United States Office of Solid Waste EPA530-R-99-019b
Environmental Protection Agency Washington, DC 20460 November 1999
svEPA Revised Risk Assessment for the
Air Characteristic Study
Volume II
Technical Background Document
-------
EPA530-R-99-019b
November 1999
Revised Risk Assessment for the Air
Characteristic Study
Volume II
Technical Background Document
Off ice of Solid Waste
U.S. Environmental Protection Agency
Washington, DC 20460
Printed on Recycled Paper
-------
Table of Contents
Section Page
1.0 Introduction 1-1
1.1 Purpose and Requirements of the Air Characteristic Study 1-1
1.2 Overview of Risk Assessment 1-2
1.3 Organization of Report 1-6
1.4 Companion Documents 1-6
2.0 Modeling Approach and Data Sources 2-1
2.1 Overview of Modeling Approach 2-1
2.2 Conducting the Analysis 2-5
2.3 Data Sources 2-10
3.0 Waste Source Characteristics 3-1
3.1 Estimation of Missing or Replacement Data for WPs, LAUs, and LFs 3-1
3.1.1 Wastepile Height 3-2
3.1.2 Landfill Depth and Capacity 3-2
3.1.3 Waste Quantity Replacement Values 3-5
3.2 Waste Management Unit Data and Size Categories Used in
Dispersion Model 3-5
3.3 Removal ofWastepiles Containing Bevill-Excluded Wastes 3-10
3.3.1 Identification of Bevill Industrial D Facilities 3-11
3.3.2 Bevill Facilities Culled from Industrial D Data 3-12
3.4 Characterization of Tanks 3-12
3.4.1 Tank Classification 3-14
3.4.2 Additional Tank Data Used for Imputation 3-17
3.4.3 Estimation of Missing Data 3-19
4.0 Source Emission Estimates 4-1
4.1 Model Selection 4-1
4.1.1 Volatile Emission Model Selection 4-1
4.1.2 Particulate Emission Model Selection 4-3
4.2 Emission Model Input Parameters 4-3
4.2.1 Chemical-Specific Input Parameters 4-3
4.2.2 Critical Input Parameters for Land-Based WMU Emission
Models 4-4
4.2.3 Critical Parameters for Tank Emissions Model 4-7
4.3 Modifications to CHEMDAT8 for Land-Based Units 4-12
4.4 Development of Volatile Emissions and Waste Concentrations for Landfills 4-15
4.5 Development of Volatile Emissions and Waste Concentrations for
Land Application Units 4-19
4.5.1 Chronic Exposure Analysis 4-19
4.5.2 Acute and Subchronic Exposure Analysis 4-25
in
-------
Table of Contents (continued)
Section Page
4.6 Development of Volatile Emissions and Waste Concentrations for
Wastepiles 4-26
4.6.1 Chronic Exposure Analysis 4-26
4.6.2 Acute and Subchronic Exposure Analysis 4-28
4.7 Development of Volatile Emissions for Tanks 4-28
4.8 Development of Particulate Emissions 4-30
4.8.1 Landfills and Land Application Units 4-30
4.8.2 Wastepiles 4-32
5.0 Dispersion Modeling 5-1
5.1 Overview of Approach 5-1
5.2 Model Selection 5-2
5.3 Critical Parameters 5-2
5.3.1 General Assumptions 5-3
5.3.2 Meteorological Stations and Data 5-4
5.3.3 Source Release Parameters 5-10
5.3.4 Receptors 5-11
5.4 Unitized Air Concentrations 5-14
6.0 Development of Inhalation Health Benchmarks 6-1
6.1 Chronic Inhalation Health Benchmarks Used in This Study 6-1
6.1.1 Alternate Chronic Inhalation Health Benchmarks Identified 6-1
6.1.2 Chronic Inhalation Health Benchmarks Derived for This Study .... 6-7
6.2 Subchronic Inhalation Health Benchmarks 6-8
6.3 Acute Inhalation Health Benchmarks 6-14
6.4 Comparison of Chronic, Subchronic, Acute RfCs 6-17
7.0 Development of Risk-Specific Waste Concentration Distribution 7-1
7.1 Overview 7-1
7.1.1 Chronic Exposures to Carcinogens 7-1
7.1.2 Chronic Exposures to Noncarcinogens 7-2
7.1.3 Acute/Subchronic Exposures 7-2
7.1.4 Adult Exposure Approach 7-2
7.1.5 Child Exposure Approach 7-3
7.1.6 Specific Steps Required in the Risk Analysis 7-3
7.2 Select Receptor Location 7-4
7.2.1 Chronic Exposure 7-4
7.2.2 Acute/Subchronic Exposures 7-4
7.3 Obtain Unitized Air Concentrations 7-5
7.4 Calculate Air Concentration 7-5
7.5 Select Exposure Factors 7-6
7.5.1 Inhalation Rate 7-7
IV
-------
Table of Contents (continued)
Section Page
7.5.2 Body Weight 7-7
7.5.3 Exposure Duration 7-7
7.5.4 Exposure Frequency 7-7
7.6 Obtain Health Benchmarks 7-8
7.7 Calculate Risk or Hazard Quotient 7-8
7.8 Backcalculate Risk-Specific Waste Concentration 7-10
7.9 Adjust Tank Results for Non-Linearity of Biodegradation 7-11
7.10 Flag Results that Exceed Soil Saturation Concentration of Solubility 7-11
7.10.1 Organic-Phase Emissions Higher than Aqueous-Phase Emissions . 7-12
7.10.2 Backcalculated Concentrations Exceed Physical Limitations on
Aqueous Phase 7-12
7.11 Derivation of Exposure Factors 7-13
7.11.1 Inhalation Rate 7-13
7.11.2 Body Weight 7-14
7.11.3 Exposure Duration 7-17
7.11.4 Exposure Frequency 7-19
7.12 Stability Analysis of the Monte Carlo Model 7-19
7.13 Modifications to Methodology for Lead 7-20
8.0 Analysis of Variability and Uncertainty 8-1
8.1 Variability 8-1
8.1.1 Source Characterization and Emissions Modeling 8-2
8.1.2 Fate and Transport Modeling 8-3
8.1.3 Exposure Modeling 8-3
8.2 Uncertainty 8-4
8.2.1 Scenario Uncertainty 8-4
8.2.2 Model Uncertainty 8-6
8.2.3 Parameter Uncertainty 8-7
9.0 References 9-1
Appendix A - Basic Dalenius-Hodges Procedure for Constructing Strata A-l
Appendix B - Chemical-Specific Data B-l
Appendix C - Sensitivity Analysis of ISC Air Model C-l
Appendix D - Derivation of Chronic Inhalation Noncancer and Cancer Health Benchmarks D-l
-------
List of Figures
Figure Page
2-1 Conceptual diagram of a waste site 2-2
2-2 Model framework 2-6
2-3 Combination of results for individual WMUs into a distribution across all WMUs . . . 2-9
3-1 Landfill characteristics from Industrial D Screening Survey 3-4
3-2 Wastepile characteristics from Industrial D Screening Survey 3-6
3-3 Land application unit characteristics from Industrial D Screening Survey 3-7
3-4 Comparison of tank depth regression lines 3-20
5-1 Meteorological station regions and landfill locations 5-6
5-2 Meteorological station regions and LAU locations 5-7
5-3 Meteorological station regions and wastepile locations 5-8
5-4 Meteorological station regions and tank locations 5-9
5-5 Air concentration vs. size of area source 5-12
6-1 Approach used to select chronic inhalation health benchmark values 6-2
6-2 Approach used to select subchronic noncancer inhalation benchmark values 6-14
6-3 Approach used to select acute noncancer inhalation health benchmark values 6-17
7-1 Linearity adjustments for tanks with biodegradation 7-11
7-2 Stability analysis results showing typical variations resulting in change in results
at one significant figure 7-20
VI
-------
List of Tables
Table Page
1-1 Constituents Modeled for All WMUs 1-3
1-2 Constituents Modeled for Tanks Only 1-5
3-1 Summary of Land-Based WMUs Removed from Industrial D Survey Database 3-3
3-2 Assignment of Wastepile Height 3-3
3-3 Area Strata Modeled for Landfills and Land Application Units 3-9
3-4 Areas and Source Heights Modeled for Wastepiles 3-10
3-5 Areas and Source Heights Modeled for Tanks 3-10
3-6 Summary of Tanks Removed from TSDR Survey Database 3-13
3-7 TSDR Survey Wastewater Treatment Codes Used in Identifying Treatment Tanks . . 3-15
3-8 Numbers of Tanks by Classification 3-17
3-9 Summary of Tank Size Information Collected by EPA Site Visits for RCRA
Air Emission Standards 3-18
3-10 Summary of Depth Imputation Techniques 3-21
4-1 CHEMDAT8 Land-Based Unit Model Input Requirements 4-5
4-2 CHEMDAT8 Tank Model Input Requirements 4-8
4-3 Summary of Mechanical Aerator Information Collected by EPA Site Visits for
RCRA Air Emission Standards 4-9
4-4 Relationship Between Frequency of Application and Waste Quantity 4-22
4-5 Estimated Frequency of Application Using Industrial D Data 4-22
4-6 Inputs and Intermediate Values Used for Wind Erosion from Landfills and LAUs . . 4-32
4-7 Calculated Paniculate Emission Rates for Landfills and LAUs 4-33
4-8 Calculated Particulate Emission Rates for Wastepiles 4-34
5-1 Air Dispersion Model Capabilities 5-2
5-2 Meteorological Stations Used in the Air Characteristic Study 5-5
5-3 Areas Modeled for Landfills and Land Application Units 5-13
5-4 Areas and Source Heights Modeled for Wastepiles 5-13
5-5 Areas and Source Heights Modeled for Tanks 5-14
6-1 Chronic Inhalation Health Benchmarks Used in the Air Characteristic Analysis 6-3
6-2 Alternate Chronic Inhalation Health Benchmarks 6-7
6-3 Chronic Inhalation Health Benchmarks Derived for This Study 6-9
6-4 Subchronic Inhalation Health Benchmarks 6-11
6-5 ATSDR Acute Inhalation MRLs 6-15
6-6 CalEPA's 1-Hour Acute Inhalation Reference Exposure Levels (RELs) 6-15
7-1 Estimated Parameters for Inhalation Rate for Residents Assuming Lognormal
Distribution 7-15
7-2 Recommended Inhalation Rates for Workers 7-15
7-3 Body Weights for Adults, Ages 18-74 Years (kg) 7-15
vii
-------
List of Tables (continued)
Table Page
7-4 Body Weights for Children, Ages 6 Months to 18 Years (kg) 7-16
7-5 Descriptive Statistics for Residential Occupancy Period by Age (years) 7-18
7-6 Estimated Weibull Parameters for Exposure Duration 7-18
7-7 Summary of Stability Analysis 7-20
7-8 Summary of Inputs for IEUBK Model 7-22
7-9 Results of IEUBK Modeling 7-23
Vlll
-------
Volume II Section 1.0
1.0 Introduction
The U.S. Environmental Protection Agency (EPA), Office of Solid Waste (OSW), has
analyzed the potential risks to human health posed by the inhalation of vapor (gaseous) and
particulate (nongaseous) air emissions from a set of chemicals and metals when managed in
certain waste management units. An analysis of these risks was initially performed in 1998 as
part of the Air Characteristic Study (U.S. EPA, 1998a). In accordance with Agency policy, the
risk assessment conducted for the 1998 Air Characteristic Study was peer reviewed to ensure that
science was used credibly and appropriately in the work performed. Based on comments made
by the peer reviewers, EPA has revised the original risk assessment.
This report presents the revised risk assessment in three volumes. This document is
Volume II, the Technical Background Document. This volume provides a detailed description of
the methodologies, data, and supporting analyses used for the risk assessment. A discussion of
the changes made from the 1998 Air Characteristic Study, a general overview of the risk
assessment, and the integration of the revised risk assessment results with the May 1998
regulatory gaps and occurrence analyses can be found in Volume I, Revised Risk Analysis for the
Air Characteristic Study: Overview. The complete results of the analysis are presented in
Volume III, Revised Risk Analysis for the Air Characteristic Study: Results (on CD-ROM).
1.1 Purpose and Requirements of the Air Characteristic Study
This report and the 1998 Air Characteristic Study are among the initial steps for EPA in
fulfilling a long-standing goal to review the adequacy and appropriateness of the hazardous waste
characteristics. The first step in achieving this goal was the Hazardous Waste Characteristic
Scoping Study (U.S. EPA, 1996) that the Agency completed November 15, 1996, under a
deadline negotiated with the Environmental Defense Fund. This study was conducted to identify
potential gaps in the current hazardous waste characteristics, as well as other modifications and
updates that are necessary to ensure that the definition of characteristics is complete, up-to-date,
and based on state-of-the-art methodologies. Based on the initial bounding analysis of potential
risks due to air emissions done as part of the Scoping Study, as well as follow-up analysis on
potential gaps in regulatory coverage under the Clean Air Act and Subpart CC of the Resource
Conservation and Recovery Act (RCRA), OSW identified air emissions from waste management
units as one of the areas meriting further analysis.
The Air Characteristic Study addresses this area by examining the potential direct
inhalation risks due to emissions from certain waste management units. On May 15, 1998, in
accordance with a consent decree, EPA completed the first portion of the study. According to the
consent decree with EOF, a second part of the Air Characteristic Study, covering surface
impoundments receiving wastewaters that never exhibited a characteristic, will be completed
~
-------
Volume II Section 1.0
March 26, 2001. The purpose of the 1998 Air Characteristic Study, as outlined by the consent
decree, was to investigate gaps in the current hazardous waste characteristics and CAA programs.
In addition, resulting potential risks to human health posed by the inhalation of air emissions
from wastes managed in certain waste management units were to be investigated.
The 1998 Air Characteristic Study has three components: an evaluation of the coverage
and potential regulatory gaps in RCRA Subtitle C and the CAA, a risk analysis of air emissions
from waste management units, and an evaluation of the occurrence of these constituents in
nonhazardous industrial waste. The risk assessment component has undergone a peer review,
and EPA has made a number of changes to the risk assessment based on peer reviewer
comments. In addition, other revisions have been made based on public comments and
improvements initiated by the Agency. Since the other components of the May 1998 Air
Characteristic Study have not been revised, those analyses are not covered in this document. The
original results of these analyses are used in Volume I of this report to present the significant
findings from the integration of the revised risk assessment results with the regulatory gaps and
occurrence analyses.
1.2 Overview of Risk Assessment
The risk assessment described in this document is a national analysis designed to assess
the potential human health risk attributable to inhalation exposures when certain chemicals and
metals are managed as waste in certain types of waste management units. The purpose of the
analysis is to determine which chemicals and waste management units are of potential national
concern purely from a risk perspective; it is not intended to draw conclusions concerning
regulatory coverage. This information, combined with preliminary information on regulatory
coverage and on the presence of these chemicals in nonhazardous waste, will be useful in
determining the need for expanded regulatory coverage. Specifically, the purpose of this study is
to provide technical information on the potential risk from WMU emissions to assist EPA in
determining the need to expand regulatory coverage in the future.
The analysis presented in this report addresses specific chemicals that when managed as a
waste may pose a risk through direct inhalation exposures. Tables 1-1 and 1-2 list the chemicals
and metals included in this analysis. The analysis is structured so that the results of the risk
assessment are the concentrations of each constituent that can be present in each type of WMU
and still be protective of human health. The protective concentrations in waste were developed
for three types of receptors: adult residents, child residents, and workers. Three risk
endpoints—chronic (1 year), subchronic (1 month), and acute (1 day)—were evaluated.
The protective waste concentrations were estimated by modeling the emissions from a
waste management unit, the transport through the ambient environment, and the exposure to a
receptor to backcalculate a threshold concentration in a waste below which the risk to human
health would fall below a pre-established threshold. The waste management scenario modeled in
this analysis is storage, disposal, or treatment of industrial waste streams in RCRA subtitle D
waste management units.
1-2
-------
Volume II
Section 1.0
Table 1-1. Constituents Modeled for All WMUs
Constituent
CAS No.
Acetaldehyde [ethanal]
Acetone [2-propanone]
Acetonitrile [methyl cyanide]
Acrolein
Acrylonitrile
Allyl chloride
Arsenic
Barium
Benzene
Beryllium
Bromodichloromethane [dichlorobromomethane]
Bromoform [tribromomethane]
Bromomethane [methyl bromide]
1,3-Butadiene
Cadmium
Carbon disulfide
Carbon tetrachloride
Chlorobenzene
Chlorodibromomethane [dibromochloromethane]
Chloroform
Chloromethane [methyl chloride]
Chloroprene [2-chloro-1,3-butadiene]
Chromium VI
Cobalt
Cumene [isopropyl benzene]
Cyclohexanol
l,2-Dibromo-3-chloropropane
1,2-Dichlorobenzene [o-dichlorobenzene]
1,4-Dichlorobenzene [p-dichlorobenzene]
Dichlorodifluoromethane [CFC-12]
1,2-Dichloroethane [ethylene dichloride]
1,1-Dichloroethylene [vinylidene chloride]
1,2-Dichloropropane [propylene di chloride]
cis-1,3 -Di chl oropropy lene
trans-1,3 -Di chl oropropy 1 ene
1,4-Dioxane [1,4-diethyleneoxide]
Epichlorohydrin [l-chloro-2,3-epoxypropane]
1,2-Epoxybutane
2-Ethoxyethanol [ethylene glycol monoethyl ether]
2-Ethoxyethanol acetate [2-EEA]
Ethylbenzene
Ethylene dibromide [1,2-dibromoethane]
75-07-0
67-64-1
75-05-8
107-02-8
107-13-1
107-05-1
7440-38-2
7440-39-3
71-43-2
7440-41-7
75-27-4
75-25-2
74-83-9
106-99-0
7440-43-9
75-15-0
56-23-5
108-90-7
124-48-1
67-66-3
74-87-3
126-99-8
7440-47-3
7440-48-4
98-82-8
108-93-0
96-12-8
95-50-1
106-46-7
75-71-8
107-06-2
75-35-4
78-87-5
10061-01-5
10061-02-6
123-91-1
106-89-8
106-88-7
110-80-5
111-15-9
100-41-4
106-93-4
(continued)
1-3
-------
Volume II Section 1.0
Table 1-1. (continued)
Constituent CAS No.
Ethylene oxide 75-21-8
Formaldehyde 50-00-0
Furfural 98-01-1
Hexachloroethane 67-72-1
w-Hexane 110-54-3
Lead 7439-92-1
Manganese 7439-96-5
Mercury 7439-97-6
Methanol 67-56-1
2-Methoxyethanol 109-86-4
2-Methoxyethanol acetate [2-MEA] 110-49-6
Methyl tert-butyl ether 1634-04-4
Methylene chloride [dichloromethane] 75-09-2
Methyl ethyl ketone [2-butanone][MEK] 78-93-3
Methyl isobutyl ketone [hexone] [4-methyl-2-pentanone] 108-10-1
Methyl methacrylate 80-62-6
Naphthalene 91-20-3
Nickel 7440-02-0
2-Nitropropane 79-46-9
N-Nitrosodi-w-butylamine 924-16-3
N-Nitrosodiethylamine 55-18-5
N-Nitrosopyrrolidine 930-55-2
Propylene oxide 75-56-9
Pyridine 110-86-1
Styrene 100-42-5
1,1,1,2-Tetrachloroethane 63 0-20-6
1,1,2,2-Tetrachloroethane 79-34-5
Tetrachloroethylene [perchloroethylene] 127-18-4
Toluene 108-88-3
1,1,1 -Trichloroethane [methyl chloroform] 71-55-6
1,1,2-Trichloroethane [vinyl trichloride] 79-00-5
Trichloroethylene 79-01-6
Trichlorofluoromethane [trichloromonofluoromethane] 75-69-4
l,l,2-Trichloro-l,2,2-trifluoroethane [freon 113] 76-13-1
Triethylamine 121-44-8
Vanadium 7440-62-2
Vinyl acetate 108-05-4
Vinyl chloride 75-01-4
Xylenes, mixed isomers [xylenes, total] 1330-20-7
1-4
-------
Volume II Section 1.0
Table 1-2. Constituents Modeled for Tanks Only
Constituent CAS No.
Acrylamide 79-06-1
Acrylic acid 79-10-7
Aniline 62-53-3
Benzidine 92-87-5
Benzo(a)pyrene 50-32-8
2-Chlorophenol [o-chlorophenol] 95-57-8
Cresols, total 1319-77-3
7,12-Dimethylbenz[a]anthracene 57-97-6
N,N-Dimethyl formamide 68-12-2
3,4-Dimethylphenol 95-65-8
2,4-Dinitrotoluene 121-14-2
1,2-Diphenylhydrazine 122-66-7
Ethylene glycol 107-21-1
Hexachlorobenzene 118-74-1
Hexachloro-1,3-butadiene [hexachlorobutadiene] 87-68-3
Hexachlorocyclopentadiene 77-47-4
Isophorone 78-59-1
3-Methylcholanthrene 56-49-5
Nitrobenzene 98-95-3
Phenol 108-95-2
Phthalic anhydride 85-44-9
2,3,7,8-TCDD [2,3,7,8-tetrachlorodibenzo-^-dioxin] 1746-01-6
o-Toluidine 95-53-4
1,2,4-Trichlorobenzene 120-82-1
Emissions, transport, and exposure were modeled somewhat differently for the three risk
endpoints (chronic, subchronic, and acute). For emissions and transport, different averaging
times were used for each endpoint (one year for chronic, one month for subchronic, and one day
for acute) to generate emission rates and dispersion factors. For exposure, subchronic and acute
exposures were modeled deterministically, using the point of maximum exposure at a specific
distance. Chronic exposures were modeled probabilistically using a Monte Carlo approach to
capture variation in receptor location and exposure factors. The waste management units
assessed are aerated treatment tanks, nonaerated treatment tanks, storage tanks, landfills, waste
piles, and land application units. The risk assessment was structured to capture national
variations in environmental settings. In addition, Monte Carlo analysis was used in the modeling
to include the variations in receptor characteristics such as exposure parameters and location
around the facility. The following section, Section 2, provides a more detailed overview of the
risk analysis framework.
1-5
-------
Volume II Section 1.0
1.3 Organization of Report
The remainder of this report is organized as follows. Section 2 presents a brief overview
of the modeling approach. Section 3 discusses waste management unit data and how they were
used in the analysis. Section 4 describes the development of emissions for each unit type.
Development of dispersion coefficients and health benchmarks are discussed in Sections 5 and 6,
respectively. Section 7 describes the development of risk-specific waste concentration
distributions. Section 8 discusses how uncertainty and variability are addressed in this analysis.
References are provided in Section 9, and supporting analyses are included in Appendixes A
through D.
1.4 Companion Documents
Volume I of this report, the Overview, contains a general discussion of the risk
assessment and provides detailed information on the changes that were made to the risk
assessment performed in 1998. In addition, Volume I presents significant findings from the
integration of the revised risk assessment results with the May 1998 regulatory gaps and
occurrence analyses.
Volume III of this report, Results (provided on CD-ROM), presents the detailed results of
the risk analysis.
1-6
-------
Volume II
Section 2.0
2.0 Modeling Approach and Data Sources
This section provides a general overview of the modeling approach and primary sources
of data and describes how the risk analysis was conducted. Further detail on the methodology,
models used, and a complete set of inputs and associated references are provided in Sections 3
through 8 of this document.
2.1 Overview of Modeling Approach
The overall goal of this risk analysis is to
estimate the concentrations of constituents that can be
present in a waste management unit (WMU) and
remain protective of human health. These protective
waste concentrations were calculated for 104
constituents1 including volatiles, semi-volatiles, and
metals. These constituents were selected for their
potential to result in risk from inhalation exposure.
Workers, adults, and children were evaluated for three
different types of exposures or risk endpoints: chronic
(over 1 year), subchronic (1 month), and acute (1 day).
Estimating protective concentrations required a
multistep modeling process that could relate the
concentrations in ambient air at a receptor point that
could create a health effect to a concentration in the
waste management unit. To achieve this, the analytical
approach for this analysis is based on three primary
components:
# Emissions modeling—characterizing
emissions from a WMU
# Dispersion modeling—describing the
transport of these emissions through the
ambient environment
The Air Characteristic Study addresses:
# 105 constituents
# 4 WMU types
- landfill
- land application unit (LAU)
- wastepile (WP)
- tank
# 5 receptors
- adult resident, exposure starting
age 19 years
- child resident, exposure starting
age 0-3 years
- child resident, exposure starting
age 4-10 years
- child resident, exposure starting
age 11-18 years
- off-site worker
# direct inhalation only
# volatiles and particulates
# 6 distances from the site
# 3 risk endpoints or averaging times
- chronic (over 1 year)
- subchronic (1 month) - LAU, WP
- acute (1 day) - LAU, WP
105 were addressed but one constituent, 3,4 dimethylphenol, did not have an inhalation benchmark.
2-1
-------
Volume II
Section 2.0
# Exposure modeling/risk estimation—estimating exposure to a receptor and then
backcalculating to arrive at a waste concentration (Cw) that presents a risk equal to
a prespecified risk level (e.g., 1 in 1 million, or 1E-6).
To illustrate the scenario that was modeled for this study, Figure 2-1 is a conceptual
diagram of a waste site. Constituents managed in the WMU can be released as gases if they
volatilize and as particulates if the constituent attaches to solid particles in the waste. Once the
constituent is released from the site, the ambient air provides a medium for the transport of the
airborne constituent. The direction the constituent travels and its concentration in the air are
determined by meteorological conditions in the surrounding area such as wind direction, air
temperature, and atmospheric stability at the time it is released. Because meteorological patterns
are dynamic, the concentration of the constituents in the air varies over time and people who live
and work at various locations around the WMU have different inhalation risks. The risk to an
individual from the release of a constituent also depends upon characteristics of that individual
such as body weight, inhalation rate, and the length of time that individual remains in the area
around the WMU. These last characteristics are the reason that this assessment considers the
exposure to multiple types of receptors: adult residents, child residents of various ages, and
workers.
In order to model the scenario described above, the preliminary requirements for the
analysis included:
# Emissions models for the various WMUs to provide estimates of gas and particle
releases from the unit
# A dispersion model capable of modeling area sources for chronic (over 1 year),
subchronic (1 month), and acute (1 day) releases
Figure 2-1. Conceptual diagram of a waste site.
2-2
-------
Volume II Section 2.0
# An exposure model for locating receptors proximate to the WMUs and estimating
their exposure
# A risk model that combines the exposure characteristics of different types of
receptors with constituent-specific toxicity benchmarks.
# The ability to backcalculate Cw from a prespecified risk level (e.g., 1E-6).
For each constituent and each WMU type, EPA wanted to be able to specify a Cw that
would not exceed a target risk level (e.g., 1 in 100,000, 1E-5) in more than a specified percentage
(e.g., 10 percent) of the cases being modeled. Therefore, a probabilistic modeling approach,
which would produce a distribution of Cw's, was needed, as opposed to a deterministic
approach, which would only produce a point estimate. A deterministic analysis produces a point
estimate because it uses a single value for each parameter in the analysis. A probabilistic
approach considers the variability in the inputs required to estimate the concentration nationally.
This type of approach produces a distribution of results because the method iterates through the
analysis more than once, allowing the input parameters in the analysis to take on different values
for each iteration from a distribution of values. For this analysis, EPA used a Monte Carlo
simulation. This is a type of probabilistic analysis that can be used when the distribution of some
or all input variables is known or can be estimated. A large number of iterations of the
calculations are performed (i.e., 1,000), with a value for each input variable selected at random
from the variable's distribution and the result (in this case, Cw) calculated for each iteration. The
results of each iteration are combined into a distribution of Cw. It was assumed that the modeled
cases represent the national distribution of risk-specific concentrations.
The probabilistic approach described above was used to model chronic exposures. A
deterministic approach designed to produce a more high-end point estimate was used to model
acute and subchronic exposures. The acute/subchronic approach uses the maximum exposure
point at any given distance, so no variability in receptor location is accounted for. It also uses the
meteorological conditions that produce the maximum air concentration for a 24-hour or 30-day
time period over 5 years of meteorological data. The results from the acute/subchronic analysis
are comparable to the 100th percentile of the distribution generated for the chronic analysis. It
should be noted that acute/subchronic exposures were only assessed for land application units
and wastepiles, which may have episodic loading events. There are a variety of other differences
in the acute/subchronic approach in how the emission rates and dispersion factors were
calculated; these are described in more detail in Sections 4 and 5.
To estimate volatile emissions from each type of WMU, EPA's CHEMDAT8 model was
used. For the landfill, LAU, and wastepile, the concentration of hazardous constituent in the
surface layer of the soil (hereafter referred to as soil concentration) was estimated using a mass
balance approach (i.e., competing pathways such as volatilization, adsorption, and
biodegradation are accounted for). Particulate emissions due to wind erosion were modeled for
land-based units (landfills, LAUs, and wastepiles). Landfills and LAUs were modeled as
ground-level sources using the Cowherd model (U.S. EPA, 1985b, 1988). Wastepiles were
modeled as elevated sources using the AP-42 model for wind erosion from aggregate storage
piles (U.S. EPA, 1985a). To obtain the emission rate of constituent sorbed to particulate matter,
2-3
-------
Volume II Section 2.0
the emission rate of particulate matter was multiplied by the soil concentration calculated by
CHEMDAT8. This was done to account for the portion of the original waste concentration that
would remain in the waste after volatilization and biodegradation losses, and so would
realistically be available for emission in the particulate phase.
The modeling assumes waste is continuously added to landfills and tanks, while LAUs
and wastepiles have noncontinuous, episodic waste loadings. To capture potential peaks in
emissions immediately after episodic loading events, acute and subchronic exposures were
evaluated for LAUs and wastepiles.
Dispersion modeling was performed for each WMU using EPA's Industrial Source
Complex Model Short-Term (ISCST3) to develop unitized air concentrations (UACs) for vapors
and particulates. UACs are dispersion coefficients based on a unit emission (i.e., 1 jug/m2-s) for
use in a backcalculation. UACs varied depending on the averaging time (i.e., chronic,
subchronic, or acute), the size of the WMU, the distance and direction of the receptor from the
WMU, and the associated meteorological station. Dispersion modeling for vapors did not
account for depletion, as sensitivity analysis showed that depletion of vapors has a negligible
impact on air concentration of vapors. Dispersion modeling for particulates accounted for dry
depletion of particles, since a sensitivity analysis showed that dry depletion has a potentially
significant impact on air concentrations of particulates. Wet depletion of particulates was not
accounted for in the dispersion modeling, as sensitivity analysis showed that wet depletion has
little impact on air concentration.
The air concentration at any specific receptor is the product of the emission rate
(in |ig/m2 -s) and appropriate UAC (in [|ig/m3]/[|ig/m2 -s]). Air concentrations were estimated
for chronic, subchronic, and acute exposures (using averaging times of 1 year, 1 month, or 1
day), based on a combination of volatile and particulate emissions.
Many previous risk analyses have used the maximum point of exposure at some
prespecified distance from the WMU as the point for analysis. Such an approach is usually
criticized as being overly conservative because it does not consider the possibility of no one
living at that exact point. Because individuals may potentially be located in any direction and at
various distances from a facility, this analysis developed an explicit way to incorporate this
consideration. First, a sensitivity analysis was conducted to determine a reasonable distance at
which to bound the analysis. This sensitivity analysis showed that, beyond 1,000 m, most air
concentrations are a small percentage (less than 10 percent) of the concentration at the point of
maximum exposure. Therefore, 1,000 m was used as the outer bound on the distance of receptors
included in this analysis. A receptor grid was set up to allow individuals to reside in any of 16
directions and at distances of 25, 50, 75, 150, 500, and 1,000 m from the edge of the unit.
For this analysis, five receptors were included: an adult resident, a child resident with
exposure starting between 0 and 3 years old, a child resident with exposure starting between 4
and 10 years old, a child resident with exposure starting between 11 and 18 years old, and an off-
site worker. These receptors could be located in any of 16 directions and at distances of 25, 50,
75, 150, 500, and 1,000 m from the edge of the unit. Each distance was evaluated separately and
the location of a receptor was allowed to vary among any of the 16 directions. The 16 directions
2-4
-------
Volume II Section 2.0
were equally weighted, so there is equal probability of a receptor being located anywhere around
the WMU. The 16 directions were equally weighted, so there is equal probability of a receptor
being located anywhere around the WMU. For acute and subchronic exposures, receptors were
modeled at 25, 50, and 75 m because it was assumed that the greatest possibility of acute
exposure would be closest to the site.
2.2 Conducting the Analysis
As discussed earlier, the analysis consists of three main parts: emissions modeling,
dispersion modeling, and exposure modeling/risk estimation. Figure 2-2 shows the model
framework. Emissions and dispersion modeling was performed first and the results used as
inputs to the exposure modeling/risk estimation. In addition, a database containing
characterizations of WMUs was used.
The goal of the analysis is to backcalculate a waste concentration that will result in a
specified risk. Because risk is assumed to be linear with waste concentration under most
circumstances, a waste concentration was generated by forward-calculating a risk associated with
a unit concentration in the waste (i.e., 1 mg/kg for land-based units and 1 mg/L for tanks), then
scaling the unit concentration using the ratio of target risk to calculated risk. The assumption of
linearity is accurate for the dispersion modeling and the exposure and risk modeling. The
emissions model is linear for land-based units and tanks without biodegradation. The emissions
model for tanks with biodegradation is nonlinear at the concentration where biodegradation shifts
from first order to zero order. The results for tanks with biodegradation were backcalculated
using first order emission rates; however, if this result exceeded the concentration at which
biodegradation becomes zero order, the result was adjusted to be based on zero order emission
rates. Even when the emissions model is linear, it is possible, using this approach, to
backcalculate waste concentrations that exceed the solubility or soil saturation concentration for
the chemical. Results that exceed the solubility or soil saturation concentration under neutral
conditions are footnoted in the result tables. Soil saturation concentration and solubility can vary
accordint to site-specific temperature and pH conditions.
Emissions modeling was performed for all WMUs and all chemicals, assuming a unit
concentration of the chemical in the waste (1 mg/kg for land-based units or 1 mg/L for tanks).
These emissions were used as inputs to Step 2 of the exposure modeling/risk estimation portion
of the model.
Dispersion modeling was performed for 76 representative WMU areas and height
combinations and 29 meteorological locations, assuming a unit emission rate of 1 //g/m2-s. This
produced vapor and particle-phase UACs for each area/height combination, meteorological
station, and receptor location, which were used as the basis from which to interpolate in Step 4 of
the exposure modeling/risk estimation portion of the model.
The analytical framework shown in Figure 2-2 consists of a series of steps and loops. In
Step 1, a chemical and WMU type (e.g., landfills) were selected (thus, all landfills were analyzed
as a group for each chemical, and so on).
2-5
-------
Volume II
Section 2.0
Emission Modeling
For each chemical and WMU,
estimate:
• Volatile emissions
• Participate emissions
• Remaining average waste/
soil concentration
1
Select constituent and waste
management unit (WMU) type
Z Select WMU site from WMU database
Z Characterize site (area, height, and met station)
Dispersion Modeling
For each of 76 WMU
area/height combinations
and 29 meteorological
stations, estimate UAC
at each of 96 receptor
locations.
Select receptor location for each distance
Interpolate UAC for site
(based on area, height, and met station)
Select exposure factors for each receptor
(chronic exposure to carcinogens only)
Determine risk-specific waste concentration (CJ for
Monte Carlo realization for each receptor
Select next receptor location for simulation
Select 85th, 90th, and 95th percentile of risk-specific waste
concentration (Cw) from all realizations for this site
Next WMU simulation
Construct histogram for national risk-specific waste
concentration (Cw) for this type of WMU using 90th
percentile result from realizations for each site
Next constituent and WMU type
Monte Carlo iterations (1,008)
Figure 2-2. Model framework.
2-6
-------
Volume II Section 2.0
In Step 2, a WMU was selected from the data file for that unit type. For example, for
landfills, the database has a data record containing the facility identification and WMU
characteristics such as surface area, depth, and waste quantity managed per year for each of 801
landfill units. The database also has a sampling weight for each facility that defines how many
facilities nationally were represented by that facility. An assigned meteorological station was
added to the database based on locational information for each WMU. The model simulation
starts with the first record and moves to each successive record. For each WMU record, the
associated emission rate for that WMU and chemical was obtained from the emission modeling
results.
In Step 3, receptor locations were selected by choosing at random one of the 16 directions
modeled in the dispersion modeling. Receptors were modeled in that direction at each of six
distances from the site.
In Step 4, a UAC was interpolated for the WMU. Due to the long run time of ISCST3 for
area sources, UACs were modeled for only 76 selected WMU area/height combinations for each
meteorological station and receptor location. To calculate a UAC corresponding to the WMU's
actual area and height, EPA first chose the modeled height closest to the actual unit height, then
interpolated between the UACs for the two closest of the areas modeled. For example, the first
three areas modeled for wastepiles were 20, 162, and 486 m2. These were modeled at heights of
1, 2, 4, 6, and 8 m. For a WMU with an actual area of 100 m2 and an actual height of 3.5 m, the
UAC was interpolated from the UACs for 20 m2/4 m high and 162 m2/4 m high. For a WMU
with an actual area of 200 m2 and an actual height of 6.9 m, the UAC was interpolated from the
UACs for 162 m2/6 m high and 486 m2/6 m high.
In Step 5, for chronic exposures to carcinogens, values of exposure factors such as body
weight, inhalation rate, and exposure duration were chosen at random from distributions of these
parameters (developed from data in the Exposure Factors Handbook, U.S. EPA, 1997c and
1997d) to capture the variability in exposure factors for a given receptor. These exposure factors
differ for different receptor types (such as adults, children, and workers). Noncarcinogens were
not assessed in this manner because the health benchmarks, such as EPA's reference
concentration (RfC), are expressed in terms of ambient concentration and cannot be adjusted for
variations in these exposure factors. Similarly, acute and subchronic health benchmarks are
expressed as ambient exposure concentrations and cannot be adjusted for variability in exposure
factors.
In Step 6, the emission rate, UACs, and, if applicable, the exposure factors, were
combined with the health benchmark for the chemical to estimate risk (for chronic exposure to
carcinogens) or hazard quotient (for acute and subchronic exposures, and chronic exposures to
noncarcinogens) associated with the unit concentration modeled. This risk was then compared to
the target risk of either 1 in 1 million, 1 in 100,000, or 1 in 10,000 (i.e., 1E-6, 1E-5, or 1E-4) for
carcinogens, and the ratio was used to scale the unit concentration to a concentration in the waste
(Cw) that would result in the target risk at that receptor. A similar technique was used for scaling
the hazard quotient for noncarcinogens.
2-7
-------
Volume II Section 2.0
Steps 3 through 6, which form the core of the Monte Carlo simulation, were then repeated
1,008 times for each WMU, resulting in a distribution of Cw for that WMU for each receptor
(adult, child, or worker) at each distance from the site (25, 50, 75, 150, 500, and 1,000 m) for a
specific risk criteria (i.e., 1E-4, 1E-5, or 1E-6 for carcinogens and 10, 1, or 0.25 for
noncarcinogens). Once 1,008 iterations had been performed for a WMU, various percentiles
were selected from the distribution to characterize it. These percentiles represent the percentage
of receptors protected at the WMU.
Steps 2 through 6 were then repeated to obtain distributions of Cw for each WMU in the
database. These distributions are somewhat different for carcinogens and noncarcinogens and for
chronic, subchronic, and acute exposures. For chronic exposure to carcinogens, they represent
both the potential variability in location around a WMU, as well as the variability in exposure
duration, inhalation rate, and body weight for each receptor type. For noncarcinogens and for
subchronic and acute exposures, variability in these exposure factors is not considered because
the measure of risk is a ratio of air concentrations. For chronic exposures to noncarcinogens, the
distributions represent the variability in location around the WMU at a specific distance. For
subchronic and acute exposures, only point estimates were made at various distances using the
receptor located at the point of maximum air concentration for that distance.
The cumulative distribution of Cw for each WMU is presented as the percentage of
receptors that are at or below the risk criteria for any Cw (see Figure 2-3, left side). For example,
90 percent of all adult residents at a distance of 150 meters have a predicted risk at or below 1 in
100,000 (1E-5) if the concentration of the chemical (e.g., cumene) in the landfill is 1 mg/kg (see
point a). A second landfill may have a 90 percent protection level for all adult residents at 150
meters at a concentration of 10 mg/kg (point b), and a third landfill at a concentration of 100
mg/kg (point c). Thus, in Step 7, for each WMU, the distribution shows the percent of potential
receptors at or below a specified risk level for each concentration of constituent in the WMU
(Cw) for each distance and each receptor.
In Step 8, once all WMUs of a certain type had been modeled, the distributions of Cw for
all individual WMUs of the same type (e.g., landfills) were combined to produce a cumulative
distribution that presents the variability in Cw across all units of a certain type. For a given
percentage of protected receptors (e.g., 90 percent) as described above, the Cw was combined
across all WMUs of a specified type (e.g., landfills) to provide a distribution of the percentage of
sites considered protective at that level, as shown in Figure 2-3 (right side). Figure 2-3, for
example, shows the cumulative distribution of Cw at a 90 percent protection level across all
landfills. From this distribution, the 90th percentile Cw value for all 90 percent protection levels
across all landfills could be estimated. As described above, three landfills that give a 90 percent
protection level (i.e., at 1E-5) for a resident at 150 meters from the unit boundary have
corresponding Cw values of 1 mg/kg, 10 mg/kg, and 100 mg/kg (see points labeled a, b, and c).
These values plus similar values from all other landfills constitute the cumulative
distribution. The Cw value that is protective of 90 percent of receptors across 90 percent of the
sites is referred to in this study as the 90/90 protection level. These distributions were developed
for each unit type, each receptor type, each risk criteria, and each distance from the WMU.
2-8
-------
Volume II
Section 2.0
WMU 1
WMU 2
100
90
I 80
,S c 70
fo 50
SE 2 30
o
3S 20
10
0.1
WMU 3
I
£
100
90
80
g 70
g'S 60
f 5 50
8-tg 40
rr~ •—
S^ 30
o
SS 20
10
0.1
10 100 1,000 10,000
Cw (ppm)
b (90th %ile)
V (50th %ile)
10 100 1,000 10,000
Cw (ppm)
(90th %ile)
i -\ Z (50h%ile)
All WMUs
100
fe 90
o-g 80
g'jj 70
j= O 60
'iln 50
^^40
pj 30
•g
-------
Volume II Section 2.0
2.3 Data Sources
The Industrial D Survey database (Shroeder et al., 1987) was the primary source of data
on WMUs used in this analysis. This database provides information on each of the WMUs
assessed, with the exception of tanks. Tank data are from EP A's National Survey of Hazaradous
Waste Treatment, Storage, Disposal and Recycling Facilities (TSDR Survey, U.S. EPA, 1987).
The Industrial D Survey database contains information on the size and capacity of a statistical
sample of each WMU type, general location information, and statistical weights for each facility
in the sample. The statistical sample was designed to represent all industrial waste management
units not regulated under the RCRA hazardous waste program at the time the survey was
conducted in 1987. The weights in the database indicate the number of facilities represented by
each facility in the sample. For this assessment, it is assumed that the data contained in this
database provide an appropriate representation of the characteristics of each WMU type and of
the general location of these types of facilities with respect to climate regions of the country.
Meteorological stations provided temperature and windspeed data as inputs to the
emissions model and a large set of inputs for the dispersion model. Although meteorological
data are available at over 200 meteorological stations in the United States (see, for example,
Support Center for Regulatory Air Models (SCRAM) Bulletin Board at http://www.epa.
gov/scramOOl), various resource constraints prevented the use of all available data sets in this
analysis. Therefore, a set of 29 stations was used that had been selected as representative of the
nine general climate regions in the contiguous United States in an assessment for EPA's
Superfund Soil Screening Level (SSL) program (EQM, 1993).
In EPA's Superfund study, it was determined that 29 meteorological stations would be a
sufficient sample to represent the population of 200 meteorological stations and predict mean
dispersion values with a high (95 percent) degree of confidence. The 29 meteorological stations
were distributed among the nine climate regions based on meteorological representativeness and
variability across each region. Large-scale regional average conditions were used to select the
actual stations.
The 29 meteorological stations are listed in Section 5. To assign each Industrial D or
TSDR facility to a meteorological station, EPA used a geographic information system (GIS) to
construct areas around each station that encompass the areas closest to each station. The
boundaries of these areas were then adjusted to ensure that each boundary encloses an area that is
most similar in meteorological conditions to those measured at the meteorological station. First,
the boundaries were adjusted to correspond to Bailey's ecological divisions (Bailey et al., 1994),
which are defined primarily on physiography and climate. The boundaries were further adjusted
for coastal (including Great Lakes) areas and the central valley of California to ensure that these
stations were used only in regions with similar meteorology. Based on zip codes in the
Industrial D Survey database and EPA IDs in the TSDR database, the sites were then overlaid on
this GIS coverage, and meteorological station assignments were then exported for use in the
modeling exercise. Several sites in Alaska, Hawaii, and Puerto Rico were deleted from the
analysis at this point because the 29 meteorological stations are limited to the continental United
States.
2-10
-------
Volume II Section 3.0
3.0 Waste Source Characteristics
Waste sources modeled in the air characteristic risk analysis are landfills, land application
units, wastepiles, aerated treatment tanks, nonaerated treatment tanks, and storage tanks. The
dimensions and operating characteristics of these waste management units (WMUs) are
important determinants of the modeled emission rates and dispersion factors used to estimate
inhalation risks to receptors. This section describes how data specific to these unit types were
collected and processed for use as emission and air model inputs. Source parameters that are
critical to emissions but not to dispersion are discussed in Section 4.0, Source Emission
Estimates.
The primary source of data used to characterize waste sources is the Industrial D
Screening Survey (Schroeder et al., 1987). These survey data provide information (for all
WMUs except tanks) on the number of each type of WMU, annual quantity of waste managed in
the WMU (in 1985), and some information on the dimensions of the WMU (area and capacity).
Along with waste concentration, these parameters are the primary waste source characteristics
that impact the emission model estimates. The dispersion model also requires WMU area for area
sources (landfills, wastepiles, and land application units) as well as the height of the source for
wastepiles and tanks. The primary source of data for tanks is EPA's National Survey of
Hazardous Waste Treatment, Storage, Disposal and Recycling Facilities (TSDR survey, U.S.
EPA, 1987).
Facility locations also were required so that facilities could be assigned to the 29
meteorological stations used to provide meteorological data for the air model. These were
derived by matching Industrial D and TSDR facilities, using the Dun and Bradstreet DUNS®
number, name, street address, and zip code, and EPA ID to EPA's Envirofacts system to obtain
the "best" location for each facility as defined by EPA's Locational Data Improvement Project
(LDIP) and represented by data in the Envirofacts Locational Reference Table (LRTs)
(http://www.epa.gov/enviro/html/lrt/lrt_over.html). For sites that could not be matched to the
LRT tables, the latitude/longitude of the facility's zip code centroid was determined and used as
the facility location.
3.1 Estimation of Missing or Replacement Data for WPs, LAUs, and LFs
The Industrial D Screening Survey provided the data needed to characterize land-based
WMUs except for wastepile height and landfill depth, which had to be estimated. In developing
these characterizations, the data set was reviewed to ensure that facilities reporting insufficient
information or outlier data were not included. In some cases, replacement values were required
for waste quantity and landfill capacity due to screened unrealistic values or missing data. Units
were culled for any of the following reasons:
~
-------
Volume II Section 3.0
# Surface area was not reported
# Address information was not provided
# Sites were located outside the contiguous United States
# Reported data were identified as outliers
# Site was identified as a Bevill facility as discussed in Section 3.3.
Table 3-1 shows the total number of facilities reporting each WMU in response to the survey,
those culled from the data set, and those remaining in the database whose data were used in
developing the source characterizations.
3.1.1 Wastepile Height
Reported data on wastepile height were not available in the Industrial D database. In
general, the height of a wastepile is a function of the area of the unit and the annual waste
quantity managed in the unit. In the absence of reported data, the relationships between pile
height and the annual waste quantity per unit area shown in Table 3-2 were developed. The
height of a wastepile depends on the annual waste quantity, the area of the pile, the waste density,
and the retention time. Given a waste density and retention time, a pile with a greater ratio of
waste quantity to area must be higher. A low height of 1 m was set, as differences in height
below 1 m make little difference to the dispersion modeling. The high end height of 10 m was
chosen as a reasonable maximum, given that waste would have to be added to the top of a pile,
so equipment adding waste would have to be able to reach the highest pile height. A height of
10 m is about the equivalent of a 3-story building. A discrete set of heights was selected to
simplify the meteorological component of the data required for conducting dispersion modeling.
