population density,
it is likely that people would be living closer to the selected
sites than to all final and proposed NPL sites, on average.
Further, one might expect that more people would be living within a
given distance of the selected sites than to all final and proposed
NPL sites, on average. These considerations indicate that the
analysis may somewhat (1) understate the average distance to the
MEI, (2) overstate the average MEI and AEI risks and, (3) inflate
the estimates of average population incidence at each distance.
Overall the effect of these biases is likely to be minor in
comparison to the uncertainties induced by the limitations discussed
previously.
64
-------
TABLE 4-1
COMPARISON OF GEOGRAPHIC DISTRIBUTION
OF SITES SELECTED FOR ANALYSIS
Distribution
Region
I
II
III
IV
V
VI
VII
VIII
IX
X
Sample
Number
2
4
3
0
10
2
0
1
2
1
Distribution
Percent
8
16
12
0
40
8
0
4
8
4
Actual NPL*
Percent
6
19
12
11
24
6
5
4
9
4
*National Priorities List (proposed and final) sites as of NPL
Update 5 (51 FR 21099, 10 June 1986).
65
-------
An additional limitation arising from the site selection
procedure derives from the use of only final and proposed NPL sites
as the sampling population. The HRS is designed to assist in
evaluating all uncontrolled waste sites, not just NPL sites. As
such, the results presented may be biased if generalized to apply to
all uncontrolled waste sites. The overall implications of this bias
are unknown. However, the structure of the HRS is such that the
presence or absence of people, and the number of people living near
a site, is an important determinant of the site HRS score. Thus,
final and proposed NPL sites will probably have more people living
near them than all uncontrolled waste sites, on average. Therefore,
the bias induced by employing only final and proposed NPL sites in
the analysis probably amplifies the geographic biases discussed
above.
4.6 Overall Implications for the Analysis
Table 4-2 summarizes the limitations of the target distance
analysis. Each of these limitations increases the uncertainty in
the final results differentially. The implications of the
uncertainties and limitations discussed above for the validity and
usefulness of the target distance analysis results are complex and
difficult to assess.
In the opinion of the author, the uncertainties in the emission
rate and potency assumptions are the most critical for the target
distance analysis. These two factors have the greatest effect in
66
-------
TABLE 4-2
SUMMARY OF STUDY LIMITATIONS
Source of Limitation
Limitation
Emission Rate
Potency Factors
Simplifying Assumptions
HEM
Site Selection Process
Chosen Emission Rate
Uniform Rate Among Sites
Constant Rate Over Time
Sustainability for 70 Years
Chosen Potency Factor
Uniform Factor Among Sites
Uniform Applicability to All
Individuals
Translation of Dose Factors to
Exposure Factors
Lack of Interactions
Flat Terrain
Urban Area
100 Kilometer Scale
Uniform Meteorology
No Pollutant Loss
Source and Applicability of
Meteorological Data
Constant Exposure/Dose
Homogenous Population
Unchanging Population
Exposure Location
Linear Multi-stage Model
Gaussian Plume Model
Possible Regional Bias
67
-------
determining the estimated risk from the site and hence the distance
at which the risk becomes acceptable given a criterion for
acceptability. There is considerable variation in the composition
of waste deposited in uncontrolled waste sites and hence there
should be considerable variation in the rate and potency of their
actual emissions.* The conflict between this suspected variation
and the simple assumptions used in the analysis underline the
importance of these assumptions in the analysis.
The effect of both the limitations associated with the
simplifying assumptions and the use of HEM on the validity of the
study and its conclusions is more problematic. Some of the
assumptions would result in overestimates (e.g., linear,
dose-response model) while others would result in underestimates
(homogenous population). Little information is available indicating
the effects that these assumptions would have on risk estimates of
this type. However, these uncertainties are common to many, if not
all, airborne carcinogenic risk assessments. Lacking further
information, the importance of these assumptions on the validity of
the conclusions and the analysis remains a question of individual
judgment.
*The actual variation in emission rates between sites is largely
unknown due to a lack of monitoring data that can be used to
estimate emission rates. Differences in emission potencies is amply
evidenced by the wide variation in contaminants detected at sites.
68
-------
The uncertainty In generalizing the results from the sample of
25 to the entire universe of sites arising from possible regional
bias in the site selection procedure is probably small.
Overall, the limitations and uncertainties indicate that a high
level of uncertainty should be ascribed to the results of the
analysis. This is particularly true for the risk estimates. The
study results probably overstate the risk posed by most sites. The
degree of this overstatement is unknown. As a result, it is not
clear that even an order of magnitude difference between the risk
values represents a significant difference in terms of the actual
risk posed by sites. This uncertainty is mitigated somewhat by the
lack of certainty as to what constitutes an acceptable risk. EPA
has not established uniform guidelines for acceptable risk that
account for uncertainties in potency and exposure factors.
However, because of the structure of the model and its
assumptions regarding population allocation, the analysis of the
distance to the maximally exposed individual is less uncertain.
Despite these limitations, the author believes that this
analysis can provide insight into the risks posed by most
uncontrolled waste sites and could provide a reasonable basis for
establishing, in part, the target distance limit in the HRS air
pathway.*
*The sub-chronic risks posed by site should also be considered in
setting the target distance limit.
69
-------
70
-------
5.0 CONCLUSIONS OF THE STUDY AND IMPLICATIONS FOR THE HRS AIR
PATHWAY TARGET DISTANCE CATEGORY
The objective of this analysis in the context of the air
pathway of the Hazard Ranking System is primarily to determine the
target distance limit in terms of the distance beyond which
emissions from uncontrolled waste sites would generally be deemed to
pose an acceptable risk to the surrounding population. A secondary
objective is to provide information useful in answering other
questions relevant to the modification of the air pathway including:
Definition of risk measures to be employed in the air
pathway.
Determination of the relative weight of carcinogenic effects
in assessing toxicity.
Determination of the relative importance of air pathway
exposures in determining overall site risks.
This chapter presents the results of the analysis of these questions
based on the risk estimates presented earlier.
5.1 Risks from Air Contaminant Releases from Waste Sites
Despite the limitations of the analysis, several general
conclusions about the nature of cancer risk from uncontrolled waste
sites can be drawn. First, both the estimated average individual
risk and the incremental incidence of cancer in the exposed
population arising from exposure to uncontrolled waste site
emissions are likely to be small. The combined emission rate/potency
factor would have to be increased by a factor of about 100 to induce
one additional cancer in the exposed population in the worst case
71
-------
examined (see Table 3-4). It must be stressed, however, that the
study assumptions may result in underestimates of the risks to those
individuals residing near the site.
Second, the analysis indicates that for most waste sites, the
MEI risks will probably be less than those that have been considered
acceptable by EPA or other Federal agencies in past instances (see,
for example, Thomas, 1984 and Travis et al., 1987). However, an
unacceptable MEI risk is possible for some sites, given (1) the
uncertainties in the emission rates and in the ranges of potency
factors, and (2) the possible underestimation of near site risks
resulting from the assumptions of the study. The use of a potency
_2
factor of 10 (e.g., arsenic) for the worst case examined would
_3
result in an MEI risk of about 10 , which is greater than the
generally acceptable range.
Third, an examination of the range of potency values, emission
rates and the results of this analysis indicate that participates
rather than volatile organic gases may pose the greater cancer
threat from releases into the air from uncontrolled waste sites.
This preliminary conclusion is based on two observations. Potency
factors for particulates (e.g., arsenic) appear to be higher, on
average, than for volatile organics. Also, it is unlikely that a
site would emit a critical amount of volatile organic contaminants
for a sustained period of time (e.g., 100 kilograms per year for
70 years) although it may be possible that some large sites may emit
72
-------
such a critical amount of contaminated particulates. Major
uncertainties in this conclusion are (1) the potentially greater
impact of deposition processes on particulates as compared with
gases (which would result in potentially lower exposure
concentrations for particulates), (2) the effect of contaminated
particle re-entrainment on increasing total exposure, (3) the impact
of transformation processes which could create more or less potent
carcinogenic organic compounds in the ambient air,* and (4) the
uncertainties in long-term emission rates.
A limited conclusion concerning non-cancer risks can also be
made based on the results of this analysis. One can conclude that
the risk of non-carcinogenic effects at distances beyond 4 miles
would be negligible, at most sites, for the contaminants with known
safe thresholds. This conclusion arises from the relationship
between AEI and MEI risk and exposure concentrations. AEI risk is
the product of the population-weighted average exposure
concentration and the potency factor. Thus, the average exposure
3 5
concentration (in units of ug/m ) is 10 times the AEI risk.
Using this fact and the results from Table 3-3, the average annual
concentration to which the average individual would be exposed at
distances beyond 4 miles would be well below 1 nanogram per cubic
meter, at all of the sites examined. At distances beyond 10 miles
*For example, benzo(a)pyrene, a common hydrocarbon transformation
product, has an estimated carcinogenic potency of 1.7 x 10~3.