(Specifically, a calculation is required to adjust the windspeed from the height of the measuring
device to the height of the wastepile.) The annual waste quantity and area were known from the
Industrial D survey data for the wastepiles modeled. The ratio of waste quantity to area ranged
from 0.08 to 1,900 Mg/yr-m2, with 77 percent of piles falling below 10 Mg/yr-m2 and less than
3 percent falling above 80 Mg/yr-m2. This information was combined with the range of pile
height selected to establish the relationship shown in Table 3-2. Given a reasonable range of
waste bulk density from 1 to 1.5 g/cm3, these heights imply retention times on the order of
months to a year.
Using this approach, wastepile height assignments were made for each wastepile in the
Industrial D data set. Approximately 77 percent of the total number of wastepiles included in the
air characteristics study data set were assigned a wastepile height of 1 meter, and approximately
95 percent of the wastepiles were assigned heights of 4 meters or less.
3.1.2 Landfill Depth and Capacity
Landfill depth was calculated consistent with previous EPA modeling efforts using the
Industrial D data (U.S. EPA, 1996a). The following equation was used, assuming the same
waste bulk densities (fixed by unit type) used in the previous modeling efforts:
depth = total waste capacity / (surface area x bulk density).
3-2
-------
Volume II
Section 3.0
Table 3-1. Summary of Land-Based WMUs Removed from Industrial D Survey Database
Description
Number of Sites Reporting WMU Type in Industrial
D Screening Survey
Culled Sites: Surface area was not reported
Culled Sites: Address information was not provided
Culled Sites: Outside contiguous United States
Culled Sites: Surface area identified as outlier
Culled Sites: Bevill facilities
Total Number of Sites included in Air
Characteristics Study Data Set
Wastepiles
853
6
15
3
0
84
745
Land Application Units
354
2
5
1
1
(area = 40,000 acres)
NA
345
Landfills
827
3
20
2
1
(area =120,000 acres)
NA
801
Table 3-2. Assignment of Wastepile Height
Annual Waste Quantity/Surface Area
(Mg/m2/y)
Pile Height (m)
> 10, <20
>20, <40
>40, <60
>60, <80
>80
1
2
4
6
8
10
The bulk density used for landfills in previous efforts was 1.1 g/cm3; this was also used in this
study for consistency.
Also in accordance with previous EPA modeling efforts using the Industrial D Screening
Survey (U.S. EPA, 1996a), landfill capacities were removed from the Industrial D data when
depth or capacity constraints were violated. These constraints were imposed to eliminate
unrealistic values for depth and capacity. The constraints were: landfill depth had to be greater
than or equal to 2 feet and less than or equal to 33 feet, and, within the Industrial D database,
landfill total capacity had to be greater than the remaining capacity (U.S. EPA, 1996a, p. 3-6).
Of the 801 landfills modeled, 103 had a depth less than 2 feet, 87 had a depth greater than 33
feet, and 21 had remaining capacity that exceeded the total capacity. In addition, 92 facilities
were missing data on total capacity, remaining capacity, or both. Thus landfill capacity was
missing or screened for 303 landfills.
3-3
-------
Volume II
Section 3.0
Landfill capacities to replace the 303 missing or removed values were estimated based on
the correlation between surface area and capacity of the waste management unit in the Industrial
D data. First, a statistical regression of log (average total capacity) versus log (average surface
area) was done on the facilities with known capacities. The regression yielded an equation for a
best fit line through the known values. This equation gave the capacity as a function of area, so
the missing or screened capacities could be estimated based on the known areas. To provide a
more probabilistic sampling of average capacities, and since the known capacities seemed to be
in a limited range above and below the best fit line, a positive or negative random number was
generated within that range and added to the calculated log (average total capacity) to replace
each missing capacity with a random value that was reasonable with respect to landfill area. This
value was then used to calculate landfill depth as described above. Figure 3-1 shows the
regression plots, including the replaced ("random capacity") values, for landfills.
Log (Average Surface Area, acres) Line Fit Plot
-101 2
Log (Average Surface Area, acres)
Landfill Regression Plot
Figure 3-1. Landfill characteristics from Industrial D Screening Survey.
3-4
-------
Volume II Section 3.0
3.1.3 Waste Quantity Replacement Values
Replacement values were calculated for waste quantity due to screened unrealistic values
or missing data. Waste quantity replacement values were calculated under three conditions: (1) if
surface area data were reported for a given unit but the waste quantity was not provided, (2) if the
minimum wastepile height constraint of 1 meter was violated, and (3) if the application rate for
land application units exceeded 10 Mg/m2-yr. The minium 1-meter wastepile height constraint
was imposed to avoid a situation in which the wastepile replacement frequency is less than once
every 5 years. The maximum application rate for LAUs was established as a cut-off in previous
EPA modeling efforts using the Industrial D data (U. S. EPA, 1996a) and was retained in this
study for consistency. The first condition applies to all unit types, while the second and third
apply only to wastepiles and land application units, respectively.
Waste quantity replacement values were calculated for 391 wastepiles and 20 LAUs,
based on the three conditions described above. Thirty wastepiles and 8 LAUs reported no waste
quantity, 361 wastepiles violated the minimum height constraint, and 12 LAUs violated the
maximum application rate constraint.
Similar to the approach applied in estimating landfill capacity, replacement values were
developed for the screened and missing waste quantities. These values were estimated based on
the correlation between surface area and waste quantity of the waste management units in the
Industrial D data. A statistical regression of log (waste quantity) versus log (average surface
area) was done on the facilities with known quantities. The regression yielded an equation for a
best fit line through the known values. This equation gave the waste quantity as a function of
area, so the missing or screened capacities could be estimated based on the known areas. To
provide a more probabilistic sampling of average waste quantities, and since the known
quantities seemed to be in a limited range above and below the best fit line (with some outliers),
a positive or negative random number was generated within that range and added to the
calculated log (average waste quantity) to replace each missing waste quantity with a random
value that was reasonable with respect to wastepile area. Figures 3-2 and 3-3 show the
regression plots, including the replaced ("random waste quantity") values for wastepiles and land
application units, respectively.
3.2 Waste Management Unit Data and Size Categories Used in Dispersion
Model
The area of a WMU is a very sensitive parameter in the dispersion model. However,
because of run-time limitations associated with the air dispersion model, it was necessary to
develop area strata for the land-based WMU types. Dispersion modeling was conducted for the
median area of each of these strata and the resulting unitized air concentrations (UACs) were
used as inputs to an interpolation routine used to determine UACs for each WMU (see
Section 5).
3-5
-------
Volume II
Section 3.0
Log (Average Surface Area, acres) Line Fit Plot
J
o
I
I
Actual waste quantity
values (438 sites)
Random waste quantity
values (391 sites)
Screened values based on
rrinimum height constraint
(361 sites)
-1 0
Log (Average Surface Area, acres)
Waste Pile Regression Plot
Figure 3-2. Wastepile characteristics from Industrial D Screening Survey.
3-6
-------
Volume II
Section 3.0
-3
Log (Average Surface Area, acres) Line Fit Plot
-10123
Log (Avg. Surface Area, acres)
Land Application Unit Plot
Figure 3-3. Land application unit characteristics from Industrial D Screening Survey.
3-7
-------
Volume II Section 3.0
Initially, a regular distribution of strata outpoints (5th, 25th, 50th, 75th, and 95th
percentiles) was considered to provide the representative WMU surface areas. However,
examination of the distribution of areas for each WMU type showed that the distributions were
all extremely skewed to the right, with a large number of facilities at the smaller end of the
distribution and a few facilities with very large areas. This raised the concern that the regular
cutpoints would not adequately cover the tail of the distribution (i.e., WMUs with large areas) for
the interpolation process.
To improve the coverage of the cutpoints, a statistical methodology called the Dalenius-
Hodges procedure was used as a starting point. This technique is designed to break the
distribution of the known variable X (in this case, facility area), which is assumed to be highly
correlated with the characteristic Y (in this case, the UACs), into L strata in an optimal way.
From the L strata, an equal number of observations, n, are picked at random (assuming each
observation has an equal weight). The process was actually employed in this case using the
facility weights in the Industrial D data. To develop the initial WMU strata, L was set at 5; that
is, 5 "optimal" strata were initially formed using this approach. This process produces a set of
percentiles with the following property: if the distribution is skewed to the right, there will be
more emphasis on characterizing the right tail. Appendix A describes the steps of the Dalenius-
Hodges procedure in detail.
Based on ISCST3 model run-time estimates, the target number of strata for all WMU
types was set at 20. Because land application units and landfills were modeled in the same
fashion using ISCST3, it was possible to combine the areas for these WMU types to cover the
distribution with as many strata as possible. Fifteen strata were initially formed for this
distribution using the Dalenius-Hodges procedure. Wastepiles could not be combined with the
other waste management units because they are modeled in ISCST3 as elevated area sources, not
as ground-level sources like the others. Five Dalenius-Hodges strata were initially developed for
this unit type.
In each case, the Dalenius-Hodges procedure yielded a good characterization of the right
tail of the distribution but a poor characterization of the lower portion of the distribution.
Because the UACs are most sensitive to area in this region of the distribution, the procedure was
modified by collapsing two or more of the highest strata into one and successively splitting the
lowest stratum into four strata. The splitting was performed by choosing stratum breakpoints as
described in Appendix A.
For the combined landfill and land application distribution, the four highest strata (12, 13,
14, and 15) were collapsed into one and the facility nearest the median of the combined stratum
was used to represent that stratum. The next two highest strata (10 and 11) were also collapsed
into one and, again, the facility nearest the median of the combined stratum was used to represent
that stratum. The first stratum (1) was then split, similar to the wastepiles, into four strata (1 A,
IB, 1C, and ID). The result was a total of 14 strata for the combined waste management units
and 21 strata overall. Table 3-3 shows the final area strata for landfills and land application
units, the medians used for the air dispersion modeling runs and the cumulative percentile of the
distribution each stratum represents. Air dispersion modeling was conducted using the median
surface area for each stratum.
3-8
-------
Volume II Section 3.0
Table 3-3. Area Strata Modeled for Landfills
and Land Application Units
Average area (m2)
Strata
1A
IB
1C
ID
2
3
4
5
6
7
8
9
10& 11
12 to 15
Low
14
310
809
2,307
7,588
27,115
60,300
120,763
210,444
303,525
554,439
753,754
1,007,703
2,521,281
Median
81
567
1,551
4,047
12,546
40,500
78,957
161,880
243,000
376,776
607,000
906,528
1,408,356
8,090,000
High
293
789
2,293
7,487
26,980
59,653
119,000
210,000
295,000
546,345
728,460
999,609
2,430,000
13,500,000
Cumulative
percentile
(Ind D data)
17
24
39
54
74
81
87
92
94
96
97
98
99
100
For wastepiles, the Dalenius-Hodges stratification resulted in most (90 percent) of the
Industrial D facilities falling into the lowest stratum where more interpolation points are
necessary to accurately estimate the UACs. To correct this, the two highest strata (4 and 5) were
collapsed into one. The first stratum (1) was then split into two (la and Ib). This still resulted in
a large number of facilities in the lowest strata, so it was split again into two strata (la' and la").
The new lowest strata (la') was further split into two strata (la'A and la'B). The result was a
total of seven strata for wastepiles.
All wastepiles in the database were assigned one of the six heights as discussed in
Section 3.2. Using wastepile surface area data from the Industrial D data set, it was possible to
identify which area strata were associated with each assigned pile height. Review of height/area
strata data revealed that not all seven strata were associated with all heights. Based on these data,
ISCST3 modeling was conducted for the 29 area/height combinations shown in Table 3-4. These
combinations represent the height/strata combinations identified based on the Industrial D data
set and a few additional combinations identified to facilitate interpolation between UACs (see
Section 5).
Tanks were modeled in the same manner as wastepiles, using the area/height
combinations shown in Table 3-5. The heights that were modeled represent the range of heights
(1 to 6 meters) imputed from depth as described in Section 3.4.2. The 10 areas modeled were
3-9
-------
Volume II
Section 3.0
Table 3-4. Areas and Source Heights Modeled for Wastepiles
Assigned
Surface Area (m2)
Height (m)
1
2
4
6
8
10
20
X
X
X
X
X
X
162
X
X
X
X
X
X
486
X
X
X
X
X
2,100
X
X
X
X
X
10,100
X
X
X
101,000
X
X
1,300,000
X
X
x = Combination modeled.
Table 3-5. Areas and Source Heights Modeled for Tanks
Assigned
Surface Area (m2)
Height (m)
1
2
4
6
2
X
X
3
X
X
X
5
X
X
X
10
X
X
X
25
X
X
X
50
X
X
X
100
X
X
X
X
400
X
X
X
X
1,500
X
X
X
X
5,000
X
X
X
X
x = Combination modeled.
selected based on a sensitivity analysis conducted with ISCST3. These areas were identified to
minimize error (not to exceed 10 percent) when interpolating between UACs.
3.3 Removal of Wastepiles Containing Bevill-Excluded Wastes
RCRA excludes from hazardous waste regulation certain wastes related to mining and
mineral processing. These wastes are commonly known as Bevill wastes. Enacted on
October 21, 1980, Public Law 96-482 amended RCRA Section 8002 to include subsection (p),
which requires the Administrator to study the adverse effects on human health and the
environment, if any, of the disposal and utilization of solid waste from the extraction,
beneficiation, and processing of ores and minerals, including phosphate rock and overburden
from the mining of uranium ore. Section 7 of these amendments (the Bevill Amendment)
amended Section 3001 of RCRA to exclude these wastes from regulation under Subtitle C of
RCRA pending completion of the studies called for in Sections 8002(f) and (p). On
November 19, 1980, EPA published an interim final amendment to its hazardous waste
regulations to reflect the mining waste exclusion (40 CFR 261.4(b)(7)). Since that time,
261.4(b)(7) has been amended to clarify which solid wastes fall under the exclusion as wastes
from processing of ores and minerals.
3-10
-------
Volume II Section 3.0
Because an air characteristic would not apply to WMUs containing Bevill Amendment
wastes, an analysis of the presence of such facilities in the Industrial D data was conducted to
determine whether such facilities would bias the WMU area distributions used in the risk
analysis. Identification of facilities that may be handling Bevill wastes was based upon Standard
Industrial Classification (SIC) codes assigned to the Bevill waste streams. SIC codes assigned to
industries producing Bevill wastes are as follows:
SIC Code Description SIC Code Description
2816 Inorganic Pigments 10-1099 Metal Mining
2819 Ind. Inorganic Chem., Nee. 12-1299 Coal Mining
2874 Phosphatic Fertilizers 13-1399 Oil & Gas Extraction
3312 Blast Furnaces & Steel Mills 14 - 1499 Mining & Quarrying
3331 Primary Copper of Nonmetallic
3339 Primary Nonferrous Metals, Nee. Minerals Except Fuels
Electric power facilities (SIC code 4911, electric services) were also included in the analysis
because their high-volume waste streams are also exempt from RCRA Subtitle C regulation.
3.3.1 Identification of Bevill Industrial D Facilities
Two approaches were used to identify Bevill facilities from the Industrial D database. The
first removed facilities using a relationship between the Bevill-related SIC codes and the industry
groups in the Industrial D database:
Industry Group Removed Group Description SIC Codes
6 Inorganic Chemicals 2812-2819
3 Fertilizer & Agr. Chem. 2873 - 2879
2 Primary Iron & Steel 3312-3321
9 Primary Nonferrous Metals 3331-3399
4 Electric Power Generation 4911
To provide additional resolution to the analysis (i.e., to reduce the number of facilities
inappropriately culled because of the breadth of the Industrial D industry groups), 1,722
Industrial D facilities were matched to EPA IDs and SIC codes in EPA's Facility Index System
(FINDS) database. The FINDS SIC codes are multiple per facility and were originally obtained
from Dun & Bradstreet. The 1,722 facilities with EPA IDs were then culled by the SIC codes
listed above for Bevill and electric power wastes; unlike the "industry group" approach, this
included the mining waste SIC codes. Because of the multiple SIC codes for many facilities, a
facility-by-facility review of 4911, 2819, and some mining SIC codes was conducted using
Resource Conservation and Recovery Information System (RCRIS) data to identify facilities that
should not be culled (i.e., facilities that could be producing and managing hazardous wastes not
covered under Bevill, typically large-quantity generators as indicated in RCRIS). Thus, some
facilities that would have been culled based on SIC code alone were not removed from the
analysis.
3-11
-------
Volume II Section 3.0
This approach is not a perfect screen. It may leave in some facilities that manage
primarily Bevill wastes, and it may exclude some facilities that manage some non-Bevill wastes.
However, there are insufficient data available to refine the screening approach further.
3.3.2 Bevill Facilities Culled from Industrial D Data
To decide whether or not to remove the Bevill WMUs, the area percentiles for each
WMU type in the entire Industrial D database were compared with those from the Industrial D
datasets with Bevill facilities removed. For landfills and land application units, the area
percentiles for the entire Industrial D universe were very similar to those with the Bevill WMUs
removed, and the Bevill facilities were not removed for these WMU types. However, removing
the Bevill wastepiles resulted in significantly lower wastepile areas for the upper percentile
ranges. Based on the SIC codes and facility names, the largest 20 wastepiles could be classified
as follows:
# 9 phosphate mines
# 5 other mines
# 6 facilities that do not appear to be mining-related.
Only 3 of the 17 sites with average wastepile area greater than 250,000 square meters were not
connected to mining activities.
Based on this assessment, the wastepiles identified as likely to be receiving Bevill waste
were identified and removed from the Industrial D WMU data used in this analysis. In total, 84
wastepiles were removed because they contained Bevill wastes, leaving 745 facilities with
wastepiles for the risk analysis.
3.4 Characterization of Tanks
The Industrial Subtitle D Survey (Schroder et al., 1987) did not include tanks. Therefore,
a tanks database was developed for this analysis that compiled flow rates and tank volumes. The
primary source for these data was the EPA's 1986 National Survey of Hazardous Waste
Treatment, Storage, Disposal, and Recycling Facilities (TSDR) Database (U.S. EPA, 1987). This
database is the result of a comprehensive survey of 2,626 hazardous waste treatment, storage,
disposal, and recycle (TSDR) facilities, requested information concerning 1986 waste
management practices and quantities. Responses were received from 2,322 facilities. The TSDR
survey included a specific questionnaire concerning tanks used at each facility. Responses to this
questionnaire provided tank information for about 18,773 tanks at 1,700 facilities.
A subset of the TSDR survey responses was available for facilities that received any
quantity of waste from off-site. This subset of data contained information on 8,510 tanks located
at 710 facilities (approximately 45 percent of all of the tanks contained in the TSDR Survey).
This reduced data set was used to characterize tanks for this analysis. Although it would have
been preferable to use the complete data set, including tanks at facilities that treat only wastes
generated on-site, these data are unfortunately no longer available. However, it is likely that the
subset of data used represents the range of tank volumes reported for all tanks. The subset data
3-12
-------
Volume II
Section 3.0
include a broad range of tank volumes ranging from less than 55 gallons to over 5 million
gallons.
Table 3-6 shows the total number of tanks in the database, those culled from the dataset,
and those remaining in the database that was used in this study. The totals add to somewhat less
than 8,510 because the database included 1,270 units that, based on process codes, were not
actually tanks. The remaining 7,240 were categorized as treatment or storage based on process
codes, as described in Section 3.4.1; there were 472 tanks that specified both treatment and
storage codes and that were therefore included in both treatment and storage datasets.
Several criteria were used in guiding the development of the tanks database. These
criteria were applied to the TSDR survey data to determine which tanks should be excluded from
the data set.
# Flow rate. Only those tanks reporting nonzero flow rates were included in the
analysis.
# Open versus covered tanks. Only open storage and treatment tanks were
considered in the analysis; closed or covered tanks were dropped because this
study is only concerned with Industrial D scenarios, and RCRA has no
requirement to cover nonhazardous tanks.
# Tank volume. All tanks with a volume of 55 gallons or less were excluded from
the analysis. These smaller-volume containers should be classified as drums and
not tanks due to their size.
Additionally, two very large tanks (approximately 30 million gallons), one aerated
treatment and one nonaerated treatment, were reviewed, because these tanks were
many times larger than the next largest tanks and appeared to be nonrepresenta-
tive. The facility that owns both tanks was contacted and it was determined that
both actually have volumes of 3 million gallons, a value within the range
Table 3-6. Summary of Tanks Removed from TSDR Survey Database
Description
Number of Sites Reporting WMU Type in TSDR
Database
Culled Sites: Zero size or flow
Culled Sites: Tank covered or cover not specified
Culled Sites: Size < 55 gallons
Culled Sites: Outside contiguous United States
Total Number of Sites included in Air
Characteristics Study Data Set
Treatment Tanks
2,346
464
979
6
4
893
Storage Tanks
5,366
793
3,885
48
2
638
3-13
-------
Volume II Section 3.0
represented by the other tanks in the database. Both values were corrected to
3 million gallons for this analysis.
# Location. Tanks located outside the continental United States were excluded
from the database, because suitable meteorological stations were not available to
model these sites.
3.4.1 Tank Classification
Industrial tanks can be used for either treatment or storage of wastes. Treatment tanks
can be either quiescent or aerated/agitated. Examples of quiescent treatment tanks are clarifiers
and filters (such as sand or mixed-media filters). Storage tanks are, by definition, quiescent
because they do not include aeration processes. In the absence of aeration, both treatment and
storage tanks are still subject to small amounts of agitation during filling and emptying
operations if they have above-surface intakes. Aeration or agitation in a wastewater treatment
system transfers air to the liquid to improve mixing or increase biodegradation. The turbulence
caused by aeration/agitation also enhances mass transfer to the air, thus increasing emissions.
Therefore, for a given treatment volume, a facility with aerated tanks will have higher emissions
than a facility with quiescent tanks.
To reflect emission characteristics associated with treatment versus storage tanks and
differences within the treatment tank category related to aeration intensity, three different tank
categories were identified and modeled:
# Storage tanks
# Aerated treatment tanks
# Nonaerated treatment tanks.
Sorting of the tanks in the database into these three categories was done using the WMU
code reported for each unit. Within the TSDR survey, the respondents were requested to provide
a WMU code to describe the type of process for which each tank was used. Those tanks with
WMU codes of either 2A (accumulation in tanks) or 2ST (storage in tanks) were classified as
storage tanks. The TSDR Survey used a broad range of WMU treatment codes (including codes
for incinerators and belt filter presses). For this reason, classification of treatment tanks was
limited to those processes listed in Table 3-7.
The treatment tank WMU codes were evaluated further to determine the level of aeration
used. HI aeration was assigned to tanks reporting processes that actively mix the liquid surface
for the purpose of aeration or that add diffused air. LO aeration was assigned to tanks reporting
processes that are likely to require mixing devices due to the addition of chemicals or other
purposes. NO aeration was used for tanks that are purposefully operated to minimize mixing or
agitation (e.g., a clarifier). The aeration level assignments for each WMU code are shown in
Table 3-7. The high- versus low-aeration classification was based on the nature of the process
description associated with the various process codes.
3-14
-------
Volume II
Section 3.0
Table 3-7. TSDR Survey Wastewater Treatment Codes Used in
Identifying Treatment Tanks
Process Code/Process Aeration Level
Equalization
1WT Equalization
LO
Cyanide oxidation
2WT Alkaline chlorination
3WT Ozone
4WT Electrochemical
5WT Other cyanide oxidation
LO
LO
LO
LO
General oxidation (including disinfection)
6WT Chlorination
7WT Ozonation
8WT UV radiation
9WT Other general oxidation
LO
LO
LO
LO
Chemical precipitation
10WT Lime
1 1WT Sodium hydroxide
12WT Soda ash
13WT Sulfide
14WT Other chemical precipitation
LO
LO
LO
LO
LO
Chromium reduction
1 5WT Sodium bisulfite
16WT Sulfur dioxide
17WT Ferrous sulfate
1 8WT Other chromium reduction
1 9WT Complexed metals treatment
LO
LO
LO
LO
LO
Emulsion breaking
20WT Thermal
21WT Chemical
22WT Other emulsion breaking
NO
LO
LO
Evaporation
31WT Solar
NO
Fuel blending
1FB Fuel blending
LO
Process Code/Process Aeration Level
Filtration
34WT Diatomaceous earth
35WT Sand
36WT Multimedia
37WT Other filtration
NO
NO
NO
NO
Sludge dewatering
38WT Gravity thickening
NO
Air flotation
43WT Dissolved air flotation
44WT Partial aeration
45WT Air dispersion
46WT Other air flotation
HI
HI
HI
HI
Oil skimming
47WT Gravity separation
48WT Coalescing plate separation
49WT Other oil skimming
NO
NO
NO
Other liquid phase separation
50WT Decanting
5 1 WT Other liquid phase separation
NO
NO
Biological treatment
52WT Activated sludge
54WT Fixed film— rotating contactor
57WT Anaerobic
58WT Other biological treatment
HI
LO
NO
HI
Other wastewater treatment
60WT Neutralization
61WT Nitrification
62WT Denitrification
63WT Flocculation and/or coagulation
64WT Settling (clarification)
66WT Other wastewater treatment
LO
LO
LO
NO
NO
LO
Other Processes
1TR Other treatment
LO
3-15
-------
Volume II Section 3.0
# Equalization, cyanide oxidation, general oxidation, chemical precipitation, and
chromium reduction all involve adding and mixing a chemical into the wastewater
followed by a quiescent period. Therefore, these tanks were classified as LO
aeration because the chemical addition and mixing involve more agitation than a
storage tank but involve no processes with intense agitation or forced air.
# Emulsion breaking included two different processes. Thermal heating simply
involves heating and letting the wastewater stand, whereas chemical emulsion
breaking involves chemical addition and mixing followed by a quiescent period.
Therefore, thermal emulsion breaking was classified as NO and chemical
emulsion breaking was classified as LO. The category "other emulsion breaking"
was classified as LO because the other processes in the emulsion breaking
category ranged from NO to LO, so this represented a conservative default
classification in the absence of more specific process data.
# Filtration processes are quiet and generally covered; therefore, these were
classified as NO aeration. Many of these, in fact, were eliminated from the
database because covered tanks, as a class, were removed.
# Air flotation processes all involve high-energy forced air operations and are
therefore all classified as HI aeration.
# Oil skimming involves liquid phase separation, which requires quiescent
conditions; therefore NO aeration was assumed for these processes. Similarly,
liquid phase separation processes were classified as NO aeration.
# Biological treatment processes are quite diverse and include HI aeration activated
sludge processes and LO aeration film processes. The "other biological
treatment" processes were classified as HI because the other processes in the
biological treatment category ranged from NO to HI and HI represents a
conservative default classification in the absence of more specific process data.
# Finally, the "other wastewater treatment" process in the "other wastewater
treatment" category and the "other processes" category were classified as LO
aeration as a default since no process information can be inferred from the
description.
The numbers of tanks included in each classification are summarized in Table 3-8. Some
tanks reported multiple process codes that included both a storage code (2A or 2ST) and a
treatment code (one of the 50 in Table 3-7). These 472 tanks were included in both the storage
tank database and one of the treatment tanks databases. However, there is no overlap between
the two treatment tank databases.
The tank database does appear to underrepresent highly aerated tanks. This may be due
to the age of the survey data, reflecting that highly aerated biological processes were in less
widespread use at that time than now. This underrepresentation introduces some uncertainty into
the analysis, the result of which is that risks from aerated tanks may be underestimated.
-------
Volume II Section 3.0
Table 3-8. Numbers of Tanks by Classification
Tank Classification
Storage tanks
Aerated treatment tanks
High aeration
Low aeration
Nonaerated treatment tanks
Total
Number
638
29
591
620
273
1,531
3.4.2 Additional Tank Data Used for Imputation
To address tank-specific data gaps in the tanks database, additional data sources were
identified. These data included information collected in 1985 and 1986 during EPA site visits to
aerated treatment systems in support of the development of RCRA Air Emission Standards. The
systems visited were selected to represent a range of aeration processes and reflect a variety of
industries and waste types. At that time, candidate facilities were identified through numerous
phone contacts with state and local environmental agencies. From those conversations,
information on wastewater treatment systems at 54 facilities was collected, and then site visits to
these facilities were conducted in 1985 and 1986. Data on the individual tanks (both aerated and
nonaerated) were provided by the facilities during the site visits, including data on tank
dimensions. The data on aerated tanks are summarized in Coburn et al. (1988) and Eichinger
(1985) (these documents are available in EPA Docket No. F-90-CESP, Hazardous Waste
Treatment, Storage, and Disposal Facilities. RCRA Air Emission Standards for Proposal:
Studies, Reports, and Background Information. Document numbers are F-90-CESP-S00251 and
F-90-CESP-S00130, respectively). The data on nonaerated tanks were collected at the same site
visits and are unpublished. Added to these data were five tanks from the TSDF BID (U.S. EPA,
1991). This resulted in a supplemental database of 49 tanks (13 with high aeration, 9 with low
aeration, and 27 with no aeration), presented in Table 3-9.
In addition to these data, several tank vendors were contacted to establish a reasonable
high end for tank capacity and depth based on design principles. As a result, a reasonable
maximum capacity for an open, partially or completely aboveground tank was defined to be
approximately 3 million gallons and the depth of such a tank would not be expected to exceed 10
m (about 32 ft) (Kendall Smith, personal communication, AO Smith Industrial, March 16, 1999).
These site visit tanks and hypothetical tanks were used only as a basis for imputing values and
were not modeled in the analysis in order to maintain the integrity of the source database.
3.4.3 Estimation of Missing Data
The TSDR Survey provided flow rate and tank volume data for use in characterizing
tanks for this analysis. However, other key parameters, including depth, surface area, and height,
also needed to be defined. In the absence of reported TSDR Survey data, these parameters were
calculated as described below. Other operating parameters (aeration parameters), which impact
emission estimates but not dispersion, are discussed in Section 4.0.
-------
Volume II
Section 3.0
Table 3-9. Summary of Tank Size Information Collected in EPA Site
Visits for RCRA Air Emission Standards
Type of unit
Aerated Trtmnt Tank
Aeration Tank
Bubbling Pit
Aerated Trtmnt Tank
Aeration Tank
Aeration Tank
Aeration Tank
Aeration Tank
Aeration Tank
Aeration Tank
Aux. Aer. Tank
Aeration Tank
Aeration Tank
Treatment Tank
Mixing Tank
Treatment Tank
Mixing Tank
No Eq. Basin
So Eq. Basin
Eq. Basin
Treatment Tank
Eq. Basin
Gravitator
Pre-Filter
Final Filter
Backwash Clarifier
Clarifier
Bio-sludge Holding Tanks
Bio-sludge Holding Tanks
Primary Clarifier
Digester
API Separator
Primary Clarifier
Biosludge Thickener
Clarifier/Thickener
Final Clarifier
Final Clarifier
Final Clarifier
Clarifier
Clarifier
Clarifier
Clarifier
Clarifier
Clarifier
Clarifier
Ship's Ballast Water
Final Clarifier
Ship's Ballast Water
Solid Waste Disposal Basin
Aeration
HI
HI
HI
HI
HI
HI
HI
HI
HI
HI
HI
HI
HI
LO
LO
LO
LO
LO
LO
LO
LO
LO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
Type of
Aerator
Mechanical
Mechanical
Diffused
Mechanical
Diffused
Mechanical
Diffused
Diffused
Mechanical
Diffused
Mechanical
Mechanical
Mechanical
Mechanical
Mechanical
Mechanical
Mechanical
Mechanical
Mechanical
Volume
(m3)
108
112
453
1,600
1,666
3,367
3,785
4,542
5,678
5,764
21 ,804
26,546
41,261
30
68
76
112
191
240
681
800
41,261
5.6
132
154
207
283
300
300
641
819
836
1,803
1,803
2,504
2,513
2,513
2,513
2,670
2,670
2,670
3,065
3,065
3,065
3,065
3,394
3,918
10,244
386,464
Area
(m2)
27
34
74
430
159
910
730
618
931
1,051
4,459
5,806
1 1 ,241
13
9.3
26
34
84
109
200
65
1 1 ,241
1.8
39
42
71
46
66
66
263
117
457
591
591
410
687
687
687
730
730
730
804
804
804
804
1,271
1,430
1,051
60,385
Depth
(m)
4.0
3.4
6.1
3.7
10.5
3.7
5.2
7.4
6.1
5.5
4.9
4.6
3.7
2.4
7.3
2.7
3.4
2.3
2.2
3.4
12.0
3.7
3.1
3.4
3.7
2.9
6.1
4.6
4.6
2.4
7.0
1.8
3.1
3.1
6.1
3.7
3.7
3.7
3.7
3.7
3.7
3.8
3.8
3.8
3.8
2.7
2.7
9.8
6.4
3-18
-------
Volume II Section 3.0
The depth of the waste was imputed from the reported tank volume (or capacity). This
was accomplished by developing a regression of log (depth) versus log (capacity) using data in
the supplemental data set discussed in Section 3.4.2 (49 tanks from the site visit/TSDF BID data-
base plus TSDF BID tanks). Because the site visit data did not include any very small tanks or
many very large tanks, a cube-shaped 55-gallon tank and a 3-million gallon/32-ft-deep tank
(based on the vendor information) was included in the regression derivation. Regression lines
were derived for aerated tanks (Equation 3-1) and nonaerated tanks (Equation 3-2), on the
assumption that these might have different volume-to-depth relationships since aerated tanks may
be shallower to facilitate aeration.
D = 1Q[0.1358 x log(V) + 0.2236] Q_^
D = lo^0'1334 x log^ + °-1657] (3-2)
where
D = depth (m)
V = volume (m3)
However, as can be seen in Figure 3-4, the two regressions were nearly identical.
Therefore, a single regression was developed using all 49 tanks from the site visit/TSDF BID
database plus the hypothetical tanks, as follows:
D = io[0-1057 x log(V) + °-2804] (3-3)
Comparisons of this regression with regressions done without one or both hypothetical
tanks indicate that the hypothetical tanks do not unduly dominate the regression.
This equation was then examined for the reasonableness of the depths predicted. Using
Equation 3-3, 60-gallon tanks (the smallest tanks in the database) are approximately 1.9 m
(6.2 ft) deep and about 39 cm (15.4 inches) in diameter. This seemed unreal!stically tall and
narrow; consequently, for very small tanks, a second equation was derived from the assumption
of a cube-shaped tank:
D = V0'333 (3-4)
At tanks of approximately 10 m3, Equation 3-3 predicts approximately cube-shaped tanks;
therefore, Equation 3-4 is used for tanks smaller than 10m3.
The largest tank in the TSDR database in the NO and LO aeration categories is
25,000 m3, and the projected depth for this tank using Equation 3-3 is 5.6 m (18 ft), which is
acceptable for mixing tanks.
The largest tank in the TSDR database in the HI aeration category is 23,000 m3, and the
projected depth for this tank is 5.5 m (18 ft). In evaluating the predicted depth of HI aeration, the
-------
Volume II
Section 3.0
100
10
B.
&
1 -
All tanks
Aerated
Nonaerated
10
100 1,000
Volume (m3)
10,000
100,000 1,000,000
• Actual Aerated
• Actual Nonaerated
Predicted - Aerated
A Hypothetical
Predicted - Nonaerated
Figure 3-4. Comparison of tank depth regression lines.
eight mechanically aerated tanks from the site visits were considered. The maximum depth from
these data was 6.1 m (20 ft) and even this appeared to be an outlier compared to the other HI
aeration, mechanically aerated tanks. Data for the other seven site visit tanks all have depths
ranging from 3.35 m (11 ft) to 4.88 m (16 ft). The mid-range of the latter depths is
approximately equivalent to a 1,000-m3 tank as evaluated using Equation 3-3. Therefore, for HI
aeration tanks greater than 1,000 m3, a random depth was assigned using a uniform distribution
with endpoints of 3.5 m and 4.8 m.
Table 3-10 summarizes the methods used to make an initial estimate of tank depth for
each type of tank. While these methods were intended to represent the actual relationship
between volume and depth as closely as possible, they imply a certain precision that is
unrealistic. In fact, there will be variation in the dimensions of tanks of the same volume. To
address that variation, a random variation was applied to these initial estimates using a normal
distribution with a mean of 1 and 90 percent of the values between 0.8 and 1.2. The initial depth
estimate was multiplied by this random factor to obtain a final depth estimate used in this
analysis.
Surface area data were not provided in the TSDR Survey. In the absence of these data,
surface area for each of the TSDR tanks was calculated by dividing tank volume by depth.
3-20
-------
Volume II Section 3.0
Table 3-10. Summary of Depth Imputation Techniques
Tank Type
HI Aeration
LO Aeration
NO Aeration
Volume
Range (m3) Imputation Technique
<10 Equation 3 -4
10-1,000 Equation 3 -3
>1,000 Uniform distribution from 3.5-4.8 m
<10 Equation 3 -4
>10 Equation 3 -3
<10 Equation 3 -4
>10 Equation 3 -3
The height of the top of the tank above the ground is needed for dispersion modeling.
Height is related to depth, but not necessarily equal to depth, as tanks may be partially in the
ground. In the absence of height data being reported in the TSDR Survey, height was imputed
from depth using a two-step process:
1. A number was selected at random from 0 to 20 (uniform distribution).
2. If this number was less than the depth in meters + 0.5, it was used as the height.
If it was greater than the depth in meters + 0.5, set height = depth + 0.5 m.
None of the tank depths imputed were greater than 9.5 m, therefore, none of the heights above
ground were more than 10 m (9.5 +0.5) using this method; 10m above the ground is the realistic
maximum height from a structural point of view, according to tank vendor contacts.
This approach establishes percentages of tanks of certain depths that will be all above-
ground vs. partially or completely in ground:
# For a 10-m tank, about half are partly or all in ground (when the random selection
is between 0 and 10; as the random number increases from 0 to 10, more and
more of the tank depth is aboveground, until at 10 all of it is), and about half are
all aboveground (when the random pick is between 10 and 20).
# For a 1-m tank, about 5 percent are partly or all in ground (random numbers from
0 to 1) and 95 percent all aboveground (random numbers from 1 to 20).
# For a 5-m tank, 25 percent are partly or all in ground (random numbers from 0
to 5) and 75 percent all aboveground (random numbers from 5 to 20).
3-21
-------
Volume II Section 4.0
4.0 Source Emission Estimates
This chapter describes the source-specific emission model and assumptions used to
develop the emission estimates for each waste management unit type. Section 4.1 discusses the
selection of a general volatilization model and a particulate emission model to use for the
emission estimates. Section 4.2 describes some of the critical model input parameters required to
run the volatilization and particulate emission models. Section 4.3 describes modifications made
to CHEMDAT8 for land-based units. Subsequent sections describe unit-specific modeling
scenarios and assumptions used for the volatilization model effort. The final section of this
chapter describes the particulate emission model estimates. Both volatile and particulate
emissions were estimated for landfills, land application units, and wastepiles (referred to as land-
based WMUs), while only volatile emissions were estimated for tanks.
4.1 Model Selection
4.1.1 Volatile Emission Model Selection
Several factors were considered in selecting emission models for assessing the potential
for contaminant exposure through inhalation. In developing acceptable contaminant limits for
wastes, the ideal emission model would provide emission estimates that are as accurate as
possible without underestimating the contaminant emissions. Because both volatile emissions
(for all WMU types) and particulate emissions due to wind erosion (for land-based WMUs) were
required in the risk analysis, the volatile emission model had to estimate both volatile emission
rates and long-term average soil concentration in the unit (for land-based WMUs). Ideally, the
model would provide a relatively consistent modeling approach (in terms of model complexity
and accuracy) for each of the different emission sources under consideration. Additionally, the
emission model would have to be reviewed both internally by EPA and externally by both state
and local agencies and industry representatives. Finally, the model would have to be publicly
available for use in more site-specific evaluations.
Based on these considerations, EPA's CHEMDAT8 model was selected as the model to
estimate volatile emission rates and long-term average soil concentrations in the WMU. The
CHEMDAT8 model was originally developed in projects funded by EPA's Office of Research
and Development (ORD) and Office of Air Quality Planning and Standards (OAQPS) to support
National Emission Standards for Hazardous Air Pollutants (NESHAPs) from sources such as
tanks, surface impoundments, landfills, wastepiles, and land application units for a variety of
industry categories including chemical manufacturers, pulp and paper manufacturing, and
petroleum refining. It also has been used to support the emissions standards for hazardous waste
treatment, storage, and disposal facilities (U.S. EPA, 1991) regulated under Subpart CC rules of
RCRA, as amended in 1984. The CHEMDAT8 model is publicly available and has undergone
-------
Volume II Section 4.0
extensive review by both EPA and stakeholder representatives. The CHEMDAT8 spreadsheet
model and model documentation may be downloaded at no charge from EPA's web page
(http://www.epa.gov/ttn/chief/software.html).
The CHEMDAT8 model considers most of the competing removal pathways that might
limit air emissions, including adsorption, hydrolysis (for tanks only), and biodegradation.
Adsorption/absorption is the tendency of a chemical or liquid media to attach or bind to the
surface or fill the pores of particles in the soil or waste and therefore not volatilize into the air.
This tendency to adsorb to or absorb in particles is an important process for estimating the
concentration of the chemical on particles emitted to the air due to wind erosion. CHEMDAT8
in its original form models adsorption for land-based units by presuming that the entered waste
concentration is in liquid phase. Because waste concentrations are more typically measured as
total concentration (liquid plus solid phase), CHEMDAT8 was modified to model adsorption
explicitly for an entered total waste concentration for land-based units. Biodegradation is the
tendency of a chemical to be broken down or decomposed into less complex chemicals by
organisms in the waste or soil. Similarly, hydrolysis is the tendency of a chemical to be broken
down or decomposed into less complex chemicals by reaction with water. Chemicals that
decompose due to either biodegradation or hydrolysis have lower potential for emission to the air
as gases or particles than those that do not. Loss of contaminant by leaching or runoff is not
included in the CHEMDAT8 model. Both leaching and runoff are a function of a chemical's
tendency to become soluble in water and follow the flow of water (e.g., due to rainfall) down
through the soil to groundwater (leaching) or downhill to surface water (runoff). These two
mechanisms would also result in less chemical being available for emission to the air as gases or
particles. As such, CHEMDAT8 is considered to provide reasonable to slightly high
(environmentally protective) estimates of air emissions from the land-based units.
The CHEMDAT8 model was used to estimate the emissions for all WMUs with some
modifications. CHEMDAT8 calculates a fraction of chemical emitted. Some additional
equations, which are described in Sections 4.4 through 4.7, were added to calculate emission
rates in g/m2-sec and remaining concentration in mg/kg from the fraction emitted and other
inputs. This document does not present the equations used by CHEMDAT8 to calculate fraction
emitted other than to show modifications made to model adsorbtion for total waste concentration
instead of just liquid-phase waste concentration. The reader interested in the CHEMDAT8
algorithms is referred to the CHEMDAT8 documentation (U.S. EPA, 1994e). Additionally,
certain equations were modified to prevent division by zero when certain volatilization
parameters (Henry's law constant or vapor pressure) were zero (e.g., for metals).1
1 Specifically, the CHEMDAT8 model was modified to prevent division by zero as follows:
• "If-statements" were added to set the biodegradation rate to a negligible level or zero and
prevented division by zero when no biodegradation rate constants were available.
• "If-statements" were added to prevent division by zero for chemicals that did not have vapor
pressure, Henry's law constant, or diffusivity inputs (e.g., metals).
4-2
-------
Volume II Section 4.0
4.1.2 Particulate Emission Model Selection
The model selection criteria for the particulate emission models were similar to those for
the volatilization model. Specifically, the particulate emission model would provide as accurate
emission estimates as possible without underestimating the contaminant emissions. The model
would provide a relatively consistent modeling approach (in terms of model complexity and
accuracy) for each of the different emission sources under consideration, and the emission model
would have to be both reviewed and publicly available for use for more site-specific evaluations.
Two different models were selected to model wind erosion: one for wastepiles (elevated
sources) and one for landfills and land application units (ground-level sources). Based on the
considerations above, the Cowherd model (U.S. EPA 1985b, 1988) was selected for modeling
wind erosion emissions from ground-level sources, and the AP-42 model for wind erosion from
aggregate storage piles (U.S. EPA, 1985a) was selected for modeling wind erosion emissions
from wastepiles. Newer versions of both of these models are available; however, the newer
versions are event-based algorithms that require extensive site-specific data that were not
available for the sites modeled in this analysis. The versions used probably result in somewhat
higher particulate emission estimates than the event-based algorithms would. This overestimation
of particulate emissions is not significant for volatile chemicals, as particulate emissions were
found to be a negligible fraction (less than 2 percent in most cases) of total emissions for the
volatile chemicals modeled in land-based units. The protective waste concentrations (Cw's) for
metals other than mercury (which do not volatilize and are therefore based solely on particulate
emissions) may be somewhat lower as a result of this overestimation of emissions.
4.2 Emission Model Input Parameters
This section discusses the various parameters that impact the estimated volatilization and
particulate emission rates. Inputs that influence these rates include input parameters specific to
the physical and chemical properties of the constituent being modeled, the physical and chemical
characteristics of the waste material being managed, input parameters specific to the process and
operating conditions of the WMU being modeled, and meteorological parameters.
A general discussion of the physical and chemical properties of the constituents is
provided in Section 4.2.1. Critical input parameters for the remaining sets of inputs are discussed
first for land-based WMUs and then for tanks. A sensitivity analysis was performed for the 1998
Air Characteristic Study to better understand the impact of certain modeling assumptions on the
model results. While the models and data have changed somewhat, those changes would not alter
the conclusions drawn; therefore, these sensitivity analyses are not included here. The interested
reader is referred to Appendix C of the May 1998 Air Characteristic Study.
4.2.1 Chemical-Specific Input Parameters
Key chemical-specific input parameters include: air-liquid equilibrium partitioning
coefficient (vapor pressure or Henry's law constant), liquid-solid equilibrium partitioning
coefficient (log octanol-water partition coefficient for organics), biodegradation rate constants,
and liquid and air diffusivities. The Hazardous Waste Identification Rule (HWIR) chemical
properties database (RTI, 1995; U.S. EPA, 1995b) was used as the primary data source for the
4-3
-------
Volume II Section 4.0
physical and chemical properties for the constituents being modeled. This chemical properties
database provided the following chemical-specific input parameters: molecular weight, vapor
pressure, Henry's law constant, solubility, liquid and air diffusivities, and log octanol-water
partition coefficient. Soil biodegradation rate constants were obtained from Howard et al.
(1991). The CHEMDAT8 chemical properties database (U.S. EPA, 1994b) was used as a
secondary data source for the physical and chemical properties for the constituents being
modeled. This chemical properties database provided the following chemical-specific input
parameters: density, boiling point, Antoine's coefficients (to adjust vapor pressure for different
temperatures), and biodegradation rate constants for tanks. The biodegradation rate constants in
the downloaded CHEMDAT8 database file were compared with the values reported in the
summary report that provided the basis for the CHEMDAT8 tank biodegradation rate values
(Coburn et al., 1988). Tank biodegradation rate constants for compounds with no data were
assigned biodegradation rates equal to the most similar compound in the biodegradation rate
database (or set to zero for metals). The specific chemical properties input database used for
emission modeling is provided in Appendix B.
4.2.2 Critical Input Parameters for Land-Based WMU Emission Models
4.2.2.1 Volatile Emissions and Waste Concentration. The input parameters used for
the CHEMDAT8 land-based unit emissions model are presented in Table 4-1. (Note: The data
entry form in the CHEMDAT8 model refers to oil rather than waste; the term waste is used here
for clarity.) Of these parameters, two are actually flags to determine which model equations to
apply (Input ID Nos. L7 and L9). The most important flag for emission estimates is probably the
aqueous waste flag (Input ID No. L7). This flag tells the CHEMDAT8 model which equilibrium
partitioning model to use between the liquid and gas phases. For organic wastes, the model uses
Raoult's law and the liquid-to-air partition coefficient becomes proportional to the contaminant's
partial vapor pressure. For aqueous wastes, the model uses Henry's law and the liquid-to-air
partition coefficient becomes proportional to the contaminant's Henry's law coefficient. All land-
based WMUs were run twice; once assuming unit concentration (concentration set to 1 mg/kg,
assuming Henry's law applies) and once assuming pure component (concentration set to 1E+6
mg/kg, assuming Raoult's law applies). The results presented in Volumes I and III are based on
the aqueous phase emission rates (unit concentration and Henry's law). The pure component
emission rates were used only to identify chemicals for which greater emissions occur from the
organic phase than from the aqueous phase (which is rare) or to identify chemicals for which the
aqueous-based results exceeded soil saturation concentrations, and note for these whether the
target risk or hazard quotient would be exceeded modeling pure component.