73
-------
the average annual concentration at the worst site examined is
estimated to be less than or equal to about 0.1 nanograms per cubic
meter. Further, a comparison of the estimates for AEI risk and MEI
risk for distances less than 4 miles indicates that the MEI risk is
generally between 10 and 100 times the AEI risk. Extrapolating this
relationship to distances greater than 4 miles indicates that the
highest average annual exposure concentration is below 100 nanograms
per cubic meter at all such sites. This concentration is below all
of the safe exposure thresholds* listed in United States
Environmental Protection Agency, 1986b for chronic inhalation
exposures. Thus, it is reasonable to conclude that it is unlikely
that safe exposure thresholds for chronic effects would be exceeded
at distances beyond four miles from most waste sites.
The analysis does not provide information on the potential
magnitude or geographic extent of health effects caused by
sub-chronic exposures or from chronic non-carcinogenic exposures to
contaminants that lack safe exposure thresholds.
*Using the acceptable daily intake for chronic inhalation effects as
the safe threshold concentration, and assuming the relationship
between dose and exposure concentration discussed previously, the
lowest safe threshold concentration for chronic effects listed in
United States Environmental Protection Agency, 1986b is about 180
nanograms per cubic meter (based on an ADI for inhalation of 5.1 x
10~5 milligrams per kilogram per day for inorganic mercury
compounds ).
74
-------
5.2 Implications for the HRS Air Pathway Targets Category
The results of the analysis have major implications for the
structure of the HRS air pathway targets category. First, the
results indicate that cancer risk should not be emphasized in the HRS
air pathway except possibly for near-site residents. Rather,
despite the lack of knowledge about non-carcinogenic risk, it is
probably reasonable to emphasize non-cancer effects in the air
pathway pending the results of study of such risks. Such analyses
to delineate the non-cancer risks arising from uncontrolled waste
site air emissions should be undertaken and modifications to the HRS
(e.g., target distance limits) developed as warranted.
This conclusion does not imply that cancer risk should be
excluded from the HRS. The results of the analysis indicate that
cancer risk from some sites may be unacceptable depending on the
characteristics of the waste contaminants on the sites and their
emission rates. The results of the analysis indicate, however, that
the only potentially significant variation between between cancer
risks from sites arises from possible differences in site
contaminant emission rates and potencies. Thus, emissions rates and
potencies provide the bulk of the discrimination between sites in
terms of long-term cancer risk. Both of these factors can be
reflected best in the HRS waste characteristics category. Emissions
rates are in fact currently reflected, albeit indirectly, in the
waste quantity factor. The carcinogenic potency of the compounds at
75
-------
the site, and hence the probable potency of emissions, can be
reflected in the toxicity factor. A method for reflecting
carcinogenic potency is discussed in DeSesso et al., 1986. It is
not unreasonable to conclude, therefore, that cancer risk in the air
pathway is best reflected via the waste characteristics category and
that no cancer risk evaluation table need be developed in the
targets category. This conclusion emphasizes the importance of
revising the waste characteristics category to better reflect cancer
risk and emission rates as discussed in DeSesso et al., 1986 and
Wolfinger, 1986.
Second, the results indicate that if cancer risk were to be
reflected in the targets category, then MEI risk should be the
principal cancer risk measure employed in the air pathway targets
category. However, the results of the analysis indicate that for
most sites, MEI risk would be considered acceptable by EPA (using
the guidance given in Thomas, 1984). Hence, the relative weight
given to MEI risk in the air pathway targets category, in comparison
to non-cancer risk measures, should be small.
Further, the analysis indicates that the overall population
incidence of cancer and the risk of developing cancer to average
exposed individuals are not of concern at most sites and should not
be reflected in the HRS target distance limit and population target
evaluation tables. Again, this conclusion is subject to the
uncertainty in near-site estimates.
76
-------
The results also have implications for the structure of any MEI
cancer risk evaluation table, if such a table were to be included in
the air pathway targets category. The magnitude of site-to-site
variation in MEI risk indicates that, assuming uniform potency
factors and emission rates, MEI risk does not differ significantly
between sites whenever the MEI lives within two miles of the site.
Also, the MEI risks that occur at two miles and beyond are probably
_o
inconsequential (e.g., less than 10 ). Further, current EPA
policy on MEI cancer risk de-emphasizes the number of maximally
exposed individuals in favor of the magnitude of the risk to each
MEI. Thus, the structure of an MEI cancer risk evaluation table, if
developed, should de-emphasize the number of MEIs in favor of their
probable exposure. This would indicate that an MEI risk target
evaluation table should be based on the distance to the nearest MEI
with a maximum value assigned whenever the MEI lives within two
miles and a minimum value (or zero) assigned whenever the MEI lives
further than three miles.