The annual waste quantity is a critical source (site-specific input) parameter. This
parameter along with assumptions concerning the frequency of contaminant addition and the
dimensions of the unit combine to influence a number of model input parameters (Input ID Nos.
L1,L2,L3,L8, andL12).
The CHEMDAT8 model is insensitive to windspeeds for long-term emission estimates
from land-based units. Temperature affects the air diffusivity, which affects the volatilization
rate and potentially affects the biodegradation rate (biodegradation rates were independent of
temperature above 5°C and were set to zero below 5°C). Consequently, temperature is the only
4-4
-------
Volume II
Section 4.0
Table 4-1. CHEMDAT8 Land-Based Unit Model Input Requirements
Input ID No.
Input Parameter
Data Source/Assumption
LI Loading (g waste/cm3 soil)
L2 Concentration in waste (ppmw)
L3 Depth of tilling (or unit) (cm)
L4 Total porosity
L5 Air porosity (0 if unknown)
L6 Molecular weight of waste
L7 For aqueous waste, enter 1
L8 Time of calculation (d)
L9 For biodegradation, enter 1
L10 Temperature (°C)
Lll Windspeed (m/s)
L12 Area (m2)
LI3 Fraction organic carbon
Waste quantity and/or density from Ind D Survey
1 for unit concentration run; 1E+6 for pure component run
Assumed or set by capacity
Assumed default value of 0.5
Assumed default value of 0.25
18 for unit concentration run; 147 for pure component run
1 for unit concentration run; 0 for pure component run
Dependent on type of WMU
Dependent on type of WMU
Set by location of WMU
Set by location of WMU
Input from Ind D Survey
Assigned randomly from distribution
meteorological data input that potentially impacts the emission results for the CHEMDAT8
model for the land-based WMU.
The total porosity and air porosity values that were used in the emission assessments were
the default CHEMDAT8 model values for these parameters. These assumed porosity values
appear to be reasonable for a waste and waste/soil matrices that have a density of 1.1 g/cm3.
For aqueous wastes, the molecular weight of the waste (Input ID No. L6) does not impact
the calculations.
The molecular weight of the waste for the "pure component" runs using Raoult's law was
set to 147 g/mol, which is the CHEMDAT8 default value for this input parameter. If the waste
were truly pure constituent, then the appropriate molecular weight input for the waste would be
the specific constituent's molecular weight. However, the pure component run is used to
backcalculate an appropriate waste concentration limit that is often considerably less than pure
component. Therefore, the scenario modeled is not actually pure constituent, and modeling the
waste at the molecular weight of the constituent is not appropriate. If the actual molecular
weight of the waste is higher than 147 g/mol, the default molecular weight used may
underestimate volatile emissions. Conversely, if the actual molecular weight of the waste is
lower than 147, the default value may overestimate volatile emissions. The magnitude of this
under- or over estimation is expected to be small over the range of likely waste molecular
weights.
The process of biodegradation is an important one because it lowers both the emission
rate and the average soil concentration. Consequently, biodegradation is an important input
parameter, and the biodegradation rate constants used in the model are critical parameters.
4-5
-------
Volume II Section 4.0
Biodegradation was treated differently for the various WMUs. Landfills are not designed for
biodegradation, and waste in wastepiles managed over short periods will not be affected
substantially. Therefore, both the landfill emission runs and the short-term wastepile emission
runs did not include biodegradation losses. First-order biodegradation was included in the LAU
emission runs and long-term wastepile emission runs. Note that the default CHEMDAT8 model
method of calculating biodegradation rates was not used. CHEMDAT8 biodegradation rates
were derived primarily from wastewater studies and applied to model biodegradation in soils
using an assumed, low-biomass concentration. Because the first-order biodegradation rate
constants obtained from Howard et al. (1991) were either based on soil studies or explicitly
evaluated for applicability to soil, they were used instead. These biodegradation rate constants
provide a more direct link to soil-based biodegradation and are considered more appropriate for
modeling biodegradation in land-based WMUs.
The fraction of organic carbon, foc, affects adsorption. The foc of interest is the foc of the
waste (assuming the waste contains sludge/solids) and not of the soil. Little data exist
concerning the foc of the waste itself; therefore, default values for this parameter were defined and
applied in the absence of data. A distribution was developed to represent foc at all sites. Because
this parameter is a fraction, it must range from 0 to 1. A beta distribution was selected for the
distributional form because the beta distribution also varies between 0 and 1. This distribution is
defined by two parameters called alpha and beta. Fraction organic carbon is waste- and site-
specific but is most often less than 0.1. Therefore, the distribution was fitted with the criteria
that half the values generated should be less than 0.1, and that 90 percent of the values should be
less than 0.5. The fitted distribution has an alpha value of 0.455 and a beta value of 2.05.
Individual foc values for each WMU were then selected randomly from the distribution.
4.2.2.2 Particulate Emissions. Particulate emissions due to wind erosion were modeled
for land-based units (landfills, land application units, and wastepiles). Landfills and LAUs were
modeled differently than wastepiles because they are ground-level sources and wastepiles are
elevated sources. Wind erosion emissions from landfills and LAUs were modeled using the
Cowherd model (U.S. EPA, 1985b, 1988). This model estimates the emission of respirable
particles (i.e., PM10) due to wind erosion from a ground-level surface with an unlimited reservoir
of erodible particles. Surfaces are defined as having a limited or unlimited reservoir based on
threshold friction velocity (U*); surfaces with a U* greater than 0.5 m/s are considered limited;
those with U* less than 0.5 m/s are considered unlimited (U.S. EPA, 1988). Threshold friction
velocity is a measure of the windspeed at the ground surface that would be required to remove
particles from the surface. Examples of limited reservoirs include nonhomogeneous surfaces
with stones, clumps of vegetation, or other nonerodible elements or crusted surfaces. Further,
wind erosion is considered unlikely to occur from surfaces with full vegetative cover.
Wind erosion emissions from wastepiles were modeled using an equation from AP-42
(U.S. EPA, 1985a) for estimating emissions from wind erosion from active storage piles. The
equation gives emissions of total suspended particulates (TSP). Typically, an equation-specific
particle size multiplier would be applied to reduce the emissions to a desired size category, in
this case, PM10. No particle size multipliers (PSM) are given for this equation in AP-42;
however, Cowherd (U.S. EPA, 1988) gives a PM10 particle size multiplier of 0.5 for use with this
equation.
4-6
-------
Volume II Section 4.0
Important input parameters for particulate emissions include silt content of waste (i.e.,
percent with small particle size), number of days with greater than 0.01 inches of rainfall, and
percent of time that windspeed exceeds 5.4 m/s. Data on the silt content of the wastes being
modeled were not available. A median silt content for miscellaneous fill material of 12 percent
(U.S. EPA, 1988) was used. The number of precipitation days and the frequency of windspeed
greater than 5.4 m/s were location-specific; values were obtained from the National Oceanic and
Atmospheric Administration (NOAA, 1992) and are summarized in Section 4.7.2.
4.2.3 Critical Parameters for Tank Emissions Model
Table 4-2 presents the required CHEMDAT8 input parameters2 along with units and
comments on the source of the parameter values. As shown in Table 4-2, only one parameter
(flow rate) has values that are taken directly from the TSDR survey data as discussed in Section
3.4 (that is, the data were provided by facility owner/operators at the time the survey was
conducted). Volume data, which were used to impute a number of model input parameter values
for tanks, were also taken directly from the TSDR survey data. The imputation procedures for
the aeration and waste characteristics parameter values are discussed in this section. The
procedures applied to estimate unit design parameters, which are also critical to air dispersion
modeling, are discussed in Section 3.4.
Factors that affect the relative surface area of turbulence and the intensity of that
turbulence are important in determining the fate of chemicals in tanks. The tank model has
several input parameters that impact the degree and intensity of the turbulence created by the
aeration (or mixing). Note that many of these parameters are determined for both aerated and
some nonaerated (both treatment and storage) tanks. For nonaerated tanks with an above-surface
intake (based on information available from the survey data), a small degree of aeration is
modeled to account for the agitation of splash loading. Although there is not actually an aerator
in these tanks, one is characterized to simulate the effects of splash loading. For nonaerated
tanks with below-surface intakes, no agitation is modeled, and the aeration parameters are all set
to zero. The values of most of the aeration parameters were estimated based on data collected in
1985 and 1986 during EPA site visits to aerated treatment systems (Coburn et al., 1988;
Eichinger, 1985) and hypothetical tanks from the TSDF BID (U.S. EPA, 1991) (see Section 3.4.2
for further details). Aerator parameters for 16 tanks and surface impoundments (which are
expected to have similar aerator properties to tanks) from these site visits and the TSDF BID are
shown in Table 4-3. These data were used to provide a sense of the range and typical values of
the aeration parameters.
The tank model is most sensitive to the fraction aerated; the total power, number of
aerators, and impeller diameter have some impact on the emission results; and the other
parameters have little to no impact on the estimated emissions.
2 Note that this table also includes one parameter, height, that is not used in CHEMDAT8 but is used in
dispersion modeling. For clarity, it is presented here as it is derived from depth.
-------
Volume II
Section 4.0
Table 4-2. CHEMDAT8 Tank Model Input Requirements
Input Parameter
Source
Imputation Method
Unit Design
Flow rate (m3/s)
Depth (m)
Average surface area (m2)
Height above ground (m)
Survey
Imputed
Imputed
Imputed
None
Based on volume
Based on volume and depth
Based on depth
Aeration Parameters
Fraction agitated
Total power (hp)
Number of impellers
Impeller diameter (cm)
Impeller speed (rad/s)
Power efficiency (unitless)
O2 transfer rate (lbO2/h-HP)
Submerged air flow (m3/s)
Estimated distribution
Imputed
Imputed
Estimated constant =61
Estimated constant =130
Estimated constant = 0.83
Estimated constant = 3
Estimated constant = 0
Based on volume
Based on total power
Waste Characteristics
Active biomass cone, (kg/m3)
Total solids in (kg/m3)
Total organics (COD) In (g/m3)
Total biorate (mg/g-h)
Estimated distribution
Estimated distribution
Estimated distribution
Estimated constant =19
Depends on treatment code
Meteorological Data
Temp(°C)
Imputed
Based on meteorological station
Windspeed (m/s)
Imputed
Based on meteorological station
The fraction aerated depends on the level of aeration. Distributions for this parameter
were developed for each aeration level. Highly aerated tanks should have a higher fraction
aerated than less-aerated tanks. No tank can have a fraction aerated greater than 1 or less than
zero, and realistically, the fraction aerated for an aerated tank should not be close to zero. A non
aerated tank may have a small fraction aerated to simulate the agitation from splash loading from
an above-surface intake. For HI aeration tanks, fraction aerated is randomly assigned from a
normal distribution with a mean of 0.75 and a standard deviation of 0.1. Values greater than 1 are
truncated to 1. For LO aeration tanks, fraction aerated is randomly assigned from a uniform
distribution with endpoints of 0.2 and 0.8. For NO aeration and storage tanks with above-
surface intakes, fraction aerated was randomly assigned from a normal distribution with a mean
of 0.08 and a standard deviation of 0.03. Values less than zero were truncated at zero, and values
that implied an area agitated more than 10m2 were truncated so that the agitated area was 10 m2.
The rationale for this is that splash loading would not affect an area greater than 10 m2. For NO
aeration tanks with below-surface intakes, the fraction aerated was set to zero.
4-8
-------
Volume II
Section 4.0
Table 4-3. Summary of Mechanical Aerator Information Collected
in EPA Site Visits for RCRA Air Emission Standards
Type of unit
Aeration tank
Aeration tank
Aeration tank
Aux. aer. tank
Aeration tank
Aeration tank
Aerated trtmnt tank
Aerated trtmnt tank
So eq. basin
No eq. basin
Mixing tank
Eq. basin
Mixing tank
Aerated lagoon
Aerated lagoon
Aerated lagoon
Aeration
ffl
ffl
ffl
ffl
ffl
ffl
ffl
ffl
LO
LO
LO
LO
LO
SI
SI
SI
Total
Power #
(hp) Aerators
300
7.5
900
450
900
150
7.5
120
30
20
3
150
1.5
30
270
1800
3
1
9
6
6
2
2
1
1
5
1
2
6
28
Aerator
Power
Oxygen for Total
Impeller Impeller transfer Power/ Power
diameter speed rate volume >100 hp
(cm) (rad/s) (Ib/hp-h) (hp/m3) (hp)
259 188
259 126
274
50 124
107 7.1
152 5.9
122 7.1
183 1.0
14 367
42 123
0.053
0.067
3.0 0.022
3.0 0.021
0.034
3.0 0.045
0.069
0.075
0.125
0.105
0.027
0.004
0.022
0.009
0.011
0.040
100
100
75
150
75
30
45
64
Total aerator power depends on the volume of the tank and the level of aeration. For HI
aeration tanks, total power per million gallons of volume was randomly assigned using a normal
distribution with a mean of 115 hp/million gallons and 90 percent of the values between 80 and
150 hp/million gallons (Metcalf and Eddy, 1979). For LO aeration tanks, total power per million
gallons of volume was randomly assigned using a normal distribution with a mean of 30
hp/million gallons and 90 percent of the values between 15 and 45 hp/million gallons This was
based on industry comments on the 1998 ACS that indicated that surface aerators used for
mixing typically have power levels between 15 and 20 hp/million gallons, and aerators used for
activated sludge have a minimum power of 20 to 30 hp/million gallons. The upper value of 30
was adjusted upward by a factor of 1.5 to provide values above the minimum, and the resulting
range of 15 to 45 hp/million gallons was presumed to encompass 90 percent of all values. For
both HI and LO aerated tanks, the power per million gallons was multiplied by the volume in
millions of gallons to determine total power. For NO aeration and storage tanks, the power per
million gallons was determined in the same way as for LO aeration tanks. However, the total
power was then calculated by multiplying the power per million gallons by fraction aerated and
volume (in million gallons) to estimate total power. A minimum value of 0.25 hp was set.
Fraction aerated is included in this calculation to account for the fact that the whole tank volume
is not affected by the agitation caused by splash loading.
The number of aerators (or impellers) for aerated tanks was derived from the total power
and the power per aerator. Typical values of power per aerator for different total power levels
were based on the data in Table 4-3 and the number of aerators set as follows:
-------
Volume II Section 4.0
# For Total Power < 25 HP, one aerator
# For 25 FTP < Total Power < 80 FTP, randomly pick one or two aerators with equal
probability
# For Total Power > 80 FTP, randomly pick a power per aerator using a uniform
distribution with endpoints of 60 and 100, then divide the Total Power by the
random power per aerator and round up to the next integer. While some of the
power per aerator values associated with the data in Table 4-3 fell outside the
range of 60 to 100, this range represented most of the values.
For nonaerated tanks, the number of impellers was set to 1 if the tank had an above-surface
intake and to zero if it had a below-surface intake.
Impeller diameter and rotational speeds appeared to be related parameters. Generally, the
longer (or high-diameter) impellers found in the site visit data (see Table 4-3) had lower
rotational speeds and the shorter impellers had faster rotational speeds. Consequently,
independent random assignments of these variables was determined to be inappropriate. Rather
than attempting to develop a correlation between the two parameters based on limited data, the
fixed values used for the model tanks developed for the Hazardous Waste TSDF air rules (U.S.
EPA, 1991) were selected. These values are reasonable central tendency values based on the
limited available data presented in Table 4-3:
# Impeller diameter fixed at 61 cm (2 ft), based on data reported by Watkins (1990)
# Impeller rotational speed fixed at 130 rad/s, based on data reported by Watkins
(1990)
# Oxygen transfer rating and power efficiency do not have much impact on the
emission, and vary over a very small range. Therefore, these values were fixed, as
follows:
# Oxygen transfer rating fixed at 3.0 Ib O2/HP-h; U.S. EPA (1991) reports a range
of2.9to31bO2/hp-h
# Power efficiency fixed at 0.83; U.S. EPA (1991) reports a range of 0.8 to 0.85.
Submerged air flow was set to zero for all tanks because aeration is modeled as
mechanical aeration, not diffused air.
Waste characteristics that influence the rate of biodegradation are important in
determining emissions from both aerated and storage tanks. As shown in Table 4-2, these
parameters include active biomass concentration, total solids in, total organics in, and total
biorate. Limited data were available on waste characteristics for tanks; most facilities only need
to measure these parameters for the final wastewater discharge and do not measure them in the
influent. From a review of the site visit reports described previously, the data estimations
4-10
-------
Volume II Section 4.0
discussed below were derived and used for aerated tanks reporting biological treatment. Aerated
tanks reporting other types of treatment, nonaerated treatment tanks, and storage tanks were
modeled with no biodegradation.
Unlike the biodegradation rate model that was used for the land-based units, the
biodegradation rate model used in CHEMDAT8 for tanks is dependent on the amount of active
biomass in the WMU. Therefore, the active biomass concentration is a critical parameter for
aerated tanks. Because this parameter can vary widely for different types of tanks, biomass
concentrations were set on a tank-by-tank basis for aerated tanks using process code information
(WMU codes) from the TSDR Survey. For biological treatment aerated tanks, the active
biomass concentration measured as mixed liquor volatile suspended solids (MLVSS) from the
site visit reports range up to 4 g/L, and one source test measured MLVSS up to 6 g/L. However,
the typically observed MLVSS concentration fell in the 1.5- to 3.0-g/L range. Many of the
biodegradation rate constants developed for CHEMDAT8 used 2 g/L as a default MLVSS
concentration to normalize the constituent disappearance rates. Therefore, 2 g/L was considered
the most appropriate central tendency value for the active biomass concentration. Consequently,
the following algorithm was used to select the active biomass concentration.
# For biological treatment units (WMU codes 52WT and 58WT), this value was
randomly assigned for each tank, using a uniform distribution with endpoints of 1
and 3 g/L
# For all other aerated tanks (as well as storage and nonaerated treatment tanks),
active biomass concentration was set to 0 g/L.
The "total biomass(solids) in," "total organics in," and "total biorate" (Input ID Nos. T6,
T8, and T9) impact the rate of biomass production and subsequently the amount of contaminant
that is absorbed onto the solids. (Note: The "biomass solids in" does not affect the
biodegradation rate and is more appropriately labeled simply "solids in."). These inputs,
however, have little or no impact on the estimated emission rates for most of the contaminants
modeled in this analysis. The value for "total solids in" was randomly assigned for each tank
using a uniform distribution with endpoints of 0.1 and 1 g/L, based on best professional
judgment. CHEMDAT8 (U.S. EPA, 1994e) suggests a range of 0.1 to 0.4 for surface
impoundments designed for biodegradation. The "total organics in" value was randomly
assigned for each tank, using a uniform distribution with endpoints of 100 and 1,000 mg/L, based
on best professional judgment. The CHEMDAT8 default value is 250 mg/L (U.S. EPA, 1994e).
The input parameter listed as "total biorate, mg/gbio-h" is used in conjunction with the
active biomass concentration to estimate the growth or replacement rate of biomass, e.g., how
much of the "old" biomass is consumed to grow "new" biomass that is now available for
contaminant adsorption. This biomass replacement rate is used in conjunction with the influent
total solids and total organic concentrations to determine the total rate at which total suspended
solids (TSS) are removed (or "wasted") from the system. Because this input parameter impacts
only tanks that have active biomass greater than zero (a small fraction of the total number of
tanks in the data set) and this biomass replacement factor will generally have only a small, if any,
contribution to the adsorptive losses, no additional research was performed for this parameter.
4-11
-------
Volume II Section 4.0
Instead, the total biorate is fixed at 19 mg/g biomass-h, which is the default value recommended
for CHEMDAT8 (U.S. EPA, 1994e).
Due to the nonlinearity of the biodegradation rate model used in the tank emission
estimates, direct backcalculation of an acceptable waste concentration may not be appropriate for
some compounds. Unlike the emission results from the land-based units, the contaminant
concentration used in the analysis may impact the predicted "normalized" emission rate (i.e., the
emission rate in g/m2-s per mg/L of contaminant). Therefore, the tanks with biodegradation were
run at a low concentration (i.e., 0.001 mg/L) and at a high concentration (i.e., the constituent's
solubility). The most appropriate backcalculated emission value was then selected based on the
concentration range of the backcalculated values and the constituent's biodegradation
characteristics (see Section 7.9 for further details).
Meteorological inputs are also important for the tank emission model. For nonaerated
treatment tanks and storage tanks, the emission estimates are impacted by both temperature and
windspeed. Because the emissions for aerated tanks are predominantly driven by the turbulent
area and associated mass transfer coefficients, the emissions from the aerated tanks are not
strongly impacted by the windspeed. Aerated tank emissions are impacted by temperature.
Annual average temperatures were used as input to the model based on tank locations. (Note
that, dependent on the residence time of the waste in the tank, the temperature of the waste in the
tank was not expected to vary significantly with changing atmospheric temperatures, and annual
average temperatures were used to estimate the average waste temperature in the tanks). The
location of each tank is available from the TSDR survey data. Based on this information, each
tank was assigned to one of the 29 meteorological stations used in the dispersion modeling. As
discussed in Section 5, these assignments were made based on both proximity and similarity of
the climatological characteristics that affect meteorological data. The windspeed and
temperature used in emissions modeling are the annual averages for the assigned meteorological
station, taken from NOAA (1992).
4.3 Modifications to CHEMDAT8 for Land-Based Units
The CHEMDAT8 model estimates emissions from land-based WMUs (such as land
application units, open landfills, and wastepiles) using a simple emissions model that accounts
for contaminant partitioning between a liquid waste matrix and the air, diffusion of vapors
through a porous media, and contaminant loss through biodegradation. The CHEMDAT8 model,
however, does not accommodate entered total waste concentrations (i.e., liquid and solid phase).
The assumption of an entered waste concentration in liquid phase was based on the petroleum
wastes for which CHEMDAT8 was originally developed and may not apply to the chemicals
considered in this analysis. Therefore, a method for including adsorptive partitioning for total
waste concentrations was developed and is presented below.
Assuming three-phase partitioning (adsorbed, dissolved, and volatile), the total
concentration of a contaminant can be expressed as the sum of the masses of the contaminant
adsorbed on the soil, dissolved in the liquid, and in the air spaces divided by the total mass of
contaminated soil as follows:
4-12
-------
Volume II Section 4.0
=CsPb + QWCW + QaCa (4-1)
where
CT = total contaminant concentration (|ig/cm3soil = g/m3soil)
Cs = concentration of contaminant adsorbed on soil (|ig/gsoil = g/Mgsoil)
pb = soil dry bulk density (gsoil/cm3soil = Mgsoil/m3soil)
6W = water-filled soil porosity (m3water/m3soil)
Cw = concentration of contaminant dissolved in liquid (|ig/cm3water = g/m3water)
6a = air-filled soil porosity (m3air/m3soil)
Ca = concentration of contaminant in air dig/cm3.^ = g/m3air).
The adsorbed contaminant concentration is assumed to be linearly related to the liquid
phase concentration as follows:
(4-2)
where
Cs = concentration of contaminant adsorbed on soil (|ig/gsoil = g/Mgsoil)
Kd = soil-water partition coefficient (cm3/g = m3/Mg)
Cw = concentration of contaminant dissolved in liquid (|ig/cm3water = g/m3water)
For organic constituents:
K, = K f (A.3\
a ocj oc \* J)
where
Koc = organic-carbon partition coefficient (cm3/g = m3/Mg)
foc = weight fraction organic carbon content of the solid matrix (g/g = Mg/Mg).
The contaminant concentration in the vapor phase is assumed to be linearly related to the
liquid phase concentration as follows:
Ca = H'Cw or Cw = —f (4-4)
where
H = dimensionless Henry's law constant = H/RT = 41 xHat25 °C where
H = Henry's law constant at 25 °C (atm-m3/mol).
4-13
-------
Volume II Section 4.0
Combining Equations 4-1, 4-2, and 4-4 by replacing Cs in Equation 4-1 with the term in
Equation 4-2 and Cw in Equation 4-1 with the term in Equation 4-4 yields the following
expression in terms of the gas phase concentration Ca:
(4-5)
H1 * H/+ a
The total contaminant concentration, CT, represents the measured soil concentration.
Equation 4-5 can be rearranged to calculate the gas phase concentration given the total
contaminant concentration as follows:
CTH
K,p,+Q + Q H
a rb w a
(4"6)
This partitioning theory, as represented by the above equations, was used to include
adsorption in CHEMDAT8, as described below.
The CHEMDAT8 land treatment emission model is based on the diffusion of a gas from
a semi-infinite slab that initially has a uniform concentration of diffusing material throughout and
that has equal concentrations of diffusing material at each surface (U.S. EPA, 1994e). The
emission equations presented in CHEMDAT8 are in terms of the mass (as opposed to
concentration) of contaminant in the gas phase.
CHEMDAT8 uses an equilibrium partitioning factor, Keq, as a multiplier to correct the
effective diffusion coefficient. The partitioning factor, Keq, represents the ratio of the mass of
organics in the vapor phase to the mass of organics in the soil/waste mixture and, therefore, is
used to estimate the amount of material that partitions into the vapor phase based on equilibrium
conditions within the soil/waste mixture. The CHEMDAT8 solution, as described in the
CHEMDAT8 documentation, requires a ratio of the total mass of contaminant in the gas phase to
the total mass of contaminant in the soil/waste matrix. The mass ratio for the partitioning
correction factor (including adsorption) that was used in the CHEMDAT8 model, Keq ads,
therefore, was defined as follows:
_ a
Keq, adS = ~ (4-7)
MT
where
Ma = mass of constituent in the air-filled soil porosity (g)
MT = total mass of constituent in the soil/waste mixture (g).
4-14
-------
Volume II Section 4.0
The masses of constituent in the air-filled soil porosity and in the soil/waste mixture are
equal to their respective concentrations times volumes:
M C V
(4-8)
MT CT VT
where
Va = total volume in the air-filled soil porosity (m3)
VT = total volume of the soil/waste mixture (m3).
Using the relationship between Ca and CT presented in Equation 4-6, and the definition of
porosity shown in the following equation,
(4-9)
the terms in Equation 4-8 can be substituted for, and Keq ads can be rewritten as follows:
6 H
b d
The CHEMDAT8 model was modified to include Equation 4-10. This equation presents
an expression for Keq ads that achieves the goal of including adsorptive partitioning of total waste
concentration in the model. This equation always yields a partitioning value of 1 or less. At high
Henry's law values, Keq ads is necessarily equal to 1, providing the same emission rate predictions
as if the total initial mass of contaminant was in the vapor phase with no partitioning.
4.4 Development of Volatile Emissions and Waste Concentrations for
Landfills
The basic assumptions used for modeling landfills are as follows:
# The landfill operates for 20 years filling 20 cells of equal size sequentially.
# The active cell is modeled as being instantaneously filled at time t=0 and remains
open for 1 year.
# Emissions are calculated only for one cell for 1 year (after 1 year, the cells are
either depleted of the constituent or capped).
# The waste is homogeneous with an initial concentration of 1 mg/kg.
-------
Volume II Section 4.0
# The waste matrix may be aqueous (Henry's law partitioning applies) or organic
(Raoult's law partitioning applies).
# Annual average temperature is used (determined by assigned meteorological
station).
# Acute and subchronic exposures were not modeled.
The results presented in Volumes I and III are based on the aqueous-phase emission rates
(i.e., assuming a concentration of 1 mg/kg and Henry's law partitioning). Most of the waste
streams managed in land-based units are expected to contain constituents in the aqueous, rather
than the organic, phase; therefore, this is the most realistic scenario. However, results based on
organic phase emissions are of interest in two circumstances:
# Most chemicals are better able to volatilize from an aqueous medium than from an
organic medium; therefore, for most chemicals, the aqueous-phase emission rates
are considerably higher than organic-phase emission rates. However, for a few
chemicals (most notably formaldehyde), the organic-phase emissions are higher
than the aqueous-phase emissions. When this is the case, the results based on
aqueous-phase emissions are footnoted to indicate that organic-phase emissions
would be higher, so a concentration based on organic-phase emissions would be
lower. This does not invalidate the aqueous-phase results, as that is still the most
likely waste matrix.
# Some of the backcalculated waste concentrations based on aqueous-phase
emissions exceed the soil saturation concentration at a neutral pH and temperature
of 20 to 25 °C. This is the maximum possible aqueous-phase concentration in
soil; once this is exceeded, free (organic-phase) product will occur in the soil.
This soil saturation concentration has been estimated for each chemical in the
analysis, but this is a somewhat site- and waste-specific value because it depends
on solubility, soil properties such as bulk density and porosity, and temperature
(see Section 7.10.2 for the equation used to calculate the soil saturation
concentration). Therefore, a backcalculated concentration may exceed it in some
situations but not in others. When the backcalculated concentration based on
aqueous-phase emissions exceeded the calculated soil saturation concentration at
a neutral pH and temperature of 20 to 25 °C, the result was footnoted to indicate
whether pure component (i.e., a concentration of 106 mg/kg) would result in a risk
exceeding the target risk when modeled using organic phase assumptions and
Raoult's law.
Table 4-1 provides the CHEMDAT8 model input requirements for land-based units with
some commentary about each input parameter. The inputs that were calculated from the
Industrial D Screening Survey data were calculated as follows:
4-16
-------
Volume II Section 4.0
# All total quantities, capacities, and areas in the Industrial D Screening Survey
were divided by the number of landfills at the facility to get landfill-specific
estimates.
# Loading = bulk density =1.1 g/cm3.
# Tilling depth (cm) = landfill depth, 1, calculated as follows:
100 x C
D..n =
where
Dtill = landfill depth (cm)
100 = unit conversion (cm/m)
C = capacity (Mg)
A = landfill area (m2)
BD = bulk density (g/cm3 = Mg/m3)
If the calculated depth was less than 2 feet or more than 33 feet, then the method
described in Section 3.1 was used.
# Total landfill surface area was divided by 20 to get surface area of landfill cell.
# The total landfill capacity was divided by 20 to get the average annual quantity of
waste, Qannual.
The landfill cell areas and depth were entered into the CHEMDAT8 input table (along
with average ambient temperature), and the emission fraction for the "intermediate time"
(365.25 days) was calculated. This emission fraction was then multiplied by the annual waste
quantity and waste concentration and divided by the area of the cell to calculate the output
emission rate as follows:
O x C1 x f
T^ _ ^--annual waste J emit
Acell x 31,557,600 ^ " '
where
E = emission rate (g/m2-s)
Qannuai = annual waste quantity (Mg/yr)
Cwaste = waste concentration (mg/kg = g/Mg)
femit = fraction emitted (unitless)
Acell = cell area (m2)
31,557,600 = unit conversion factor (s/yr).
4-17
-------
Volume II
Section 4.0
The average concentration of the waste in the landfill cell was estimated from the
emission fraction and the biodegradation fraction (although the biodegradation fraction was
zero - no biodegradation - for the landfill) by assuming first-order contaminant (concentration)
disappearance. Assuming first-order kinetics with respect to the contaminant concentration in
the landfill cell, an exponential decay can be written in terms of the apparent overall first-order
decay rate. The concentration at a given time is equal to the initial concentration as follows:
'-'waste,! '-'waste.o
t)
(4-13)
where
'-'waste, t
waste, 0
, all
waste concentration at time t (mg/kg)
waste concentration at time 0 (mg/kg)
apparent first-order decay rate (yr"1)
time period of calculation (yr).
At the end of 1 year, Cwaste t/Cwaste 0 = 1 - emission fraction - biodegraded fraction.
Therefore, the KM1 t term, at the time period for which the fraction loss terms were calculated, is
simply:
K, „ t = -In (\-f .-/",. ) (4-141
l,a// ^ J emit J bio' \* A~/
where
KI, an = apparent first-order decay rate (yr"1)
t = time period of calculation (yr)
femit = fraction emitted (unitless)
fbio = fraction biodegraded (unitless).
The concentration versus time profile (Equation 4-13) can then be integrated to calculate
the average waste concentration, Cwasteave, over the time period of the calculation:
C = C
waste,ave waste,®
(4-15)
where
r
v' waste, ave
waste, 0
, all
average waste concentration (mg/kg)
waste concentration at time 0 (mg/kg)
apparent first-order decay rate (yr"1)
time period of calculation (yr).
The input parameters required for the landfill are presented in Table 4-1. The annual
waste quantity and unit dimensions are the critical source parameters. For landfills, the loading
4-18
-------
Volume II Section 4.0
rate is pure waste material so that loading (Input ID No. LI) is basically the waste density. A
waste density of 1.1 g/cm3 was used for the landfill to be consistent with the waste densities used
in the analysis of the Industrial D Screening survey data. The annual waste quantity is also
combined with the area of the landfill to calculate the depth of the landfill.
Temperature and porosities have some impact on predicted emissions. The
biodegradation flag was set to zero (no biodegradation) for landfills. Therefore, temperature
variations should have less of an impact on the annual emission rates from landfills than from
land application units. The model is insensitive to molecular weight of the waste (for aqueous
wastes) and windspeed (for long-term emission estimates).
4.5 Development of Volatile Emissions and Waste Concentrations for Land
Application Units
4.5.1 Chronic Exposure Analysis
Because the same basic CHEMDAT8 model was used for landfills and land application
units, the emission estimates for land application units have some similarities to the landfill
emission estimates, but there are also a number of differences. The basic modeling assumptions
used for modeling land application units are as follows:
# The land application unit emissions are modeled as pseudo-steady-state.
Emissions are actually time-dependent (depending on how recently waste has
been added) but are modeled as a series of steady-state emissions for short time
intervals, which are then averaged to produce a long-term emission rate.
# Emissions in year 40 are used to estimate long-term emissions. This does not
reflect an assumed operating life of the unit but is simply a sufficiently long
period to ensure that steady state has been reached, if it is ever going to be,
(typically, steady state is reached in 1 or 2 years) and to exceed most of the
exposure durations used in the modeling.
# Waste application occurs twice monthly (i.e., 24 times per year).
# The waste is homogeneous with an initial concentration of 1 mg/kg.
# The waste matrix may be aqueous (Henry's law partitioning applies) or organic
(Raoult's law partitioning applies).
# Temperature is determined by assigned meteorological station; monthly average
temperature was used.
# Biodegradation occurs at temperatures greater than 5 °C.
The results presented in Volumes I and III are based on the aqueous-phase emission rates
(i.e., assuming a concentration of 1 mg/kg and Henry's law partitioning). Most of the waste
-------
Volume II Section 4.0
streams managed in land-based units are expected to contain constituents in the aqueous, rather
than the organic, phase; therefore, this is the most realistic scenario. However, results based on
organic phase emissions are of interest in two circumstances:
# Most chemicals are better able to volatilize from an aqueous medium than from an
organic medium; therefore, for most chemicals, the aqueous-phase emission rates
are considerably higher than organic-phase emission rates. However, for a few
chemicals (most notably formaldehyde), the organic-phase emissions are higher
than the aqueous-phase emissions. When this is the case, the results based on
aqueous-phase emissions are footnoted to indicate that organic-phase emissions
would be higher, so a concentration based on organic-phase emissions would be
lower. This does not invalidate the aqueous-phase results, as that is still the most
likely waste matrix.
# Some of the backcalculated waste concentrations based on aqueous-phase
emissions exceed the soil saturation concentration at a neutral pH and temperature
of 20 to 25 °C. This is the maximum possible aqueous-phase concentration in
soil; once this is exceeded, free (organic-phase) product will occur in the soil.
This soil saturation concentration has been estimated for each chemical in the
analysis, but this is a somewhat site- and waste-specific value because it depends
on solubility, soil properties such as bulk density and porosity, and temperature
(see Section 7.10.2 for the equation used to calculate the soil saturation
concentration). Therefore, a backcalculated concentration may exceed it in some
situations but not in others. When the backcalculated concentration based on
aqueous-phase emissions exceeded the calculated soil saturation concentration at
a neutral pH and temperature of 20 to 25 °C, the result was footnoted to indicate
whether pure component (i.e., a concentration of 106 mg/kg) would result in a risk
exceeding the target risk when modeled using organic phase assumptions and
Raoult's law.
The inputs that were calculated for the land application units were calculated as follows:
# The total annual waste quantities and surface areas for each facility, as reported in
the Industrial D Screening Survey, were divided by the number of LAUs at the
facility to get LAU-specific estimates.
# Tilling depth (cm) = 20 cm if Qannual (Mg/yr)/Area(m2) < 0.2 . If Qannual/Area > 0.2,
then depth (cm) = 100 x Qannual/Area.
# Loading rate, L, is calculated as follows:
Q i x t
T _ ^annual
L — (4-16)
A x D
4-20
-------
Volume II Section 4.0
where
L = loading rate (Mg/m3 = g/cm3)
Qannuai = annual waste quantity (Mg/yr)
t = time period of calculation (yr)
A = LAU area (m2)
Dtill = tilling depth (m).
# Biodegradation is assumed to occur if temperature is greater than 5 °C. If the
temperature is 5°C or lower, biodegradation is turned off.
# Time of calculation = 365.25/24 applications per year =15.2 days.
# Monthly temperature and windspeeds were calculated by averaging the hourly
temperature and windspeeds.
A sensitivity analysis was conducted that investigated the impact that application
frequency (monthly versus quarterly) and the averaging period for meteorological data (monthly
versus yearly) had on emission estimates. Results from this analysis indicated that both
application frequency and averaging period do impact emissions. Therefore, modeling was
conducted using monthly average meteorological data. In the absence of reported unit-specific
data on application frequency in the Industrial D database, it was assumed that waste was applied
to the unit two times during each month of the year (i.e., 24 times per year).
Land treatment generally involves the application of wastes to the land in either a liquid
or a semi-solid form with treatment occurring through the biological degradation of the
hazardous constituents. Waste is assumed to be delivered by tank trucks and is applied uniformly
across the entire unit area. Based on the literature reviewed, it appears that the frequency at
which waste is to be applied is dependent on a number of variables, including waste
characteristics (e.g., constituent concentrations and oil content), soil type, vegetation, and
climatic conditions. The application frequencies found in literature range from yearly to an
extreme of 260 times per year. U.S. EPA (1989), which cites data presented in Land Treatment
Practices in the Petroleum Industry (EnvironmentalResearch & Technology, 1983), reports a
typical range for refineries of 2 to 52 applications per year. Martin et al. (1986) presents
frequencies for 13 operating petroleum refineries that represent the geographical distribution of
refineries. The reported frequencies ranged from 1 to 260 times per year, with the majority of the
sites reporting monthly or yearly frequencies. The Handbook of Land Treatment Systems for
Industrial and Municipal Wastes (Reed and Crites, 1984) reports that an application of once per
week is commonly used but suggests that determination of application frequency should consider
site-specific conditions. One of the site-specific variables that can impact the frequency of
application is the number of months out of the year the unit is active. Martin et al. (1986)
reported that the number of months facilities were actively used varied from 6 months (colder
climates) to 12 months (warmer climates).
Based on the information from the literature review described above, the relationship
between number of applications per year and waste quantity shown in Table 4-4 was developed.
4-21
-------
Volume II
Section 4.0
The relationships shown in Table 4-4 were applied to the Industrial D data set for land
application units. Table 4-5 summarizes the frequency of waste application estimated using the
Industrial D data. As shown, approximately 86 percent of the land application units were
predicted to have application frequencies 52 times or less per year and only 26 percent were
estimated to have waste applied 4 times or less per year. The median number of applications per
year is 16.
It was not feasible from a modeling perspective to assign different frequencies of
application to each unit. Also, the use of monthly meteorological data suggested an application
frequency that could be expressed as an integer on an applications per month basis. Therefore,
an application frequency of 24 applications per year (2 applications per month) was selected as
best representing the data available within these modeling constraints.
Table 4-4. Relationship Between Frequency of Application and Waste Quantity
Annual Waste Quantity at Unit
(Mg/yr)
Number of applications/year
<1,500
>1,500, <15,000
>15,000, <150,000
>150,000
Annual waste quantity at unit/ 15
Annual waste quantity at unit/150
Annual waste quantity at unit/ 1,5 00
Annual waste quantity at unit/ 15, 000
Table 4-5. Estimated Frequency of Application Using Industrial D Data
Frequency of Waste
Application
-------
Volume II Section 4. 0
The CHEMDAT8 model was run for each monthly meteorological condition using half-
month (15.2 days) "time of calculation."
The fraction emitted and the fraction biodegraded for each of the monthly meteorological
conditions, as calculated by CHEMDAT8, were stored and used to project the disappearance and
accumulation of contaminant due to sequential applications. Given the initial concentration just
after a waste application, the ending concentration (just prior to the next application) is
calculated as follows:
^end,n ~ ^ start,n ^ J emit ~ J bio) (4-17)
where
Cend,n = ending soil concentration after nth application (mg/kg)
Cstart,n = initial soil concentration after nth application (mg/kg)
femit = fraction emitted (unitless)
fbio = fraction degraded (unitless).
It is assumed that the volume of the land application unit remains constant. Therefore, as more
waste is applied, it is assumed that an equal volume of waste/soil mixture becomes buried or
otherwise removed from the active tilling depth. Including the contaminant loss by
burial/removal, the initial soil concentration for any given application number is calculated as
follows:
CStart,n = C end, «-l X I * ~ + CWaSte X (4-18)
where
Cstart,n = initial soil concentration after nth application (mg/kg)
Cend,n-i = ending soil concentration after nth-1 application (mg/kg)
L = loading rate (g/cm3)
BD = bulk density (g/cm3)
CWaste = waste concentration (mg/kg).
Assuming no background contaminant concentration, Cendo = 0 mg/kg. Equations 4-17
and 4-18 were solved sequentially using the fraction emitted and fraction degraded for Month 1
(January) twice (once for each 15.2-day interval in Month 1), then twice using the fraction
emitted and fraction degraded calculated for Month 2, etc., until semimonthly concentrations
had been calculated for 40 years (the assumed life of the unit). For the 40th year, the starting and
ending concentrations of the month were used to calculate a log-mean average soil concentration
for each month. These 12 monthly soil concentrations were then arithmetically averaged to yield
an average long-term concentration, which was then output for each of the 40 years modeled.
4-23
-------
Volume II Section 4.0
Long-term emission rates were calculated by multiplying the initial soil concentration for
each application rate by the fraction emitted, the bulk density, the tilling depth, and then by
dividing by time period of the calculation as follows:
C x f x BD x D „
j^ _ start, n J emit till ,. * Qx
" t x 86,400 "
where
En = emission rate over n* application period (g/m2-s)
start,n
Cstart,n = initial soil concentration after nth application (mg/kg)
= fraction emitted (unitless)
BD = bulk density (g/cm3 = Mg/m3)
Dtill = tilling depth (m)
t = time period of calculation (d)
86,400 = unit conversion factor (s/d).
The long-term emission rates were calculated for all 24 waste applications during the 40th year,
then the arithmetic average of these emission rates was calculated and output.
The input parameters required for the land application unit are presented in Table 4-1.
The annual waste quantity and unit dimensions are the critical source parameters. The area of the
land application unit is a site-specific input parameter. The tilling depth was assumed to be 20
cm unless the annual quantity of waste added to the land application unit was greater than the
volume of the land application unit with a set depth of 20 cm. If the annual waste quantity
exceeded the volume of the land application unit given a depth of 20 cm, the depth of the land
application unit was calculated so that the land application unit could hold the entire annual
waste quantity. Twice-monthly waste applications were employed (time of calculation, Input ID
No. L8, was set to 15.2 days).
A waste density of 1 . 1 g/cm3 was used to test the unit size and to calculate the tilling
depth, if needed. This waste density was used for the land application unit to be consistent with
the waste densities used in the analysis of the Industrial D Screening survey data. Air porosity
has the greatest influence on the predicted emissions of these parameters. Nonetheless, because
these parameters typically do not vary over wide ranges, they are considered secondary
parameters for the emission estimates.
Because biodegradation both lowers the emission rate and the average soil concentration,
the biodegradation flag is an important input parameter, and the biodegradation rate constants
used in the model are also critical parameters. The biodegradation flag was set to 1 for all LAU
unit runs; however, the default CHEMDAT8 model method of calculating the biodegradation
rate was not used. Instead, values for first-order biodegradation rate constants were obtained
from Howard et al. (1991). These biodegradation rate constants are based, for the most part, on
contaminant half-lives in soil. The biodegradation rates used in CHEMDAT8 were derived
4-24
-------
Volume II
Section 4.0
primarily from wastewater studies, and then a low biomass concentration was used in
CHEMDAT8 to model biodegradation in soils. Because the Howard et al. biodegradation rate
constants were more directly linked to soil-based WMUs, they were used for all land-based
WMUs that included biodegradation.
The Taylor series approximation used to estimate emissions in CHEMDAT8 is inaccurate
for chemicals of very low volatility and low biodegradability. For a few chemicals with the
lowest Henry's law constants, this resulted in chronic emissions greater than acute emissions
(which are instantaneous rates based on an exact solution rather than the Taylor series
approximation). For chemicals with very low volatility and biodegradability, the instantaneous
emission rates are not expected to change much over time. Therefore, for these chemicals, the
acute (instantaneous) emission rates and soil concentrations were also used for chronic exposure
estimates.
4.5.2 Acute and Subchronic Exposure Analysis
Acute (1-day) and subchronic (1-month) emission estimates were made for LAUs. These
emission estimates used the same general modeling approach as used for the chronic approach;
however, the emission estimates were made for the first 24 hours immediately after a waste
application (during the 40th year, to ensure steady-state had been reached) or averaged over 30
days. As with the chronic exposure emission estimates, biodegradation was included in the
model.
The acute emission estimates were
projected from the CHEMDAT8
"instantaneous" emission values by
weighting the instantaneous emission rate
with a corresponding time interval (see
box). The "instantaneous" emission rates
were then normalized by the initial
concentration used for the calculation, and
normalized emission rates were output for
each of the 12 monthly meteorological
conditions. The appropriate normalized
emission rate was then multiplied by the
corresponding initial soil concentration for
each of the applications occurring during
the 40th year. The maximum (1-day)
emission rate and the maximum initial
concentration occurring during the 40th year
were output.
The subchronic emission rates used the same emission fractions as calculated for the
chronic exposure. Average monthly emissions were estimated using the paired application
emission rates for January, February, and so on. The maximum monthly emission rate and the
maximum average monthly concentration occurring during the 40th year were output.
Time at Which
Instantaneous
Emissions were
Calculated
0.25 h
Ih
4h
10 h
18h
Period of Time
Represented by
Instantaneous
Emissions
0-0.5h
0.5 -2 h
2-6h
6-14h
14-24h
4-25
-------
Volume II Section 4.0
Acute and subchronic emission estimates were made for both unit and pure component
concentrations. The Taylor series approximation used to estimate short-term (i.e., subchronic)
emissions in CHEMDAT8 is inaccurate for chemicals of very low volatility and low
biodegradability. For six chemicals with the lowest Henry's law constants, this resulted in
subchronic emissions greater than acute emissions (which are instantaneous rates based on an
exact solution rather than the Taylor series approximation). For chemicals with very low
volatility and biodegradability, the instantaneous emission rates are not expected to change much
over time. Therefore, for these six chemicals, the acute (instantaneous) emission rates and soil
concentrations were also used for subchronic exposure estimates.
4.6 Development of Volatile Emissions and Waste Concentrations for
Wastepiles
4.6.1 Chronic Exposure Analysis
The same basic CHEMDAT8 model that was used for landfills and land application units
was also used for wastepiles. Similar to landfills, waste is applied, and the basic modeling
equations used for landfills apply. However, subsequent calculations of the wastepile results are
required to account for monthly variations, similar to the land application units. The basic
modeling assumptions used for modeling land application units are as follows:
# The wastepile operates with a fixed volume; depths (heights) ranging from 1 to 10
meters were assigned based on waste quantity and surface area; modeling was
conducted for various wastepile height/surface area strata combinations (see
Section 3.2).
# The waste is homogeneous with an initial concentration of 1 mg/kg.
# The waste matrix may be aqueous (Henry's law partitioning applies) or organic
(Raoult's law partitioning applies).
# Monthly average temperatures are used; temperature determined by assigned
meteorological station.
# Biodegradation occurs at temperatures greater than 5 °C.
The results presented in Volumes I and III are based on the aqueous-phase emission rates
(i.e., assuming a concentration of 1 mg/kg and Henry's law partitioning). Most of the waste
streams managed in land-based units are expected to contain constituents in the aqueous, rather
than the organic, phase; therefore, this is the most realistic scenario. However, results based on
organic phase emissions are of interest in two circumstances:
# Most chemicals are better able to volatilize from an aqueous medium than from an
organic medium; therefore, for most chemicals, the aqueous-phase emission rates
are considerably higher than organic-phase emission rates. However, for a few
chemicals (most notably formaldehyde), the organic-phase emissions are higher
4^26
-------
Volume II Section 4.0
than the aqueous-phase emissions. When this is the case, the results based on
aqueous-phase emissions are footnoted to indicate that organic-phase emissions
would be higher, so a concentration based on organic-phase emissions would be
lower. This does not invalidate the aqueous-phase results, as that is still the most
likely waste matrix.
# Some of the backcalculated waste concentrations based on aqueous-phase
emissions exceed the soil saturation concentration at a neutral pH and temperature
of 20 to 25 °C. This is the maximum possible aqueous-phase concentration in
soil; once this is exceeded, free (organic-phase) product will occur in the soil.
This soil saturation concentration has been estimated for each chemical in the
analysis, but this is a somewhat site- and waste-specific value because it depends
on solubility, soil properties such as bulk density and porosity, and temperature
(see Section 7.10.2 for the equation used to calculate the soil saturation
concentration). Therefore, a backcalculated concentration may exceed it in some
situations but not in others. When the backcalculated concentration based on
aqueous-phase emissions exceeded the calculated soil saturation concentration at
a neutral pH and temperature of 20 to 25 °C, the result was footnoted to indicate
whether pure component (i.e., a concentration of 106 mg/kg) would result in a risk
exceeding the target risk when modeled using organic phase assumptions and
Raoult's law.
Inputs that were calculated for the wastepile were calculated as follows:
# The total annual waste quantities and surface areas for all wastepiles at each
facility, as reported in the Industrial D Screening Survey, were divided by the
number of wastepiles at the facility to get wastepile-unit-specific estimates.
# Loading = waste bulk density =1.1 g/cm3.
# Biodegradation is assumed to occur if temperature is greater than 5 °C. If the
temperature is 5 °C or lower, biodegradation is turned off.
# Time of calculation = average residence time of waste in the wastepile, tave, as
follows:
A x D,.,, x BD x 365.25
'™ = ^ (4-20)
*£ annual
where
tave = average residence time of waste in the wastepile (d)
A = wastepile areas (m2)
Dtill = tilling depth (m)
BD = bulk density (g/cm3 = Mg/m3)
365.25 = unit conversion factor (d/yr)
= annual waste quantity (Mg/yr).
4-27
-------
Volume II Section 4.0
# The same monthly temperature and windspeeds calculated for land application
units were used for wastepiles.
The average waste concentration and emission rate for wastepiles can be calculated using
the equations presented for the landfill model (Equations 4-12 and 4-15).
The same input parameters required for land application units and landfills were used for
wastepiles (see Table 4-1). The annual waste quantity and unit dimensions are still the critical
source parameters. As with landfills, the loading for wastepiles is pure waste material so that the
loading is basically the waste density. A waste density of 1.1 g/cm3 was used for the wastepile to
be consistent with the waste densities used in the analysis of the Industrial D Screening survey
data. Wastepile emission (and concentration) estimates were made for wastepiles representing
various height/surface area combinations (see Section 3.2). The annual waste quantity combined
with the capacity (dimensions) of the wastepile were used to calculate residence time for the
waste in the wastepile (i.e., time of calculation - Input ID No. L8).
Temperature and porosities have some impact on emissions. Biodegradation was
assumed to occur for wastepiles. The model is insensitive to molecular weight of the waste (for
aqueous wastes) and windspeed (for long-term emission estimates).
4.6.2 Acute and Subchronic Exposure Analysis
Acute (1-day) and subchronic (1-month) emission estimates were made for wastepiles
following the same approach used for LAUs as described in Section 4.5.2, with the following
differences:
# The wastepile was assumed to be completely filled with new waste at time t = 0.
# The time of calculation used for subchronic exposure was 1 month (365/12 days).
# Biodegradation was not included (a lag period is generally associated with
biodegradation, especially for the assumed "all new waste" scenario. This is
because wastepiles are not designed for biodegradation, so it takes time for
enough acclimated biomass to accumulate for biodegradation to occur. This time
is typically several months, longer than the 1-day or 30-day periods modeled for
acute and subchronic).
# Since there was no biodegradation, only maximum-temperature runs were needed,
which maximize volatile emissions.
4.7 Development of Volatile Emissions for Tanks
The basic modeling assumptions used (or inherent in CHEMDAT8) for the tank model
emission estimates include:
# The WMU operates at steady state.
4^28
-------
Volume II Section 4.0
# The WMU is well mixed.
# Emission estimates were performed for three different influent concentrations: an
influent concentration of either 0.001 (for aerated tanks) or 1 mg/L (for
nonaerated and storage tanks; the low concentration modeled was set to a
different, lower, value for aerated tanks to ensure first-order biodegradation), an
influent concentration equal to the constituent's solubility (to estimate aqueous-
phase emissions with zero-order biodegradation), and an influent concentration of
1E+6 mg/L (pure component).
# The waste matrix may be aqueous (Henry's law partitioning applies) or organic
(Raoult's law partitioning applies).
# Annual average temperatures are used; temperature is determined by assigned
meteorological stations.
# Biodegradation rate is first order with respect to biomass concentrations. (Only
some aerated tanks were modeled with biodegradation.)
# Biodegradation rate follows Monod kinetics with respect to contaminant
concentrations.
# Hydrolysis rate is first order with respect to contaminant concentrations.
# Acute and subchronic exposures were not modeled.
For tanks with biodegradation (some aerated tanks), biodegradation shifts from first order
at low concentrations to zero order at high concentrations. To account for this, two runs were
done using aqueous phase waste: one at 0.001 mg/L (a concentration low enough to ensure first-
order biodegradation for all chemicals) and one at the chemical's solubility (the maximum
possible for aqueous phase, which will produce zero-order biodegradation for all chemicals that
shift to zero-order at concentrations below solubility). Backcalculated concentrations were
initially calculated using the first-order biodegradation emissions. If this concentration exceeded
the half-saturation concentration for the chemical, suggesting that zero-order biodegradation
would be more appropriate, it was replaced with a backcalculated concentration based on the
zero-order emission rate.
For tanks, the volume and flow were taken directly from the TSDR Survey. Other unit
design parameters, such as surface area and depth, were imputed from the survey data. Modeling
was conducted for both treatment (aerated and nonaerated) and storage tanks.
The CHEMDAT8 model is used to calculate the emission fractions for each tank. The
emission rate, in g/m2-s, is calculated from the fraction emitted, the flow rate, waste
concentration, and the surface area as follows:
4-29
-------
Volume II Section 4. 0
v /~* x T^1
infl emit (4 71">
A
where
E = emission rate (g/m2-s)
Qflow = flow rate (m3/s)
Cjnfi = waste concentration in influent (mg/L = g/m3)
Femit = fraction emitted (unitless)
A = tank surface area (m2).
4.8 Development of Particulate Emissions
Two different models were used to model wind erosion - one for wastepiles (elevated
sources) and one for landfills and land application units (ground-level sources). The Cowherd
model (U.S. EPA, 1985b, 1988) was selected for modeling wind erosion emissions from ground-
level sources, and the AP-42 model for wind erosion from aggregate storage piles (U.S. EPA,
1985a) was selected for modeling wind erosion emissions from wastepiles.
For both types of WMU, the models described in this section predict the emission rate of
particulate matter released from a site due to wind erosion. To obtain the emission rate of
constituent sorbed to particulate matter, the emission rate of particulate matter must be multiplied
by the soil or waste concentration. That calculation is described in Section 7.4.
4.8.1 Landfills and Land Application Units
Wind erosion emissions from landfills and LAUs were modeled using the Cowherd
model (U.S. EPA, 1985b). A newer version of Cowherd's model is available in U.S. EPA
(1988). However, the newer version is an event-based model that requires detailed site-specific
information unavailable for this analysis. Therefore, it was not used. The older Cowherd model
tends to slightly overestimate emissions relative to the event-based version. Although the degree
to which it overestimates is not known, it is expected to be relatively small. Because particulate
emissions are negligible compared to volatile emissions for the volatile chemicals modeled, this
is only of concern for the metals (other than mercury), which are based only on particulate
emissions.
The Cowherd model estimates the emission of respirable particles (i.e., PM10) due to wind
erosion from a ground-level surface with an unlimited reservoir of erodible particles. Surfaces
are defined as having a limited or unlimited reservoir based on threshold friction velocity (U* );
surfaces with a U* greater than 0.5 m/s are considered limited, while those with U* less than 0.5
m/s are considered unlimited (U.S. EPA, 1988). Threshold friction velocity is a measure of the
windspeed at the ground surface that would be required to remove particles from the surface.
Examples of limited reservoirs would include nonhomogeneous surfaces with stones, clumps of
4-30
-------
Volume II Section 4.0
vegetation, or other nonerodible elements or crusted surfaces. Further, wind erosion is
considered unlikely to occur from surfaces with full vegetative cover.
Wind erosion emissions were calculated as follows (U.S. EPA, 1985b):
/ r ,
Ew = 0.036 x (i-V) x M x F(X) (4.22)
where
E10 = emission rate of PM10 (g/m2-h)
V = vegetative cover (fraction)
[u] = average annual windspeed (m/s)
Ut = threshold windspeed (m/s).
F(x) is a complex function of x, which is a ratio of threshold windspeed and average
annual windspeed. The following is a linear approximation of F(x) from a graph of F(x) in U.S.
EPA (1985b, Figure 4-3), except for the equation for values of x greater than 2, which is given in
Appendix B of U.S. EPA (1985b).
1.91 x<0.5
2.06-0.33* 0.52
x = 0.886 x —L (4-24)
[u]
The threshold windspeed, Ut, is a function of threshold friction velocity and roughness
height, as follows:
(4.25)
where
U, = threshold windspeed (m/s)
Z = anemometer height (cm)
z0 = roughness height (cm)
U* = threshold friction velocity (m/s).
4-31
-------
Volume II Section 4.0
The inputs used for these equations (other than average annual windspeed and
anemometer height, which are location-specific determined by the meteorological station) are
summarized in Table 4-6.
The vegetative cover, V, is an important modifying factor that can range from 0 (bare
ground) to 1 (100 percent vegetative cover). Because this analysis focuses on active industrial
landfills and land application units, the assumption of no vegetative cover was applied (i.e.,
vegetative cover was set to 0). For active industrial landfills, it was assumed that the constant
filling and dumping taking place would not allow for vegetative growth to occur. Active land
application units were assumed to be tilled and thus were also assumed to have no vegetative
cover.
The roughness height, z0, accounts for the size and spacing of surface roughness elements
(which may include vegetation, or even buildings). U.S. EPA (1985b) provides a range of
roughness heights for various surfaces ranging from 0.1 cm (natural snow) to 1,000 cm (high-rise
buildings). A value of 1 cm, corresponding to a plowed field, was selected as most
representative of the conditions expected to occur in an active landfill or LAU.
The threshold friction velocity, U*, is a function of the physical condition of the soil and
would normally be considered a site-specific parameter. In the absence of data, a value of 0.5
m/s was selected based on the assumption that the WMU is an unlimited reservoir.
The following inputs and intermediate values are location-specific: anemometer height,
threshold windspeed, average annual windspeed, x, and F(x). These are shown in Table 4-7
along with the calculated particulate emissions rate for each meteorological location used in the
analysis. The anemometer heights and average annual windspeed values are from NOAA (1992).
Values for Ut, x, F(x), and E10 have been rounded to two significant figures.
4.8.2 Wastepiles
Wind erosion emissions from wastepiles were modeled using an equation from AP-42
(U.S. EPA, 1985a) for estimating emissions from wind erosion from active storage piles. The
equation gives emissions of total suspended particulates. Typically, an equation-specific particle
size multiplier would be applied to reduce the emissions to a desired size category, in this case,
PM10. No particle size multipliers are given for this equation in AP-42; however, Cowherd
(U.S. EPA, 1988) gives a PM10 particle size multiplier of 0.5, which was used with this equation.
Table 4-6. Inputs and Intermediate Values Used for Wind
Erosion from Landfills and LAUs
Symbol
V
zo
U*
Parameter
Vegetative cover
Roughness height
Threshold friction velocity
Units
fraction
cm
m/s
Value
0
1
0.5
Source
Assumption
U.S. EPA (1985b)
Assumption
4-32
-------
Volume II
Section 4.0
Table 4-7. Calculated Particulate
Location
Albuquerque
Atlanta
Bismarck
Boise
Casper
Charleston
Chicago
Cleveland
Denver
Fresno
Harrisburg
Hartford
Houston
Huntington
Las Vegas
Lincoln
Little Rock
Los Angeles
Miami
Minneapolis
Philadelphia
Phoenix
Portland, ME
Raleigh-Durham
Salem, OR
Salt Lake City
San Francisco
Seattle
Winnemucca
Emissions
Z(cm)
700
610
610
610
610
610
610
1010
1010
610
610
1010
610
610
610
610
610
910
700
1010
610
1010
610
610
610
610
1010
610
1010
were calculated
E 19
U, (m/s)
8.2
8.0
8.0
8.0
8.6
8.0
8.0
8.6
8.6
8.0
8.1
8.6
8.0
8.0
8.0
8.0
8.0
8.5
8.2
8.6
8.0
8.6
8.0
8.0
8.0
8.0
8.6
8.0
8.6
as follows:
x PSM x
Emission Rates for Landfills and LAUs
[U] X
(m/s) (unitless)
4.1
4.6
6.2
4.6
7.2
4.1
4.6
5.1
4.1
3.6 :
4.6
4.1
4.1
3.6 :
5.1
5.1
3.6 :
4.1
4.6
5.7
4.6
3.1 :
4.6
4.1
4.6
4.6
6.2
5.1
4.1
.8
.5
.1
.5
.1
.7
.5
.5
.9
..0
.6
.9
.7
..0
.4
.4
..0
.8
.6
.3
.5
..5
.5
.7
.5
.5
.2
.4
.9
S 365 -p f
1.5 235 15
F(x)
(unitless)
0.06
0.89
1.4
0.89
1.5
0.65
0.89
0.95
0.47
0.34
0.86
0.47
0.65
0.34
1.1
1.1
0.34
0.51
0.85
1.2
0.89
0.060
0.89
0.65
0.89
0.89
1.3
1.1
0.47
(g/m'-h)
0.0027
0.0061
0.023
0.0061
0.032
0.0031
0.0061
0.0070
0.0018
0.0011
0.0056
0.0018
0.0031
0.0011
0.010
0.010
0.0011
0.0020
0.0054
0.012
0.0061
0.00010
0.0061
0.0031
0.0061
0.0061
0.017
0.010
0.0018
(4-26)
where
Mo
PSM
P
f
emission rate of PM10 (kg/ha-d)
particle size multiplier for PM10 (unitless) = 0.5
silt content of waste (%)
number of days with >0.25 mm (0.01 in) of precipitation per year (d/yr)
percentage of time that the unobstructed windspeed exceeds 5.4 m/s
(12 mph) (%).
4-33
-------
Volume II Section 4.0
The emission rate was converted to g/m2-h as follows:
E x 103
E = — (4-27)
104 x 24
where
E' = emission rate (g/m2-s)
E = emission rate (kg/ha-d)
103 = unit conversion factor (g/kg)
104 = unit conversion factor (m2/ha)
24 = unit conversion factor (h/d).
Silt content varies depending on the waste stored in the wastepile. AP-42 contains a table
of typical silt and moisture contents of materials at various industries. Based on these data, a
default value of 12 percent was used. This value was identified in the AP-42 table as the mean
silt content for miscellaneous fill materials.
The number of precipitation days and the frequency of windspeed greater than 5.4 m/s
were location specific. These data were collected at the meteorological stations and not at the
waste management unit; values were obtained from NOAA (1992) and are summarized in
Table 4-8 for each of the 29 meteorological locations modeled. Also included in Table 4-8 are
the paniculate emission rates in both kg/ha-d and g/m2-h that were calculated using Equations
4-26 and 4-27; these values have been rounded to two significant figures. The particle size
multiplier value of 0.5 used in calculating these estimates was chosen based on 10-|im particles
and was taken directly from U.S. EPA (1988).
Table 4-8. Calculated Particulate Emission Rates for Wastepiles
Location
Albuquerque
Atlanta
Bismarck
Boise
Casper
Charleston
Chicago
Cleveland
Denver
Fresno
Harrisburg
Hartford
Houston
P
(d/yr)
58
116
96
91
95
113
125
157
89
89
125
126
101
f
(%)
22.0
21.0
33.2
21.0
47.3
19.1
31.7
33.6
20.9
7.4
16.6
21.9
16.3
E
(kg/ha-d)
15
11
19
12
28
10
16
15
12
4.4
8.6
11
9.3
E
(g/m2-h)
0.061
0.047
0.080
0.052
0.11
0.043
0.068
0.063
0.052
0.018
0.036
0.047
0.039
(continued)
4O4
-------
Volume II Section 4.0
Table 4-8. (continued)
Location
Huntington
Las Vegas
Lincoln
Little Rock
Los Angeles
Miami
Minneapolis
Philadelphia
Phoenix
Portland, ME
Raleigh-Durham
Salem, OR
Salt Lake City
San Francisco
Seattle
Winnemucca
P
(d/yr)
142
27
91
104
33
128
113
117
37
129
110
146
92
63
157
67
f
(%)
8.2
25.9
31.1
14.4
14.7
25.5
35.2
25.6
6.8
23.0
14.5
15.2
20.1
37.4
22.1
17.2
E
(kg/ha-d)
3.9
19
18
8.1
11
13
19
14
4.8
12
8.0
7.2
12
24
9.9
11
E
(g/m2-h)
0.016
0.079
0.077
0.034
0.044
0.054
0.080
0.057
0.020
0.049
0.033
0.030
0.049
0.10
0.041
0.046
4-35
-------
Volume II Section 5.0
5.0 Dispersion Modeling
Dispersion describes the transport of chemical emissions through the air to a receptor. In
this risk analysis, dispersion modeling was used to estimate air concentrations associated with a
unit emission (1 |ig/m2-s) (unitized air concentrations, or UACs) at a variety of potential receptor
locations. The following sections discuss overview of the approach, model selection, the critical
parameters of the model, and the model results or UACs.
5.1 Overview of Approach
A dispersion model (ISCST3) was run to calculate air concentrations associated with a
standardized unit emission rate (1 |ig/m2-s) to obtain a unitized air concentration (UAC), also
called a dispersion factor, which is measured in |ig/m3 per |ig/m2-s. Total air concentration
estimates are then developed by multiplying the constituent-specific emission rates derived from
CHEMDAT8 with this dispersion factor.
Running ISCST3 to develop a dispersion factor for each of the approximately 3,400
individual WMUs modeled in this study would have been very time consuming due to the run
time of the area source algorithm in ISCST3. In addition, modeling for many different locations
requires extensive preprocessing to generate the detailed meteorological data needed for each
location modeled. Therefore, a database of dispersion factors was developed by running ISCST3
for many separate scenarios designed to cover a broad range of unit characteristics, including:
# both ground-level and elevated sources;
# 14 surface area sizes for landfills and land application units, 29 surface area-
height combinations for waste piles, and 33 surface area height combinations for
tanks;
# 6 receptor distances from the unit (25, 50, 75, 150, 500, and 1,000 meters) placed
in 16 directions in relation to the edge of the unit.
Based on the size and location of a specific unit, an appropriate dispersion factor was
interpolated from the database of dispersion factors using the closest location and the two closest
unit sizes.
In addition, WMUs were assigned to and dispersion modeling was performed for 29
meteorological stations. These were chosen from the more than 200 available to represent the
nine general climate regions of the continental U.S.
5-1
-------
Volume II
Section 5.0
Each UAC in the database is specific to one meteorological station, one area-height
combination, one distance from the unit, and one direction from the unit.
5.2 Model Selection
A number of dispersion models are available on the EPA Support Center for Regulatory
Air Models (SCRAM) Bulletin Board (http://www.epa.gov/scram001/). These dispersion
models were developed for a variety of applications and each has its own strengths and
weaknesses. This analysis required a model with the capability to model the dispersion of vapors
and particulates from landfills, land application units, wastepiles, and tanks to off-site receptors
for chronic, subchronic, and acute averaging times. Therefore, a dispersion model was needed
that could model (1) area sources; (2) ground-level and elevated sources; (3) off-site impacts;
(4) vapors and particulates; and (5) annual, monthly, and daily averaging times.
Five models were considered for this analysis:
# Industrial Source Complex - Short Term v.3 (ISCST3) - U.S. EPA, 1995a
# Industrial Source Complex - Long Term v.3 (ISCLT3) - U.S. EPA, 1995a
# Toxic Modeling System - Short Term (TOXST) - U.S. EPA, 1994c
# Fugitive Dust Model (FDM) -U.S. EPA, 1992
# COMPDEP - U.S. EPA, 1990.
Table 5-1 summarizes the capabilities of these commonly used air dispersion models with
respect to the requirements of this analysis. The ISCST3 (U.S. EPA, 1995a) was selected for all
aspects of this analysis because it met all the criteria. This model, however, requires
considerable run time, which limited the number of meteorological stations included in this
analysis.
5.3 Critical Parameters
This section discusses the critical parameters of the selected model, ISCST3, and the
results of sensitivity analyses performed to investigate several of the model parameters. Results
of the sensitivity analyses are presented in Appendix C.
Table 5-1. Air Dispersion Model Capabilities
Model
ISCST3
ISCLT3
TOXST
FDM
Source
Geometry Receptor Location
Area Source Height Off-site Chemical Phase
Source Elevated Ground Air Cone. Vapor Particulate
/ / / / / /
/ / / / / /
a
Averaging Period
Annual Monthly Daily
/ / /
/
/ /
COMPDEP
a Minimum height of source for modeling is 0.5 meters.
5-2
-------
Volume II
Section 5.0
5.3.1 General Assumptions
This section discusses depletion,
rural vs. urban, and terrain assumptions.
#
5.3.1.1 Depletion. Air
concentrations can be calculated in ISCST3
with or without wet depletion for vapors
and particulates, and with or without dry
depletion for particulates only (ISCST3
cannot model dry depletion of vapors.)
Modeled concentrations without depletions
are higher than those with depletions. The
run time for calculating concentrations
using the ISCST3 model with depletion
options, however, is 15 to 30 times longer
than the run time without depletions. Even
using the database approach described in
Section 5.1 rather than modeling all 3,400
WMUs separately, the additional run time
needed to model depletion was
considerable. To be as efficient as
possible, a sensitivity analysis was
conducted to determine the sensitivity of model
modeled for vapors and particulates.
#
Assumptions Made for Dispersion Modeling
Dry depletion was activated in the dispersion
modeling for particulates. Depletion was not
considered for vapors.
The rural option was used in the dispersion modeling
since the types of WMUs being assessed are
typically in nonurban areas.
# Flat terrain was assumed.
# An area source was modeled for all WMUs.
#
#
#
To minimize error due to site orientation, a square
area source with sides parallel to X- and Y- axes was
modeled.
Receptor points were placed on 25, 50, 75, 150, 500,
and 1,000 m receptor squares starting from the edge
of the source with 16 receptor points on each square.
Modeling was conducted using a unit emission rate
of 1
results to depletion processes that could be
For particulates, the sensitivity analysis, which is presented in Appendix C, shows that
the differences in the maximum concentrations with and without wet depletion are small at close-
to-source receptors, increasing only slightly as the distance from the source increases. The
sensitivity analysis, however, also shows that very large differences in concentration can occur
when results with and without dry depletion are compared for particulates. (The difference is
greater for larger sources; see sensitivity analysis in Appendix C for details.) Therefore,
concentrations for particulates were calculated with dry depletion but without wet depletion in
this analysis.
A similar sensitivity analysis for vapors showed that wet depletion of vapors is negligible
(differences in concentration with and without depletion were mostly less than 1 percent).
Therefore, wet depletion of vapors was not modeled. ISC cannot model dry depletion of vapors,
so this was also not modeled; however, dry depletion of vapors is also expected to be negligible.
5.3.1.2 Rural vs. Urban. ISCST3 may be run in rural or urban mode, depending on land
use within a 3-km radius from the source. These modes differ with respect to wind profile
exponent and potential temperature gradients. Unless the site is located in a metropolitan area,
the rural option is generally more appropriate. Because the types of WMUs being assessed are
typically in nonurban areas, the rural option was used in this analysis.
5-3
-------
Volume II Section 5.0
5.3.1.3 Terrain. Flat terrain for both the source and the surrounding area was assumed
in the modeling analysis for two reasons: (1) ISCST3 models all area sources as flat, and (2)
complex terrain simulations in the surrounding area result in air concentrations that are highly
dependent upon site-specific topography. A specific WMU's location in relation to a hill or
valley produces results that would not be applicable to other locations. Because complex terrain
applications are extremely site-specific, model calculations from one particular complex terrain
location cannot be applied to another. Conversely, simulations from flat terrain produce values
that are more universally applicable.
5.3.2 Meteorological Stations and Data
Meteorological data at over 200 meteorological stations in the United States are available
on the SCRAM Bulletin Board (http://www.epa.gov/scram001/) and from a number of other
sources. To develop the dispersion factor database described in Section 5.1, a set of 29
meteorological stations selected in an assessment for EPA's Superfund program Soil Screening
Levels (SSLs) (EQM, 1993) as being representative of the nine general climate regions of the
continental United States was used.
In EPA's SSL study, it was determined that 29 meteorological stations would be a
sufficient sample to represent the population of 200 meteorological stations and predict mean
dispersion values with a high (95 percent) degree of confidence. The 29 meteorological stations
were distributed among nine climate regions based on meteorological representativeness and
variability across each region. These climate regions were:
# North Pacific Coastal # Northwest Mountains # Midwest
# South Pacific Coastal # Central Plains # Northern Atlantic
# Southwest # Southeast # South Florida.
Large-scale regional average conditions were used to select the actual stations (EQM, 1993).
The 29 meteorological stations are listed in Table 5-2. To assign each Industrial D
facility to a meteorological station, EPA used a geographic information system (GIS) to construct
Thiessen polygons around each station that enclose the areas closest to each station. The
boundaries of these areas were then adjusted to ensure that each boundary encloses an area that is
most similar in meteorological conditions to those measured at the meteorological station. To
assist in this process, a GIS coverage of Bailey's ecoregion divisions and provinces (Bailey et al.,
1994) was used to conflate the boundaries to correspond to physiographic features likely to
influence climate or boundaries corresponding to changes in temperature or precipitation.
General wind regimes were also considered in the conflation process.
Key factors considered in the conflation process include: defining coastal regimes as
narrow polygons, which generally stretched about 25 to 50 miles inland, to capture regions
dominated by coastal climate effects; maintaining tropical/subtropical and arid/semiarid divisions
in the southwestern United States; and using the ecoregion boundaries in Washington, Oregon,
and California to separate the more humid marine/redwood or Mediterranean mountain regimes
5-4
-------
Volume II
Section 5.0
Table 5-2. Meteorological Stations Used
Met Station
City
Albuquerque
Atlanta
Bismarck
Boise
Casper
Charleston
Chicago
Cleveland
Denver
Fresno
Harrisburg
Hartford
Houston
Huntington
Las Vegas
Lincoln
Little Rock
Los Angeles
Miami
Minneapolis
Philadelphia
Phoenix
Portland
Raleigh-Durham
Salem
Salt Lake City
San Francisco
Seattle
Winnemucca
State
NM
GA
ND
ID
WY
SC
IL
OH
CO
CA
PA
CT
TX
WV
NV
NE
AR
CA
FL
MN
PA
AZ
ME
NC
OR
UT
CA
WA
NV
#
23050
13874
24011
24131
24089
13880
94846
14820
23062
93193
14751
14740
12960
03860
23169
14939
13963
23174
12839
14922
13739
23183
14764
13722
24232
24127
23234
24233
24128
in the Air Characteristic Study
Latitude
Degree
35
33
46
43
42
32
41
41
39
36
40
41
29
38
36
40
34
33
25
44
39
33
43
35
44
40
37
47
40
Minute
3
39
46
34
55
54
59
25
46
46
13
56
58
22
5
51
44
56
49
53
53
26
39
52
55
47
37
27
54
Longitude
Degree
106
84
100
116
106
80
87
81
104
119
76
72
95
82
115
96
92
118
80
93
75
112
70
78
123
111
122
122
117
Minute
37
25
45
13
28
2
54
52
52
43
51
41
21
33
10
45
14
24
17
13
15
1
19
47
0
57
23
18
48
Source; EQM (1993).
from the deserts to the east. In general, Thiessen polygons were used to define the
meteorological station areas for the remainder of the country.
Based on facility locations derived from Industrial D or TSDR survey data, the WMU
sites were then overlaid on the GIS coverage of the conflated meteorological boundaries and
meteorological station assignments were then exported for use in the modeling exercise. Several
sites in Alaska, Hawaii, and Puerto Rico were deleted from the analysis because the 29
meteorological stations are limited to the continental United States. Figures 5-1 to 5-4 show the
final meteorological station boundaries used for the study along with the locations of the
Industrial D or TSDR facility sites for landfills, LAUs, wastepiles, and tanks, respectively.
5-5
-------
jj Met. Regions (29)"
| 1 US States
Bailey's Ecoregion Divisions
ill] Hot Continental Division
Hot Continental Regime Mountains
Marine Division
Marine Regime Mountains Redwood Forest Province
Mediterranean Division
f' ' Mediterranean Regime Mountains
Prairie Division
; Savanna Division
iLSr- Subtropical Division
*--,._/
Subtropical Regime Mountains
Temperate Desert Division
Temperate Desert Regime Mountains
Temperate Steppe Division
Temperate Steppe Regime Mountains
Tropical/Subtropical Desert Division
Tropical/Subtropical Regime Mountains
Tropical/Subtropical Steppe Division
Warm Continental Division
Warm Continental Regime Mountains
Industrie! D Landfills,
29 Meteorological Regions,
and Bailey's Ecoregions
#Met. station regions created from thiessen polygons
and Bailey's Ecoregion boundaries.
created 7/23/93
File: ecci__29m ei__al!s lies. apr -gio
Figure 5-1. Meteorological station regions and landfill locations.
O'
s
-------
jf J£&j^*
Map Legend
o Industrial D Facilities with Land Application Units
«Ar Met. Stations (29)
TlJ. Met Regions (29)*
| 1 US States
Bailey's Ecoregion Divisions
F - Hot Continental Division
Hot Continental Regime Mountains
Marine Division
Marine Regime Mountains Redwood Forest Province
Mediterranean Division
Mediterranean Regime Mountains
Prairie Division
Savanna Division
E--"J Subtropical Division
Subtropical Regime Mountains
§^ Temperate Desert Division IndustriaS D Land Application Units,
Temperate Desert Regime Mountains «jn |ll|otor»rr»lr»nir»al Bonirtnc
m Temperate Steppe Division /ij IVieieOFOIOgiCai ItegiOnS,
ELir; Temperate Steppe Regime Mountains and Bailey's ECOFCgiOnS
Tropic al/Subtropical Desert Division
Tropical/Subtropical Regirne_Mountains *Met. station regions created from thiessen polygons
and Bailey's Ecoregion boundaries.
Tropical/Subtropical Steppe Division
Warm Continental Division
Warm Continental Regime Mountains
Created 7ffi3.«8
Fiie: ecom29met__albites.apr -gic
Figure 5-2. Meteorological station regions and LAU locations.
o'
s
-------
oo
V
Map Legend
o Industrial D Facilities with Waste Piles
T% Met. Stations (29)
IZji Met. Regions (295*
| 1 US States
Bailey's Ecoregion Divisions
; ._ Hot Continental Division
Hot Continental Regime Mountains
Marine Division
Marine Regime Mountains Redwood Forest Province
Mediterranean Division
Mediterranean Regime Mountains
Prairie Division
Savanna Division
l^_=} Subtropical Division
Subtropical Regime Mountains
Temperate Desert Division
Temperate Desert Regime Mountains
Temperate Steppe Division
Temperate Steppe Regime Mountains
Tropical/Subtropical Desert Division
Tropical/Subtropical Regime Mountains *Met. station regions created from thiessen polygons
Tropical/Subtropical Steppe Division anc) Bailey's Ecoregion boundaries.
Warm Continental Division
Warm Continental Regime Mountains
Industrial D Waste Piles,
29 Meteorological Regions,
and Bailey's Ecoregions
Created 7C3/88
File: eoo_29m ei_al!s lies, ipr -gio
Figure 5-3. Meteorological station regions and wastepile locations.
O'
s
-------
Map Legend
0 Industrial D Facilities with Tanks
(aerated, nonaerated, arid storage)
Tflr Met- Stations (29)
Met- Regions (29)*
| 1 US States
Bailey's Ecoregion Divisions
''""-'-:' Hot Continental Division
Hot Continental Regime Mountains
Marine Division
Marine Regime Mountains Redwood Forest Province
Mediterranean Division
I Mediterranean Regime Mountains
Prairie Division
; Savanna Division
•TV.V. Subtropical Division
Subtropical Regime Mountains
3 Temperate Desert Division
I Temperate Desert Regime Mountains
| Temperate Steppe Division
:] Temperate Steppe Regime Mountains
Tropical/Subtropical Desert Division
Tropical/Subtropical Regime Mountains
Tropical/Subtropical Steppe Division
Warm Continental Division
Warm Continental Regime Mountains
Industrial D Tanks,
29 Meteorological Regions,
and Bailey's Ecoregions
*Met. station regions created from thiessen polygons
and Bailey's Ecoregion boundaries-
created 7/23(88
File: eco_29iTiet_alisites.apr -gic
Figure 5-4. Meteorological station regions and tank locations.
o'
s
-------
Volume II
Section 5.0
The modeling analysis was conducted
using 5 years of representative meteorological
data from each of the 29 meteorological
stations. Five-year wind roses representing the
frequency of wind directions and windspeeds
for the 29 meteorological stations were
analyzed. These show that the 29
meteorological stations represent a variety of
wind patterns and are presented in Appendix C.
Shape of Wind Rose for
29 Meteorological Stations
Shape of Wind Rose No. of Stations
Narrowly distributed
Moderately distributed
Evenly distributed
Bimodally distributed
10
4
6
9
Meteorological Data for
the ISCST3 Model
without Depletion
Wind Direction (or Flow Vector)
Windspeed
Ambient Temperature
Stability Class
Mixing Height
Wind direction and windspeed are typically the most
important meteorological inputs for dispersion modeling
analysis. Wind direction determines the direction of the
greatest impacts. Windspeed is inversely proportional to
ground-level air concentrations, so that the lower the
windspeed, the higher the air concentration.
Mixing height determines the heights to which
pollutants can be diffused vertically. Stability class is
also an important factor in determining the rate of lateral
and vertical diffusion. The more unstable the air, the greater the diffusion. This increase would
lower concentration at the centerline of the plume. Note that air concentration is highest at the
centerline.
5.3.3 Source Release Parameters
This section describes the source parameters and assumptions used in the dispersion
modeling, including source type and elevation, source shape and orientation, and source areas.
5.3.3.1 Source Type and Elevation. All WMU types modeled in this analysis were
modeled as area sources. Landfills and land application units were modeled as ground-level
sources, and wastepiles and tanks were modeled as elevated sources.
5.3.3.2 Source Shape and Orientation. The ISCST3 models an area source as a
rectangle or combination of rectangles. The user may also specify an angle of rotation relative to
a north-south orientation. A sensitivity analysis was conducted to compare the air concentrations
from a square area source, a rectangular area source oriented east to west, and a rectangular area
source oriented north to south to determine what role source shape and orientation play in
determining dispersion coefficients of air pollutants. The results show that the differences in
unitized air concentration between the square area source and the two rectangular area sources
are less than the differences between the two rectangular sources. In addition, a square area
source has the least amount of impact on orientation. Because information on source shapes or
orientations is not available, a square source was chosen to minimize the errors caused by source
shapes and orientations. (See sensitivity analysis in Appendix C for details.)
5-10
-------
Volume II Section 5.0
5.3.3.3 Source Areas Modeled. In the modeling analysis, four types of WMUs were
considered (i.e., landfill, land application unit, wastepile, and tank). Because the ISCST3 model
is sensitive to the size of the area source, the relationship between air concentrations and size of
the area source was analyzed. As illustrated in Figure 5-5, the results show that, for relatively
small area sources, air concentrations increase significantly as the size of the area source
increases. For large area sources, this increase in air concentrations is not as significant.
To address this model sensitivity yet avoid modeling approximately 3,400 separate
WMUs, EPA developed area strata that represented the distribution of the surface area for each
of the WMU types. Landfills and land application units were modeled as ground-level area
sources, while wastepiles and tanks were treated as elevated area sources. Separate area strata
were developed for ground-level and elevated sources. Tables 5-3, 5-4, and 5-5 present the
source areas and heights used in the modeling analysis. Fourteen area strata were selected for
landfills and land application units; all 14 were modeled at zero height, since landfills and LAUs
are not elevated.
For wastepiles, seven areas and six heights were modeled. All wastepiles in the database
were assigned one of the seven heights as described in Section 3. Each wastepile has a different
area and UACs were extrapolated between areas modeled; therefore, for any given wastepile, a
UAC is needed for each of the two areas bracketing the actual area, at the assigned height. An
analysis of the wastepile database revealed that only 29 of the possible 42 combinations of the
seven areas and six heights were needed; therefore, UACs were modeled only for those 29 area-
height combinations. Tanks were modeled in the same manner as wastepiles, using 10 areas and
4 heights. Analysis of the tank databases revealed that only 33 of the possible 40 combinations
of 10 areas and 4 heights were needed.
These areas or area-height combinations were modeled for each of the 29 meteorological
locations. This provided a set of UACs for use in the analysis. For any specific WMU, a UAC
was then estimated using an interpolation routine that used the UACs associated with modeled
areas immediately above and below the actual area of the unit. The interpolation routine
provides a technique for minimizing the number of ISCST3 runs required for a WMU while also
minimizing the error associated with the difference between the UACs for preselected areas and
the UAC for the actual area of the WMU. The interpolation is described in more detail in
Section 7.
5.3.4 Receptors
The ISCST3 model allows the user to specify receptors with Cartesian receptor grid
and/or polar receptor grid. In general, Cartesian receptors are used for near-source receptors and
polar grid receptors for more distant receptors. The number of receptors modeled greatly impacts
run time. However, if too few receptors are modeled, the location of peak concentration may be
missed. A sensitivity analysis was conducted to determine receptor locations and spacings that
would provide adequate resolution without modeling an excessive number of receptors. (See
Appendix C for details.) The results of the sensitivity analysis show that the maximum
concentrations are generally higher for a dense receptor grid (i.e., 64 or 32 receptors on each
5-11
-------
Volume II
Section 5.0
Air Concentrations vs. Surface Area
(Landfills)
• Little Rock
• Los Angeles
400,000
800,000 1,200,000 1,600,000
Surface Area (m)
Air Concentrations vs. Surface Area
(2m High Waste Piles)
• Little Rock
• Los Angeles
0 20,000 40,000 60,000 80,000 100,000
Surface Area (m)
Note: Largest areas modeled for each WMU type have been omitted from the chart to improve clarity.
Figure 5-5. Air concentration vs. size of area source.
5-12
-------
Volume II
Section 5.0
Table 5-3.
Source Area
81
567
1,551
4,047
12,546
40,500
78,957
161,880
243,000
376,776
607,000
906,528
1,408,356
8,090,000
Areas Modeled for Landfills and
Land Application Units
(m2) Source Height (m)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Table 5-4. Areas and Source Heights Modeled for Wastepiles
Assigned
Height
(m)
1
2
4
6
8
10
Surface Area (m2)
20
X
X
X
X
X
X
162
X
X
X
X
X
X
486
X
X
X
X
X
2,100
X
X
X
X
X
10,100
X
X
X
101,000
X
X
1,300,000
X
X
x = Combination modeled.
5-13
-------
Volume II
Section 5.0
Table 5-5. Areas and Source Heights Modeled for Tanks
Assigned
Height
(m)
1
2
4
6
Surface Area (m2)
2
X
X
3
X
X
X
5
X
X
X
10
X
X
X
25
X
X
X
50
X
X
X
100
X
X
X
X
400
X
X
X
X
1,500
X
X
X
X
5,000
X
X
X
X
x = Combination modeled.
square) than for a scattered receptor grid (i.e., 16 receptors on each square). However, the
differences of the maximum receptor concentrations are not significant between a dense and a
scattered receptor grid. Therefore, 16 evenly spaced receptor points on each square were used in
the modeling. The sensitivity analysis also shows that the maximum downwind concentrations
decrease sharply from the edge of the area source to about 1,000 meters from the source. After
the first 1,000 meters from the edge of the area source, concentrations decrease very slowly as the
downwind distance increases. Therefore, for annual average concentrations, the receptor points
were placed on 25, 50, 75, 150, 500, and 1,000 meter receptor squares starting from the edge of
the source, with 16 receptor points on each square.
5.4 Unitized Air Concentrations
Unitized air concentrations were calculated by running ISCST3 with a unit emission rate
(i.e., 1 jug/m2-s). The selected areas for each type of WMU were modeled with 29 representative
meteorological locations in the continental United States to estimate UACs. The 5-year average
UACs at all receptor points were calculated for the long-term or chronic exposure scenario.
They were used as input to the Monte Carlo analysis and as input to the interpolation routine
discussed above.
A similar methodology and assumptions were used to model dispersion for acute and sub-
chronic exposures. Since the ISCST3 model uses hourly meteorological data, the outputs from
the model can be used to develop any averaging times equal to or greater than 1 hour. One set of
ISCST3 runs (for the 76 area-height combinations and 29 meteorological stations) was done for
both acute and subchronic, resulting in 5 years of hourly average concentrations at each receptor.
For each area, meteorological location, and receptor location, the maximum air concentration for
any 24-hour period over the 5 years was selected. Then, for each area and meteorological station,
the maximum 24-hour air concentration among all receptor locations at each distance modeled
was selected, and this value was used as the UAC for that area and meteorological station for
acute exposure. The same method was used to determine the subchronic UAC, except that the
maximum 30-day period over the 5 years was used instead of the maximum 24-hour period.
5-14
-------
Volume II Section 5.0
Typically, the location of maximum impacts with respect to the source are determined by
the prevailing wind direction. For ground-level area sources (i.e., landfills and land application
units), maximum annual average UACs are always located on the first receptor square (closest to
source). For elevated area sources, the maximum annual average UACs are usually located on
the first receptor square and occasionally located on the second or third receptor square. Annual
average UACs increase with the increasing area size of the sources.
Maximum UACs vary with meteorological location. For landfills and land application
units, the maximum UACs at some meteorological locations can be twice as much as those at
other locations. For wastepiles and tanks, the maximum UACs at some meteorological locations
are more than twice those at other meteorological locations.
5-15
-------
Volume II Section 6.0
6.0 Development of Inhalation Health
Benchmarks
Inhalation health benchmarks for chronic, subchronic, and acute exposure durations were
needed for the risk characterization model. This section describes the noncancer and cancer
benchmarks used in this study.
6.1 Chronic Inhalation Health Benchmarks Used in This Study
Chronic inhalation health benchmarks used in the Air Characteristic study include
inhalation reference concentrations (RfCs) for noncarcinogens and inhalation unit risk factors
(UKFs) and inhalation cancer slope factors (CSFs) for carcinogens. UKFs and CSFs were used
in the analysis for carcinogenic constituents, regardless of the availability of an RfC. Inhalation
health benchmarks were identified in the Integrated Risk Information System (IRIS) and the
Health Effects Assessment Summary Tables (HEAST) (U.S. EPA, 1997a, 1999). IRIS and
HEAST are maintained by the Agency, and values from IRIS and HEAST were used in the
analysis whenever available. Provisional EPA benchmarks and Agency for Toxic Substances
and Disease Registry (ATSDR) minimal risk levels (MRLs) were used to fill in data gaps (see
Section 6.1.1). Additional chronic inhalation health benchmarks were derived for use in this
analysis for constituents lacking EPA or ATSDR values (see Section 6.1.2).
Figure 6-1 describes the approach used to identify or develop the chronic inhalation
health benchmarks used in this analysis. The benchmarks are summarized in Table 6-1.
6.1.1 Alternate Chronic Inhalation Health Benchmarks Identified
If IRIS or HEAST chronic inhalation health benchmarks were not available, benchmarks
from alternative sources were sought. Provisional EPA benchmarks, ATSDR inhalation MRLs,
and California EPA noncancer chronic reference exposure levels (CalEPA, 1997a) were included
whenever available. Alternate RfCs were identified for
# Acetone
# Cyclohexanol
# Isophorone
# Phenol
# Pyridine
# Tetrachloroethylene
# 1,1,1 -Tri chl oroethane
# Xylenes.
-------
Volume II
Section 6.0
Used IRIS value
if not available
Used HEAST value
if not available
Obtained value from other
sources (e.g., ATSDR,
EPA, CalEPA)
if not available
Derived value for
Noncarcinogen
1. Conducted literature search
and calculated RfC using
standard RfC methodology
or
2. If no appropriate inhalation
toxicity studies available,
developed RfC by using
alternative methodology,
including route-to-route
extrapolation
Carcinogen
Calculated inhalation URF
using route-to-route
extrapolation from oral CSF
Figure 6-1. Approach used to select chronic
inhalation health benchmark values.