The best method of identifying the locations of maximally
exposed individuals, and hence estimating the distance to the
closest MEI, is to employ a dispersion model to estimate
concentrations at receptor locations corresponding to actual
residences (based, for example, on aerial photographs). The use of
such models is currently beyond the current scope of site
inspections and HRS evaluation. However, given information on the
77
-------
latitude and longitude of a site, the cost of employing HEM to
determine the distance to the MEI would be negligible, assuming EPA
made the model available.*
Alternately, the MEI cancer risk evaluation table could be
based on a surrogate for this distance, such as the distance to the
nearest residence. The use of this particular surrogate measure
would probably not provide sufficient discrimination among the sites
as probably few sites lack a residence within two miles.
Finally, the analysis provides little indication of the air
pathway target distance limit that should be employed for evaluating
non-cancer risks or for evaluating combined cancer and non-cancer
risks. Thus, the question of the target distance limit to be used
in the HRS air pathway is not resolved by this analysis. As stated
earlier, studies of non-carcinogenic risks are needed to establish
the target distance limit completely.
*The estimated cost is about two dollars per site in computer costs
and well less than two hours per site analyst time (Dusetzina,
1986).
78
-------
6.0 REFERENCES
American Conference of Governmental Industrial Hygienists, TLVs;
Threshold Limit Values and Biological Exposure Indices for
1985-1986, American Conference of Governmental Industrial
Hygienists, Cincinnati, OH, 1985.
American Meteorological Society, Air Quality Modeling and the Clean
Air Act: Recommendations to EPA on Dispersion Modeling for
Regulatory Applications, American Meteorological Society, Boston, MA,
1981.
Baker, Lynton W., An Evaluation of Screening Models for Assessing
Toxic Air Pollution Downwind of Hazardous Waste Landfills, Masters
Thesis, Office of Graduate Studies and Research, San Jose State
University, San Jose, CA, May 1985.
Battye, William et al., Preliminary Source Assessment for Hazardous
Waste Air Emissions from Treatment, Storage and Disposal Facilities
(TSDFs), (Draft Final Report), GCA Corporation, Bedford, MA,
February 1985.
Breton, Marc et al., Assessment of Air Emissions from Hazardous
Waste Treatment, Storage, and Disposal Facilities (TSDFs) -
Preliminary National Emissions Estimates, (Draft Final Report),
(GCA-TR-83-70-G), GCA Corporation, Bedford, MA, August 1983.
Breton, Marc et al., Evaluation and Selection of Models for
Estimating Air Emissions from Hazardous Waste Treatment, Storage,
and Disposal Facilities, (EPA-450/3-84-020), United States
Environmental Protection Agency, Research Triangle Park, NC,
December 1984.
Brubaker, K. L., P. Brown and R. R. Cirillo, Addendum to User's
Guide for Climatological Dispersion Model, (EPA-450/3-77-015),
United States Environmental Protection Agency, Research Triangle
Park, NC, 1977.
Busse, A. D. and J. R. Zimmerman, User's Guide for the
Climatological Dispersion Model, (EPA-R4-73-024), United States
Environmental Protection Agency, Research Triangle Park, NC, 1973.
Caravanos, Jack and Thomas T. Shen, "The Effect of Wind Speed on the
Emission Rates of Volatile Chemicals from Open Hazardous Waste Dump
Sites," Proceedings of the Fifth National Conference on Management
of Uncontrolled Hazardous Waste Sites, Held on November 7-9, 1984 in
Washington, DC, Hazardous Materials Control Research Institute,
Silver Spring, MD, 1984, pp. 68-71.
79
-------
Cupitt, Larry T., Fate of Toxic and Hazardous Materials in the Air
Environment, (EPA-600/3-80-084), United States Environmental
Protection Agency, Research Triangle Park, NC, August 1980.
DeSesso, John et al., Hazard Ranking System Issue Analysis; Toxicity
as a Ranking Factor, (MTR-86W128 Draft Report), The MITRE
Corporation, McLean, VA, September 1986.
Dusetzina, Michael, Strategies and Air Standards Division, Office of
Air Quality Planning and Standards, Environmental Response Team,
personal communications, December 1985 through August 1986.
Haemisegger, Elaine et al., The Air Toxics Problem in the United
States; An Analysis of Cancer for Selected Pollutants,
(EPA-450/1-85-001), United States Environmental Protection Agency,
Washington, DC, May 1985.
Hanna, Steven B., Gary A. Briggs, and Rayford P. Hosker, Jr.,
Handbook on Atmospheric Diffusion, (DOE/TIC11223), U.S. Department
of Energy, Washington, DC, 1982.