6-2
-------
Volume II
Section 6.0
Table 6-1. Chronic Inhalation Health Benchmarks Used in the
Air Characteristic Analysis
CAS# Name
75-07-0
67-64-1
75-05-8
107-02-8
79-06-1
79-10-7
107-13-1
107-05-1
62-53-3
7440-38-2
7440-39-3
71-43-2
92-87-5
50-32-8
7440-41-7
75-27-4
75-25-2
106-99-0
7440-43-9
75-15-0
56-23-5
126-99-8
108-90-7
124-48-1
67-66-3
95-57-8
7440-47-3
7440-48-4
1319-77-3
98-82-8
108-93-0
Acetaldehyde
Acetone
Acetonitrile
Acrolein
Acrylamide
Acrylic acid
Acrylonitrile
Allyl chloride
Aniline
Arsenic
Barium
Benzene
Benzidine
Benzo(a)pyrene
Beryllium
Bromodichloromethane
Bromoform
(Tribromomethane)
Butadiene, 1,3-
Cadmium
Carbon disulfide
Carbon tetrachloride
Chloro-l,3-butadiene, 2-
(Chloroprene)
Chlorobenzene
Chlorodibromomethane
Chloroform
Chlorophenol, 2-
Chromium VI
Cobalt
Cresols (total)
Cumene
Cyclohexanol
J
RfC
(mg/m3)
9.0E-03
3.1E+01
6.0E-02
2.0E-05
NA
l.OE-03
2.0E-03
l.OE-03
l.OE-03
NA
5.0E-04
NA
NA
NA
2.0E-05
NA
NA
NA
NA
7.0E-01
NA
7.0E-03
2.0E-02
NA
NA
1.4E-03
l.OE-04
l.OE-05
4.0E-04
4.0E-01
2.0E-05
^oncarcinogens
RfC Target Organ
Respiratory
Neurological
Mortality
Respiratory
Respiratory
Respiratory
Neurological
Spleen
Reproductive
Respiratory
Neurological
Respiratory
Kidney and liver
Repro/developmental
Respiratory
Respiratory
Hematological
Kidney and adrenal
Muscle
C
Inhal URF
Ref (jig/m3)1
I
A
I
I
I
I
I
I
H
I
I
H
H
D
I
D
D
I
FR
2.2E-06
NA
NA
NA
1.3E-03
NA
6.8E-05
NA
NA
4.3E-03
NA
8.3E-06
6.7E-02
8.8E-04
2.4E-03
1.8E-05
1.1E-06
2.8E-04
1.8E-03
NA
1.5E-05
NA
NA
2.4E-05
2.3E-05
NA
1.2E-02
NA
NA
NA
NA
ircinogens
Inhal CSF
(mg/kg/d)1 ReP
7.7E-03
NA
NA
NA
4.6E+00
NA
2.4E-01
NA
NA
1.5E+01
NA
2.9E-02
2.3E+02
3.1E+00
8.4E+00
6.2E-02
3.9E-03
1.8E+00
6.3E+00
NA
5.3E-02
NA
NA
8.4E-02
8.1E-02
NA
4.2E+01
NA
NA
NA
NA
I
I
I
I
I
I
N
I
D
I
I
I
I
D
I
I
(continued)
6-3
-------
Volume II
Section 6.0
Table 6-1. (continued)
CAS# Name
96-12-8
95-50-1
106-46-7
75-71-8
107-06-2
75-35-4
78-87-5
10061-01-5
10061-02-6
57-97-6
68-12-2
95-65-8
121-14-2
123-91-1
122-66-7
106-89-8
106-88-7
111-15-9
110-80-5
100-41-4
106-93-4
107-21-1
75-21-8
50-00-0
98-01-1
87-68-3
118-74-1
77-47-4
67-72-1
110-54-3
Dibromo-3 -chloropropane,
1,2-
Dichlorobenzene, 1,2-
Dichlorobenzene, 1,4-
Dichlorodifluoromethane
Dichloroethane, 1,2-
Dichloroethylene, 1,1-
Dichloropropane, 1,2-
Dichloropropene, cis-1,3-
Dichloropropene, trans- 1,3-
Dimethylbenz(a)anthracene,
7,12-
Dimethylformamide, N,N-
Dimethylphenol, 3,4-
Dinitrotoluene, 2,4-
Dioxane, 1,4-
Diphenylhydrazine, 1,2-
Epichlorohydrin
Epoxybutane, 1,2-
Ethoxyethanol acetate, 2-
Ethoxyethanol, 2-
Ethylbenzene
Ethylene dibromide
Ethylene glycol
Ethylene oxide
Formaldehyde
Furfural
Hexachloro- 1 , 3 -butadiene
Hexachlorobenzene
Hexachlorocyclopentadiene
Hexachloroethane
Hexane, n-
Noncarcinogens Carcinogens
RfC
(mg/m3)
2.0E-04
2.0E-01
8.0E-01
2.0E-01
NA
NA
4.0E-03
2.0E-02
2.0E-02
NA
3.0E-02
NA
NA
8.0E-01
NA
l.OE-03
2.0E-02
3.0E-01
2.0E-01
l.OE+00
2.0E-04
6.0E-01
NA
NA
5.0E-02
NA
NA
7.0E-05
NA
2.0E-01
RfC Target Organ
Reproductive
Body weight
Liver
Liver
Respiratory
Respiratory
Respiratory
Liver
NA
No liver, kidney, or
hemato effects
Respiratory
Respiratory
Reproductive (male),
hemato
Reproductive (male),
hemato
Developmental
Reproductive (male)
Respiratory
Respiratory
Respiratory
Respiratory and
neurological
Inhal URF
ReP (Mg/m3)1
I
H
I
H
I
I
I
I
D
I
I
D
I
I
H
D
H
H
I
6.9E-07
NA
NA
NA
2.6E-05
5.0E-05
NA
3.7E-05
3.7E-05
2.4E-02
NA
NA
1.9E-04
NA
2.2E-04
1.2E-06
NA
NA
NA
NA
2.2E-04
NA
l.OE-04
1.3E-05
NA
2.2E-05
4.6E-04
NA
4.0E-06
NA
Inhal CSF
(mg/kg/d)1 ReP
2.4E-03
NA
NA
NA
9.1E-02
1.8E-01
NA
1.3E-01
1.3E-01
8.4E+01
NA
NA
6.8E-01
NA
7.7E-01
4.2E-03
NA
NA
NA
NA
7.7E-01
NA
3.5E-01
4.6E-02
NA
7.7E-02
1.6E+00
NA
1.4E-02
NA
H
I
I
H
H
D
D
I
I
I
H
I
I
I
I
(continued)
6-4
-------
Volume II
Section 6.0
Table 6-1. (continued)
CAS# Name
78-59-1
7439.96-5
7439.97-6
67-56-1
110-49-6
109-86-4
74-83-9
74-87-3
78-93-3
108-10-1
80-62-6
1634-04-4
56-49-5
75-09-2
91-20-3
7440-02-0
98-95-3
79-46-9
55-18-5
924-16-3
930-55-2
108-95-2
85-44-9
75-56-9
110-86-1
100-42-5
1746-01-6
630-20-6
127-18-4
79-34-5
Isophorone
Manganese
Mercury
Methanol
Methoxyethanol acetate, 2-
Methoxyethanol, 2-
Methyl bromide
(bromomethane)
Methyl chloride
(chloromethane)
Methyl ethyl ketone
Methyl isobutyl ketone
Methyl methacrylate
Methyl fer/-butyl ether
Methylcholanthrene, 3-
Methylene chloride
Naphthalene
Nickel
Nitrobenzene
Nitropropane, 2-
Nitrosodiethylamine
Nitrosodi-w-butylamine
w-Nitrosopyrrolidine
Phenol
Phthalic anhydride
Propylene oxide
Pyridine
Styrene
TCDD, 2,3,7,8-
Tetrachloroethane, 1,1,1,2-
Tetrachloroethylene
Tetrachloroethane, 1,1,2,2-
Noncarcinogens Carcinogens
RfC
(mg/m3)
1.2E-02
5.0E-05
3.0E-04
1.3E+01
3.0E-02
2.0E-02
5.0E-03
NA
l.OE+00
8.0E-02
7.0E-01
3.0E+00
NA
3.0E+00
3.0E-03
NA
2.0E-03
2.0E-02
NA
NA
NA
6.0E-03
1.2E-01
3.0E-02
7.0E-03
l.OE+00
NA
NA
3.0E-01
NA
RfC Target Organ
Body weight
Neurological
Neurological
Developmental
Reproductive (male)
Reproductive (male)
Respiratory
Developmental
Kidney and liver
Respiratory
Kidney and liver
Liver
Respiratory
Kidney, liver,
hematological,
adrenal
Liver
No effects
Respiratory
Respiratory
Liver
Neurological
Neurological
Inhal URF
ReP (Mg/m3)1
FR
I
I
D
D
I
I
I
H
I
I
H
I
H
I
FR
H
I
O
I
A
NA
NA
NA
NA
NA
NA
NA
1.8E-06
NA
NA
NA
NA
2.1E-03
4.7E-07
NA
2.4E-04
NA
2.7E-03
4.3E-02
1.6E-03
6.1E-04
NA
NA
3.7E-06
NA
NA
3.3E+01
7.4E-06
5.8E-07
5.8E-05
Inhal CSF
(mg/kg/d)1 ReP
NA
NA
NA
NA
NA
NA
NA
6.3E-03
NA
NA
NA
NA
7.4E+00
1.6E-03
NA
8.4E-01
NA
9.4E+00
1.5E+02
5.6E+00
2.1E+00
NA
NA
1.3E-02
NA
NA
1.6E+05
2.6E-02
2.0E-03
2.0E-01
H
D
I
I
H
I
I
I
I
H
I
SF
I
(continued)
6-5
-------
Volume II
Section 6.0
Table 6-1. (continued)
Noncarcinogens
Carcinogens
CAS# Name
108-88-3
95-53-4
76-13-1
120-82-1
71-55-6
79-00-5
79-01-6
75-69-4
121-44-8
7440-62-2
108-05-4
75-01-4
1330-20-7
Toluene
Toluidine, o-
Trichloro- 1,2,2-
trifluoroethane, 1,1,2-
Trichlorobenzene, 1,2,4-
Trichloroethane, 1,1,1-
Trichloroethane, 1,1,2-
Trichloroethylene
Trichlorofluoromethane
Triethylamine
Vanadium
Vinyl acetate
Vinyl chloride
Xylenes (total)
RfC
(mg/m3)
4.0E-01
NA
3.0E+01
2.0E-01
l.OE+00
NA
NA
7.0E-01
7.0E-03
7.0E-05
2.0E-1
NA
4.0E-01
RfC Target Organ
Respiratory and
neurological
Body weight
Liver
Neurological
Kidney and
respiratory
Respiratory
Respiratory
Respiratory
Neurological
Inhal URF
ReP (Mg/m3)1
I
H
H
SF
H
I
D
I
A
NA
6.9E-05
NA
NA
NA
1.6E-05
1.7E-06
NA
NA
NA
NA
8.4E-05
NA
Inhal CSF
(mg/kg/d)1 ReP
NA
2.4E-01
NA
NA
NA
5.6E-02
6.0E-03
NA
NA
NA
NA
3.0E-01
NA
D
I
SF
H
CAS = Chemical Abstract Service.
CSF = Cancer slope factor.
NA = Not available.
RfC = Reference concentration.
URF = Unit risk factor.
a Sources:
I = IRIS (U.S. EPA, 1999)
H = HEAST (U.S. EPA, 1997a).
A = Agency for Toxic Substances Disease Registry (ATSDR) minimal risk levels (MRLs).
SF = Superfund Risk Issue Paper (U.S. EPA, 1996b; U.S. EPA, n.d.).
FR = 63 FR 64371-0402 (U.S. EPA, 1998b)
N = NCEA Risk Assessment Issue Paper (U.S. EPA, 19941).
D = Developed for this study.
O = Other source (see Sections 6.1.1 and 6.1.2).
For acetone, tetrachloroethylene, and total xylenes, ATSDR's chronic inhalation MRLs
were used. Provisional RfCs were identified for cyclohexanol, isophorone, and phenol in a
Federal Register notice (63 FR 64371) of the final listing rules for solvents (U.S. EPA, 1998b).
An acceptable daily intake (ADI) was identified for pyridine (U.S. EPA, 1986). An RfC for
1,1,1-trichloroethane was identified in a Superfund risk issue paper (U.S. EPA, 1996c).
Table 6-2 summarizes the alternate RfCs identified for this analysis, as well as the target organs,
sources, and critical studies.
6-6
-------
Volume II
Section 6.0
Table 6-2. Alternate Chronic Inhalation Health Benchmarks
CAS#
67-64-1
108-93-0
78-59-1
108-95-2
110-86-1
127-18-4
71-55-6
1330-20-7
Chemical Name
Acetone
(2-propanone)
Cyclohexanol
Isophorone
Phenol
Pyridine
Tetrachloroethylene
1,1,1-
Trichloroethane
Xylenes (total)
Inhalation Benchmark
and Benchmark Value
RfC=13ppm(31
mg/m3)
Provisional RfC =
0.00002 mg/m3
Provisional RfC= 0.012
mg/m3
Provisional RfC =
0.006 mg/m3
Inhalation ADI= 0.002
mg/kg/d; converts to
0.007 mg/m3
RfC = 0.04 ppm (0.3
mg/m3)
RfC= 1.0 mg/m3
RfC = 0.1 ppm (0.4
mg/m3)
Target
Organ
Neurological
Muscle
Body weight
No effects
Liver
Neurological
Neurological
Neurological
Source
AT SDR chronic inhal MRL based
on Stewart et al., 1975. Acetone:
Development of a Biological
Standard for the Industrial Worker
by Breath Analysis, Cincinnati,
OH:NIOSH. NTIS PB82-172917.
63 FR 64376 (U.S. EPA 1998b);
standard RfC methodology
63 FR 64376 (U.S. EPA 1998b);
standard RfC methodology
63 FR 64376 (U.S. EPA 1998b);
standard RfC methodology
Cited in Health and Environmental
Effects Profile (HEEP) for Pyridine
(U.S. EPA, 1986)
AT SDR chronic inhal MRL based
on Ferroni et al., 1992.
Neurobehavioral and
neuroendocrine effects of
occupational exposure to
perchloroethy lene .
Neurotoxicology 12: 243-247.
Superfund risk issue paper
(U.S. EPA, 1996b)
AT SDR chronic inhal MRL based
onUchida et al., 1993. Symptoms
and signs in workers exposed
predominantly to xylenes. Int Arch
Occup Environ Health 64:597-605.
6.1.2 Chronic Inhalation Health Benchmarks Derived for This Study
Chronic inhalation health benchmarks were developed for constituents lacking IRIS,
HEAST, alternative EPA, or ATSDR values. RfCs were developed for
# 2-Chlorophenol
# Cobalt
# Cresols
# 1,4-Dioxane
# 2-Ethoxyethanol acetate
# Ethylene glycol
# Methanol
6-7
-------
Volume II Section 6.0
# 2-Methoxyethanol acetate
# Vanadium.
For cobalt, cresols, 1,4-dioxane, ethylene glycol, and methanol, appropriate inhalation
studies were identified and RfCs were developed using EPA's standard RfC methodology as
detailed in Methods for Derivation of Inhalation Reference Concentrations and Application of
Inhalation Dosimetry (U.S. EPA, 1994d). For vanadium, the study that the ATSDR acute
inhalation MRL is based on was used but was adjusted for chronic exposure. For
2-chlorophenol, an RfC was developed using route-to-route extrapolation of the oral RfD (U.S.
EPA, 1999). RfCs were derived for 2-ethoxyethanol acetate and 2-methoxyethanol acetate based
on RfCs for 2-ethoxyethanol and 2-methoxyethanol, respectively.
EPA examined the toxicity data for 3,4-dimethylphenol and determined that it was not
appropriate to derive an RfC. Very little toxicity or metabolism data specific to
3,4-dimethylphenol are available. Although an RfD has been developed by EPA (U.S. EPA,
1999), route-to-route extrapolation is not recommended because of the potential for respiratory
tract effects following inhalation exposure (i.e., "portal-of-entry" effects) and "first-pass" effects
following ingestion exposure (due to extensive metabolism in the liver which may result in
differences in toxicity between oral and inhalation exposures).
Inhalation unit risk factors and inhalation cancer slope factors were developed for
# Bromodichloromethane
# Chlorodibromomethane
# 7,12-Dimethylbenz[a]anthracene
# 2,4-Dinitrotoluene
# 3-Methylcholanthrene
# o-Toluidine.
For bromodichloromethane, Chlorodibromomethane, 2,4-dinitrotoluene, and o-toluidine,
the oral CSFs (U.S. EPA, 1997a, 1999) were used to develop inhalation URFs for the
compounds. For 7,12-dimethylbenz[a]anthracene and 3-methylcholanthrene, inhalation URFs
developed by California's EPA (CalEPA 1997b) were used as the cancer benchmarks.
Table 6-3 summarizes the RfCs, inhalation unit risk factors, and inhalation cancer slope
factors derived for use in the air characteristic analysis, the method of development and critical
studies used, and the target organs identified. Details on the derivation of these inhalation
benchmark values are provided in Appendix D.
6.2 Subchronic Inhalation Health Benchmarks
Information on intermediate or subchronic inhalation noncancer benchmark values for
constituents considered in this study is summarized in Table 6-4. Data collected include
subchronic RfCs and ATSDR intermediate inhalation MRLs.
6-8
-------
Table 6-3. Chronic Inhalation Health Benchmarks Derived for This Study
CAS#
75-27-4
124-48-1
95-57-8
7440-48-4
1319-77-3
57-97-6
95-65-8
121-14-2
123-91-1
Chemical Name
Bromodichloromethane
(dichlorobromomethane)
Chlorodibromomethane
(dibromochloromethane)
2-Chlorophenol (o-)
Cobalt
Cresols, total
7,12-
Dimethylbenz[a]anthracene
3,4-Dimethylphenol
2,4-Dinitrotoluene
1,4-Dioxane
(1,4-diethyleneoxide)
Inhalation Benchmark
and Benchmark Value
Inhal CSF = 6.2E-02 per mg/kg/d
Inhal URF = 1.8E-05 per ug/m3
Inhal CSF = 8.4E-02 per mg/kg/d
Inhal URF = 2.4E-05 per ug/m3
RfC = 0.0014mg/m3
RfC = 0.00001 mg/m3
RfC = 0.0004 mg/m3
Inhal CSF = 8.4E+01 per mg/kg/d
Inhal URF = 2.4E-02 per ug/m3
NA - RfC derivation is inappropriate
Inhal CSF = 6.8E-01 per mg/kg/d
Inhal URF= 1.9E-04 per ug/m3
RfC = 0.8 mg/m3
RfC Target
Organ
Repro/
developmental
Respiratory
Hematological
Liver, kidney,
hematological
Method of Derivation
Inhal CSF and URF based on IRIS oral CSF
(renal)
Inhal CSF and URF based on IRIS oral CSF
(hepatocellular adenoma/carcinoma)
Route-to-route extrapolation of IRIS RfD
(0.005 mg/kg/d for reproductive effects)
Standard RfC derivation based on NTP TR-471
(1996a); supported by: ATSDR intermediate
inhal MRL = 3E-05 mg/m3 based on NTP TR-
TOX-5 (1991) and CalEPA RfC = 5E-06 mg/m3
based on Bucher et al., 1990. Inhalation toxicity
studies of cobalt sulfate in F344/N rats and
B6C3F1 mice. FundamAppl Toxicol 15:357-
372.
Standard RfC derivation based on: Uzhdavini
ER, Astaf'yeva K, Mamayeva AA, Bakhtizina
GZ. 1972. [Inhalation toxicity of o-cresol].
Trudy Ufimskogo Nauchno-Isseldovatel'skogo
Institute Gigiyeny Profzabolevaniya, 7:115-9.
(Russian)
Inhal CSF and URF derived by CalEPA
(1997b) based on TD50 approach
Inhal CSF and URF based on IRIS oral CSF
(liver, mammary gland)
Standard RfC derivation based on Torkelson et
al. 1974. 1,4-Dioxane. II. Results of a 2-year
inhalation study in rats. Toxicol Appl
Pharmacol 30:287-298.
(continued)
o'
s
-------
Table 6-3. (continued)
CAS#
111-15-9
107-21-1
67-56-1
110-49-6
56-49-5
95-53-4
7440-62-2
Chemical Name
2-Ethoxyethanol acetate
Ethylene glycol
Methanol
2-Methoxyethanol acetate
3 -Methy Icholanthrene
o-Toluidine
Vanadium
Inhalation Benchmark
and Benchmark Value
RfC= 0.3 mg/m3
RfC= 0.6 mg/m3
RfC = 13 mg/m3
RfC= 0.03 mg/m3
Inhal CSF = 7.4E+00 per mg/kg/d
Inhal URF = 2.1E-03 per ug/m3
Inhal CSF = 2.4E-01 per mg/kg/d
Inhal URF = 6.9E-05 per ug/m3
RfC = 0.00007 mg/m3
RfC Target
Organ
Repro (male)/
hematological
Respiratory
Developmental
Reproductive
(male)
Respiratory
Method of Derivation
Derived from RfC for ethoxyethanol
Derived using standard RfC methodology
Standard RfC derivation based on Rogers et al.,
1993. The developmental toxicity of inhaled
methanol in the CD-I mouse, with quantitative
dose-response modeling for estimation of
benchmark doses. Teratology 47(3): 175-188.
Derived from RfC for methoxyethanol
Inhal CSF and URF derived by CalEPA
(1997b) based on TD50 approach
Inhal CSF and URF based on HEAST oral CSF
(skin fibroma)
RfC based on ATSDR acute inhal MRL study,
but adjusted for chronic exposure.
Zenz and Berg, 1967. Human responses to
controlled vanadium pentoxide exposure. Arch
Environ Health 14:709-12.
(acute inhal MRL = 0.0002 mg/m3)
o'
s
-------
Volume II
Section 6.0
Table 6-4. Subchronic Inhalation Health Benchmarks
CAS#
75-07-0
67-64-1
75-05-8
107-02-8
79-10-7
107-13-1
107-05-1
62-53-3
7440-39-3
71-43-2
75-15-0
56-23-5
126-99-8
108-90-7
67-66-3
7440-47-3
7440-48-4
1319-77-3
98-82-8
108-93-0
96-12-8
95-50-1
106-46-7
75-71-8
107-06-2
75-35-4
78-87-5
10061-01-5
Name
Acetaldehyde
Acetone
Acetonitrile
Acrolein
Acrylic acid
Acrylonitrile
Allyl chloride
Aniline
Barium
Benzene
Carbon disulfide
Carbon tetrachloride
Chloro- 1,3 -butadiene, 2-
(chloroprene)
Chlorobenzene
Chloroform
Chromium VI
Cobalt
Cresols (total)
Cumene
Cyclohexanol
Dibromo-3-chloropropane, 1,2-
Dichlorobenzene, 1,2-
Dichlorobenzene, 1,4-
Dichlorodifluoromethane
Dichloroethane, 1,2-
Dichloroethylene, 1,1-
Dichloropropane, 1,2-
Dichloropropene, cis-1,3-
Subchronic
RfC (mg/m3)
9.0E-02
3.1E+01
6.0E-02
2.0E-04
3.0E-03
2.0E-02
l.OE-02
l.OE-02
5.0E-03
1.3E-02
7.0E-01
3.1E-01
7.0E-02
2.0E-01
2.4E-01
5.0E-04
3.0E-05
1.2E-03
4.0E+00
2.0E-04
2.0E-03
2.0E+00
2.5E+00
2.0E+00
8.1E-01
7.9E-02
1.3E-02
2.0E-02
Target Organ
Respiratory
Neurological
Death
Respiratory
Respiratory
Respiratory
Neurological
Spleen
Developmental
Neurological
Neurological
Liver
Respiratory
Liver, kidney
Liver
Respiratory
Respiratory
Hematological
Kidney, adrenal
Muscle
Reproductive
Body weight
Liver
Liver
Liver,
immunological
Liver
Respiratory
Respiratory
Source
RfC based on subchronic study -
removed UF
ATSDR intermediate MRL
IRIS
RfC based on subchronic study -
removed UF
HEAST
RfC based on chronic study -
adjusted w/ AF
HEAST
HEAST
HEAST
ATSDR intermediate MRL
HEAST
ATSDR intermediate MRL
HEAST
Superfund Risk Assessment
Issue Paper
ATSDR intermediate MRL
ATSDR intermediate MRL
ATSDR intermediate MRL
RfC based on subchronic study -
removed UF
RfC based on subchronic study -
removed UF
RfC based on subchronic study -
removed UF
RfC based on subchronic study -
removed UF
HEAST
HEAST
HEAST
chronic & acute MRLs are equal;
subchronic = chronic also
ATSDR intermediate MRL
HEAST
HEAST
(continued)
6-11
-------
Volume II
Section 6.0
Table 6-4. (continued)
CAS#
10061-02-6
68-12-2
123-91-1
106-89-8
106-88-7
110-80-5
100-41-4
106-93-4
107-21-1
75-21-8
50-00-0
98-01-1
77-47-4
67-72-1
110-54-3
78-59-1
7439-96-5
7439-97-6
109-86-4
74-83-9
74-87-3
78-93-3
108-10-1
80-62-6
1634-04-4
75-09-2
Name
Dichloropropene, trans- 1,3-
Dimethylformamide, TV, 7V-
Dioxane, 1,4-
Epichlorohydrin
Epoxybutane, 1,2-
Ethoxyethanol, 2-
Ethylbenzene
Ethylene dibromide
Ethylene glycol
Ethylene oxide
Formaldehyde
Furfural
Hexachlorocyclopentadiene
Hexachloroethane
Hexane, n-
Isophorone
Manganese
Mercury
Methoxyethanol, 2-
Methyl bromide
(bromomethane)
Methyl chloride
(chloromethane)
Methyl ethyl ketone
Methyl isobutyl ketone
Methyl methacrylate
Methyl fer/-butyl ether
Methylene chloride
Subchronic
RfC (mg/m3)
2.0E-02
3.0E-02
8.0E+00
l.OE-02
2.0E-01
2.0E+00
8.7E-01
2.0E-03
6.0E+00
1.6E-01
1.2E-02
5.0E-01
7.0E-04
5.8E+01
2.0E-01
1.2E-01
5.0E-04
3.0E-04
2.0E-01
1.9E-01
4.1E-01
l.OE+00
8.0E-01
7.0E+00
2.1E+00
3.0E+00
Target Organ
Respiratory
Liver, GI
No liver,
kidney, or
hematological
effects
Respiratory
Respiratory
Hematological
Developmental
Reproductive
Respiratory
Kidney
Respiratory
Respiratory
Respiratory
Neurological
Neurological
Body weight
Neurological
Neurological
Reproductive
Neurological
Liver
Developmental
Liver, kidney
Respiratory
Neurological
Liver
Source
HEAST
HEAST
RTI-derived RfC based on
chronic study - adjusted w/ AF
HEAST
RfC based on chronic study -
adjusted w/ AF
HEAST
ATSDR intermediate MRL
HEAST
RfC based on subchronic study -
removed UF
ATSDR intermediate MRL
ATSDR intermediate MRL
HEAST
HEAST
ATSDR intermediate MRL
HEAST
RfC based on subchronic study -
removed UF
RfC based on chronic study -
adjusted w/ AF
HEAST
HEAST
ATSDR intermediate MRL
ATSDR intermediate MRL
HEAST
HEAST
RfC based on chronic study -
adjusted w/ AF
ATSDR intermediate MRL
HEAST
(continued)
6-12
-------
Volume II
Section 6.0
Table 6-4. (continued)
CAS#
91-20-3
7440-02-0
98-95-3
79-46-9
108-95-2
85-44-9
75-56-9
100-42-5
79-34-5
127-18-4
108-88-3
76-13-1
120-82-1
71-55-6
79-01-6
75-69-4
121-44-8
7440-62-2
108-05-4
75-01-4
1330-20-7
Name
Naphthalene
Nickel
Nitrobenzene
Nitropropane, 2-
Phenol
Phthalic anhydride
Propylene oxide
Styrene
Tetrachloroethane, 1,1,2,2-
Tetrachloroethylene
Toluene
Trichloro- 1 ,2,2-trifluoroethane,
1,1,2-
Trichlorobenzene, 1,2,4-
Trichloroethane, 1,1,1-
Trichloroethylene
Trichlorofluoromethane
Triethylamine
Vanadium
Vinyl acetate
Vinyl chloride
Xylenes (total)
Subchronic
RfC (mg/m3)
3.0E-02
2.0E-03
2.0E-02
2.0E-02
6.0E-02
1.2E-01
3.0E-02
3.0E+00
2.7E+00
3.0E+00
4.0E+00
3.0E+01
2.0E+00
3.8E+00
5.4E-01
7.0E+00
7.0E-02
7.0E-05
2.0E-01
7.7E-02
3.0E+00
Target Organ
Respiratory
Respiratory
Hematological,
adrenal, kidney,
liver
Liver
No effects
Respiratory
Respiratory
Neurological
Liver
Neurological
Respiratory,
neurological
Body weight
Liver
Neurological
Neurological
Kidney,
respiratory
No effects
Respiratory
Respiratory
Liver
Developmental
Source
RfC based on chronic study -
adjusted w/ AF
chronic MRL avail. - adjusted w/
AF
HEAST
HEAST
RfC based on subchronic study -
removed UF
HEAST
HEAST
HEAST
ATSDR intermediate MRL
RTI-derived RfC based on
chronic study - adjusted w/ AF
RfC based on chronic study -
adjusted w/ AF
HEAST
HEAST
ATSDR intermediate MRL
ATSDR intermediate MRL
HEAST
RfC based on subchronic study -
removed UF
RfC based on acute study
(subchronic = chronic)
HEAST
ATSDR intermediate MRL
ATSDR intermediate MRL
Figure 6-2 describes the approach used to select the subchronic inhalation health
benchmarks used in this analysis. Subchronic inhalation RfCs were identified in HEAST (U.S.
EPA, 1997a) for 33 constituents. If subchronic RfCs were not available from HEAST, then they
were derived from existing chronic RfCs or chronic inhalation MRLs. Ten chronic RfCs cited in
6-13
-------
Volume II
Section 6.0
Used HEAST
subchronic RfC value
if not available
Adjusted chronic RfC
(IRIS, HEAST, or derived for study)
1. Value based on subchronic
study - removed UF
or
2. Value based on chronic
study - adjusted with AF
if no chronic RfC available
Used ATSDR intermediate
inhalation MRL
Figure 6-2. Approach used to select subchronic noncancer
inhalation benchmark values.
IRIS or HEAST were based on subchronic studies; for these constituents, the uncertainty factor
(UF) applied to extrapolate from subchronic to chronic duration (usually 10) for the derivation of
the chronic RfC was removed, resulting in subchronic RfCs. Six RfCs cited in IRIS were based
on chronic studies; for these constituents, an adjustment factor (AF) of 10 was applied to
extrapolate from chronic to subchronic duration to derive subchronic RfCs.
ATSDR intermediate inhalation MRLs are for use with exposure durations of 15 to 364
days and are derived from subchronic toxicological or epidemiological studies. ATSDR
intermediate inhalation MRLs were sought if subchronic RfCs from HEAST were not available
or if chronic RfCs could not be adjusted to derive subchronic RfCs; intermediate MRLs were
identified for 19 constituents. In addition, a chronic MRL was available for nickel (2.0E-4
mg/m3); an adjustment factor was used to extrapolate from chronic to subchronic duration,
resulting in an interim subchronic RfC of 2.0E-3 mg/m3. RfCs derived for this study for
1,4-dioxane and tetrachloroethylene were based on chronic inhalation studies; an adjustment
factor was applied to these RfCs to extrapolate from chronic to subchronic duration.
Seventy-five subchronic RfCs or intermediate inhalation MRLs were identified for the
105 selected constituents; however, because subchronic exposures were modeled only for LAUs
and wastepiles for volatile chemicals and metals, only 64 of these benchmarks were used. The
remaining 11 were for semivolatiles not needed for assessing LAUs and wastepiles.
6.3 Acute Inhalation Health Benchmarks
Acute noncancer inhalation benchmark values for constituents considered in this study
are summarized in Tables 6-5 and 6-6. Acute benchmarks identified include ATSDR acute
6-14
-------
Volume II
Section 6.0
Table 6-5. ATSDR Acute Inhalation MRLs
CAS#
67-64-1
107-02-8
107-13-1
71-43-2
56-23-5
67-66-3
106-46-7
107-06-2
78-87-5
107-21-1
50-00-0
67-72-1
74-83-9
74-87-3
1634-04-4
75-09-2
127-18-4
108-88-3
71-55-6
79-01-6
7440-62-2
75-01-4
1330-20-7
ATSDR acute inhal
Name MRL (mg/m3) Target Organ
Acetone
Acrolein
Acrylonitrile
Benzene
Carbon tetrachloride
Chloroform
Dichlorobenzene, 1,4-
Dichloroethane, 1,2-
Dichloropropane, 1,2-
Ethylene glycol
Formaldehyde
Hexachloroethane
Methyl bromide (bromomethane)
Methyl chloride (chloromethane)
Methyl fer/-butyl ether
Methylene chloride
Tetrachloroethylene
Toluene
Trichloroethane, 1,1,1-
Trichloroethylene
Vanadium
Vinyl chloride
Xylenes (total)
6.2E+01
1.1E-04
2.2E-01
1.6E-01
1.3E+00
4.9E-01
4.8E+00
8.1E-01
2.3E-01
1.3E+00
6.1E-02
5.8E+01
1.9E-01
l.OE+00
6.1E+00
l.OE+01
1.4E+00
1.5E+01
1.1E+01
1.1E+01
7.0E-04
1.3E+00
4.3E+00
Neurological
Ocular
Neurological
Immunological
Liver
Liver
Developmental
Immunological
Respiratory
Kidney
Respiratory
Neurological
Neurological
Neurological
Neurological
Neurological
Neurological
Neurological
Neurological
Neurological
Respiratory
Developmental
Neurological
Table 6-6. CalEPA's 1-Hour Acute Inhalation Reference Exposure Levels (RELs)
CAS#
79-10-7
7440-38-2
75-15-0
123-91-1
106-89-8
111-15-9
110-80-5
7439.97-6
Name
Acrylic acid
Arsenic
Carbon disulfide
Dioxane, 1,4-
Epichlorohydrin
Ethoxyethanol acetate, 2-
Ethoxyethanol, 2-
Mercury
CalEPA 1-h REL
(mg/m3)
6.0E+00
4.0E-04
2.0E+01
6.0E+00
3.0E+00
3.0E-01
9.0E-01
2.0E-03
Effect
Level
I
II
II
I
I
II
II
II
Effect
Respiratory irritation
Reproductive/developmental
Reproductive/developmental
Eye irritation
Eye & respiratory irritation
Reproductive/developmental
Reproductive/developmental
Reproductive/developmental
(continued)
6-15
-------
Volume II
Section 6.0
Table 6-6. (continued)
CAS#
67-56-1
109-86-4
78-93-3
7440-02-0
108-95-2
75-56-9
100-42-5
Name
Methanol
Methoxyethanol, 2-
Methyl ethyl ketone
Nickel
Phenol
Propylene oxide
Styrene
CalEPA 1-h REL
(mg/m3)
3.0E+01
2.0E-02
3.0E+01
l.OE-02
6.0E+00
6.0E+00
2.0E+01
Effect
Level
Effect
Mild neurological
Hematological
Eye & respiratory irritation
Respiratory irritation; immunological
Eye & respiratory irritation
Eye & respiratory irritation
Eye & respiratory irritation
I = mild; II = severe
inhalation MRLs, EPA acute exposure guideline levels (AEGLs), and CalEPA (1998) 1-h acute
inhalation reference exposure levels (RELs). Twenty-three acute ATSDR MRLs were identified.
ATSDR acute inhalation MRLs are for use with exposure durations of 14 days or less and are
derived from acute toxicological or epidemiological studies (see Table 6-5). AEGLs have been
derived for aniline and ethylene oxide. CalEPA (1998) derived 1-h acute inhalation RELs for 29
of the 105 constituents. ATSDR acute MRLs were considered most appropriate for use in
modeling, therefore CalEPA acute inhalation RELs were used only when MRLs were unavailable
(for 15 constituents, see Table 6-6). Figure 6-3 describes the approach used to select the acute
inhalation health benchmarks used in this analysis.
CalEPA's RELs can be applied in risk characterization for routine industrial emissions
and planned releases as well as unplanned releases. These acute RELs are similar to EPA's
AEGLs; both incorporate the recommendations of the National Academy of Science (NAS)
published in Guidelines for Developing Community Emergency Exposure Levels for Hazardous
Substances (NAS, 1993). AEGLs represent short-term threshold or ceiling exposure values
intended for the protection of the general public, including susceptible or sensitive individuals.
The AEGLs represent biological reference values for each of four different exposure periods of
30 minutes, 1 hour, 4 hours, and 8 hours. CalEPA's acute RELs and EPA's AEGLs have been
established for three grades of effect:
# Level I and AEGL-1: discomfort or mild effect level
# Level II and AEGL-2: disability or serious effect level
# Level III and AEGL-3: life-threatening effect level.
Only AEGL-ls were used in this study. The 8-h AEGL-1 for aniline is 3.8 mg/m3; no AEGL-1
levels have been derived for ethylene oxide (U.S. EPA, 1997b).
A total of 39 acute benchmarks were identified for the 105 constituents in this study;
however, because acute exposures were modeled only for LAUs and wastepiles for volatile
chemicals and metals, only 35 of these benchmarks were used. The remaining four were for
semivolatiles not modeled in LAUs and wastepiles.
6-16
-------
Volume II
Section 6.0
Used ATSDR acute
inhalation MRLs
if not available
Used EPA
AEGL values
if not available
UsedCalEPA1-h
acute inhalation RELs
Figure 6-3. Approach used to select acute noncancer
inhalation health benchmark values.
6.4 Comparison of Chronic, Subchronic, and Acute RfCs
RfCs for chronic, subchronic, and acute exposure durations were compared to verify the
appropriate progression from chronic (lowest) to acute (highest). Acrolein, 1,2-dichloroethane,
1,4-dioxane, 2-ethoxyethanol, ethylbenzene, ethylene glycol, 2-methoxyethanol, methyl tert-
butyl ether, and tetrachloroethylene have an inconsistent progression between their chronic,
subchronic, and/or acute RfCs. Discrepancies in this progression are due to the fact that RfCs of
different durations were derived by different agencies (i.e., U.S. EPA, ATSDR, CalEPA,
Superfund). Some of these differences are due to the different dates on which values were
derived (e.g., less complete data available to EPA for the derivation of an IRIS health benchmark
in 1987 as compared to ATSDR for the derivation of an MRL in 1998), the use of different
methodologies, or the use of different toxicity endpoints (e.g., for nickel, a respiratory effect is
the endpoint used for the subchronic RfC and an immunological effect is the endpoint used for
the acute RfC).
The chronic and subchronic RfCs for vanadium are both 7E-5 mg/m3 because they were
derived from an acute-duration study and had an uncertainty factor of 10 applied to account for
the extrapolation from acute to subchronic and chronic exposure duration (the acute RfC is 7E-4
mg/m3).
The chronic and subchronic RfCs for acetonitrile are both 6E-2 mg/m3 because the
chronic RfC is based on a subchronic-duration study that does not require the application of an
uncertainty factor to adjust for extrapolation to longer exposures according to EPA. That is, EPA
would have derived the same value for both chronic and subchronic RfCs based on this study.
EPA's rationale for not applying an uncertainty factor to account for the use of a subchronic
study was that "there was no mortality in the longer term mouse study. Therefore, although this
endpoint is of concern based on the subchronic study, increased exposure would not be expected
to increase the sensitivity to this endpoint" (U.S. EPA, 1999).
6-17
-------
Volume II Section 7.0
7.0 Development of Risk-Specific Waste
Concentration Distribution
This section describes the models used to combine results of the emissions modeling
(Section 4.0) and dispersion modeling (Section 5.0) with exposure factor distributions to
calculate air concentration, risk or hazard quotient, and risk-specific waste concentration (Cw) for
chronic, subchronic, and acute exposures. In addition, it describes the stability analysis
performed for the Monte Carlo model used to calculate chronic risk-specific waste
concentrations for carcinogens and modifications made to the methodology for lead. The section
begins with an overview of the modeling, then describes the modeling in detail step by step.
Following that are subsections describing the development of exposure factors, the Monte Carlo
model stability analysis, and modifications for lead.
7.1 Overview
The risk model calculates risk-specific waste concentrations for carcinogens and
noncarcinogens for three averaging times (chronic, subchronic, and acute) for adults and
children. The model structure varies somewhat among these different categories.
7.1.1 Chronic Exposures to Carcinogens
For carcinogens, a Monte Carlo analysis was performed in which the location of the
receptor at a given distance and various exposure factors (body weight, inhalation rate, and
exposure duration) were varied (exposure frequency was kept fixed as it has little variability).
For each constituent and waste management unit type combination, a separate Monte Carlo
simulation was run for each WMU in the Industrial D Survey database (Shroeder et al., 1987) or
TSDR tank database (U.S. EPA, 1987). The emission rate for the specific constituent from the
specific WMU was used as an input to the Monte Carlo simulation and was not varied across
iterations within a simulation.
Approximately 1,000 iterations were performed for each WMU, resulting in a distribution
of waste concentrations (Cw) that would result in the specified risk criterion. This distribution
captures the range in waste concentration attributable to the variability in the potential location
and to the exposure factors associated with each receptor. From this distribution, the 85th, 90th,
and 95th percentiles were selected to characterize the distribution. These percentiles represent the
percentage of receptors that are protected at the risk criterion for a specific WMU.
When the Monte Carlo simulation had been run for all the WMUs, a cumulative
distribution across all facilities for each protection level (85, 90, or 95 percent of receptors) was
~
-------
Volume II Section 7.0
obtained for each receptor at each distance. This distribution reflects the variability across
facilities. In developing this distribution, the results were weighted using the facility weights
from the Industrial D Survey data. These weights indicate the number of U.S. facilities
represented by a particular facility in the Industrial D Survey database. The resulting cumulative
distribution accounts for variability across all facilities represented, not just those actually
modeled. The TSDR survey data used to characterize tanks did not include facility weights;
therefore, the tank distributions are not weighted.
7.1.2 Chronic Exposures to Noncarcinogens
Hazard quotients for noncarcinogens depend only on air concentration and the health
benchmark (a reference concentration). Therefore, exposure factors are not used and only the
location of the receptor is relevant of the variables varied in the Monte Carlo analysis. Because
the location of the receptor is such a simple distribution, a Monte Carlo analysis was unnecessary
for noncarcinogens; the distribution of hazard quotient (and therefore Cw) based on the
distribution of the location of the receptor can be obtained analytically by calculating hazard
quotient (and Cw) for each of the 16 receptor locations (or directions around the site) and taking
the desired percentiles from those 16 values.
7.1.3 Acute/Subchronic Exposures
Exposure and risk modeling for subchronic and acute exposures differed somewhat from
the modeling for chronic exposures in several respects. All acute and subchronic health
benchmarks are analagous to chronic noncarcinogen benchmarks, so exposure factors were not
used. Neither receptor location around the site was varied for acute and subchronic exposures.
Therefore, a Monte Carlo analysis was not performed for subchronic and acute exposures.
Instead, a point estimate of Cw was calculated for each WMU in the Industrial D database using
the single sector that resulted in the maximum air concentration. This point estimate represents
the maximum, or 100th percentile, concentration, and therefore is most comparable to the 100th
percentile of the distribution generated by the Monte Carlo model for chronic exposures. The
point estimate can be interpreted as the level at which 100 percent of receptors are protected at a
particular WMU. A distribution across all WMUs of a specific type was generated from these
point estimates, and the 90th percentile of that distribution is presented in the results for
subchronic and acute exposures.
7.1.4 Adult Exposure Approach
Adult receptors modeled include adult residents and off-site workers. Risk for adults is
calculated using long-term average air concentration that is constant over the entire exposure
duration.
The inhalation rate, exposure frequency, and exposure duration differ for residents and
workers. Body weight is the same for all adults, whether resident or worker. All exposure
factors for adults are held constant over the entire exposure duration. The actual point estimates
or distributions of the exposure factors used are described in Section 7.11.
7-2
-------
Volume II Section 7.0
7.1.5 Child Exposure Approach
Three child age groups, or cohorts, were used to model child exposures: 0 to 3, 4 to 10
and 11 to 18 years of age. These cohorts reflect the age cohorts for which inhalation rate data are
available. Results were calculated and saved for receptors falling into each of these three age
cohorts at the start of exposure and presented as child 1 (0-3), child 2 (4-10), and child 3 (11-18).
An exposure duration was selected randomly for each of the three starting age cohorts from a
distribution specific to each starting age cohort. For each age cohort, exposure begins at a
starting age selected at random within the cohort and then continues through succeeding age
cohorts and into adulthood as necessary until the exposure duration selected for that starting age
cohort is reached.
Annual risk for each year of exposure (from starting age to starting age plus exposure
duration) was calculated and summed over the exposure duration for each child receptor. If the
child reached age 19 before the exposure duration ended, adult exposure factors were used for the
remainder of the exposure duration. This approach requires both body weight and inhalation rate
distributions by year of age; however, only body weight is available by year. Inhalation rate is
available only for the age groups used to define the cohorts (0-3, 4-10, and 11-18 years).
Because inhalation rate data could not be disaggregated to individual years of age, we retained
year-by-year body weights and used the inhalation rate for the cohort associated with each year of
age for that year. Thus, the inhalation is a constant for all ages within an age cohort and changes
only when the receptor ages from one cohort to the next. Both EPA and a statistician
experienced in working with EFH exposure factor data (L. Myers, RTI, personal communication
with Anne Lutes, RTI, March 16, 1998) preferred this approach over the alternative of pooling
body weights to the age cohort age ranges as retaining the most detail from the available data
without sacrificing statistical rigor.
7.1.6 Specific Steps Required in the Risk Analysis
Most of the following steps are common to the risk model structure for carcinogens and
noncarcinogens, all three averaging times, and both adults and children:
# Select a receptor location
# Obtain the appropriate unitized air concentration (UAC)
# Calculate air concentration
# Select exposure factors from distributions
# Obtain health benchmarks
# Calculate risk or hazard quotient
# Backcalculate risk-specific waste concentration (Cw)
# Adjust tank results for nonlinearity in biodegradation
# Flag results exceeding soil saturation concentration or solubility.
For chronic exposure to carcinogens, these steps were repeated for each iteration in the
Monte Carlo simulation. For chronic exposure to noncarcinogens, these steps were performed
once for each of the 16 receptor locations. For acute and subchronic exposures, these steps were
performed once. Sections 7.2 through 7.10 describe these above steps in more detail.
7-3
-------
Volume II Section 7.0
7.2 Select Receptor Location
7.2.1 Chronic Exposure
Dispersion coefficients were modeled for each of 6 distances and 16 directions as
described in Section 5.0. The 6 distances were 25, 50, 75, 150, 500, and 1,000 meters. The 16
directions were equally spaced around the WMU. Calculations were performed for each
distance, and direction was allowed to vary.
Five receptors were modeled: an adult resident, a child resident ages 0 to 3 years at the
start of exposure, a child resident ages 4 to 10 years at the start of exposure, a child resident ages
11 to 18 years at the start of exposure, and an off-site worker. These receptors were assessed at
all six distances. All receptors could be located in any of the 16 directions.
Direction was modeled as a uniform distribution so that a receptor had a probability of 1
in 16 (or 0.0625) of being located in any one of the 16 directions. The same direction was used
for all receptors for a particular iteration.
For carcinogens, Crystal Ball® was used to randomly sample the distributions and
generate values of the exposure factors for each iteration. However, receptor locations (i.e.,
direction from the site) were not generated using Crystal Ball®. Preliminary analysis suggested
that random sampling from this distribution would not achieve stability unless considerably more
than 1,000 iterations were used. However, it is not necessary to sample from this distribution,
because the distribution is discrete and is completely described by the uniform probability
(l/16th) of a receptor being located in any particular direction. Thus, this distribution is
computationally most efficiently represented in the Monte Carlo simulation, not by random
sampling but rather by simply "visiting" each of the 16 directions an equal number of times
during the course of the Monte Carlo simulation. That is, if one visits each of the 16 directions
(deterministically) an equal number of times, there is no remaining direction-related variability
from which to "sample," because the entire population has been visited. Visiting the directions
an equal number of times for each Monte Carlo simulation required a number of iterations that
was a multiple of 16 (e.g., 1,008). We then associated each direction with one set of exposure
factors generated by the Monte Carlo simulation. In this way, we ensured that an equal selection
of the directions was made and that the directions were associated randomly with exposure
factors.
For noncarcinogens, exposure factors are not used, so results were calculated once for
each of the 16 directions.
7.2.2 Acute/Subchronic Exposures
Dispersion coefficients were modeled for each of 3 distances and 16 directions as
described in Section 5.0. The 3 distances were 25, 50, and 75 meters. The 16 directions were
equally spaced around the WMU. Calculations were performed for each distance, and the
direction that gave the maximum air concentration was used. The direction of maximum air
concentration varied from WMU to WMU depending on prevailing wind directions.
7-4
-------
Volume II Section 7.0
7.3 Obtain Unitized Air Concentrations
After a receptor location had been selected, vapor-phase and particulate-phase UACs for
that location were calculated. The UACs were estimated by interpolating between the UACs
developed for areas immediately above and below the actual area of the unit, as follows:
A-A.
(UACr UAC^ + UACi (7-1)
UAC =
where
UAC = unitized air concentration for specific WMU ([|ig/m3]/[|ig/m2-s])
A = area of specific WMU (m2)
A; = area modeled in dispersion modeling immediately below area of specific
WMU (m2)
Aj = area modeled in dispersion modeling immediately above area of specific
WMU (m2)
UACj = unitized air concentration developed for area j ([|ig/m3]/[|ig/m2-s])
UAQ = unitized air concentration developed for area i ([|ig/m3]/[|ig/m2-s]).
A few WMUs had areas that fell either below the smallest area or above the largest area modeled
in the dispersion modeling. If the area was less than the smallest area modeled, Aj and UACj
were set to the values for the smallest area modeled, and A; and UAQ were set to zero. If the
area was greater than the largest area modeled, the A;, UAQ, Aj, and UACj were set to
correspond to the two largest areas modeled, based on the assumption that the UAC continues to
increase with the same slope above the largest area modeled.
For WMUs that were elevated (i.e., wastepiles and tanks), UACs were developed for
several different heights for each modeled area. The UACs for the modeled height closest to
each unit's actual height were used in the interpolation.