Hwang, Seong T., "Model Prediction of Volatile Emissions,"
Environmental Progress, Vol. 4, No. 2, May 1985, pp. 141-144.
Moon, Charles, Bureau of the Census, United States Department of
Commerce, personal communications, Fall 1985.
National Research Council, Risk Assessment in the Federal Government;
Managing the Process, National Academy of Sciences, Washington, DC,
1983.
Office of Technology Assessment, Assessment of Technologies for
Determining Cancer Risks from the Environment, Office of Technology
Assessment, Washington, DC, June 1981.
Shen, Thomas T., "Estimating Hazardous Air Emissions from Disposal
Sites," Pollution Engineering, Vol. 13, No. 8, August 1981,
pp. 31-34.
Shen, Thomas T., "Estimation of Organic Compound Emissions from
Waste Lagoons," Journal of the Air Pollution Control Association,
Vol. 32, No. 1, January 1982a, pp. 79-82.
Shen, Thomas T., "Air Quality Assessment for Land Disposal of
Industrial Wastes," Environmental Management, Vol. 6, No. 4, 1982b,
pp. 297-305.
80
-------
Smith, Maynard E., "Transport and Diffusion Modeling and Its
Application1980," Air Quality Modeling and the Clean Air Act;
Recommendations to EPA on Dispersion Modeling for Regulatory
Applications, American Meteorological Society, Boston, MA, 1981,
pp. 228-270.
Springer, Charles, Louis J. Thibodeaux and Shrikrisna Chatrathi,
"Simulation Study of the Volatilization of Polychlorinated Biphenyls
from Landfill Sites," Environment and Solid Wastes Characterization,
Treatment and Disposal, Francis, C. W. and S. E. Auerbach, eds.,
Proceedings of the Fourth Life Sciences Symposium, Environment and
Solid Wastes, Held in Gatlinburg, TN, on October 4-8, 1981,
Butterworth Publishers, Woburn, MA, 1983, pp. 209-222.
Thibodeaux, Louis J., "Estimating The Air Emissions of Chemicals
from Hazardous Waste Landfills," Journal of Hazardous Materials,
Vol. 4, 1981, pp. 235-244.
Thibodeaux, Louis J. et al., "Air Emission Monitoring of Hazardous
Waste Sites," Proceedings of the National Conference on Management
of Uncontrolled Hazardous Waste Sites, Held on November 29-
December 1, 1982 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1982, pp. 70-75.
Thomas, Lee M., EPA Memorandum on Determining Acceptable Risk Levels
for Carcinogens in Setting Alternate Concentration Levels Under RCRA
(Dated November 19, 1984), Bureau of National Affairs, Washington,
DC, November 23, 1984.
Travis, Curtis C. et al., "Cancer Risk Management," Environmental
Science and Technology, Vol. 21, No. 5, May 1987, pp. 415-420.
United States Environmental Protection Agency, Technical Facts
Concerning "Pool C" Adjoining the Kin-Buc Landfill, Edison, New
Jersey, United States Environmental Protection Agency, Washington,
DC, 1982.
United States Environmental Protection Agency, Standard Operating
Safety Guides, United States Environmental Protection Agency,
Washington, DC, November 1984a.
United States Environmental Protection Agency, Risk Assessment and
Management; Framework for Decision Making, United States
Environmental Protection Agency, Washington, DC, December 1984b.
81
-------
United States Environmental Protection Agency, Field Standard
Operating Procedures for Air Surveillance F.S.O.P. 8, (DraftT,
United States Environmental Protection Agency, Environmental
Response Team, Washington, DC, 1985.
United States Environmental Protection Agency, Guideline on Air
Quality Models (Revised), (EPA-450/2-78-027R), United States
Environmental Protection Agency, Research Triangle Park, NC, July
1986a.
United States Environmental Protection Agency, Superfund Public
Health Evaluation Manual, (EPA 540/1-86/060, OSWER Directive
9285.4-1), United States Environmental Protection Agency,
Washington, DC, October 1986b.
Wolfinger, Thomas F., Hazard Ranking System Issue Analysis; Options
for Revising the Air Pathway, (MTR-86W53 Draft Report), The MITRE
Corporation, McLean, VA, August 1986.
82
-------
APPENDIX A
SELECTED INFORMATION ON EMISSION RATES
Little data are available on contaminant emission rates for
uncontrolled waste sites. This appendix presents a compilation of
selected information on waste disposal site emission rates. This
information is presented to show the wide variation in emission
rates under differing assumptions and conditions. It is intended to
provide a context for the emission rate assumptions used in the
target distance analysis. This information is an amalgam of
sampling studies, laboratory simulation studies, and simple analyses
using emission rate equations that have been conducted by others.