7.4 Calculate Air Concentration
Air concentration was calculated from the WMU emission rate (which remains constant
through all iterations of the Monte Carlo simulation for a particular WMU) and the interpolated
UACs. Recall that the emission rates are emissions associated with a unit waste concentration of
1 mg/kg (or 1 mg/L for tanks), so the resulting air concentration is also associated with a unit
waste concentration. Air concentration was calculated as follows:
7-5
-------
Volume II Section 7. 0
UACvapor + C^^UACpartic)^mimg^g (7-2)
where
Cair = air concentration associated with a unit waste concentration
([mg/m3]/[mg/kg] or [mg/m3/[mg/L])
Evapor = emission rate of constituent in vapor phase ([g/m2-s]/[mg/kg] or
[g/m2-s]/mg/L])
CSOii = average annual soil concentration in unit ([mg/kg]/[mg/kg] =
[|ig/g]/[mg/kg] or
Epartic = emission rate of parti culates (g/m2-h)
UACvapor = vapor-phase unitized air concentrations ([|ig/m3]/[|ig/m2-s]).
UACpartic = particulate-phase unitized air concentrations ([|ig/m3]/[|ig/m2-s]).
Note that Csoil and Epartic are both zero for tanks. Vapor emissions (Evapor) are zero for nonvolatile
constituents (all metals except mercury).
7.5 Select Exposure Factors
The development of exposure factor distributions is described in Section 7.11. Values for
each exposure factor (inhalation rate, body weight, and exposure duration) were selected
independently for each receptor. Specifically, for child receptors, exposure factors were selected
independently for a specific age or cohort for each starting age cohort. For example, one set of
exposure factors for a 9-year-old child was selected for use with the child 1 receptor (a child
whose exposure starts between 0 and 3 years of age) and a second set was selected independently
for use with the child 2 receptor (a child whose exposure starts between 4 and 10 years of age).
Risk results are not calculated separately by gender. Therefore, we considered pooling
distributions for males and females. However, pooling these distributions across genders risks
losing the slightly bimodal character of the overall distribution created by differences in exposure
factors for males vs. females. Therefore, it was considered preferable and more statistically
rigorous to retain the separate distributions for males and females. We then used the male
distribution for half the Monte Carlo iterations and the female distribution for the other half.
The health benchmarks used for chronic exposure to noncarcinogens and subchronic and
acute exposure to all chemicals are expressed as ambient air concentrations, and are compared
directly to the modeled air concentration, without the use of exposure factors. Therefore, this
step is not needed for noncarcinogens or acute and subchronic exposures.
7-6
-------
Volume II Section 7.0
7.5.1 Inhalation Rate
To assess chronic exposures, an average daily inhalation rate is needed. Such a rate is
based on inhalation values for a variety of activities averaged together.
The inhalation rate is linearly related to exposure and risk. A twofold change in the
inhalation rate results in a twofold change in the risk estimate for a carcinogen. Thus, the
sensitivity of this parameter in the analysis is a function of its variability. An examination of data
presented in the draft EFH on daily inhalation rates for adults, workers, and children provides a
general idea of the variability in this factor. In general, the range for both adults and workers
would be around a factor of 2 or 3. Children's rates are more variable.
For children, inhalation rates were held constant over each age cohort because there were
insufficient data to disaggregate them to specific years of age.
7.5.2 Body Weight
Distributions for body weight were needed that were consistent with the distributions
used for inhalation rate. Therefore, body weight distributions for adult males and females (to be
used for both adult residents and workers) and male and female children were needed. For
children, body weights specific to each year of age were used.
7.5.3 Exposure Duration
A starting age within each child cohort was selected at random, assuming equal
distributions of ages within the cohort. The assumption of equal distribution by age within each
cohort is supported by Census data showing that the population is roughly equally divided among
each year of age (see http://www.census.gov/cdrom/lookup/922975351; 100% count, national
totals, Table PI 1-Age).
When an exposure duration extended beyond the end of the age cohort, exposure was
continued in the next age cohort (or adulthood) until the exposure duration was reached. For
example, if exposure began at 0 years, and exposure duration was 13 years, then 3 years were
modeled as age 0 to 3, 7 years were modeled as age 4 to 10, and 3 years were modeled as age 11
to 18. An exposure duration was picked at random for each starting age cohort.
7.5.4 Exposure Frequency
Exposure frequency is held constant throughout the analysis.
7.6 Obtain Health Benchmarks
Standard health benchmarks (cancer slope factors for carcinogens and reference
concentrations for noncarcinogens) were obtained for each constituent (see Section 6).
7-7
-------
Volume II Section 7. 0
7.7 Calculate Risk or Hazard Quotient
The risk or hazard quotient associated with a unit waste concentration was calculated
based on the calculated air concentration and, if relevant, the exposure factors selected.
Risk for carcinogens for adults was calculated as follows:
C x CSF xIR
Risk , ,, = — - (7-3)
calcd BWxATx365 d/yr ^ }
where
Riskcak'd = individual risk associated with unit waste concentration (per mg/kg)
Cair = air concentration associated with a unit waste concentration
([mg/m3]/[mg/kg])
CSF = cancer slope factor (per mg/kg-d)
IR = inhalation rate (m3/d)
ED = exposure duration (yr)
EF = exposure frequency (d/yr)
BW = body weight (kg)
AT = averaging time (yr) = 70.
The air concentration is a long-term average and is constant over the entire exposure duration.
Averaging time is a fixed input to this equation (with a value of 70) because it must be consistent
with the averaging time used to develop standard cancer slope factors.
For children, annual risk for each year of exposure (from starting age to starting age plus
exposure duration) was calculated and summed over the exposure duration for each child
receptor. Ideally, we would have body weight and inhalation rate data specific to each single
year of age, and risk would be calculated as follows for each starting age cohort:
s+EDj C
Risk = Y — - - (7-4)
7 ^ BWxATx365d/yr
7-8
-------
Volume II Section 7. 0
where
Riskj = individual risk associated with unit waste concentration for child starting
exposure in cohort j (unitless)
i = age (yr)
s = age at start of exposure (yr)
EDj = total exposure duration for starting age cohort j (yr)
Cair = long-term average air concentration associated with a unit waste concentration
([mg/m3] [mg/kg])
IRj = inhalation rate for age i (m3/d)
CSF = cancer slope factor (per mg/kg-d)
ED; = exposure duration for age i (yr). This is usually 1, but for the last year of
exposure, may be less than 1 to accommodate an ED that is not an integer
number of years, or greater than 1 if the child has aged into adulthood. The
sum of all ED =
EF = exposure frequency (d/yr)
BW; = body weight for age i (kg)
AT = averaging time (yr) = 70.
As for adults, the air concentration is a long-term average and is constant over the entire
exposure duration.
As described in Section 7.1.5, inhalation rate data could not be disaggregated to
individual years of age; therefore, the inhalation rate for the cohort associated with age i is
substituted in the above equation for IR^. Thus, IR^ becomes a constant for all ages within an age
cohort and changes only when the receptor ages from one cohort to the next.
The hazard quotient for noncarcinogens (including all chemicals for acute and subchronic
exposures) was calculated as follows:
HQcaic'd = J^ (7-5)
7-9
-------
Volume II Section 7. 0
where
HQCaic'd = hazard quotient associated with unit waste concentration (per mg/kg)
Cgjr = air concentration associated with a unit waste concentration
([mg/m3]/[mg/kg])
RfC = chronic reference concentration or acute or subchronic benchmark (mg/m3).
Because the hazard quotient equation does not consider exposure factors, there is no
difference in results for different receptors at the same location (e.g., adult resident, child
resident, and offsite worker). Therefore, only an adult resident was modeled for noncarcinogens
and for acute and subchronic exposures.
7.8 Backcalculate Risk-Specific Waste Concentration
The next step was to backcalculate the risk-specific waste concentration from the risk or
hazard quotient corresponding to a unit waste concentration. Because risk is linear with respect
to waste concentration,1 in the models used in this analysis, this may be done by a simple ratio
technique:
Risk , HO ,
L (7-6)
calc'd
where
Cw = risk-specific waste concentration (mg/kg)
Riskcrit = risk criterion (unitless)
Riskcak'd = risk associated with unit waste concentration (per mg/kg)
HQcrit = hazard quotient criterion (unitless)
HQcaic>d = hazard quotient associated with unit waste concentration (per mg/kg).
When a particular constituent had both carcinogenic and noncarcinogenic effects, the
carcinogenic risk was used to continue the calculations, because it generally results in a more
protective concentration.
7.9 Adjust Tank Results For Non-Linearity of Biodegradation
As mentioned, risk is assumed to be linear with waste concentration. The assumption of
linearity is accurate for the dispersion modeling and the exposure and risk modeling. However,
the emissions model is linear only for land-based units and tanks with no biodegradation; for
tanks with biodegradation, the emissions model is nonlinear with respect to biodegradation. At
The volatile emission model for tanks with biodegradation can be an exception to this assumption of
linearity. Section 7.9 describes how those situations were identified and resolved.
-------
Volume II
Section 7.0
low concentrations, biodegradation in tanks is first order. However, at concentrations in excess
of the half-saturation level, biodegradation becomes zero order. In order to address this,
emissions were modeled in the aqueous phase at 0.001 mg/L to capture first-order biodegradation
and at the solubility to capture zero-order biodegradation. These emission rates then were
normalized to a unit concentration by dividing by 0.001 or the solubility. When the
backcalculated waste concentration based on first order biodegradation exceeded the half-
saturation constant, suggesting that biodegradation would be zero order, it was recalculated based
on the normalized solubility limit emission rate. Figure 7-1 summarizes this approach for tanks
with biodegradation (i.e., some aerated tanks).
7.10 Flag Results That Exceed Soil Saturation Concentration or Solubility
The results for all WMU types presented in Volumes I and III were calculated as
described above using aqueous-phase emission rates. Most of the waste streams managed in the
types of units modeled are expected to contain constituents in the aqueous, rather than the
organic, phase; therefore, this is the most realistic scenario. However, results based on organic-
phase emissions are of interest in two circumstances: when organic-phase emissions are higher
than aqueous-phase emissions, and when backcalculated results based on aqueous-phase
emissions exceed physical limitations on the aqueous phase such as the soil saturation
concentration or solubility. These are discussed below.
7.10.1 Organic-Phase Emissions Higher than Aqueous-Phase Emissions
Most chemicals are better able to volatilize from an aqueous medium than from an
organic medium; therefore, for most chemicals, the aqueous-phase emission rates are
Calculate
r
w, aqueous, 1st order'
p
w, aqueous, zero order
I o f^ ** f^ O
w, aqueous, 1st order — 14sat'
w w, aqueous, 1st order
w w, aqueous, zero order
Figure 7-1. Linearity adjustments for tanks with biodegradation.
7-11
-------
Volume II Section 7. 0
considerably higher than organic-phase emission rates. However, for a few chemicals (most
notably formaldehyde), the organic-phase emissions are higher than the aqueous-phase emissions
and waste concentrations based on organic-phase emissions would be lower than waste
concentrations based on aqueous-phase emissions. When this is the case, the results based on
aqueous-phase emissions are footnoted to indicate this. This does not invalidate the aqueous-
phase results, as that is still the most likely waste matrix.
7.10.2 Backcalculated Concentrations Exceed Physical Limitations on Aqueous Phase
For land-based units, some of the backcalculated waste concentrations based on aqueous-
phase emissions exceed the soil saturation concentration. This is the maximum possible aqueous-
phase concentration in soil; once this is exceeded, free (organic-phase) product will occur in the
soil. This soil saturation concentration has been estimated for each chemical in the analysis,
using the following equation:
P, + , +Xa (7-7)
b
where
Csat = soil saturation limit (mg/kg)
S = solubility limit (mg/L)
pb = bulk density of soil / waste matrix (kg/L)
Kd = soil-water partition coefficient (L/kg)
6W = water-filled soil porosity (unitless)
FT = dimensionless Henry's law constant (unitless, = H/RT)
6a = air-filled soil porosity (unitless).
The soil saturation concentration is a somewhat site- and waste-specific value. Therefore,
a backcalculated concentration may exceed it in some situations but not in others. When the
backcalculated concentration based on aqueous-phase emissions exceeded the typical soil
saturation concentration calculated for this analysis as shown in Equation 7-7, the result was
footnoted to indicate whether pure component (i.e., a concentration of 106 mg/kg) would result in
a risk exceeding the target risk when modeled using organic-phase emission rates.
Similarly, for tanks, some of the backcalculated waste concentrations based on aqueous-
phase emissions exceed the solubility. This is the maximum possible aqueous-phase
concentration; once this is exceeded, free (organic-phase) product will occur in the tank and
either sink, yielding aqueous-phase emissions from a concentration equal to the solubility, or
float on the surface, yielding emissions from the organic phase at a concentration of pure
component. The solubility under standard temperature and pH conditions (20-25 C and neutral
pH) has been estimated for each chemical in the analysis, but this is a somewhat site- and waste-
specific value. Therefore, a backcalculated concentration may exceed it in some situations but
not in others. When the backcalculated concentration based on aqueous-phase emissions
exceeded the typical solubility calculated for this analysis, the result was footnoted to indicate
7-12
-------
Volume II Section 7.0
whether pure component (i.e., a concentration of 106 mg/kg) would result in a risk exceeding the
target risk when modeled using organic-phase emission rates.
7.11 Derivation of Exposure Factors
This section describes the derivation of exposure factor distributions used in the model.
All data in this section are from the Exposure Factors Handbook (U.S. EPA, 1997'c and 1997d;
hereafter, the EFH).
All exposure factor distributions and point estimates were developed for the following
subpopulations:
# Adult residents
# Children ages 0-3 years
# Children ages 4-10 years
# Children ages 11-18 years
# Workers.
The age ranges for children were chosen for consistency with the data on inhalation rate
in the EFH.
7.11.1 Inhalation Rate
Single values for inhalation rates are presented as recommended values for long-term
dose assessments in the EFH. Values of 11.3 m3/d for women and 15.2 m3/d for men are
representative of average inhalation rates for adults. Upper percentile values are not
recommended. These values differ significantly from the 20 m3/d commonly assumed in past
EPA risk assessments. Additional data are presented for other activity patterns that can be used
for special subpopulations (athletes, outdoor workers), as well as to more accurately reflect the
actual exposed population. The EFH does not make any recommendation on the use of
distributional data for this parameter for any receptor.
Although the EFH provides no guidance for developing a distribution of inhalation rates,
EPA is continuing to work on the development of distributions for exposure factors from the data
in the EFH. As part of this effort, Myers et al. (1998) present two fitted distributions, one
lognormal and one gamma, for inhalation rate for male and female residents in six age ranges (0-
3 years, 4-10 years, 11-18 years, 19-30 years, 31-60 years, and >60 years) based on data from the
EFH. Myers et al. found that the difference between the two distributions is negligible and
recommend using the lognormal distribution (L. Myers, RTI, personal communication with Anne
Lutes, RTI, March 16, 1998).
The parameters of the lognormal distribution (from Table 5.2 in Myers et al.) are
presented in Table 7-1. For adult residents, we used ages 30 to 60 years (there is little difference
between this and other adult age cohorts, which are 18 to 30 and over 60). It is simpler to use
one representative age group than to pool the data and develop an overall distribution. For
children, we used the age cohorts 0 to 3, 4 to 10, and 11 to 18 years of age as presented. The
-------
Volume II Section 7.0
Table 7-1. Estimated Parameters for Inhalation Rate for
Residents Assuming Lognormal Distribution
Age (yr)
0-3
0-3
4-10
4-10
11-18
11-18
19-30
19-30
31-60
31-60
>60
>60
Gender
Male
Female
Male
Female
Male
Female
Male
Female
Male
Female
Male
Female
Mean
(m3/d)
7.52
5.75
9.30
8.65
14.58
10.76
16.75
11.14
16.32
10.95
12.69
10.44
CV
(%)
73
71
30
31
36
31
31
30
32
29
34
29
Std Dev
(m3/d)
5.49
4.08
2.79
2.68
5.25
3.34
5.19
3.34
5.22
3.18
4.31
3.03
50%ile
(m3/d)
6.1
4.7
8.9
8.3
13.7
10.3
16.0
10.7
15.6
10.5
12.0
10.0
90%ile
(m3/d)
14.1
10.7
13.0
12.1
21.4
15.1
23.7
15.6
23.2
15.1
18.4
14.5
95%ile
(m3/d)
17.8
13.4
14.5
13.5
24.2
16.8
26.5
17.3
25.9
16.7
20.7
16.0
99%ile
(m3/d)
27.9
20.8
17.7
16.6
30.6
20.7
32.6
21.2
32.0
20.3
25.9
19.5
Std Dev = Mean x CV.
Note: Bolded values were used in this analysis.
Source: Myers et al., 1998, Table 5.2.
specific values that were used to define the distribution for the Monte Carlo analysis are shown
in bold.
The data in the EFH are not adequate to develop a distribution of inhalation rates for
workers. Therefore, a point estimate was used for the inhalation rate for workers. Table 7-2
summarizes the values for inhalation rate for workers presented in the EFH. The recommended
hourly average of 1.3 m3/h was used. To convert this to a daily value, an 8-h workday was
assumed, yielding a daily inhalation rate for workers of 10.4 m3/d.
To calculate risk for children on a year-by-year basis, inhalation rates must be generated
in each Monte Carlo iteration for each age cohort through which the child might age. Inhalation
rate in one cohort is strongly correlated with inhalation rate in the previous and subsequent
cohorts: a child in the 80th percentile for inhalation rate from 0 to 3 years of age is not likely to
be in the 20th percentile for inhalation rate from 4 to 10 years. Selecting the inhalation rates for
each cohort independently presumes a correlation coefficient of 0 between cohorts. It would be
preferable to account for the correlation. Unfortunately, no data documenting such a correlation
quantitatively were readily available. The true correlation coefficient should be between 0
(meaning no correlation) and +1 (meaning a perfect correlation, in which the value of one
parameter increases as the other increases) and is probably closer to +1 than 0; therefore, in the
absence of data to support a more precise estimate of the correlation coefficient, we set it to +1.
This ensures that inhalation rates remain roughly the same percentile from cohort to cohort. A
correlation of less than +1 would result in more variation between the percentiles of the two
correlated values.
7-14
-------
Volume II Section 7.0
Table 7-2. Recommended Inhalation Rates for Workers
Activity Type
Slow activities
Moderate activities
Heavy activities
Hourly average
Mean
(m3/h)
1.1
1.5
2.5
1.3
Upper Percentile
(m3/h)
NA
NA
NA
3.3
NA = Not available.
Note: Bolded values were used in this analysis.
Source: U.S. EPA, 1997c, Table 5-23.
7.11.2 Body Weight
The EFH provides data on body weight for males and females by age. For children, body
weights are provided for individual years (e.g., age 5, age 6, etc.). For adults, data are provided
by 10-year age groups as well as for all adults ages 18 to 75. The EFH presents mean and
standard deviation for each gender and age group. In addition, means (but no standard
deviations) are presented by age for both genders pooled.
To develop distribution statistics from the EFH, we used EFH Tables 7-2 (adults,
including workers) and 7-3 (children). These data are summarized here in Tables 7-3 and 7-4. A
lognormal distribution was found to best fit these data and was used.
For adults (residents and workers), we used the mean and standard deviation for adults
ages 18 to 75 years from Table 7-2. For children, we used year-by-year means and standard
deviations from Table 7-3; for age 0, we used the statistics for 6- to 11-month-olds. These two
parameters are sufficient to define a lognormal distribution.
To calculate risk for children on a year-by-year basis, body weights must be generated in
each Monte Carlo iteration for each age from 0 to 18. Body weight in one year is strongly
correlated with body weight in the previous and subsequent years: a child in the 80th percentile
for weight one year is not likely to be in the 20th percentile for weight the next year. Selecting
these annual body weights independently presumes a correlation coefficient of 0 between years.
It would be preferable to account for the correlation. Unfortunately, no data documenting such a
correlation quantitatively were readily available. The true correlation coefficient should be
between 0 (meaning no correlation) and +1 (meaning a perfect correlation in which the values of
one parameter increases as the other increases) and is probably closer to +1 than 0; therefore, in
the absence of data to support a more precise estimate of the correlation coefficient, we set it to
+1. This ensures that body weights remain roughly the same percentile from year to year. A
correlation of less than +1 would result in more variation between the percentiles of the two
correlated values.
7-15
-------
Volume II
Section 7.0
Table 7-3. Body Weights for Adults, Ages 18-74 Years (kg)
Age
(years)
18-74
18-24
25-34
35-44
45-54
55-64
65-74
Males
Mean
78.1
73.8
78.7
80.9
80.9
78.8
74.8
Std Dev
13.5
12.7
13.7
13.4
13.6
12.8
12.8
Females
Mean
65.4
60.6
64.2
67.1
68.0
67.9
66.6
Std Dev
14.6
11.9
15.0
15.2
15.3
14.7
13.8
Note: Bolded values were used in this analysis.
Source: U.S. EPA, 1997c, Table 7-2.
Table 7-4. Body Weights for Children, Ages 6 Months to 18 Years (kg)
Males
Age (years)
6-11 months
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Mean
9.4
11.8
13.6
15.7
17.8
19.8
23.0
25.1
28.2
31.1
36.4
40.3
44.2
49.9
57.1
61.0
67.1
66.7
71.1
Std Dev
1.3
1.9
1.7
2.0
2.5
3.0
4.0
3.9
6.2
6.3
7.7
10.1
10.1
12.3
11.0
11.0
12.4
11.5
12.7
Females
Mean
8.8
10.8
13.0
14.9
17.0
19.6
22.1
24.7
27.9
31.9
36.1
41.8
46.4
50.9
54.8
55.1
58.1
59.6
59.0
Std Dev
1.2
1.4
1.5
2.1
2.4
3.3
4.0
5.0
5.7
8.4
8.0
10.9
10.1
11.8
11.1
9.8
10.1
11.4
11.1
Source: U.S. EPA, 1997c, Table 7-3.
Body weights and inhalation rates were not correlated with each other. While there may
be some correlation between these variables, there are no data to provide a quantitative estimate.
Therefore, these variables were not correlated.
7-16
-------
Volume II Section 7.0
7.11.3 Exposure Duration
Exposure duration was determined using data on residential occupancy from the EFH,
Table 15-168. These data are presented here in Table 7-5. These data represent the total time a
person is expected to live at a single location, based on age. This table presents males and
females combined, and there are not sufficient data in the EFH to develop separate distributions
for males and females. Therefore, the pooled distribution was used for all iterations of the Monte
Carlo analysis.
For adult residents, we pooled age groups from 21 to 90. For children, we used the 3-yr-
old age group for the 0 to 3 cohort. We pooled the 6- and 9-yr-old age groups for the 4 to 10
cohort. We pooled the 12-, 15-, and 18-yr-old age groups for the 11 to 18 cohort.
Myers et al. (1998) found that the residential occupancy data were best fit by a Weibull
distribution. The Weibull distribution as implemented in Crystal Ball® is characterized by three
parameters: location, shape, and scale. Location is the minimum value, and in this case was
presumed to be 0. Shape and scale were determined by fitting a Weibull distribution to the
pooled data, as follows.
To pool residential occupancy data for each of the four desired age cohorts, we formed an
arithmetic mean of data means for each of the age groups that we desired to pool (i.e., for
children ages 4-10, we took the mean of the data means for the 6-yr-old age group and the 9 yr-
old age group). Then, assuming a Weibull distribution, we calculated the variance within each
age group (e.g., 6-yr-olds) in the age cohort. These variances in turn were pooled over the age
cohort using equal weights. This is not the usual type of pooled variance, which would exclude
the variation in the group means. However, we wanted the overall variance to reflect the
variance of means within the age groups (e.g., within the 6-yr-old age group). The standard
deviation was estimated as the square root of the variance. The coefficient of variation (CV) was
calculated as the ratio of the standard deviation divided by the Weibull mean. For each cohort,
we then took the Weibull distribution that agreed with the given calculated data mean for the age
cohort and the CV as calculated above. The Weibull parameters used are presented in Table 7-6.
Exposure duration for adults was capped at 51 years. This is the maximum exposure
duration, starting at age 19, that will not exceed the 70-year lifetime assumption implicit in the
averaging time used. For landfills, exposure duration was further limited to 20 years, the
assumed operating life of landfills in this analysis.
For workers, the typical default exposure values used by EPA in the past were an 8-h
shift, 240 d/wk, for 40 years. The EFH presents data on occupational mobility that are in stark
contrast to the assumed value of 40 years at a single place of employment. As presented in the
EFH, the median occupational tenure of the working population (109.1 million people) ages 16
years of age and older in January 1987 was 6.6 years. Since the occupational tenure varies
significantly according to age, the EFH recommends using age-dependent values. When age
cannot be determined, use of the median tenure value of 6.6 years for working men and women
16 years and older is recommended.
7-17
-------
Volume II
Section 7.0
Table 7-5. Descriptive Statistics for Residential Occupancy Period by Age (years)
Percentile
Current Age (years)
3
6
9
12
15
18
21
24
27
30
33
36
39
42
45
48
51
54
57
60
63
66
69
72
75
78
81
84
87
90
25
o
6
4
5
5
5
4
2
2
o
J
3
4
5
5
6
7
8
9
9
10
11
11
12
12
13
13
12
11
11
10
8
50
5
7
8
9
8
7
4
4
5
6
7
8
9
11
13
14
15
16
17
18
19
20
20
20
20
19
20
19
18
15
75
8
10
12
13
12
11
8
6
8
9
11
13
15
18
20
22
24
25
26
27
27
28
29
29
29
29
29
28
29
27
90
13
15
16
16
16
16
13
11
12
14
17
21
24
27
31
32
33
34
35
35
36
36
37
37
38
38
39
37
39
40
95
17
18
18
18
18
19
17
15
16
19
23
28
31
35
38
39
39
40
41
40
41
41
42
43
43
44
45
44
46
47
99
22
22
22
23
23
23
23
25
27
32
39
47
48
49
52
52
50
50
51
51
51
50
50
53
53
53
55
56
57
56
Source: U.S. EPA, 1997d, Table 15-168.
Table 7-6. Estimated
Age (years)
0-3
4-10
12-18
Adult
Weibull
Parameters
Location Scale
0
0
0
0
7
for Exposure
(Alpha)
.059
9.467
9
.949
17.38
Duration
Shape (Beta)
1.32
1.69
1.73
1.34
For this study, it was not possible to find age distribution data for people employed in the
waste management sector. Nor was there sufficient information in the EFH to develop a
distribution for exposure duration. Since full-time exposure was assumed (8 hr/day, 250
7-18
-------
Volume II Section 7.0
day/year) a point estimate for all full-time workers (male and female) of 7.2 years was used (EFH
Table 15-160).
7.11.4 Exposure Frequency
Exposure frequency is the number of days per year that a receptor is exposed. It is not
expected to vary much, so no distributions were developed for exposure frequency. A value of
350 d/yr was used for adult and child residents, and a value of 250 d/yr was used for workers.
These are based, respectively, on 7 d/wk and 5 d/wk for 50 wk/yr to account for the receptor
being elsewhere on vacation for 2 wk/yr.
7.12 Stability Analysis of the Monte Carlo Model
A systematic evaluation of the stability of the chronic risk Monte Carlo model was
undertaken. The model was found to be stable at 1,000 iterations.
The chronic risk model varies exposure factors and receptor location in the Monte Carlo
iterations. Exposure factors varied include body weight, inhalation rate, and exposure duration.
The receptor location is varied among 16 directions from the site, each of which may occur with
equal probability. Different receptor distances from the site are assessed in separate model runs,
and the results maintained separately, so distance is not varied in the Monte Carlo analysis.
It was not feasible to run the stability analysis for all chemicals, WMUs, and sites due to
the long runtime of the risk model. Therefore, a selection of chemicals, WMUs, and sites were
chosen that represented all of the variables that could influence the stability. These included
different receptors (because they have different exposure factors), and different meteorological
locations, source areas, and source heights (because these could influence the variability of air
concentration with sector). In all, we ran the stability analysis for three carcinogens
(acrylonitrile, methylene chloride, and nitrosodi-w-butylamine), all six distances from the source,
all five receptors, and a selection of sites (25 LAUs and 28 wastepiles). The sites were selected
to represent all meteorological stations at which there were wastepiles or LAUs, a range of
source areas, and a range of source heights for wastepiles. Source area, height, and
meteorological location all potentially affect how air concentration varies by sector.
To assess the stability of the Monte Carlo model, we ran the model for the scenarios
described above and extracted the 90th percentile results after 1,008, 2,000, 3,008, 4,000, 5,008,
6,000, 7,008, and 8,000 iterations. We then compared the 90th percentile results across number
of iterations. In all cases, the number of iterations was chosen to be a multiple of 16 because of
the approach used to select the sector values.
In all, we ran 4,764 chemical/site/distance/receptor combinations through the stability
test. We then compared the 90th percentile results for each run, rounded to one significant figure,
across the different numbers of iterations. Figure 7-2 shows some typical results for which the
result varied slightly over 8,000 iterations. The 3 lower lines all round to a value of 40 at 1,008
iterations; the 3 upper lines all round to a value of 60 at 1,008 iterations. For all lines shown, the
7-19
-------
Volume II
Section 7.0
rounded value at 8,000
iterations is 50; note that
any value falling between
45 and 55 will round to
50.
Table 7-7
summarizes the results. In
88 percent of the runs, the
result did not change at all
from 1,008 iterations to
8,000 iterations. For
another 12 percent of the
runs, we saw minor
fluctuations in the value
(e.g., between 2.48 and
2.52) that result in
fluctuations in the result
when rounded to one
significant figure. These
fluctuations never resulted in a change of more than 1 digit in the single significant figure. This
was not considered significant.
4000 5000
Number of Iterations
Figure 7-2. Stability analysis results showing typical variations
resulting in change in results at one significant figure.
Table 7-7. Summary of Stability Analysis
Amount of change from
1008 to 8000 iterations
None
1 digit
Total
Number
of runs
4,173
591
4,764
Percent
of runs
88
12
7.13 Modifications to Methodology for Lead
Human health risk assessment for lead is unique. Instead of developing an RfC in the
traditional manner, all identified sources of lead exposure (including background) are used to
predict blood lead (PbB) levels in the exposed individuals. The predicted PbB levels are
compared to a target PbB. PbB levels have long been used as an index of body lead burdens and
as an indicator of potential health effects.
The Integrated Exposure Uptake Biokinetic Model (IEUBK) (U.S. EPA, 1994a) was
developed to predict PbB levels for an individual child or a population of children. The model
was specifically designed to evaluate lead exposure in young children (birth to 7 years of age)
because this age group is known to be highly sensitive to lead exposure.
7-20
-------
Volume II Section 7.0
The IEUBK is a versatile risk assessment tool that allows the user to make rapid
calculations from a complex array of intake, absorption, distribution, and elimination equations.
Default values representing urban background exposures to lead from soil and dust ingestion, air,
food, and water are built into the model.
The IEUBK has four primary components: exposure, uptake, biokinetic, and probability
distribution. The exposure component integrates media-specific (e.g., air, soil, food, water) lead
concentrations and age-dependent media intake rates to calculate age- and media-specific lead
intake rates. Thus, the exposure component determines how much lead enters the child's body
over the exposure period. The uptake component calculates how much of the lead that is
ingested or inhaled is actually absorbed into the blood and the biokinetic component models the
distribution of lead from the blood to other body tissues and/or elimination from the body. The
final component calculates a plausible probability distribution of PbB for a hypothetical child.
The geometric mean PbB is calculated and the probability of exceeding a target PbB is
determined.
For this analysis, the IEUBK model was used to identify air concentrations that would
result in a probability of less than 5 percent of having a PbB level higher than the target PbB and
that could be used in place of an RfC in the calculations. Because the IEUBK model cannot be
run in a backcalculation mode, different air concentrations were modeled until one was found
that satisfied the 95 percent protection level desired.
Only two receptors were modeled for lead: children ages 0 to 3 years and children ages 3
to 7 years. Adults (including workers) and older children were excluded from the analysis for
lead because those age groups are considered less sensitive to lead than 0- to 7-year-olds (and, in
fact, the pharmacokinetic relationships in the IEUBK model are only valid for 0- to 7-year-olds).
The IEUBK model inputs are summarized in Table 7-8. These include inhalation rate,
body weight, media concentrations (including soil, indoor dust, water, and food), and indoor air
concentration as a percentage of outdoor air concentration. The IEUBK model does not support
Monte Carlo analysis, so these inputs must be entered as point estimates. To be as consistent as
possible with the distributions developed for exposure factors, the median body weight and
inhalation rates from the distributions for 0- to 3- and 3- to 10-year-olds were used for 0- to 3-
and 3- to 7-year-olds, respectively. Ingestion rate for water was left at the default age-specific
values recommended in the IEUBK guidance document (U.S. EPA, 1994a).
Media concentrations were left at the standard background levels. Two background soil
concentrations were modeled: 75 mg/kg and 200 mg/kg. These values correspond to the range of
typical soil lead levels in urban soil (U.S. EPA, 1994a). Indoor dust concentrations were set to
equal soil concentrations and indoor air lead concentrations were equal to outdoor concentrations
(the normal default is indoor air = 30 percent of outdoor air). Therefore, it was assumed that
children would be exposed 24 h/d to constant air lead concentrations. This was done for
consistency with other constituents for which exposure was assumed to occur 24 h/d with no
distinction between indoor and outdoor air made.
7-21
-------
Volume II
Section 7.0
Table 7-8. Summary of Inputs for IEUBK Model
Input
Age
(yr)
Value
Source
Exposure Factors
Inhalation rate (m3/d)
Body weight (kg)
0-3
3-7
0-3
3-7
6.1
8.9
12.63
25.9
Median of distribution for 0- to 3-yr-olds
Median of distribution for 3- to 10-yr-olds
Median of distribution for 0- to 3-yr-olds
Median of distribution for 3- to 10-yr-olds
Media Concentrations
Soil (ug/g)
Indoor dust (ug/g)
Water (ug/L)
Food (ug/d)
Indoor Air Concentration
(% of outdoor air cone.)
75 to 200
75 to 200
4
Varies by age
100%
U.S.EPA(1994a)
U.S.EPA(1994a)
U.S.EPA(1994a)
U.S.EPA(1994a)
Assumption/consistency with other constituents
Model outputs were produced for three target blood lead levels (10, 8, and 5 |ig/dL) and
two assumed background soil concentrations (75 and 200 |ig/g); these are shown in Table 7-9.
Depending on the target PbB and the background soil concentration, acceptable air lead
concentrations ranged from background to 3.3 |ig/m3. For comparison, the default background
air lead concentration used by the IEUBK is 0.1 |ig/m3.
For some combinations of target blood lead level, background soil concentration, and age,
the background exposures alone led to a probability greater than 5 percent that the blood lead
target would be exceeded. This would mean that no exposure to lead in air could be allowed.
A single air concentration for each age range was needed for use in the Monte Carlo
analysis as a surrogate for the RfC. The blood lead level of 10 |ig/dL was selected because that
level has been identified as a level of concern by the Centers for Disease Control (CDC) (U.S.
EPA, 1994a). The IEUBK guidance recommends a background soil concentration of 200 |ig/g;
therefore, that concentration was used. This resulted in air concentrations of 0.5 and 1.8 |ig/m3
for the 0- to 3- and 3- to 7-year-old, respectively. These air concentrations were used in the
Monte Carlo analysis described above in lieu of a health benchmark.
7-22
-------
Volume II Section 7.0
Table 7-9. Results of IEUBK Modeling
Target Blood Level
ftig/dL)
10
8
5
10
8
5
Assumed Soil/Dust
Background
Concentration (mg/kg)
200
200
200
75
75
75
Air Concentration3
0- to 3-yr-old
0.5
Ob
Ob
2.4
1.25
Ob
ftig/m3)
3- to 7-yr-old
1.8
0.6
Ob
3.3
2.1
0.35
a Air concentration at which there is a 5% probability of exceeding the target blood lead level.
b Background exposures alone result in >5% probability of blood lead exceeding target level.
7-23
-------
Volume II Section 8.0
8.0 Analysis of Variability and Uncertainty
Variability arises from true heterogeneity in
characteristics such as body weight differences
within a population or differences in
contaminant levels in the environment.
Uncertainty represents lack of knowledge
about factors, such as the nature of adverse
effects from exposure to constituents, which
may be reduced with additional research.
The purpose of this section is to
discuss the methods that are used in this study
to capture variability and uncertainty.
Variability and uncertainty are discussed
separately because they are fundamentally
different. Variability represents true
heterogeneity in characteristics such as body
weight differences within a population or
differences in contaminant levels in the
environment. It accounts for the distribution
of risk within the exposed population.
Uncertainty, on the other hand, represents lack of knowledge about factors such as adverse
effects from contaminant exposure, which may be reduced with additional research to improve
data or models.
This discussion describes the treatment of variability in some parameters used to describe
human receptors and their behavior. Treatment of variability using a Monte Carlo simulation
forms the basis for the risk distributions. Uncertainty necessitated the use of assumptions,
default values, and imputation techniques in this study. These are discussed so that decision
makers can understand the limitations of the analysis. Taken as a whole, this discussion focuses
on how this treatment of variability and uncertainty affects the representativeness and accuracy of
the results.
8.1 Variability
In conducting a national risk assessment, numerous parameters will vary across the
nation. Variability is often used interchangeably with the term "uncertainty," but this is not
strictly correct. Variability is tied to variations in physical, chemical, and biological processes
and cannot be reduced with additional research or information. Although variability may be
known with great certainty (e.g., age distribution of a population may be known and represented
by the mean age and its standard deviation), it cannot be eliminated and needs to be treated
explicitly in the analysis. Spatial and temporal variability in parameter values used to model
exposure and risk account for the distribution of risk in the exposed population.
An example is meteorological parameters used in dispersion modeling, such as wind
speed and wind direction. These parameters are measured hourly by the National Weather
Service at many locations throughout the United States, and statistics about these parameters are
well documented. While the distributions of these parameters may be known, their actual values
-------
Volume II Section 8.0
vary spatially and temporally and cannot be predicted exactly. Thus, the concentration calculated
by a dispersion model for a particular receptor for a particular time period will provide
information on average conditions that may over- or underpredict actual concentrations. Use of
models such as ISCST3 that calculate concentrations hourly and sum these hourly values to
provide annual concentrations estimates account for much of the temporal variation.
Additionally, using meteorological data from multiple monitoring stations located throughout the
U.S. accounts for some but not all spatial variability.
In planning this analysis, it was important to specifically address as much of the
variability as possible, either directly in the Monte Carlo analysis or through disaggregation of
discrete parts of the analysis. For example, use of a refined receptor grid accounts for spatial
variability in concentrations around an WMU. Variability in WMU characteristics is accounted
for using a large database of individual WMUs that represent the range of possible WMU
characteristics. Previous sections describe how distributions and point value estimates were
made for input parameters and how these were combined through design of the analysis to
provide facility-level and national level distributions of Cw. Variability is discussed in this
section to provide an overview of the modeling techniques used to address variability. The risk
assessment components discussed included:
# Source characterization and emissions modeling
# Fate and transport modeling
# Exposure modeling.
8.1.1 Source Characterization and Emissions Modeling
Biodegradation in WMUs affects emissions by reducing the amount of waste constituent
that could be released to the atmosphere. Biodegradation rates are affected by site-specific
parameters. Of the many site specific-values that could affect biodegredation rates, only
temperature was considered in their analysis. Biodegradation was assumed to occur at
temperatures greater than 5°C. In order to turn biodegradation on and off, seasonal temperatures
were calculated for each site based on region-specific meteorological data. The use of seasonal
temperature data was intended to reduce the error in both the biodegradation rate and the
temperature-dependent volatilization rate in land-based WMUs. The landfill scenario assumed
that no biodegradation occurred, so temperature was a less important parameter for this WMU
than for the land application unit or wastepile. In the land application units, temperature was
varied monthly, as were other meteorological parameters. Monthly variation was used to capture
the interaction between meteorology, waste application, and tilling.
A single biodegradation rate was used for each constituent. This biodegradation rate was
applied to all WMUs in all locations. Site-specific conditions may cause the biodegradation rate
to be greater or less than the default value used, but this variability was not captured by this
analysis.
The effect of incorporating seasonal temperature variations in the analysis was chemical-
dependent. For most chemicals in the analysis, the difference between using seasonal
temperatures or an annual average temperature resulted in little or no change in the overall
annual emissions. For a few chemicals, however, there was a very large impact for certain
-------
Volume II Section 8.0
locations in the United States. These chemicals had rapid biodegradation rates, and turning the
biodegradation off for the winter months resulted in significantly higher annual emissions.
8.1.2 Fate and Transport Modeling
To capture geographic variation, dispersion modeling was conducted using
meteorological datasets from 29 different meteorological stations around the continental United
States. This provided regional representation of the variability in meteorological data. For
landfills and land treatment units, these datasets were combined with 14 surface areas
representing the distribution of WMU size. Combined, this provides 406 different sets of unit air
concentrations (UACs) to use with emissions data to estimate air concentrations. As discussed
below, a set of UACs consists of all UACs at 96 combinations of distances and directions at one
meteorological location. Similarly, wastepiles used 812 different sets of UACs based on
29 surface area/wastepile height combinations and 29 meteorological stations. Tanks used the
same 29 meteorological stations and 33 surface area/tank height combinations, providing 957
sets of UACs.
The location of receptors was an important source of variability addressed in the exposure
modeling. The data used to identify and characterize WMUs contained no information on the
location and types of receptors relative to the facility. Many previous risk analyses have used the
maximum point of exposure at some prespecified distance from the WMU as the point for
analysis. Such an approach is usually criticized as being overly conservative because it does not
consider the possibility of no one's living at that exact point. Because individuals may
potentially be located in any direction and at various distances from a facility, this analysis
developed an explicit way to incorporate this consideration. First, a sensitivity analysis was
conducted to determine a reasonable distance at which to bound the analysis. This sensitivity
analysis showed that, beyond 1,000 meters, most air concentrations are a small percentage (less
than 10 percent) of the concentration at the point of maximum exposure. Therefore, 1,000 meters
was used as the outer bound on the distance of receptors included in this analysis. Next, a
receptor grid was set up to allow individuals to reside in any of 16 directions and at distances of
25, 50, 75, 150, 500, and 1,000 meters from the edge of the unit. The Monte Carlo analysis uses
a uniform distribution that gives equal probability to a receptor's being located in any of 16
directions at each distance (96 combinations).
Obviously, 29 meteorological stations do not represent every site-specific condition that
could exist in the continental United States. Based on EQM (1993), however, it is believed that
these stations provide a reasonable representation of the variability in meteorological conditions
for the U.S. climate regions. Quantifying the error associated with the use of this limited number
of meteorological datasets would require comparison of results with more site-specific data,
which has not been done. A detailed analysis of the wind rose for the 29 meteorological stations
is presented in Appendix C.
8.1.3 Exposure Modeling
The exposure factors used in the analysis, including inhalation rate, body weight, and
duration of exposure, are quite variable. To include this variability explicitly in the analysis,
distributions for these variables were used for each receptor in the analysis: workers, adult
-------
Volume II Section 8.0
residents, and child residents. For adults, male and female exposure factors were developed
separately. For child exposures, three age groups (0 to 3, 4 to 10, and 11 to 18), representing age
at the start of exposure, were considered. Body weight was developed for each age and
inhalation rate for each child cohort group based on the data available. Thus, for each Monte
Carlo realization for each receptor type, values for each of these variables were selected from the
specified distribution.
At each WMU modeled, a distribution of Cw was developed that incorporated the
variability in the location of the receptor and the variability in the exposure factors used to
characterize the receptor. Note that the Monte Carlo analysis for noncarcinogenic effects
incorporates variability only about receptor location since exposure factors are not used in the
calculation.
The overall output of the analysis includes specific consideration of the variability in site-
specific WMU information, regional-specific meteorological conditions, location of receptors,
and exposure factors for each receptor to provide national distributions of a risk-specific waste
concentration across all facilities of a specified type. These distributions are presented in
Volume III.
8.2 Uncertainty
Uncertainty is a description of the imperfection in knowledge of the true value of a
particular parameter. In contrast to variability, uncertainty is reducible by additional information
gathering or analysis activities (better data, better models). EPA typically classifies the major
areas of uncertainty in risk assessments as scenario uncertainty, model uncertainty, and parameter
uncertainty. Scenario uncertainty refers to missing or incomplete information needed to fully
define exposure and dose. Model uncertainty is a measure of how well the model simulates
reality. Finally, parameter uncertainly is the lack of knowledge regarding the true value of a
parameter that is used in the analysis.
While some aspects of uncertainty were directly addressed in the analysis, much of the
uncertainty associated with this analysis could only be addressed qualitatively. Significant
sources of uncertainty are presented in this section and are discussed with regard to how
uncertainty was addressed. If the analysis directly addressed uncertainty, the approach used is
described. If the analysis did not directly address uncertainty, a qualitative discussion of its
importance is provided.
8.2.1 Scenario Uncertainty
Sources of scenario uncertainly include the assumptions and modeling decisions that are
made to represent an exposure scenario. Because we lack information or resources to define and
model actual exposure conditions, uncertainty is introduced into the analysis. Despite the
complexity of this analysis, it was necessary to exclude or simplify actual exposure conditions.
For example, this analysis only addresses inhalation exposures; indirect exposure pathways were
excluded. Professional judgement, often coupled with using the results of sensitivity analysis, is
used to decide which parameters to include in describing exposure conditions and behaviors.
8-4
-------
Volume II Section 8.0
These judgements are imperfect and uncertainty is introduced. Some of the uncertainties that are
important to understand in interpreting the results of this study are discussed below.
Emissions from WMUs that were modeled for this study included emissions from the
WMU itself (e.g., the surface of an open tank). Emissions from ancillary operations associated
with the WMU were not considered. For land-based WMUs, one source of particulate emissions
that was not included is vehicular traffic associated with trucks and other facility operations.
These vehicles entrain particulate material to the air from utility roads at WMUs, which was not
included in the model used in this analysis.
This risk analysis addresses only inhalation exposure to humans, a direct exposure
pathway. Indirect exposures include exposure pathways such as contaminated produce or meats
due to the uptake and bioaccumulation of contaminants in the food chain. Although this analysis
focused only on releases of the contaminants to air, it is possible that these contaminants could
accumulate in plants through air-to-plant transfer mechanisms or by direct deposition onto plants.
Animals could then come into contact with contaminated plants, or even contaminated soil, and
accumulate the contaminant in their bodies (bioaccumulation). Because of the short time frame
of this project and the complexity associated with modeling indirect risks, the consent decree
with EDF did not require the Agency to evaluate this pathway.
EPA conducted a preliminary analysis to determine if any constituents included in this
analysis have the potential to be taken up by plants and accumulated to the point of presenting a
potential concern for humans or domestic animals, such as beef or dairy cows that might graze in
fields adjacent to a WMU. EPA looked at a waste stream in which poly cyclic aromatic
hydrocarbons (PAHs) were the constituents of concern from a risk perspective. Releases from an
land application units (LAUs) were modeled with receptors located randomly between 75 and
1,000 m. The concentration of constituents in contaminated produce could be attributed to three
mechanisms: vapor transfers, direct deposition from air, and uptake from soil, with vapor
transfers being the predominant mechanism of contamination. The risk from ingestion of
contaminated produce as a percent of total risk from all ingestion pathways is quite significant
for the farmer scenario and less significant for the home gardener.
Two important points need to be made regarding this modeling exercise. First, the vapor
transfer coefficients used for PAHs are very high and are thought to have a high level of
uncertainty. Studies are currently under way to address this uncertainty, but the results of such
studies are not available at this time. One criticism is that these vapor transfer coefficients
greatly overestimate the concentration of PAHs in plants and thus lead to overestimates of risk.
Second, the farmer scenario has much higher risk from homegrown produce than the home
gardener due to differences in ingestion rates for these two groups. At this time, it is not known
which scenario would be the most appropriate to use for the air characteristic analysis, nor is
information available about the prevalence of farmers or home gardeners in the vicinity of these
types of facilities.
Although this sensitivity analysis suggests that indirect pathways may be important to
consider in the air characteristic risk analysis, this analysis is very limited in terms of
constituents, WMUs, settings, and scenarios. In a sense, this sensitivity analysis should be
8-5
-------
Volume II Section 8.0
considered a bounding analysis because it is based on a set of inputs that are biased toward
determining if the air-to-plant pathway could be a concern for any of the constituents.
8.2.2 Model Uncertainty
To reduce model uncertainty, EPA generally selected models that are considered state-of-
the-art. Model uncertainty is associated with all models used in all phases of a risk assessment.
These include the animal models used as surrogates for testing human carcinogenicity, dose-
response models used in extrapolations, and computer models used to predict the fate and
transport of chemicals in the environment. Computer models are simplifications of reality,
requiring exclusion of some variables that influence predictions but cannot be included in models
due either to increased complexity or to a lack of data on a particular parameter. The risk
assessor needs to consider the importance of excluded variables on a case-by-case basis because
a given variable may be important in some instances and not in others. A similar problem can
occur when a model that is applicable under average conditions is used when conditions differ
from the average. In addition, choosing the correct model form is often difficult when
conflicting theories seem to explain a phenomenon equally well. Modeling uncertainty is not
addressed directly in this study but is discussed qualitatively in this section.