Two characteristics of the information presented are important
when comparing emission rates. First, tne variation in the data is
large. The emission rate data range over eight orders of magnitude
-5 3
between contaminants, source types and studies (10 to 10
2
equivalent ng/cm /s). Benzene emission rates, for example, vary
over six orders of magnitude between studies. Second, all of the
available information relate only to relatively short-term emission
rates, at most one day in duration. None of the studies examined
addressed long term emissions sustainable for a period of 70 years,
the assumption required in the target distance analysis.
Table A-l presents conversion factors for use in comparing
rates in different units. Tables A-2 through A-10 present the
emission estimates and other supporting information.
83
-------
TABLE A-l
CONVERSION FACTORS
One Kilogram Per Year is Equivalent to:
1,000 Grams per year
0.1142 Grams per hour
3.17 x 10~5 Grams per second
31.71 Micrograms per sec
o o
3.17 x 10~° Micrograms per m^ per sec*
3.17 x 10~^ Nanograms per cm^ per sec*
*Assumes source size is 10,000 m -
84
-------
TABLE A-2
LANDFILL STUDY 1
Source; Shen, Thomas T., "Estimating Hazardous Air Emissions
from Disposal Sites," Pollution Engineering, Vol. 13,
No. 8, August 1981, pp. 31-34.
Source Type; Landfill
Estimation
Approach; Calculations; simplified emission rate estimation
equation
Assumptions; Aroclor soil concentration = 5,000 ppm
Wind speed = 4 m/s
Size = 40m x 40m
Cover depth = 10cm
Cover porosity = 0.4
Emission
Estimates; Aroclor 1242 emission rate
uncovered 184 ug/s
covered 8 ug/s
85
-------
TABLE A-3
LANDFILL STUDY 2
Source:
Estimation
Approach;
Assumptions;
Springer, Charles, Louis J. Thibodeaux and Shrikrisna
Chatrathi, "Simulation Study of the Volatilization of
Polychlorinated Biphenyls from Landfill Sites,"
Environment and Solid Wastes Characterization,
Treatment and Disposal, Francis, C. W. and S. E.
Auerbach, eds., Proceedings of the Fourth Life
Sciences Symposium, Environment and Solid Wastes, Held
in Gatlinburg, TN, on October 4-8, 1981, Butterworth
Publishers, Woburn, MA, 1983, pp. 209-222.
Source Type; Landfill
Simulation study employing emission estimation
equations
Liquid Aroclor 1248
Cell porosity =0.2
Cell depth =3.0 meters
Cell density =2.0 g/cm3
Aroclor 1248 partial pressure
1.3 E-6 atm.
Aroclor 1248 in sludge
Cell porosity =0.2
Cell depth = 10.0 meters
Aroclor 1248 partial pressure
5.3 E-7 atm.
Emission
Estimates:
Liquid Aroclor 1248
Cap Parameters
Porosity Depth (m)
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.2
0.08
0.06
3.0
3.0
3.0
3.0
1.0
0.5
3.0
3.0
3.0
3.0
Gas Rate
(cnrVg/s)
3 E-7
2 E-7
1 E-7
3 E-7
3 E-7
3 E-7
3 E-7
3 E-7
3 E-7
3 E-7
Flux
(mg/m2/d)
2.42
1.6
0.8
2.42
2.46
2.51
2.42
2.65
2.40
2.31
86
-------
TABLE A-3 (Concluded)
Emission
Estimates;
(Concluded)
Cap Parameters
Porosity Depth (m)
0.06
0.06
0.1
0.08
0.2
0,
0,
1.0
0.5
3.0
3.0
3.0
.0
.0
3.
1.
Liquid Aroclor 1248
Gas Rate Flux
0.1
0.5
3 E-7
3 E-7
0.0
0.0
0.0
0.0
0.0
0.0
(mg/m2/d)
2.36
2.39
0.08
0.05
0.33
0.08
0.24
0.45
Cap Parameters
Porosity Depth (m)
Aroclor 1248 in sludge
Gas Rate Flux
(cm3/g/s) (mg/m2/d)
0.1
0.08
0.2
0.1
0.1
0.1
1.
1,
1.