With regard to model error, the CHEMDAT8 model was used for all volatile emissions
estimates including acute and subchronic releases and as the basis for estimating soil
concentrations that are used in the paniculate emissions model. As discussed in Section 4.1,
there are many features of this model that meet the needs of this analysis. The model, however,
was developed to address only volatile emissions from the waste management units considered in
this study. Competing mechanisms such as runoff and erosion and leaching are not included in
the model. Insomuch as these competing processes actually occur, the model would tend to
slightly overestimate the volatile emissions and waste/soil concentrations in the WMU. On the
other hand, one could interpret this situation as being representative of WMUs that have leachate
controls, such as liners, or erosion and runoff controls. Such controls would tend to inhibit these
processes and result in more volatile emissions. Similar lines of argument hold for the calculation
of a waste/soil concentration that is the basis for estimating particulate emissions of a
contaminant.
The ISCST3 model was used to calculate the dispersion of particle and gas emissions
from a WMU. This model has many capabilities needed for this assessment, such as the ability
to model area, volume, or point sources for chronic, subchronic, and acute averages. For
dispersion modeling of this type, it is considered to be a fairly accurate model, with error within
about a factor of 2. It does not include photochemical reactions or degradation of a chemical in
the air, which results in additional model uncertainty for some chemicals. In addition, this
analysis did not use wet depletion but did use dry plume depletion for particulates. As presented
in Section 5 and discussed in the sensitivity analysis in Appendix C, wet deposition and
associated depletion accounts for only about 2 percent of total depletion. Wet depletion,
therefore, was not used in this analysis because of the time required to properly process the
requisite precipitation data required for the wet depletion option. Dry deposition and associated
plume depletion are important for particulates and are explicitly incorporated into this analysis.
8-6
-------
Volume II Section 8.0
Dispersion modeling is highly sensitive to meteorological data and the surface area and
dimensions of the WMU. Meteorological data used in the dispersion modeling include
windspeed and direction, temperature, precipitation type and amount, and stability class. The
ISCST3 model uses hourly data as inputs, and this analysis used 5 years of hourly data to develop
chronic (i.e., 1-year), subchronic (i.e., 30-day) and acute (i.e., 1-day) UACs, as discussed in
Section 5.
8.2.3 Parameter Uncertainty
Parameter uncertainty occurs when (1) there is a lack of data about the parameters used in
the equations, (2) the data that are available are not representative of the particular instance being
modeled, or (3) parameter values cannot be measured precisely and/or accurately either because
of equipment limitations or because the quantity being measured varies spatially or temporally.
Random, or sample errors, are a common source of parameter uncertainty that is especially
critical for small sample sizes. More difficult to recognize are nonrandom or systematic errors
that result from bias in sampling, experimental design, or choice of assumptions.
Source characterization can introduce parameter uncertainty into the analysis through
survey techniques used to characterize waste management units and associated parameter values.
As discussed in Section 3.1, existing databases were used to identify WMUs and as a basis for
determining important emissions and dispersion model input parameter values. The Industrial
Subtitle D Survey (Schroder et al., 1987) was used to characterize landfills, LAUs, and
wastepiles. Because the Industrial D database did not survey tanks, characterization of tanks was
based on the 1986 National Survey of Hazardous Waste Treatment, Storage, Disposal, and
Recycling Facilities (U.S. EPA, 1987).
These databases were used to determine physical and operating characteristics for the
WMUs modeled. The impact of the uncertainty associated with the information contained in
these databases is unknown. There are several sources of this uncertainty, including age of the
data, representativeness, missing data on waste volumes or capacity, multiple WMUs of the same
type associated with a combined surface area and waste volume, accuracy of the reported data
(i.e., measurement error), and limited information on WMU operating characteristics. Because
these surveys were completed in 1987, uncertainty exists concerning changes in waste
management practices since 1987. This is especially true for the tanks data, which may result in
an underestimation of highly aerated biological treatment tanks. Underestimation of the number
of highly aerated treatment tanks would result in lower emissions estimates and higher Cws.
The data were reviewed to assess the representativeness of the survey data for the WMUs
evaluated in this analysis. For example, the tank database contained many small "tanks," or
containers that were not representative of the tanks that are the target of this assessment. Tanks
less than or equal to 55 gallons were deleted from the database and not included in this study.
Similarly, closed tanks were deleted from the database because this analysis was concerned only
with open tanks.
Any survey results will introduce uncertainty due to inaccuracy in the way the data were
collected and compiled. Inaccuracy can result from respondents' misinterpreting the survey
questions. This uncertainty can be compounded by the imputation techniques used to develop
-------
Volume II Section 8.0
parameter values for parameters not directly surveyed. The imputation techniques used in this
study that increase the uncertainty include:
# The tank survey provides data on tank volume and flow rate. Tank depth and
surface area have to be imputed based on surveyed volume data using
supplemental data and engineering judgment about the relationship between tank
volume, surface area, and depth.
# When more than one WMU of the same type existed at a facility, the combined
area of those WMUs was reported in the Industrial D database. This analysis used
the average area of WMUs of a single type when more than one unit existed at the
same facility. This is not a problematic assumption except for a few facilities
with large total surface areas. In such cases, it is unknown whether one large
WMU dominates this surface area or the area is evenly distributed across the
WMUs.
# The Industrial D survey does not report depth for landfills. Therefore, depth had
to be imputed from the total capacity and total area data that were reported. To do
this, an assumption about the bulk density of waste deposited in the landfill was
made. To the extent that actual bulk density of waste differs from the assumed
bulk density, depth would be over- or underestimated.
# Like landfill depth, waste pile height was not reported in the Industrial D survey.
Hence, waste pile height was imputed from total capacity and total area data that
were reported. To do this, an assumption was made about the bulk density of
waste deposited in the waste pile. To the extent that actual bulk density of waste
differs from the assumed bulk density, waste pile height would be over- or
underestimated.
Source characterization also requires making assumptions about the way WMUs are
operated. For example, LAU surface area is reported in the Industrial Subtitle D Survey but
details about operating practices are not. Operators of LAUs may spread waste over the entire
surface area, or they may choose to spread waste in cells, rotating from cell to cell until the entire
surface area is used and then repeating the process. Tilling practices for LAUs will also vary
from operator to operator. Therefore, assumptions made for LAUs about active surface area and
tilling frequency and depth introduce uncertainty into the parameter values used to model these
units. For landfills, no information was reported on how much of the landfill total area was
operational in a given year. Therefore, it was assumed that the landfill operated for 20 years,
filling one cell per year, so that cell area was one-twentieth of the total area. In general, however,
because this analysis captured wide variations in WMUs, the effect of uncertainties in these
parameters should be mitigated.
There was a lack of site-specific values such as waste characteristics that influence
emissions calculation. For particular emissions default values were used to characterize all waste
streams. Use of default or average values instead of value more specific to the type of waste
managed introduces uncertainty into the analysis. In addition, data were not available for the
actual molecular weight of the waste, which is a required input to the CHEMDAT8 emissions
-------
Volume II Section 8.0
model that affects organic-phase emissions (but not aqueous-phase emissions). A default value
of 147 g/mol was used. If the actual molecular weight of waste is greater than 147 g/mol,
organic phase emissions would be underestimated. Conversely, if the actual molecular weight of
waste is less than 147 g/mol, organic-phase emissions would be overestimated.
Biodegradation rates were based on data contained in Howard et al. (1991). Some of
these rates have been determined for aqueous systems and adjusted for use in soil systems.
While they are often used in this manner, there is some degree of uncertainty associated with
making the adjustment for use in soil systems.
To minimize error associated with the use of discrete heights and surface areas in the
dispersion modeling, an interpolation routine was used to estimate the UAC. To determine the
UAC value used in the risk model, a two-step process was used. First, the UAC values for a
height closest to the actual WMU height (for wastepiles and tanks) were used. Using the closest
(and not actual) height introduces uncertainty that can over- or underpredict actual
concentrations. Then a UAC value was interpolated between the surface areas that bracketed the
actual WMU surface area (all WMUs). Areas were chosen to maintain an error of less than 10
percent for tanks. For all other WMUs, areas were chosen to best cover the distribution of areas;
although minimizing interpolation error was not an explicit factor, that error is generally less than
25 percent.
Other uncertainties introduced into the analysis in dispersion modeling are related to
WMU shape (for land-based WMUs). A sensitivity analysis was conducted (see Appendix C) to
determine the sensitivity of different shapes on ambient ground-level air concentrations
calculated by ISCST3. A square shape was selected because it minimized the error introduced by
not knowing the orientation of the WMU shape to wind direction. The number of receptors
selected was also based on sensitivity analyses (see Appendix C). A sparse grid would increase
the likelihood of missing peak concentrations, whereas a dense receptor grid would be too
resource-intensive for this analysis.
Cancer Slope Factors. Cancer slope factors (CSFs) were derived as the 95 percent lower
confidence limit of the slope of the dose-response curve using a linear, no-threshold
dose-response model. The cancer slope factor is, therefore, an upper-bound estimate of the
cancer risk per unit dose and, for this reason, may overstate the magnitude of the risk. In
addition, the use of CSFs in projecting excess individual cancer risk introduces uncertainty
stemming from a number of factors including:
# Limited understanding of cancer biology
# Variability in the response of animal models
# Differential response in animal models versus humans
# Difference between animal dosing protocols and human exposure patterns.
A key step in CSF development is high- to low-dose extrapolation. Depending on the
model used to fit the data, extrapolations to the low dose range can vary by several orders of
magnitude, reflecting the potential uncertainty associated with the cancer slope factor.
8-9
-------
Volume II Section 8.0
Reference Doses and Reference Concentrations. Uncertainty in the toxicological and
epidemiological data from which reference doses (RfDs) and reference concentrations (RfCs) are
derived are accounted for by applying uncertainty factors. An RfD (or RfC) is "an estimate (with
uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human population
(including sensitive subgroups) that is likely to be without an appreciable risk of deleterious
effects during a lifetime" (U.S. EPA, 1999). RfDs and RfCs are based on the no adverse effects
level (NOAEL) or lowest observed adverse effects level (LOAEL) for the most sensitive effect in
the most sensitive or most relevant species. A series of standard uncertainty factors are applied to
the NOAEL or LOAEL to derive the RfD or RfC. The following uncertainty factors account for
areas of scientific uncertainty:
# Intraspecies variation; accounts for variation in sensitivity among humans
(including sensitive individuals such as children, the elderly, or asthmatics)
# Interspecies variation; accounts for extrapolating from animals to humans
# LOAEL to NOAEL extrapolation
# Subchronic to chronic; accounts for extrapolating from a subchronic NOAEL or
LOAEL to a chronic NOAEL or LOAEL
# Incomplete database; accounts for the lack of data for critical endpoints (e.g.,
reproductive and developmental).
Uncertainty factors of 1, 3, or 10 are used. The default value is 10; however, an
uncertainty factor of 3 may be used if appropriate pharmacokinetic data (or models) are available.
In addition, a modifying factor may be applied to account for additional uncertainties in
accordance with professional judgment. The default value for the modifying factor is 1. All
uncertainty factors (UFs) and the modifying factor (MF) are multiplied together to derive the
total uncertainty factor, with 3,000 being the maximum recommended value (U.S. EPA, 1994b).
Therefore, the RfD (or RfC) is derived by using the following formula:
RfD = NOAEL/(UF x MF).
The effect of applying uncertainty and modifying factors is to lower the estimate of the
reference dose and increase the hazard quotient (HQ) for a given exposure.
While most of the uncertainty associated with health benchmarks is common to all risk
assessment studies, some uncertainties are specific to this analysis. Data to support the
development of subchronic and acute effects were limited. Hence, health benchmark values were
drawn from a variety of sources based on a quality hierarchy. This resulted in some chronic
health benchmark values' being higher than the corresponding subchronic or acute value. This
counterintuitive result comes about because different studies were used that were from different
years, used different methods, and were based on different studies. This result occurred for about
14 chemicals in this study.
8-10
-------
Volume II Section 9.0
9.0 References
Agency for Toxic Substances and Disease Registry (ATSDR). 1999. Minimal Risk Levels
(MRLs) for Hazardous Substances. http://atsdrl.atsdr.cdc.gov:8080/mrls.html
American Conference of Governmental Industrial Hygienists (ACGIH). 1997. Threshold Limit
Values for Chemical Substances and Physical Agents Biological Exposure Indices.
Bailey, Robert G., Peter E. Avers, Thomas King, W. Henry McNab, eds. 1994. Ecoregions and
subregions of the United States (map). Washington D.C.; U.S. Geological Survey. Scale
1:7,500,000; colored. Accompanied by a supplementary table of map unit descriptions
compiled and edited by McNab, W. Henry, and Bailey, Robert G. Prepared for the U.S.
Department of Agriculture, Forest Service, http://www.epa.gov/docs/grdwebpg/bailey/
Bucher J.R., M.R. Elwell, M.B. Thompson, et al. 1990. Inhalation toxicity studies of cobalt
sulfate in F344/N rats and B6C3F1 mice. FundamAppl Toxicol 15:357-372.
CalEPA (California Environmental Protection Agency). 1997a. Technical Support Document
for the Determination ofNoncancer Chronic Reference Exposure Levels. Draft for
Public Review. Office of Environmental Health Hazard Assessment, Air Toxicology and
Epidemiology Section, Berkeley, CA.
CalEPA (California Environmental Protection Agency). 1997b. Air Toxics Hot Spots Program
Risk Assessment Guidelines: Technical Support Document for Determining Cancer
Potency Factors. Draft for Public Comment. Office of Environmental Health Hazard
Assessment, Berkeley, CA.
CalEPA (California Environmental Protection Agency). 1998. Technical Support Document for
the Determination of Acute Reference Exposure Levels for Airborne Toxicants, Scientific
Review Panel Draft. Office of Environmental Health Hazard Assessment, Air
Toxicology and Epidemiology Section, Berkeley, CA.
http ://www. oehha. org/scientific/getdr els. html
Coburn, J., C. Allen, D. Green, and K. Leese. 1988. Site Visits of Aerated and Nonaerated
Impoundments. Summary Report. U.S. EPA, Contract No. 68-03-3253, Work
Assignment 3-8. April 1988. Research Triangle Institute, Research Triangle Park, NC.
Eichinger, J. 1985. Memorandum from J. Eichinger, GCA Corporation, to K.C. Hustvedt,
EPA/CPB: TSDF model source parameters and operating practices database,
September 6, 1985.
-------
Volume II Section 9.0
EQM (Environmental Quality Management, Inc.,) and E.H. Pechan & Associates. 1993.
Evaluation of Dispersion Equations in Risk Assessment Guidance for Super fund (RAGS):
Volume I - Human Health Evaluation Manual. Prepared for U. S. Environmental
Protection Agency, Office of Emergency and Remedial Response, Toxics Integration
Branch, Washington, DC.
Ferroni C., L. Sellis, A. Mutti, et al. 1992. Neurobehavioral and neuroendocrine effects of
occupational exposure to perchloroethylene. Neurotoxicology 12: 243-247.
Howard, P.H., R.S. Boethling, W.F. Jarvis, W.M. Meylan, and E.M. Michalenko. 1991.
Handbook of Environmental Degradation Rates. Lewis Publishers, Chelsea, MI.
Martin, J. P., R. C. Sims, and J. Matthews. 1986. Review and Evaluation of Current Design and
Management Practices for Land Treatment Units Receiving Petroleum Wastes.
EPA/600/J-86/264. U.S. Environmental Protection Agency, Robert S. Ken-
Environmental Research Laboratory, Ada, OK.
Myers, L., J. Lashley, andR. Whitmore. 1997. Development of Statistical Distributions for
Exposure Factors. Draft Report. Research Triangle Institute, Research Triangle Park,
NC, for U.S. EPA Office of Research and Development. EPA Contract 68D40091.
October.
Myers, L. (RTI) 1998. Personal communication with A. Lutes, RTI. March 16.
Myers, L., J. Lashley, andR. Whitmore. 1998. Development of Statistical Distributions for
Exposure Factors. Final Report. Research Triangle Institute, Research Triangle Park,
NC, for U.S. EPA Office of Research and Development. EPA Contract 68D40091.
March.
NAS (National Academy of Sciences). 1993. Guidelines for Developing Community Emergency
Exposure Levels for Hazardous Substances. National Academy Press, Washington, DC.
NIOSH (National Institute for Occupational Safety and Health). 1997. NIOSHPocket Guide to
Chemical Hazards. U.S. Department of Health and Human Services, Cincinnati, OH.
NOAA (National Oceanic and Atmospheric Administration). 1992. International Station
Meteorological Climate Summary, Version 2.0. CD-ROM. National Climatic Data
Center. Asheville, NC. June.
NTP (National Toxicology Program). 1996. NTP Technical Report on the toxicology and
carcinogenesis studies of cobalt sulfate heptahydrate (CAS No. 10026-24-1) in F344/N
rats and B6C3Fj mice (inhalation studies). Draft. U.S. Department of Health and Human
Services, Public Health Service, National Institutes of Health. NIH Publication no. 96-
3961. NTPTR-471.
9-2
-------
Volume II Section 9.0
NTP (National Toxicology Program). 1991. NTP report on the toxicity studies of cobalt sulfate
heptahydrate in F344/N rats and B6C3FJ mice (inhalation studies). Research Triangle
Park, NC: U.S. Department of Health and Human Services, Public Health Service,
National Institutes of Health. NIH Publication no. 91-3124. NTP TOX-5.
OSHA (Occupational Safety and Health Administration). 1997. Occupational Safety and Health
Standards: SubpartZ - Toxic and Hazardous Substances. 29 CFR 1910. 7/1/97 edition
of Code of Federal Regulations.
Reed, S.C., andR. W. Crites. 1984. Handbook ofLand Treatment Systems for Industrial and
Municipal Wastes. Noyes Publications, Park Ridge, New Jersey.
Rogers, J.M., M.L. Mole, N. Chernoff, et al. 1993. The developmental toxicity of inhaled
methanol in the CD-I mouse, with quantitative dose-response modeling for estimation of
benchmark doses. Teratology 47(3): 175-188.
RTI (Research Triangle Institute). 1995. Technical Support Document for Hazardous Waste
Identification Rule (HWIR): Risk Assessment for Human and Ecological Receptors.
Prepared for U.S. Environmental Protection Agency, Office of Solid Waste, Washington,
DC.
Shroeder, K., R. Clickner, and E. Miller. 1987. Screening Survey of Industrial Subtitle D
Establishments. Draft Final Report. Westat, Inc., Rockville, MD., for U.S. EPA Office of
Solid Waste. EPA Contract 68-01-7359. December.
Smith, Kendall, Liquid Sales Manager of AO Smith Industrial. 1999. Personal communication,
March 16, 1999.
Stewart R.D., C.L. Hake, A. Wu, et al. 1975. Acetone: Development of a Biological Standard
for the Industrial Worker by Breath Analysis. Cincinnati, OH: National Institute for
Occupational Safety and Health. NTIS PB82-172917.
Torkelson, T.R.,B.KJ. Leong, RJ. Kociba, etal. 1974. 1,4-Dioxane. II. Results of a 2-year
inhalation study in rats. ToxicolApplPharmacol 30:287-298.
Uchida Y., H. Nakatsuka, H. Ukai, et al. 1993. Symptoms and signs in workers exposed
predominantly to xylenes. IntArch Occup Environ Health 64:597-605.
U.S. EPA (Environmental Protection Agency). 1985a. Compilation of Air Pollutant Emission
Factors, Volume I: Stationary Point and Area Sources. Office of Air Quality Planning
and Standards, Research Triangle Park, NC. AP-42.
U.S. EPA (Environmental Protection Agency). 1985b. Rapid Assessment of Exposures to
Particulate Emissions from Surface Contamination Sites. Office of Health and
Environmental Assessment, Washington, DC.
9-3
-------
Volume II Section 9.0
U.S. EPA (Environmental Protection Agency). 1986. Health and Environmental Effects Profile
for Pyridine. Environmental Criteria and Assessment Office, Office of Research and
Development, Cincinnati, OH. EPA/600/X-86-168.
U.S. EPA (Environmental Protection Agency). 1987. 1986 National Survey of Hazardous
Waste Treatment, Storage, Disposal, and Recycling Facilities (TSDR) Database.
U.S. EPA (Environmental Protection Agency). 1988. Control of Open Fugitive Dust Sources.
Office of Air Quality Planning and Standards, Research Triangle Park, NC. EPA-450/3-
88-008.
U.S. EPA (Environmental Protection Agency). 1989. Hazardous Waste TSDF (Treatment,
Storage, and Disposal Facilities): Fugitive Particulate Matter Air Emissions Guidance
Document. EPA/450/3-89/019. Office of Air Quality Planning and Standards, Research
Triangle Park, NC.
U.S. EPA (Environmental Protection Agency). 1990. Methodology for Assessing Health Risks
Associated with Indirect Exposure to Combustion Emissions (COMPDEP). EPA/600/6-
90/003. Office of Health and Environmental Assessment, Washington, D.C.
U.S. EPA (Environmental Protection Agency). 1991. Hazardous Waste TSDF - Background
Information for Proposed Air Emissions Standards. Appendix C. EPA-450/3-89-023a.
Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina.
p. C-19 through C-30.
U.S. EPA (Environmental Protection Agency). 1992. User's Guide for the Fugitive Dust Model
(FDM) (revised) Volume I: User's Instructions. EPA-910/9-88-202R. Region 10,
Seattle, Washington.
U.S. EPA (Environmental Protection Agency). 1994a. Guidance Manual for the Integrated
Exposure Uptake Biokinetic Model for Lead in Children. Office of Emergency and
Remedial Response, Washington, DC. EPA/540/R-93/081 NTIS PB93-963510
U.S. EPA (Environmental Protection Agency). 1994b. Air Emissions Models for Waste and
Wastewater. EPA-453/R-94-080A. Appendix C. Office of Air Quality Planning and
Standards, Research Triangle Park, North Carolina.
U.S. EPA (Environmental Protection Agency). 1994c. Toxic Modeling System Short-Term
(TOXST) User's Guide: Volume I. EPA-453/R-94-058A. Office of Air Quality Planning
and Standards, Research Triangle Park, North Carolina.
U.S. EPA (Environmental Protection Agency). 1994d. Methods for Derivation of Inhalation
Reference Concentrations and Application of Inhalation Dosimetry. EPA/600/8-90-
066F. Environmental Criteria and Assessment Office, Office of Health and
Environmental Assessment, Office of Research and Development, Research Triangle
Park, NC.
9-4
-------
Volume II Section 9.0
U.S. EPA (Environmental Protection Agency). 1994e. CHEMDAT8 User's Guide. EPA-453/C-
94-080B. Office of Air Quality Planning and Standards, Research Triangle Park, NC.
U.S. EPA (Environmental Protection Agency). 1994f. Risk Assessment Issue Paper: Status of
Inhalation Cancer Unit Risk for Benzo(a)pyrene. National Center for Environmental
Assessment. Superfund Technical Support Center, Cincinnati, OH.
U.S. EPA (Environmental Protection Agency). 1995a. User's Guide for the Industrial Source
Complex (ISC3) Dispersion Models. EPA-454/B-95-003a. Office of Air Quality
Planning and Standards, Research Triangle Park, North Carolina.
U.S. EPA (Environmental Protection Agency). 1995b. Identification and listing of hazardous
wastes: hazardous waste identification rule (HWTR); Proposed rule. Federal Register
60 FR 66344-66469. December 21, 1995.
U.S. EPA (Environmental Protection Agency). 1996a. EPACMTP Background Document.
Office of Solid Waste. Washington, DC.
U.S. EPA (Environmental Protection Agency). 1996b. Risk Assessment Issue Paper for:
Derivation of a Chronic RfCfor 1,1,1-Trichloroethane (CASRN 71-55-6). 96-007d/8-
09-96. National Center for Environmental Assessment. Superfund Technical Support
Center, Cincinnati, OH.
U.S. EPA (Environmental Protection Agency). 1997a. Health Effects Assessment Summary
Tables (HEAST). EPA-540-R-97-036. FY 1997 Update. Office of Solid Waste and
Emergency Response, Washington, DC.
U.S. EPA (Environmental Protection Agency). 1997b. National Advisory Committee for Acute
Exposure Guideline Levels for Hazardous Substances; Notices. Federal Register 62 FR
58839-58851. October 30.
U.S. EPA (Environmental Protection Agency). 1997c. Exposure Factors Handbook, Volume I,
General Factors. EPA/600/P-95/002Fa. Office of Research and Development,
Washington, DC.
U.S. EPA (Environmental Protection Agency). 1997d. Exposure Factors Handbook, Volume III,
Activity Factors. EPA/600/P-95/002Fc. Office of Research and Development,
Washington, DC.
U.S. EPA (Environmental Protection Agency). 1998a. Air Characteristic Study, Volume I
Overview and Volume II Technical Background Document. EPA530-R-98-009a /
EPA530-R-98-009b. Office of Solid Waste, Washington, DC.
U.S. EPA (Environmental Protection Agency). 1998b. Hazardous waste management system;
identification and listing of hazardous waste; solvents; final rule. Federal Register 63 FR
64371-402.
9-5
-------
Volume II Section 9.0
U.S. EPA (Environmental Protection Agency). 1999. Integrated Risk Information System (IRIS)
- online. Duluth, MN. http://www.epa.gov/iris/
U.S. EPA (Environmental Protection Agency), (no date available). Risk Assessment Issue Paper
for: Carcinogenicity Information for Trichloroethylene (TCE)(CASRN 79-01-6).
National Center for Environmental Assessment, Superfund Technical Support Center,
Cincinnati, OH.
Uzhdavini, E.R., K. Astaf yeva, A. A. Mamayeva, and G.Z. Bakhtizina. 1972. [Inhalation
toxicity of o-cresol]. Trudy Ufimskogo Nauchno-Isseldovatel'skogo Institute Gigiyeny
Profzabolevaniya, 7:115-9. (Russian) [as cited in CalEPA, 1997, and U.S. EPA 1985,
1986]
Watkins, S.L. 1990. Background Document for the Surface Impoundment Modeling System,
Version 2.0 Documentation. Radian Corporation, Research Triangle Park, NC.
Zenz, C., B.A. Berg. 1967. Human responses to controlled vanadium pentoxide exposure. Arch
Environ Health 14:709-712.
9-6
-------
Appendix A
Basic Dalenius-Hodges Procedure for
Constructing Strata
-------
Volume II Appendix A
A. Basic Dalenius-Hodges Procedure for Constructing Strata
(applies to census data)
1. Select an available variable X correlated with the variable Y of interest. (Both X and Y
are assumed to be continuous.)
2. Form a relative frequency histogram of X with respect to M prespecified intervals with
breakpoints a0, al3 a2 ...,aM. Let f; denote the frequency count for the ith interval. (Note that
M should be fairly large, e.g., at least 50 and preferably 100, since the ultimate strata will
be formed as the union of a subset of contiguous intervals. The intervals do not have to
be the same length but it may be convenient to use one or two standard lengths.)
3. For each interval i (i= 1,2,...M), compute Z. = JLfi where Li = ai~ai_l is the length
of the ith interval.
i
4. Compute CUMZt = Y,Zj •
7 = 1
5. Determine stratum breakpoints by breaking CUMZ into equal intervals of length Q =
CUMZM/H, where H is the desired number of strata (e.g., if H=5, then stratum boundaries
would be formed at Q, 2Q, 3Q, and 4Q).
6. Sequentially cluster the intervals into strata in order to redefine breakpoints in terms of X
variable.
7. Select random samples of size n within each stratum.
In practice, H is normally about 4, 5, or 6. This procedure produces a sample that is near optimal
for estimating the population mean of Y if X and Y are highly correlated.
Modification of the Basic Procedure for the Current Context
At Step 2, determine f; as the sum of the weights of the units falling into the ith interval.1
At Step 7, select a unit (or facility, depending on whether X and associated weights relate to units
or facilities) near the midpoint of each stratum as the "representative" unit. Its percentile in the
population of units would be estimated as (sum of weights for all units having Xs <= X of given
unit)/(sum of all weights).
Note that the current weights relate to the population of facilities; to get weights appropriate to units (which one should do
if x is defined as a unit-specific measure rather than a facility-specific measure), one should multiply the number of units at a
facility by the facility's weight. As a check, the sum of the current weights should produce an estimate of the number of facilities
in the population and the sum of the (weight x number of units) should produce an estimate of the number of units in the
population.
-------
Volume II Appendix A
The above procedure was applied to distributions for landfills (LFs), land application
units (LAUs), surface impoundments (Sis), and wastepiles (WPs). It was also applied to the
combined LF/LAU/SI distribution. In each case, the procedure yields a good characterization of
the right tail of the distribution, but a poor characterization of the lower portion of the
distribution. As a result, the procedure was modified by collapsing two or more of the highest
strata into one and successively splitting the lowest stratum into two or more strata. The splitting
was performed by choosing stratum breakpoints within the lowest initial at CUMZ
-------
Appendix B
Chemical-Specific Data
-------
Volume II Appendix B
B. Chemical-Specific Data
Key chemical-specific input parameters include: air-liquid equilibrium partitioning
coefficient (vapor pressure or Henry's law constant), liquid-solid equilibrium partitioning
coefficient (log octanol-water partition coefficient for organics), biodegradation rate constants,
and liquid and air diffusivities. The HWIR chemical properties database (RTI, 1995) was used as
the primary data source for the physical and chemical properties for the constituents being
modeled. This chemical properties database provided the following chemical-specific input
parameters: molecular weight, vapor pressure, Henry's law constant, solubility, liquid and air
diffusivities, and log octanol-water partition coefficient. The soil biodegradation rate constants
were obtained from Howard et al (1991). The CHEMDAT8 chemical properties database (U.S.
EPA, 1994) was used as a secondary data source for the physical and chemical properties not
included in the HWIR data base. The CHEMDAT8 chemical properties database primarily
provided the following chemical-specific input parameters: density, boiling point, Antoine's
coefficients (for adjusting vapor pressure to temperature), and biodegradation rate constants for
tanks. Hydrolysis rates were taken from Kollig et al. (1993). The biodegradation rate constants
in the downloaded CHEMDAT8 data base file were compared with the values reported in the
summary report that provided the basis for the CHEMDAT8 tank biodegradation rate values
(Coburn et al., 1988). Tank biodegradation rates constants for compounds with no data were
assigned biodegradation rates equal to the most similar compound in the biodegradation rate data
base (or set to zero for metals). The chemical specific input parameters used for the emission
model estimates are presented in Table B-l.
References
Coburn, J., C. Allen, D. Green, and K. Leese. 1988. Site Visits of Aerated and Nonaerated
Impoundments. Summary Report. U.S. EPA, Contract No. 68-03-3253, Work
Assignment 3-8. Research Triangle Institute, Research Triangle Park, NC.
Howard, P.H., R.S. Boethling, W.F. Jarvis, W.M. Meylan, and E.M. Michalenko. 1991.
Handbook of Environmental Degradation Rates. Lewis Publishers, Chelsea, MI.
Kollig, H.P., J.J. Ellington, S.W. Karickhoff, B.E. Kitchens, J.M. Long, E.J. Weber, andN.L.
Wolfe. 1993. Environmental Fate Constants for Organic Chemicals Under
Consideration for EPA 's Hazardous Waste Identification Projects. U.S. Environmental
Protection Agency, Office of Research and Development, Athens, GA.
RTI (Research Triangle Institute). 1995. Technical Support Document for the Hazardous Waste
Identification Rule: Risk Assessment for Human Health and Ecological Receptors
Volumes I & II. Research Triangle Park, NC.
U.S. Environmental Protection Agency. 1994. Air Emissions Models for Waste and Wastewater.
EPA-453/R-94-080A. Appendix C. Office of Air Quality Planning and Standards,
Research Triangle Park, NC.
B-l
-------
Table B-l. Chemical Specific Input Parameters
CAS # COMPOUND NAME
50000 Formaldehyde
50328 Benzo(a)pyrene
55185 N-Nitrosodiethylamine
56235 Carbon tetrachloride
56495 3-Methylcholanthrene
57976 7,12-Dimethylbenz[a]anthracene
62533 Aniline
67561 Methanol
67641 Acetone
67663 Chloroform
67721 Hexachloroethane
68122 N,N-Dimethylformamide
71432 Benzene
71556 1,1,1-Trichloroethane
74839 Methyl bromide
74873 Methyl chloride
75014 Vinyl chloride
75058 Acetonitrile
75070 Acetaldehyde
75092 Methylene chloride
75150 Carbon disulfide
75218 Ethylene oxide
75252 Tribromomethane
75274 Bromodichloromethane
75354 1,1-Dichloroethylene
75569 Propylene oxide
75694 Trichlorofluoromethane
75718 Dichlorodifluoromethane
76131 1,1,2-Trichloro-1,2,2-trifluoroethane
77474 Hexachlorocyclopentadiene
78591 Isophorone
78875 1 ,2-Dichloropropane
78933 Methyl ethyl ketone
79005 1,1,2-Trichloroethane
7901 6 Trichloroethylene
79061 Acrylamide
79107 Acrylic acid
Mol.
Wt.
(a/mol)
30.03
252.32
102.14
153.82
268.36
256.35
93.13
32.04
58.08
119.38
236.74
73.09
78.11
133.4
94.94
50.49
62.5
41.05
44.05
84.93
76.14
44.06
252.73
163.83
96.94
58.08
137.37
120.91
187.38
272.77
138.21
112.99
72.11
133.4
131.39
71.08
72.06
Density
(alec)
0.97
1.11
1.59
1.02
1.02
1.02
0.79
0.79
1.49
2.09
0.9445
0.87
1.33
1.41
0.95
0.91
0.78
0.788
1.34
1.26
0.87
2.89
1.97
1.213
0.83
1.49
1.41
1.41
1.7
0.92
1.156
0.82
1.435
1.4
0.84
1.12
VAP.
Press.
(mmHa)
5240
5.5E-09
0.86
115
7.7E-09
5.6E-09
0.49
126
230
197
0.21
4
95
124
1620
4300
2980
91.1
902
433
359
1094
5.51
50
600
532.1
803
4850
332
0.0596
0.438
52
95.3
23.3
73.5
0.007
4
HLaw
Const.
(atm-
m3/mon
3.4E-07
1.1E-06
3.6E-06
0.0304
9.4E-07
3.1E-08
1 .9E-06
4.6E-06
3.9E-05
0.00367
0.00389
1 .9E-07
0.00558
0.0172
0.00624
0.00882
0.027
3.5E-05
7.9E-05
0.00219
0.03022
0.00012
0.00054
0.0016
0.0261
8.5E-05
0.097
0.343
0.4815
0.027
6.6E-06
0.0028
5.6E-05
0.00091
0.0103
1E-09
1 .2E-07
Diffusivity
in Water
(cm2/sec)
1 .98E-05
9.00E-06
8.00E-06
8.80E-06
5.36E-06
4.98E-06
8.30E-06
1 .64E-05
1.14E-05
1 .OOE-05
6.80E-06
1 .92E-05
9.80E-06
8.80E-06
1 .21 E-05
6.50E-06
1 .23E-05
1 .66E-05
1 .41 E-05
1.17E-05
1 .OOE-05
1 .45E-05
1 .03E-05
1 .06E-05
1 .04E-05
1 .OOE-05
9.70E-06
8.00E-06
8.20E-06
6.16E-06
6.76E-06
8.73E-06
9.80E-06
8.80E-06
9.10E-06
1 .06E-05
1 .06E-05
Diffusivity
in Air
(cm2/sec)
1 .78E-01
4.30E-02
8.00E-02
7.80E-02
2.09E-02
4.61 E-02
7.00E-02
1 .50E-01
1 .24E-01
1 .04E-01
2.49E-03
9.39E-02
8.80E-02
7.80E-02
7.28E-02
1 .26E-01
1 .06E-01
1 .28E-01
1 .24E-01
1 .01 E-01
1 .04E-01
1 .04E-01
1 .49E-02
2.98E-02
9.00E-02
1 .04E-01
8.70E-02
8.00E-02
7.80E-02
5.61 E-02
6.23E-02
7.82E-02
8.08E-02
7.80E-02
7.90E-02
9.70E-02
9.60E-02
Antoines' Vapor
Pressure Coefficients
ABC
7.195
9.246
6.934
8.164
6.955
6.950
7.897
7.117
6.493
7.228
6.928
6.905
6.827
7.566
7.093
6.991
7.119
8.005
6.968
6.942
7.128
7.988
7.966
6.972
7.067
6.884
7.590
8.784
8.415
7.963
6.980
7.112
7.192
6.518
1 1 .293
5.652
971
3724
1242
3364
2163
1467
1474
1211
929
1348
1401
1211
1147
1301
949
969
1314
1600
1074
1169
1055
2159
1847
1099
1133
1043
1329
1894
2835
2481
1380
1305
1480
1019
3940
649
244
273
273
230
273
171
177
229
230
196
133
196
221
219
273
249
251
230
292
223
242
238
273
273
237
236
237
273
273
273
273
223
229
229
193
273
155
log
Oct
Water Kmax
Part. mgVO/g-
Coeff. hr.
-0.05
6.11
0.48
2.73
6.42
6.62
0.98
-0.71
-0.24
1.92
4
-1.01
2.13
2.48
1.19
0.91
1.5
-0.34
1.25
1.25
2
-0.3
2.35
2.1
2.13
0.03
2.53
2.16
3.16
5.39
1.7
1.97
0.28
2.05
2.71
-0.96
0.35
5
0.001
4.4
1.5
0.001
0.001
7.1
18
1.3
28
0.001
9.7
19
3.5
10.76
10.76
10.76
9.7
82.42
18
15.3
4.2
10.76
10.76
10.76
17.56
1.076
1.076
0.001
0.001
15.3
17
2
3.5
3.9
9.7
17.56
K1
L/d-hr.
0.25
0.31
0.45
1.50
0.31
0.31
21.00
0.20
1.15
0.79
0.03
0.13
1.40
0.74
0.35
0.72
0.14
0.10
0.20
0.38
0.89
0.91
1.01
0.70
0.90
0.17
0.12
0.07
0.03
0.03
0.60
1.40
0.20
0.74
0.88
0.27
0.18
Hydrol.
Rate
sec-1
0
0
0
0
0
0
0
0
0
0
0
0
2E-08
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Soil
Biodeg.
Rate
sec-1
6.08E-10
4.61 E-08
1 .56E-08
3.13E-08
1 .22E-07
2.43E-09
6.95E-10
6.08E-10
6.08E-10
2.43E-09
1 .56E-08
1 .OOE-20
1 .39E-09
2.37E-08
2.43E-09
2.43E-09
1 .56E-08
2.43E-09
1 .OOE-20
2.43E-09
1 .OOE-20
1 .OOE-20
1 .56E-08
1 .OOE-20
1 .56E-08
1 .OOE-20
3.13E-08
1 .56E-08
1 .OOE-20
2.43E-09
2.43E-09
1.12E-07
6.08E-10
3.17E-08
3.13E-08
6.52E-11
1 .OOE-20
Solubility
ma/L
5.50E+05
2.50E-02
9.30E+04
7.93E+02
3.23E-03
2.50E-02
3.61 E+04
1 .OOE+06
1 .OOE+06
7.92E+03
5.00E+01
1 .OOE+06
1 .75E+03
1 .33E+03
1 .52E+04
5.33E+03
2.76E+03
1 .OOE+06
1 .OOE+06
1 .30E+04
1.19E+03
3.83E+05
3.10E+03
6.74E+03
2.25E+03
4.76E+05
1.10E+03
2.80E+02
1 .70E+02
1 .80E+00
1 .20E+04
2.80E+03
2.23E+05
4.42E+03
1.10E+03
6.40E+05
1 .OOE+06
-------
Table B-l. Chemical Specific Input Parameters
td
CAS # COMPOUND NAME
79345 1,1,2,2-Tetrachloroethane
79469 2-Nitropropane
80626 Methyl methacrylate
85449 Phthalic anhydride
87683 Hexachloro-1 ,3-butadiene
91203 Naphthalene
92875 Benzidine
95501 o-Dichlorobenzene
95534 o-Toluidine
95578 2-Chlorophenol
95658 3,4-Dimethylphenol
96128 1 ,2-Dibromo-3-chloropropane
98011 Furfural
98828 Cumene
98953 Nitrobenzene
100414 Ethylbenzene
100425 Styrene
1 06467 p-Dichlorobenzene
1 06887 1 ,2-Epoxybutane
106898 Epichlorohydrin
106934 Ethylene dibromide
106990 1 ,3-Butadiene
1 07028 Acrolein
107051 Allyl chloride
107062 1 ,2-Dichloroethane
107131 Acrylonitrile
107211 Ethylene glycol
108054 Vinyl acetate
108101 Methyl isobutyl ketone
108883 Toluene
1 08907 Chlorobenzene
1 08930 Cyclohexanol
108952 Phenol
109864 2-Methoxyethanol
110496 2-Methoxyethanol acetate
110543 n-Hexane
110805 2-Ethoxyethanol
110861 Pyridine
Mol.
Wt.
(d/mol)
167.85
89.09
100.12
148.12
260.76
128.17
184.24
147
107.16
128.56
122.17
236.33
96.09
120.19
123.11
106.17
104.15
147
72.11
92.53
187.86
54.09
56.06
76.53
98.96
53.06
62.07
86.09
100.16
92.14
112.56
100.2
94.11
76.09
130.15
86.18
90.12
79.1
Density
(d/cc)
1.59
0.9876
0.95
1.33
1.67
1.14
1.02
1.31
0.989
1.26
0.97
1.41
1.16
0.86
1.2
0.87
0.9
1.29
0.826
1.18
2.7
0.76
0.84
0.94
1.26
0.97
1.11
0.93
0.8
0.87
1.11
0.95
1.07
0.66
0.9
0.98
VAP.
Press.
(mmHa)
4.62
18
38.4
0.00052
0.221
0.085
8E-09
1.36
0.32
2.34
0.05836
0.58
2.21
4.5
0.245
9.6
6.12
1
207.912
16.4
13.3
2110
274
368
78.9
109
0.092
90.2
19.9
28.4
12
1.22
0.276
2.55697
9.28503
151
5.31
20.8
HLaw
Const.
(atm-
m3/mon
0.00035
0.00012
0.00034
1 .6E-08
0.00815
0.00048
3.9E-11
0.0019
2.7E-06
0.00039
2.3E-07
0.00015
4E-06
1.16
2.4E-05
0.00788
0.00275
0.0024
0.00046
3E-05
0.00074
0.0736
0.00012
0.011
0.00098
0.0001
6E-08
0.00051
0.00014
0.00664
0.0037
4.5E-06
4E-07
2.6E-07
1 .6E-06
0.0143
3.5E-07
8.9E-06
Diffusivity
in Water
(cm2/sec)
7.90E-06
1 .01 E-05
8.60E-06
9.60E-06
6.16E-06
7.50E-06
1 .50E-05
7.90E-06
9.12E-06
9.46E-06
8.33E-06
7.02E-06
1 .04E-05
7.10E-06
8.60E-06
7.80E-06
8.00E-06
7.90E-06
1 .03E-05
9.80E-06
1.19E-05
1 .08E-05
1 .22E-05
1 .08E-05
9.90E-06
1 .34E-05
1 .22E-05
9.20E-06
7.80E-06
8.60E-06
8.70E-06
8.31 E-06
9.10E-06
8.00E-06
8.00E-06
7.77E-06
9.57E-06
7.60E-06
Diffusivity Antoines' Vapor
in Air Pressure Coefficients
(cm2/sec) ABC
7.10E-02
9.23E-02
7.70E-02
7.10E-02
5.61 E-02
5.90E-02
8.00E-02
6.90E-02
7.14E-02
5.01 E-02
6.02E-02
2.12E-02
6.72E-02
8.60E-02
7.60E-02
7.50E-02
7.10E-02
6.90E-02
1 .35E-01
8.60E-02
2.17E-02
2.49E-01
1 .05E-01
1.17E-01
1 .04E-01
1 .22E-01
1 .08E-01
8.50E-02
7.50E-02
8.70E-02
7.30E-02
2.14E-01
8.20E-02
8.00E-02
8.00E-02
2.00E-01
9.47E-02
9.10E-02
6.894
7.272
6.517
8.022
7.485
7.373
7.542
6.883
7.197
6.877
7.504
8.073
6.575
6.963
7.115
6.975
6.945
7.199
6.832
8.229
7.345
7.217
7.213
7.576
7.068
7.110
8.091
7.210
6.672
6.954
6.978
6.255
7.133
6.876
7.874
7.041
1355
1531
1052
2869
1956
1968
2626
1538
1683
1472
1940
2436
1199
1461
1747
1424
1437
1690
1141
2087
1675
1145
1297
1494
1293
1336
2089
1296
1168
1345
1431
913
1517
1171
1844
1374
192
229
188
273
215
223
163
205
191
193
197
273
163
208
202
213
208
218
228
273
245
269
247
273
225
238
204
227
192
219
218
109
175
273
273
224
234
215
log
Oct
Water Kmax
Part. mgVO/g-
Coeff. hr.
2.39
0.87
1.38
-0.62
4.81
3.36
1.66
3.43
1.34
2.15
2.23
2.34
0.41
3.58
1.84
3.14
2.94
3.42
1.441
0.25
1.96
1.99
-0.01
1.45
1.47
0.25
-1.36
0.73
1.19
2.75
2.86
1.577
1.48
-0.77
0
4
-0.1
0.67
6.2
9.7
17.56
17.56
0.001
42.47
31.1
2.5
31.1
15
5.5
10.76
17.56
31.1
11
6.8
31.1
6.4
10.76
10.76
10.76
15.3
7.8
10.76
2.1
18
17.56
17.56
0.74
6.7
0.39
17.56
97
19.8
19.8
15.3
19.8
35.03
K1
L/d-hr.
0.68
0.42
4.30
0.08
0.03
1.00
0.66
0.58
0.86
0.89
1.05
0.16
0.54
2.88
2.30
2.10
0.11
2.30
0.48
0.14
0.55
0.69
0.34
0.31
0.98
0.75
0.06
0.30
0.45
2.40
10.00
0.54
13.00
1.00
1.00
1.47
1.00
0.24
Hydrol.
Rate
sec-1
0
0
0
0.0155
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Soil
Biodeg.
Rate
sec-1
3.82E-09
1 .56E-08
2.43E-09
1 .OOE-20
1 .56E-08
4.17E-09
6.95E-10
1 .56E-08
6.08E-10
1 .OOE-20
1 .OOE-20
1 .56E-08
1 .OOE-20
6.95E-10
1 .71 E-08
8.69E-10
2.43E-09
1 .OOE-20
1 .OOE-20
2.43E-09
1 .56E-08
1 .OOE-20
2.43E-09
1 .21 E-09
1 .56E-08
2.00E-09
1 .OOE-20
1 .OOE-20
6.08E-10
1 .91 E-09
1 .30E-08
1 .OOE-20
8.69E-10
1 .OOE-20
1 .OOE-20
1 .OOE-20
2.43E-09
6.08E-10
Solubility
ma/L
2.97E+03
1 .70E+04
1 .50E+04
6.20E+03
3.23E+00
3.10E+01
5.00E+02
1 .56E+02
1 .66E+04
2.20E+04
4.00E+04
1 .23E+03
1.10E+05
6.13E+01
2.09E+03
1 .69E+02
3.10E+02
7.38E+01
4.28E+04
6.59E+04
4.18E+03
7.35E+02
2.13E+05
3.37E+03
8.52E+03
7.40E+04
1 .OOE+06
2.00E+04
1 .90E+04
5.26E+02
4.72E+02
3.60E+04
8.28E+04
1 .OOE+06
1 .OOE+06
1 .24E+01
1 .OOE+06
1 .OOE+06
-------
Table B-l. Chemical Specific Input Parameters
CAS # COMPOUND NAME
1 1 1 1 59 2-Ethoxyethanol acetate
1 1 8741 Hexachlorobenzene
120821 1 ,2,4-Trichlorobenzene
121142 2,4-Dinitrotoluene
121448 Triethylamine
122667 1 ,2-Diphenylhydrazine
123911 1,4-Dioxane
124481 Chlorodibromomethane
1 26998 Chloroprene
127184 Tetrachloroethylene
630206 1,1,1,2-Tetrachloroethane
924163 N-Nitrosodi-n-butylamine
930552 N-Nitrosopyrrolidine
1319773 Cresols (total)
1 330207 Xylenes
1634044 Methyl tert-butyl ether
174601 6 2,3,7,8-TCDD
7439921 Lead
7439965 Manganese
7439976 Mercury
7440020 Nickel
7440382 Arsenic
7440393 Barium
7440417 Beryllium
7440439 Cadmium
7440473 Chromium (total)
7440484 Cobalt
7440622 Vanadium
10061015 cis-1,3-Dichloropropylene
10061026 trans-1 ,3-Dichloropropylene
16065831 Chromium (III)
18540299 Chromium (VI)
* This is the log Kd from the Mercury Report to
Mol.