1.0
0.5
2.5
5 E-8
5 E-8
5 E-8
5 E-8
5 E-8
5 E-8
0.53
0.30
0.91
0.53
0.62
0.48
87
-------
TABLE A-4
LANDFILL STUDY 3
Source:
Estimation
Approach;
Assumptions;
Emissions
Estimates:
Thibodeaux, Louis J., "Estimating The Air Emissions of
Chemicals from Hazardous Waste Landfills," Journal of
Hazardous Materials, Vol. 4, 1981, pp. 235-244.
Source Type; Landfill
Calculations using emission estimation equations
Temperature = 25 degrees C
Wind speed =8.2 mi/h
Friction velocity =0.5 m/s
Cover porosity = 0.51
Tortuosity = 1.71
Size = 100m x 100m
Cover depth = 1m
Gas velocity = 1.63 E-3 cm/s
Contaminant concentration in air spaces (g/m^)
Benzene = 399.3
Chloroform = 1282
Vinyl Chloride = 2556
Aroclor 1248 =7.76 E-3
Chemical
Benzene
Chloroform
Vinyl Chloride
Aroclor 1248
Emissions Rate (g/m2/d)
Without Gas Generation With Gas Generation
89.4
340
826
9.5 x lO
563
1820
3650
109 x 10~4
88
-------
TABLE A-5
LANDFILL STUDY 4
Source; Shen, Thomas T., "Air Quality Assessment for Land Disposal
of Industrial Wastes," Environmental Management, Vol. 6,
No. 4, 1982, pp. 297-305.
Source Type; Landfill
Estimation
Approach; Calculations using emission estimation equations
Assumptions; Location; Caputo Landfill, New York
Aroclor 1242 soil concentration = 5,000 ppm
Cover depth = 50.8 cm
Cover layers; manure, paper mill sludge and topsoil
Emission
Estimates; Aroclor 1242 emission rate
before cover 5678 ug/s
after cover 34.6 ug/s
89
-------
TABLE A-6
LANDFILL STUDY 5
Source:
Baker, Lynton W., An Evaluation of Screening Models for
Assessing Toxic Mr Pollution Downwind of Hazardous Waste
Landfills, Masters Thesis, Office of Graduate Studies and
Research, San Jose State University, San Jose, CA, May
1985.
Source Type; Landfill
Estimation
Approach;
Assumptions;
Emission
Estimates:
Calculations using emission emission equations under
different atmospheric conditions
Location: BKK Landfill, California
Temperature =20-30 degrees C
Size = 583 acres, 228 acres containing wastes
Exposed area = 0.31 x 108 - 2.6 x 108 cm2
Cover depth = 10 ft
Cover porosity = 0.2
Vinyl Chloride Emission Rate (g/s)
Atmospheric Conditions
Date
3/84
8/84
Site
A
B
A
B
Hillside
Drainage
0.197
0.023
0.279
0.033
Sea
Breeze
0.126
0.179
Valley
Drainage
0.141
0.199
90
-------
TABLE A-7
LANDFILL STUDY 6
Source:
Hwang, Seong T., "Model Prediction of Volatile Emissions,1
Environmental Progress, Vol. 4, No. 2, May 1985,
pp. 141-144.
Source Type; Landfill, Evaporation Pond, and Land Treatment Facility
Estimation
Approach;
Measurements and calculations
Assumptions; Parameters
Wind Speed (mph)
Size
Surface area (ft^)
Depth (ft)
Waste Composition*
Toluene
1,1,1-Trichloroethane
Methylene Chloride
Tetrachloroethylene
Benzene
Chi or ob enz en e
Landfill
11-19
231,000
34
0.749
0.276
0.456
Pond
4-5
5400
7.5
1.05
2.53
Land Treatment
7-10
10 (acres)
1.22 E-2
6.25 E-3
1.04 E-3
Emission
Estimates:
Parameters
Toluene
measured
predicted
1,1,1-Trichloroe thane
measured
predicted
Methylene Chloride
measured
predicted
Rates (ug/m^/s)
Landfill Pond Land Treatment**
1.48
0.93
0.4
4.04
18.4
13.5
15.3
15.2
4.7
11
*Units: Landfill, ug/g; Pond, ug/ml; Land Treatment, g/cm3.
**Application rate 3.5 gr/cm^, emission rate estimated at 70 hours
after waste application.
91
-------
TABLE A-7 (Concluded)
Emission Rates (ug/m2/s)
Estimates; Parameters Landfill Pond Land Treatment*
(Concluded)
Tetrachloroethylene
measured 0.76
predicted 0.32
Benzene
measured 1.1
predicted 7
Chlorobenzene
measured 2.4
predicted 6
*Application rate 3.5 gr/cm2, emission rate estimated at 70 hours
after waste application.
92
-------
TABLE A-8
SURFACE IMPOUNDMENT STUDY 1
Source; Shen, Thomas T., "Estimation of Organic Compound Emissions from
Waste Lagoons," Journal of the Air Pollution Control
Association, Vol. 32, No.l, January 1982, pp. 79-82.
Source Type; Surface Impoundment
Estimation
Approach; Calculations; simplified emission rate estimation equation
Assumptions; Temperature =20-30 degrees C
Lagoon Size = 25m x 40m x 3.5m
Benzene concentration in lagoon = 100 mg/1
Wind speed = 3 m/s
Emission
Estimates; Peak Short Term Benzene Emissions Rate = 5.5 g/s
93
-------
TABLE A-9
SURFACE IMPOUNDMENT STUDY 2
Source:
Thibodeaux, Louis J. et al., "Air Emission Monitoring of
Hazardous Waste Sites," Proceedings of the National Conference
on Management of Uncontrolled Hazardous Waste Sites, Held on
November 29-December 1, 1982 in Washington, DC, Hazardous
Materials Control Research Institute, Silver Spring, MD, 1982,
pp. 70-75.
Source Type: Surface Impoundments
Estimation
Approach;
Assumptions;
Pond 1,2
Pond 6
Back-calculation of emission rates from ambient data using
concentration profile technique
Size = 85 meter x 253 meter with a 1.2 - 1.5 meter berm
Eight surface aerators
Surface area = 21,600 m2
Volume = 68,400 nH
Depth =3.2 meters
Wind speed = 151 - 360 cm/s
Contaminant concentrations in waste (ug/1):
Benzene
Toluene
Total Hydrocarbons
1,1-Dichloroe thane
Total Chlorinated
Hydrocarbons
0.31
4.1 _
8.4 + 2
34+7.7
0.31
4.7
8
207 + 37
No surface aerators
Surface area = 4650 m2
Volume = 4630 m3
Depth = 1 meter
Wind speed = 159 - 325 cm/s
Contaminant concentrations in waste (ug/1):
Benzene
Toluene
Total Hydrocarbons
1,1-Dichloroe thane
Total Chlorinated
Hydrocarbons
16 + 9.5
43 + 3.0
125
22 + 19
267
94
-------
TABLE A-9 (Concluded)
Emission Rate:
Contaminant
Pond 1,2
Benzene
Toluene
Total Hydrocarbons
1,1-Dichloroethane
Total Chlorinated
Hydrocarbons
Pond 6
Benzene
Toluene
Total
Hydrocarbons
1,1-Dichloroethane
Total Chlorinated
Hydrocarbons
Calculated Flux
+0.0051
+0.062
+0.14
+0.48
+2.7
+0.047
+0.12
+0.37
+0.058
+0.064
Measured Flux* Rate
(ng/cm2/s) (kg/d)
-0.29 - +0.06 0.095
-0.038 - +0.015 1.2
-1.0 - +1.1 2.6
-0.022 - +0.04 9.0
-0.33 - +0.15 50.0
0.095 0.38**
0.014 0.48
1.3 5.2**
0.028 0.11**
0.28 1.1**
*Negatlve flux values Indicate migration of the contaminant out of
the air and into the pond.
**Based on measured rates. All others based on calculated rates.
95
-------
TABLE A-10
CONTAMINATED SOIL STUDY 1
Source:
Caravanos, Jack and Thomas T. Shen, "The Effect of Wind Speed on
the Emission Rates of Volatile Chemicals from Open Hazardous
Waste Dump Sites," Proceedings of the Fifth National Conference
on Management of Uncontrolled Hazardous Waste Sites, Held on
November 7-9, 1984 in Washington, DC, Hazardous Materials
Control Research Institute, Silver Spring, MD, 1984, pp. 68-71.
Source Type; Contaminated soil
Estimation
Approach!
Assumptions;
Laboratory experiment employing saturated soil in steel
evaporation pans
Soil porosity (%)
Soil density (g/ml)
Clay
48
1.34
Sand
32
1.59
Tops oil
51
0.96
Emission
Estimates:
Contaminant
Benzene
Carbon
Tetrachloride
Trichloro-
ethylene
Wind
Speed (mph)
0.5
2.5
5.0
0.5
2.5
5.0
0.5
2.5
5.0
Emission Rate (g/min)
Clay Sand Topsoil
0.9
2.2
4.5
1.5
4.2
9.5
0.9
2.7
6.0
0.9
2.2
3.7
1.7
4.0
7.3
0.8
2.6
5.3
0.7
2.3
4.6
1.4
4.8
11.1
0.8
3.1
7.9
96
------- |