Wt.
(a/mol)
143.01
284.78
181.45
182.14
101.19
184.24
88.11
208.28
88.54
165.83
167.85
158.24
100.12
108.1
106.17
88
322
207.2
54.938
200.59
58.69
74.92
137.33
9.012
112.41
51 .996
50.94
110.97
110.97
51 .996
51 .996
Congress.
Density
(a/cc)
2.04
1.41
1.31
0.7326
1.19
1.03
2.451
0.958
1.624
1.59
1.03
0.86
0.97
1.41
1.2
1.2
VAP.
Press.
(mmHa)
11.6912
1 .8E-05
0.431
0.00015
57.07
0.00043
38.1
4.9
213.658
18.6
12.03
0.03
0.092
0.3
8.04178
185.949
7.4E-10
0
0
0.00196
0
0
0
0
0
0
0
0
32.8
23.3
0
0
H Law Diffusivity Diffusivity
Const, in Water in Air
(atm-
m3/mon (cm2/sec) (cm2/sec)
2.2E-06 8.00E-06 8.00E-02
0.00132 5.91 E-06 5.42E-02
0.00142 8.23E-06 3.00E-02
9.3E-08 7.06E-06 2.03E-01
0.00014 7.88E-06 8.81 E-02
1.5E-06 7.36E-06 3.17E-02
4.8E-06 1.02E-05 2.29E-01
0.00078 1 .05E-05 1 .96E-02
0.0143 1.00E-05 1.04E-01
0.0184 8.20E-06 7.20E-02
0.00242 7.90E-06 7.10E-02
0.00032 8.00E-06 8.00E-02
1.2E-08 1.04E-05 7.36E-02
1.6E-06 9.30E-06 6.94E-02
0.00604 9.34E-06 7.14E-02
0.00056 1 .05E-05 1 .02E-01
1.6E-05 8.00E-06 4.70E-02
0 NA NA
0
0.0071 6.30E-06 5.5E-02
0 NA NA
0 NA NA
0 NA NA
0 NA NA
0 NA NA
0 NA NA
0 NA NA
0.00176 1.10E-05 5.85E-02
0.00125 1.10E-05 5.85E-02
0
0 1.41E-05 2.00E-01
Antoines' Vapor
Pressure Coefficients
ABC
9.554
7.706
7.981
6.959
13.836
7.351
8.220
6.161
6.976
6.894
8.850
7.940
6.852
6.977
6.807
6.807
3249
2243
3074
1272
5403
1518
2100
783
1387
1355
2795
2090
1104
2377
1328
1328
273
203
253
280
223
273
238
273
180
218
192
273
273
273
273
223
159
273
273
273
273
273
273
273
273
273
273
273
230
230
273
273
log
Oct
Water Kmax
Part. mgVO/g-
Coeff. hr.
0
5.89
4.01
2.01
1.45
2.94
-0.39
2.17
2.08
2.67
2.63
2.41
-0.19
0
3.17
1.901
6.64
5.447
3*
1.914
1.462
2.724
1.845
2.21
1.255
1.699
2
2
1.255
1.255
19.8
0.001
1.076
9.7
9.7
19
17.56
10.76
10.76
6.2
6.2
0.0001
0.0001
23
40.8
17.56
0.001
10.76
10.76
Hydrol.
K1 Rate
L/a-hr. sec-1
1.00
0.03
0.44
0.78
1.06
1.91
0.39
0.04
0.22
0.68
0.68
1.00
1.00
17.00
1.80
0.71
0.03
0.76
0.76
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Soil
Biodeg.
Rate
sec-1
1 .OOE-20
1 .82E-07
1 .56E-08
1 .56E-08
1 .OOE-20
1 .OOE-20
1 .56E-08
1 .56E-08
1 .56E-08
3.13E-08
5.81 E-09
1 .OOE-20
1 .56E-08
1 .OOE-20
2.43E-09
1 .OOE-20
1 .OOE-20
1 .OOE-20
1 .OOE-20
1 .OOE-20
1 .OOE-20
1 .OOE-20
1 .OOE-20
1 .OOE-20
1 .OOE-20
1 .OOE-20
1 .OOE-20
1 .OOE-20
9.81 E-10
9.81 E-10
1 .OOE-20
1 .OOE-20
Solubility
ma/L
1 .OOE+06
6.20E+00
3.00E+02
2.70E+02
5.50E+04
6.80E+01
1 .OOE+06
2.60E+03
1 .74E+03
2.00E+02
1.10E+03
1 .27E+03
1 .OOE+06
2.20E+04
1 .86E+02
3.88E+04
1 .90E-05
5.62E-02
2.72E+03
2.72E+03
-------
Appendix C
Sensitivity Analysis of ISC Air Model
-------
Volume II Appendix C
C. Sensitivity Analysis of ISC Air Model
This appendix describes sensitivity analyses on depletion options, source shape and
orientation, and receptor location and spacing.
C.I Options With and Without Depletions
The Air Characteristic Study relies on the output of the air dispersion model ISCST3 to
determine atmospheric concentrations of chemical constituents released by various waste
management units. The ISCST3 model has several options for modeling deposition and
depletion. Although the Air Characteristic Study is an inhalation-pathway-only analysis and does
not require modeling deposition to estimate indirect pathway exposure, depletion is important.
Depletion, or removal of chemical constituents from the plume due to deposition processes, can
affect the estimate of air concentrations. Using ISCST3 with depletion, however, requires
substantially more computer power and time to complete the extensive computer runs.
Therefore, for the current study, depletion options are used only if the risk results are sensitive to
the depletion option.
To determine ISCST3 model output sensitivity to depletion options for the Air
Characteristic study, the magnitude of wet depletion of vapors and particulates and dry depletion
of particulates were examined. ISCST3 cannot model dry depletion of vapors.
In this appendix, the setup of the sensitivity analysis is described, results are discussed,
and recommendations are presented.
C.I.I Setup of Sensitivity Analysis
Wastepiles and land application units (LAUs) were chosen to evaluate the effects of
modeling air concentration of vapors and parti culates with and without depletion. These two
WMU types were selected to represent elevated and ground-level sources, respectively. Two
areas were selected for each type of unit, one large and one small to capture differences in
depletion due to source area. The two wastepile sizes were 20.25 m2 and 1,300,056 m2. The
two land application unit sizes were 81m2 and 8,090,043 m2. These sizes represent the smallest
and largest size strata of each waste management unit type. Wastepiles were set to a source
height of 5 meters and land application units were set to a source height of 0 meters. Receptors
were placed at 25, 50, 75, 150, 500, and 1,000 meters from the edge of the source in concentric
squares.
Meteorological data from Las Vegas, Nevada, and Miami, Florida, were selected for these
analyses because these locations have the lowest and highest long-term average precipitation,
respectively, of the 29 meteorological stations used in the Air Characteristic Study. Las Vegas
averages 4.0 inches of precipitation per year and Miami averages 57.1 inches per year.
C-l
-------
Volume II Appendix C
C.1.2 Analysis of Wet Depletion and Its Magnitude for Vapors and Particulates
The first sensitivity analysis was performed to evaluate the significance of wet depletion
on calculated air concentrations of both vapors and particulates. The significance was
determined by examining the magnitude of the difference between annual average air
concentrations when wet depletion alone was selected vs. when both wet and dry depletion were
selected. Both large and small wastepiles and land application units were run in this analysis,
using the Hazardous Waste Identification Rule (HWIR 98) meteorological data and the 97363
version of ISCST3, which is the version of the model used for the May 1998 Air Characteristic
Study. Because retention of the air dispersion model runs from the May 1998 study was
desirable, the performance of this version of the model was important.
Because precipitation data in the SAMSON surface data set provided by the National
Climatic Data Center are generally not complete, a specially processed meteorological data set
was developed for the HWIR98 study based on precipitation data from the National
Climatological Data Center (NCDC) Cooperative Station Summaries of the Day data set.
Therefore, the new HWIR98 data set contains more complete precipitation data than were used in
the May 1998 Air Characteristic Study and would yield a higher and more representative amount
of wet depletion. The ISCST3 model was run for vapors and parti culates with no depletion, wet
depletion, and wet and dry depletion (particulates only) for both Miami and Las Vegas using
those meteorological data.
Vapor Results. Vapors were evaluated for no depletion and wet depletion only because
ISCST3 does not perform dry depletion calculations for vapor. In all cases, little difference in air
concentration was observed when wet depletion was included versus excluded. Most differences
were less than 1 percent, with lower air concentrations always resulting from the inclusion of wet
depletion. Differences between air concentrations with and without wet depletion slightly
exceeded 1 percent in a few cases when distances around 1 kilometer were examined. Only one
receptor showed a difference of more than 2 percent. This receptor was for a small LAU in
Miami at a distance of 1 kilometer. The difference was about 10 percent. The concentration
values, however, are extremely small in magnitude due to the small size of the site. This
difference is not considered significant.
Particulate Results. Large differences were observed between air concentrations with
and without wet and dry depletion for particles for both LAUs and wastepiles. Maximum
differences in air concentration ranged from 2 percent near the waste management unit to 57
percent at 1 kilometer from the waste management unit (see Tables C-la and C-lb). The percent
difference increases with increasing distance because the loss of material due to depletion is
cumulative in nature. A comparison of air concentrations modeled with wet depletion only and
those modeled with no depletion indicates that wet depletion plays a very small role in the annual
average depletion of particles. Wet depletion alone yields less than 2 percent difference in
concentration compared to no depletion.
C-2
-------
Volume II
Appendix C
Table C-la. Maximum Percent Differences in Air Concentration with and without Wet
Depletion Versus with and without Wet and Dry Depletion of Particles for
Land Application Units
Distance
(m)
25
50
75
150
500
1000
Miami
Large LAU
Wet
Depletion
0.9
1.0
1.1
1.3
1.8
2.2
Wet/Dry
Depletion
36.4
39.1
40.5
43.1
47.8
51.0
Small LAU
Wet
Depletion
0.1
0.1
0.2
0.9
6.7
20.0
Wet/Dry
Depletion
5.2
16.7
22.0
30.1
40.7
50.0
Las Vegas
Large LAU
Wet
Depletion
0.2
0.2
0.2
0.3
0.3
0.4
Wet/Dry
Depletion
16.3
17.8
18.7
20.4
24.7
29.1
Small LAU
Wet
Depletion
0.0
0.1
0.2
0.0
0.0
0.0
Wet/Dry
Depletion
1.6
6.6
9.4
14.3
22.2
33.3
Table C-lb. Maximum Percent Differences in Air Concentration with and without Wet
Depletion Versus with and without Wet and Dry Depletion of Particles for
Wastepiles
Distance
(m)
25
50
75
150
500
1000
Miami
Large WP
Wet
Depletion
0.7
0.7
0.8
1.0
1.5
2.0
Wet/Dry
Depletion
43.3
44.7
45.7
47.5
51.9
57.0
Small WP
Wet
Depletion
0.1
0.3
0.3
0.5
1.0
1.6
Wet/Dry
Depletion
20.6
30.5
35.7
43.0
52.4
56.3
Las Vegas
Large WP
Wet
Depletion
0.1
0.1
0.1
0.1
0.3
0.4
Wet/Dry
Depletion
19.8
20.8
21.6
23.6
30.4
34.8
Small WP
Wet
Depletion
0.0
0.0
0.0
0.1
0.2
0.3
Wet/Dry
Depletion
8.0
13.3
16.6
22.2
30.2
34.0
C-3
-------
Volume II Appendix C
C.1.3 Analysis of Different Versions of ISCST3 and Their Effects on Dry Depletion of
Particles
A second analysis was performed to determine the difference between air concentration
of particulates with dry depletion and without depletion for each of three different versions of
ISCST3. The three versions were 96113 (1996), 97363 (1997), and 98356 (1998). The 1996
version of the model was used for the sensitivity analysis conducted as part of the May 1998 Air
Characteristic Study. The 1997 version of the model was used for the production of the final
results in May 1998 and is the model version used for the current study. The 1998 version, the
latest release, was evaluated to determine if it should be adopted for the current study.
The algorithm for dry depletion differs in each of these versions of the model and the
1998 version contains a completely different area source integration routine from the other two
versions. This analysis was performed using the large land application unit and the May 1998
Air Characteristic Study meteorological data. As shown in Section C.I.2., wet depletion is not
significant for particles and thus is not included in this portion of the sensitivity analysis.
Furthermore, there is no difference in the wet depletion algorithm among the three versions of
ISCST3.
Results of Dry vs. No Depletion of Particles for Different Versions of ISCST3. As
part of this analysis, differences in air concentration results produced by different versions of the
ISCST3 model also were examined. When the three versions of ISCST3 were compared without
depletion, no difference was found between the 1996 and 1997 versions of ISCST3. Small
differences were found between these versions and the 1998 version and are not consistent in
direction. These insignificant differences are due to changes in the area source integration
technique1 used by the model and modifications to some portions of the code.
There are, however, significant differences between the three versions of ISCST3 when
dry depletion is included in the model run. The three versions were run for the largest land
application unit strata (8,090,043 m2). Large land application units were selected because larger
differences between concentrations with and without depletion are generally expected at most
receptor distances for a larger source. The dry depletion sensitivity analysis was conducted using
all three versions of the ISCST3 model and the same meteorological data that were used in the
May 1998 Air Characteristic Study.
The difference in air concentrations of particles with and without dry depletion for the
1996 version of ISCST3 was about 6 percent at 0 meters and increased to 25 percent at 1,000
meters for Las Vegas meteorological data. For Miami, these differences were 7 and 31 percent,
respectively. The 1997 version showed differences of 12 and 34 percent at Las Vegas and 15
percent and 38 percent at Miami. The difference between the 1996 and 1997 models stems from
the change in the deposition reference height from 20 times the roughness height to 1 meter.
lrThe integration technique arrives at a concentration value for a given receptor by performing a series of
iterations, each representing the area source as consecutively increasing numbers of point emission sources, until
there is little change in the result.
-------
Volume II Appendix C
The difference in air concentrations of particles with and without depletion in the 1998
version of the model was larger. At Las Vegas, the difference was 22 percent at 0 meters and 46
percent at 1,000 meters. Miami showed 28 percent and 52 percent differences, respectively.
These differences stem from changes to the algorithm used. These include a change in the
integration technique and modifications that affected the vertical dispersion coefficient for area
sources. Table C-lc shows the maximum differences with and without depletion for each
version of ISCST3 at each distance from the source.
In addition to differences of the effect of depletion between the 1997 and 1998 versions
of ISCST3, computer run time between these two versions of the model differed substantially.
The 1998 version of the model required eight times the number of hours needed to complete a
run than did the 1997 version of the model. The model developers were contacted for their
recommendation about which version of the model to use. While they confirmed the increase in
run time for the 1998 version, they offered no opinion as to which version is technically superior.
Therefore, the 1997 version was chosen based on run time considerations and consistency with
the previous Air Characteristic analysis.
C.1.4 Summary of Depletion Conclusions
Based on the results of these sensitivity analyses, the following decisions were made
concerning the dispersion modeling:
# Use meteorological data used in 1998 Air Characteristic Study since wet depletion
will not be modeled for vapors or particulates.
# Model vapors with no wet depletion because it is significant for vapors.
# Model particulates with dry depletion only because wet depletion is not significant
for particulates.
# Use 1997 version of ISCST3 because it is technically valid, compatible with the
May 1998 Air Characteristic Study, and the run time is substantially shorter than
the run time for the 1998 version.
C.2 Source Shape and Orientation
A sensitivity analysis was conducted using the ISCST3 air model to determine what role
source shape and orientation play in determining dispersion coefficients of air pollutants. A
discussion of this analysis follows.
Three different sources were chosen for this analysis. The sources were a square (source
No. 1), a rectangle oriented east to west (source No. 2), and a rectangle oriented north to south
(source No. 3). All three sources had an area of 400 m2 in order to ensure that equal emission
rates were compared. The rectangles were selected to be exactly two times longer and half as
wide as the square (see Figure C-2).
C-5
-------
Volume II
Appendix C
Table C-lc. Comparison of Maximum Percent Differences in Air Concentration of
Particles with Versus Without Dry Depletion Between Three Versions of
ISCST3
Distance
(m)
25
50
75
150
500
1000
Miami
ISCST3
1996
13
15
16
19
26
31
ISCST3
1997
25
27
29
31
36
39
ISCST3
1998
43
46
47
49
51
53
Las Vegas
ISCST3
1996
10
12
13
15
21
25
ISCST3
1997
20
22
23
25
29
34
ISCST3
1998
34
35
36
38
43
47
Two meteorological stations at Little Rock, Arkansas, and Los Angeles, California, were
selected for this modeling analysis in order to compare two different meteorological regimes.
Little Rock was selected because of its evenly distributed wind directions and Los Angeles was
selected because it has a predominantly southwest wind direction (see Figure C-3). Five years of
meteorological data were used for this analysis.
Each area source was modeled with similar receptor grids to ensure consistency. Sixteen
receptors were placed on the edge of each of the area sources and another 16 were placed 25
meters out from the edge. Each of these two receptor groups were modeled as a Cartesian
receptor grid. Two receptor rings were also placed at 50 and 100 meters out from the center of
the source. This polar receptor grid consisted of 16 receptors with a 22.5 degree interval between
receptors. See Figures C-4a through C-4c for receptor locations.
The ISCST3 model was run using the meteorological data from Little Rock, Arkansas,
and Los Angeles, California, and the results are shown in Tables C-2a and C-2b. The results
indicated that the standard deviation of the differences in air concentrations is greatest between
source No. 2 and source No. 3. This difference is due to the orientation of the source. This
occurs for both the Cartesian receptor grid and the polar receptor grid at both meteorological
locations. This shows that the model is sensitive to the orientation of the rectangular area source.
Standard deviations are significantly smaller when source No. 1 is compared to source
Nos. 2 or 3. This shows that the differences in Unitized Air Concentration (UAC) between the
square source and the two rectangular sources are less than the differences between the two
rectangular sources. A square area source also contributes the least amount of impact of
orientation. Since no information on source shape or orientation is available, a square source
will minimize the errors caused by different source shapes and orientations.
C-6
-------
Volume II
Appendix C
-30 -20
30-
-1 0
10 20
3 0
-30
20-
-20
Src3
1 0-
-1 0
0-
Src2
-0
1 0-
Srcl
— 1 0
20-
—20
30-
-30 -20
-1 0
10 20
-30
3 0
(meters)
Figure C-2. Source Shapes and Orientations.
C-7
-------
Volume II
Appendix C
Los Angeles, California
NNW
NW
NE
WWW
ENE
W
ESE
sw
SE
SSW
SSE
Little Rock, Arkansas
NNW
NNE
NW
NE
WNW
W
WSW
ENE
ESE
SW
SE
Figure C-3. Wind Roses
-------
Volume II
Appendix C
-i oo
1 GO-
'S o
5 0
1 00
-1 00
50-
-50
0—
-0
-5 0-
--5 0
1 00-
-1 00
-1 00
-5 0
5 0
1 00
(meters)
Figure C-4a. Receptor Locations (Source No. 1).
C-9
-------
Volume II
Appendix C
-i oo
1 GO-
'S o
5 0
1 00
-1 00
5 0-
-5 0
0—
Src2
--0
-50-
—50
1 00-
-1 00
-5 0
5 0
-1 00
1 00
C-10
(meters)
Figure C-4b. Receptor Locations (Source No. 2).
-------
Volume II
Appendix C
-i oo
1 GO-
'S o
5 0
1 00
-1 00
5 0-
-5 0
Src3
--0
-50-
—50
1 00-
-1 00
-5 0
5 0
-1 00
1 00
(meters)
Figure C-4c. Receptor Locations (Source No. 3).
C-ll
-------
o
to
Table C-2a. Comparisons of Unitized Air Concentrations (ug/m3/ug/s-m2) for Different Source Shapes and Orientations
(Little Rock, Arkansas)
Source
No. 1 (20m
x20m)
Source
No. 2 (40m
xlOm)
Source
No. 3 (10m
x40m)
Polar Receptor Grid
X(m)
19
38
35
71
46
92
50
100
46
92
35
71
19
38
0
0
-19
-38
-35
-71
-46
-92
-50
-100
-46
-92
-35
-71
-19
-38
0
0
Y(m)
46
92
35
71
19
38
0
0
-19
-38
-35
-71
-46
-92
-50
-100
-46
-92
-35
-71
-19
-38
0
0
19
38
35
71
46
92
50
100
UAC X(m)
0.190
0.050
0.249
0.067
0.321
0.095
0.124
0.030
0.085
0.023
0.106
0.030
0.117
0.033
0.122
0.035
0.134
0.038
0.161
0.043
0.159
0.044
0.103
0.027
0.126
0.035
0.152
0.041
0.173
0.047
0.224
0.068
19
38
35
71
46
92
50
100
46
92
35
71
19
38
0
0
-19
-38
-35
-71
-46
-92
-50
-100
-46
-92
-35
-71
-19
-38
0
0
Y(m)
46
92
35
71
19
38
0
0
-19
-38
-35
-71
-46
-92
-50
-100
-46
-92
-35
-71
-19
-38
0
0
19
38
35
71
46
92
50
100
UAC X(m)
0.199
0.051
0.243
0.067
0.361
0.098
0.128
0.030
0.096
0.024
0.109
0.030
0.113
0.032
0.117
0.033
0.128
0.036
0.158
0.043
0.185
0.046
0.114
0.027
0.145
0.036
0.160
0.042
0.179
0.047
0.191
0.061
19
38
35
71
46
92
50
100
46
92
35
71
19
38
0
0
-19
-38
-35
-71
-46
-92
-50
-100
-46
-92
-35
-71
-19
-38
0
0
Y(m)
46
92
35
71
19
38
0
0
-19
-38
-35
-71
-46
-92
-50
-100
-46
-92
-35
-71
-19
-38
0
0
19
38
35
71
46
92
50
100
UAC
0.211
0.051
0.278
0.069
0.256
0.088
0.147
0.033
0.084
0.023
0.103
0.029
0.128
0.034
0.143
0.037
0.150
0.038
0.170
0.045
0.140
0.043
0.107
0.027
0.118
0.034
0.153
0.041
0.187
0.048
0.276
0.074
Standard Deviation:
Differences
Sources No. 1
Diff. In UAC
0.010
0.001
-0.007
-0.001
0.041
0.003
0.004
0.000
0.011
0.001
0.003
0.000
-0.005
-0.001
-0.005
-0.002
-0.006
-0.002
-0.003
0.000
0.026
0.002
0.011
0.000
0.019
0.001
0.008
0.001
0.007
0.000
-0.032
-0.008
0.012
in UACs
and No. 2
Differences
Sources No. 1
% of Diff. Diff. In UAC
5%
1%
-3%
-1%
13%
3%
3%
-1%
12%
2%
3%
0%
-4%
-4%
-4%
-5%
-4%
-4%
-2%
1%
16%
4%
11%
2%
15%
4%
5%
3%
4%
0%
-14%
-11%
7%
0.021
0.001
0.028
0.001
-0.065
-0.007
0.023
0.003
-0.001
-0.001
-0.003
0.000
0.011
0.001
0.021
0.002
0.016
0.001
0.009
0.001
-0.019
-0.002
0.004
0.000
-0.008
-0.001
0.001
0.001
0.014
0.001
0.052
0.006
0.018
in UACs
and No. 3
Differences
Sources No. 2
% of Diff. Diff. In UAC
11%
2%
11%
2%
-20%
-7%
19%
9%
-1%
-2%
-3%
-1%
9%
2%
17%
5%
12%
2%
6%
3%
-12%
-4%
4%
1%
-6%
-4%
0%
2%
8%
3%
23%
9%
9%
0.012
0.000
0.035
0.002
-0.105
-0.010
0.020
0.003
-0.011
-0.001
-0.006
-0.001
0.016
0.002
0.026
0.004
0.022
0.002
0.012
0.001
-0.045
-0.004
-0.007
0.000
-0.027
-0.003
-0.007
-0.001
0.008
0.001
0.085
0.014
0.028
in UACs
and No. 3
% of Diff.
6%
1%
14%
3%
-29%
-10%
15%
11%
-12%
-5%
-6%
-2%
14%
7%
22%
11%
17%
6%
8%
3%
-24%
-8%
-6%
0%
-18%
-7%
-5%
-2%
4%
3%
44%
22%
14%
(continued)
I
*'
n
-------
Table C-2a (continued)
Source No
Cartesion
X(m)
-10
-5
0
5
10
10
10
10
10
5
0
-5
-10
-10
-10
-10
-35
-17.5
0
17.5
35
35
35
35
35
17.5
0
-17.5
-35
-35
-35
-35
. 1 (20m x 20m)
Receptor
Y(m)
-10
-10
-10
-10
-10
-5
0
5
10
10
10
10
10
5
0
-5
-35
-35
-35
-35
-35
-17.5
0
17.5
35
35
35
35
35
17.5
0
-17.5
S ource
Grid
UAC X (m)
3.014
4.266
4.354
3.961
2.175
5.211
5.968
6.012
4.946
6.804
6.846
6.157
3.245
4.923
5.169
4.809
0.164
0.219
0.243
0.186
0.108
0.141
0.277
0.503
0.254
0.315
0.417
0.272
0.155
0.211
0.213
0.265
-20
-10
0
10
20
20
20
20
20
10
0
-10
-20
-20
-20
-20
-45
-22.5
0
22.5
45
45
45
45
45
22.5
0
-22.5
-45
-45
-45
-45
No. 2 (40m
Y(m)
-5
-5
-5
-5
-5
-2.5
0
2.5
5
5
5
5
5
2.5
0
-2.5
-30
-30
-30
-30
-30
-15
0
15
30
30
30
30
30
15
0
-15
x 10m)
Source
UAC X (m)
2.675
4.219
4.307
4.069
1.899
3.875
4.704
4.918
4.468
6.758
6.830
6.353
2.793
3.801
4.032
3.727
0.158
0.247
0.284
0.192
0.088
0.105
0.164
0.396
0.263
0.373
0.445
0.286
0.131
0.155
0.145
0.193
-5
-2.5
0
2.5
5
5
5
5
5
2.5
0
-2.5
-5
-5
-5
-5
-30
-15
0
15
30
30
30
30
30
15
0
-15
-30
-30
-30
-30
No. 3 (10m
Y(m)
-20
-20
-20
-20
-20
-10
0
10
20
20
20
20
20
10
0
-10
-45
-45
-45
-45
-45
-22.5
0
22.5
45
45
45
45
45
22.5
0
-22.5
x 40m)
UAC
2.673
3.451
3.526
3.152
2.011
5.567
5.913
5.834
4.344
5.550
5.604
4.954
3.052
5.166
5.287
4.991
0.132
0.167
0.179
0.147
0.100
0.160
0.401
0.466
0.200
0.234
0.341
0.214
0.146
0.232
0.298
0.264
Standard Deviation
Differences
Sources No. 1
Diff. In UAC
-0.339
-0.047
-0.047
0.109
-0.276
-1.337
-1.264
-1.094
-0.477
-0.047
-0.016
0.196
-0.451
-1.121
-1.137
-1.081
-0.006
0.027
0.041
0.006
-0.020
-0.036
-0.113
-0.107
0.009
0.058
0.028
0.014
-0.024
-0.056
-0.068
-0.073
0.463
in UACs
and No. 2
Differences
Sources No. 1
% of Diff. Diff. In UAC
-11%
-1%
-1%
3%
-13%
-26%
-21%
-18%
-10%
-1%
0%
3%
-14%
-23%
-22%
-22%
-4%
12%
17%
3%
-19%
-25%
-41%
-21%
3%
18%
7%
5%
-15%
-27%
-32%
-27%
15%
-0.341
-0.815
-0.827
-0.809
-0.164
0.355
-0.055
-0.178
-0.602
-1.254
-1.242
-1.203
-0.193
0.244
0.118
0.182
-0.032
-0.052
-0.063
-0.039
-0.008
0.019
0.124
-0.037
-0.054
-0.081
-0.076
-0.057
-0.009
0.022
0.084
-0.002
0.435
in UACs
and No. 3
Differences
Sources No. 2
% of Diff. Diff. In UAC
-11%
-19%
-19%
-20%
-8%
7%
-1%
-3%
-12%
-18%
-18%
-20%
-6%
5%
2%
4%
-19%
-24%
-26%
-21%
-7%
14%
45%
-7%
-21%
-26%
-18%
-21%
-6%
10%
40%
-1%
17%
-0.002
-0.769
-0.781
-0.918
0.112
1.692
1.209
0.916
-0.125
-1.208
-1.226
-1.399
0.259
1.365
1.255
1.264
-0.026
-0.079
-0.104
-0.045
0.012
0.055
0.236
0.070
-0.063
-0.139
-0.104
-0.071
0.015
0.078
0.153
0.071
0.747
in UACs
and No. 3
% of Diff.
0%
-18%
-18%
-23%
6%
44%
26%
19%
-3%
-18%
-18%
-22%
9%
36%
31%
34%
-16%
-32%
-37%
-23%
14%
52%
144%
18%
-24%
-37%
-23%
-25%
11%
50%
106%
37%
41%
O
I
'
n
-------
o
Table C-2b. Comparisons of Unitized Air Concentrations (ug/m3/ug/s-m2) for Different Source Shapes and Orientations
(Los Angeles, California)
Source
No. 1 (20m
x20m)
Source
No. 2 (40m
x 10m)
Source
No. 3 (10m
x 40m)
Polar Receptor Grid
X(m)
19
38
35
71
46
92
50
100
46
92
35
71
19
38
0
0
-19
-38
-35
-71
-46
-92
-50
-100
-46
-92
-35
-71
-19
-38
0
0
Y(m)
46
92
35
71
19
38
0
0
-19
-38
-35
-71
-46
-92
-50
-100
-46
-92
-35
-71
-19
-38
0
0
19
38
35
71
46
92
50
100
UAC X (m)
0.059
0.016
0.188
0.046
0.582
0.172
0.278
0.068
0.061
0.015
0.062
0.016
0.080
0.023
0.086
0.023
0.099
0.028
0.122
0.033
0.218
0.060
0.320
0.093
0.264
0.074
0.137
0.037
0.063
0.017
0.067
0.020
19
38
35
71
46
92
50
100
46
92
35
71
19
38
0
0
-19
-38
-35
-71
-46
-92
-50
-100
-46
-92
-35
-71
-19
-38
0
0
Y(m)
46
92
35
71
19
38
0
0
-19
-38
-35
-71
-46
-92
-50
-100
-46
-92
-35
-71
-19
-38
0
0
19
38
35
71
46
92
50
100
UAC X (m)
0.065
0.016
0.168
0.045
0.607
0.174
0.293
0.067
0.062
0.015
0.068
0.017
0.076
0.022
0.084
0.024
0.092
0.027
0.119
0.032
0.223
0.061
0.378
0.098
0.273
0.075
0.123
0.035
0.066
0.017
0.058
0.018
19
38
35
71
46
92
50
100
46
92
35
71
19
38
0
0
-19
-38
-35
-71
-46
-92
-50
-100
-46
-92
-35
-71
-19
-38
0
0
Y(m)
46
92
35
71
19
38
0
0
-19
-38
-35
-71
-46
-92
-50
-100
-46
-92
-35
-71
-19
-38
0
0
19
38
35
71
46
92
50
100
UAC
0.069
0.016
0.284
0.052
0.461
0.161
0.293
0.074
0.087
0.016
0.062
0.017
0.087
0.024
0.096
0.024
0.108
0.028
0.143
0.034
0.226
0.061
0.278
0.087
0.260
0.073
0.164
0.039
0.073
0.018
0.080
0.021
Standard Deviation:
Differences
Sources No. \
Diff. In UAC
0.006
0.000
-0.020
-0.001
0.025
0.003
0.014
-0.001
0.002
0.000
0.006
0.001
-0.004
-0.001
-0.003
0.000
-0.006
-0.001
-0.003
0.000
0.005
0.001
0.057
0.005
0.009
0.001
-0.014
-0.002
0.003
0.000
-0.008
-0.002
0.013
in UACs
and No. 2
Differences
Sources No. 1
% of Diff. Diff. In UAC
9%
-1%
-11%
-3%
4%
2%
5%
-2%
3%
0%
10%
4%
-4%
-5%
-3%
1%
-7%
-2%
-2%
-1%
2%
1%
18%
6%
3%
1%
-10%
-5%
4%
-2%
-12%
-9%
6%
0.010
0.000
0.096
0.006
-0.121
-0.011
0.015
0.005
0.026
0.002
0.000
0.001
0.007
0.001
0.009
0.001
0.009
0.000
0.021
0.001
0.008
0.001
-0.042
-0.006
-0.005
-0.001
0.027
0.002
0.010
0.001
0.014
0.001
0.030
in UACs
and No. 3
Differences
Sources No. 2
% of Diff. Diff. In UAC
17%
3%
51%
13%
-21%
-6%
5%
8%
43%
10%
0%
3%
9%
3%
11%
3%
9%
1%
18%
4%
4%
1%
-13%
-6%
-2%
-2%
20%
4%
15%
3%
21%
6%
14%
0.005
0.001
0.116
0.007
-0.146
-0.014
0.001
0.007
0.025
0.002
-0.006
0.000
0.011
0.002
0.012
0.000
0.016
0.001
0.024
0.002
0.003
0.000
-0.099
-0.011
-0.013
-0.002
0.041
0.003
0.007
0.001
0.022
0.003
0.040
in UACs
and No. 3
% of Diff.
7%
4%
69%
16%
-24%
-8%
0%
10%
40%
11%
-9%
-1%
14%
8%
15%
2%
17%
3%
20%
5%
2%
0%
-26%
-11%
-5%
-2%
33%
9%
11%
5%
37%
15%
18%
(continued)
I
*'
n
-------
Table C-2b (continued)
Source No
Cartesion
X(m)
-10
-5
0
5
10
10
10
10
10
5
0
-5
-10
-10
-10
-10
-35
-17.5
0
17.5
35
35
35
35
35
17.5
0
-17.5
-35
-35
-35
-35
. 1 (20m x 20m)
Receptor
Y(m)
-10
-10
-10
-10
-10
-5
0
5
10
10
10
10
10
5
0
-5
-35
-35
-35
-35
-35
-17.5
0
17.5
35
35
35
35
35
17.5
0
-17.5
Source
Grid
UAC X (m)
3.225
4.025
3.952
3.431
1.683
5.931
6.636
6.640
5.600
6.893
6.860
6.031
3.393
5.649
5.944
5.663
0.124
0.158
0.172
0.123
0.064
0.095
0.592
0.829
0.192
0.109
0.125
0.113
0.139
0.387
0.603
0.318
-20
-10
0
10
20
20
20
20
20
10
0
-10
-20
-20
-20
-20
-45
-22.5
0
22.5
45
45
45
45
45
22.5
0
-22.5
-45
-45
-45
-45
No. 2 (40m
Y(m)
-5
-5
-5
-5
-5
-2.5
0
2.5
5
5
5
5
5
2.5
0
-2.5
-30
-30
-30
-30
-30
-15
0
15
30
30
30
30
30
15
0
-15
xlOm)
Source
UAC X (m)
3.241
4.333
4.297
3.871
1.592
4.787
5.882
6.294
5.866
8.126
8.285
7.442
3.497
5.102
5.373
5.028
0.139
0.183
0.199
0.124
0.053
0.076
0.377
0.739
0.304
0.195
0.144
0.160
0.166
0.335
0.472
0.275
-5
-2.5
0
2.5
5
5
5
5
5
2.5
0
-2.5
-5
-5
-5
-5
-30
-15
0
15
30
30
30
30
30
15
0
-15
-30
-30
-30
-30
No. 3 (10m
Y(m)
-20
-20
-20
-20
-20
-10
0
10
20
20
20
20
20
10
0
-10
-45
-45
-45
-45
-45
-22.5
0
22.5
45
45
45
45
45
22.5
0
-22.5
x40m)
UAC
2.674
3.119
3.050
2.564
1.511
5.570
5.644
5.524
4.325
4.939
4.913
4.156
2.702
5.015
5.167
5.104
0.095
0.123
0.121
0.100
0.063
0.119
0.696
0.683
0.101
0.072
0.100
0.077
0.089
0.370
0.603
0.316
Standard Deviation:
Differences
Sources No. 1
Diff. In UAC
0.016
0.308
0.345
0.440
-0.091
-1.143
-0.754
-0.346
0.266
1.232
1.424
1.411
0.103
-0.547
-0.572
-0.635
0.014
0.025
0.028
0.001
-0.011
-0.019
-0.215
-0.090
0.112
0.086
0.019
0.047
0.026
-0.053
-0.131
-0.043
0.542
in UAC s
and No. 2
Differences
Sources No. 1
% of Diff. Diff. In UAC
1%
8%
9%
13%
-5%
-19%
-11%
-5%
5%
18%
21%
23%
3%
-10%
-10%
-11%
11%
16%
16%
0%
-17%
-20%
-36%
-11%
58%
78%
15%
42%
19%
-14%
-22%
-13%
24%
-0.551
-0.906
-0.902
-0.867
-0.172
-0.360
-0.992
-1.116
-1.275
-1.955
-1.947
-1.875
-0.691
-0.634
-0.777
-0.559
-0.029
-0.035
-0.050
-0.024
-0.001
0.024
0.104
-0.146
-0.091
-0.037
-0.025
-0.035
-0.050
-0.017
0.000
-0.002
0.614
in UACs
and No. 3
Differences
Sources No. 2
% of Diff. Diff. In UAC
-17%
-23%
-23%
-25%
-10%
-6%
-15%
-17%
-23%
-28%
-28%
-31%
-20%
-11%
-13%
-10%
-23%
-22%
-29%
-19%
-2%
25%
18%
-18%
-47%
-34%
-20%
-31%
-36%
-4%
0%
-1%
15%
-0.567
-1.214
-1.247
-1.307
-0.081
0.783
-0.238
-0.770
-1.541
-3.187
-3.371
-3.286
-0.794
-0.088
-0.205
0.076
-0.043
-0.060
-0.078
-0.024
0.010
0.043
0.319
-0.055
-0.203
-0.122
-0.044
-0.082
-0.077
0.036
0.131
0.041
1.026
in UACs
and No. 3
% of Diff.
-17%
-28%
-29%
-34%
-5%
16%
-4%
-12%
-26%
-39%
-41%
-44%
-23%
-2%
-4%
2%
-31%
-33%
-39%
-20%
19%
57%
85%
-7%
-67%
-63%
-31%
-52%
-46%
11%
28%
15%
33%
o
IS
I
*'
n
-------
Volume II Appendix C
C.3 Receptor Locations and Spacings
A sensitivity analysis was conducted using the ISCST3 model to determine what receptor
locations and spacings should be used in the risk analysis for five types of waste management
units (WMUs). A discussion of the analysis follows.
Because it takes a substantial amount of time for the ISCST3 model to execute, it was
necessary to choose a limited number of receptors to be used in the dispersion modeling analysis.
The larger the number of receptor points, the longer the run time. However, modeling fewer
receptors may result in the omission of the maximum point for assessing exposure impacts.
Therefore, a sensitivity analysis was conducted to determine the number of receptors needed for
the model run and to locate ideal receptor placements.
A wind rose was plotted for each of the 29 meteorological stations to be used in the risk
analysis for a 5-year time period in order to choose two meteorological stations for this
sensitivity analysis. Little Rock, Arkansas, and Los Angeles, California, meteorological stations
were selected for the sensitivity analysis. The wind roses show that Little Rock has very evenly
distributed wind directions, and Los Angeles has a predominant southwest to west wind
(Figure C-3). Little Rock and Los Angeles were chosen to determine if a higher density of
receptors should be placed downwind of a site near Los Angeles, as compared to a site near Little
Rock. Similarly, the 5th, 50th, and 95th percentile of sizes of LAUs were used in the sensitivity
analysis to determine whether sizes of units can affect receptor locations and spacings. The areas
of the 5th, 50th, and 95th percentile of sizes of LAUs are 1,200 m2, 100,000 m2, and 1,700,000
m2, respectively.
The dispersion modeling was conducted using two sets of receptor grids. The first set of
receptor points (Cartesian receptor grid) was placed around the modeled source with distances of
25, 50, 75, and 150 meters from the edge of the unit. Square-shaped ground-level area sources
were used in the modeling. Therefore, these receptors are located on five squares surrounding
the source. The second set of receptor points (polar receptor grid) was placed outside of the first
set of receptors to 10 kilometers from the center of the source. Since the ISCST3 model's area
source algorithm does not consider elevated terrain, receptor elevations were not input in the
modeling.
In this sensitivity analysis, both downwind and lateral receptor spacings were investigated
for three unit sizes using 5 years of meteorological data from Little Rock and Los Angeles. For
the first set of receptor points (i.e., Cartesian receptor grid), five downwind distances of 25, 50,
75, and 150 meters from the edge of the source were used. For lateral receptor spacing, choices
of 64, 32, and 16 equally spaced receptor points for each square were used in the modeling to
determine the number of receptors needed to catch the maximum impacts. (See Figures C-5a
through C-5c for Cartesian receptor locations and spacings [50th percentile]). For the second set
of receptor points (i.e., polar receptor grid), about 20 downwind distances (i.e., receptor rings)
were used. Receptor lateral intervals of 22.5° and 10° were used to determine whether 22.5°
spacing can catch the maximum impacts. With a 22.5° interval, there are 16 receptors on each
ring. There are 36 receptors on each ring for the 10° interval. See Figures C-6a and C-6b for
polar receptor locations (5th percentile).
C-16
-------
Volume II Appendix C
The results (Figures C-7a through C-7f) show that the maximum downwind
concentrations decrease sharply from the edge of the area source to 150 meters from the source.
The maximum concentrations decrease more sharply for a smaller area source than for a larger
one. This means that more close-to-source receptors are generally needed for a small area source
than for a large one.
The results also show that the maximum impacts are generally higher for a dense receptor
grid (i.e., 64 or 32 receptors on each square) than for a scattered receptor grid (i.e., 16 receptors
on each square). However, the differences of the maximum receptor impacts are not significant
between a dense and a scattered receptor grid (Figures C-7a through C-7f). It should be noted
that the above conclusions apply to both Little Rock and Los Angeles. This means that the
distribution of wind directions does not play an important role in determining receptor lateral
spacings.
Figures C-8a through C-8f compare the maximum concentrations at each ring for 22.5°
and 10° intervals. The results show that the differences of the maximum concentrations are
greater for close-to-source receptors than for further out receptors, and the differences are greater
for larger area sources than for smaller area sources. The differences of the maximum
concentrations for 22.5° and 10° intervals are generally small, and the concentrations tend to be
the same at 10 kilometers. The conclusions were drawn from both Little Rock and Los Angeles
meteorological data.
C-17
-------
Volume II
Appendix C
-4
A 00
3 00-
200-
1 00-
o-
1 00-
700-
3 00-
A nn
00 -3 00 -2 00
+
+
+
+
+
-1 00 0 1 00
Land Application Unit
+ + + + + + + + +
2 00 3 00 4
+
+
+
+
+
30
A 00
-3 00
-200
-1 00
-0
— 1 00
—200
—3 00
A nn
-4 00 -3 00 -2 00
2 00 3 00 4 00
-1 00 0 1 00
(meters)
Figure C-5a. Cartesian Receptor Grid (64 receptors each square).
C-18
-------
Volume II
Appendix C
-4
A nn
3 00-
200-
1 00-
0-
1 00-
200-
3 00-
A nn
00 -300 -200 -100 0 100 200 300 4
+
+ + + + + + + + +
+
Load I'nppli onion Ll OIL
+
+ + + + + + + + +
30
A nn
-3 00
-200
-1 00
-0
— 1 00
—200
—3 00
A nn
-400 -300 -200 -100 0 100 200 300 400
(meters)
Figure C-5b. Cartesian Receptor Grid (32 receptors each square).
C-19
-------
Volume II
Appendix C
-4
400-
3 00-
2 00-
1 00-
0-
1 00-
2 00-
3 00-
A nn
00 -300 -200 -100 0 100 200 300 4
+
+
+ + + .
+
+ + +
Load i-H[i|nli CDUDD LlniL
+ + +
+
+ + +
+
+
30
-400
-3 00
-2 00
-1 00
-0
— 1 00
—2 00
—300
A nn
-400 -300 -200 -100 0 100 200 300 400
(meters)
Figure C-5c. Cartesian Receptor Grid (16 receptors each square).
C-20
-------
Volume II
Appendix C
-1 000 -8 00 -6 00 -4 00 -2 00 0 2 00 4 00 6 00 8 00 1 000
1 000^ ' ' ' ' 1 ' ' ' ' 1-1 000
600-
400-
2 00-
0—
-2 00-
-400-
-600-
1 000-
D
(meters)
Figure C-6a. Polar Receptor Grid (22.5 degree).
-600
-400
-2 00
-0
-2 00
-00
-600
; oo
-1 000
-1 000 -8 00 -6 00 -4 00 -2 00 0 2 00 4 00 6 00 8 00 1 000
C-21
-------
Volume II
Appendix C
-i ooo
i ooo-
-8 00 -6 00 -4 00 -2 00 0 2 00 4 00 6 00 8 00
600-
4 00-
2 00-
-2 00-
-400-
-6 00-
1 000-
i— +
+
— +
+ + + + +
-1 000 -8 00 -6 00 -4 00 -2 00 0 2 00 4 00 6 00
(meters)
Figure D-6b. Polar Receptor Grid (10 degree).
1 000
-1 000
-600
- 00
-2 00
-0
-2 00
-00
-6 00
-1 000
8 00 1 000
C-22
-------
S3
o
10
8
6
4
2
0
0
64 receptors
32 receptors
16 receptors
50 100 150
Distance from the edge of the unit (m)
200
p
to
Figure C-7a. Maximum Concentrations (5th Percentile, LAU, Los Angeles, CA)
I
X'
n
-------
o
to
o
U
_20
| 15
!lO
&JD
0
0
64 receptors
32 receptors
16 receptors
50
100
150
200
Distance from the edge of the unit (m)
Figure C-7b. Maximum Concentrations (50th Percentile, LAU, Los Angeles, CA)
1
n
-------
a
o
30
20
10
0
\
0
64 receptors
32 receptors
16 receptors
50
100
150
200
Distance from the edge of the unit (m)
p
to
Figure C-7c. Maximum Concentrations (95 Percentile, LAU, Los Angeles, CA)
I
*'
n
-------
o
to
10
VI
a a.
o
s 2
0
64 receptors
16 receptors
0
50
100
150
200
Distance from the edge of the unit (m)
Figure C-7d. Maximum Concentrations (5th Percentile, LAU, Little Rock, AR)
1
n
-------
^20
-^ IW
I
5
0
0
\
64 receptors
32 receptors
16 receptors
50
100
150
200
Distance from the edge of the unit (m)
o
to
Figure C-7e. Maximum Concentrations (50 Percentile, LAU, Little Rock, AR)
I
X'
n
-------
o
to
-------
0.6
0.4
0.2
0.0
0
22.5° Interval
10° Interval
2000 4000 6000 8000 10000
Distance from the edge of the unit (m)
p
to
VO
Figure C-8a. Maximum Concentrations (5th Percentile, LAU, Lose Angeles, CA)
I
X'
n
-------
o
OJ
o
5.0
4.0
3.0
2.0
0.0
0
22.5° Interval
10° Interval
2000 4000 6000 8000 10000
Distance from the edge of the unit (m)
Figure C-8b. Maximum Concentrations (50th Percentile, LAU, Los Angeles, CA)
1
n
-------
o
U
10
8
6
4
2
0
0
22.5° Interval
10° Interval
2000 4000 6000 8000 10000
Distance from the edge of the unit (m)
Figure C-8c. Maximum Concentrations (95th Percentile, LAU, Los Angeles, CA)
I
X'
n
-------
o
OJ
to
0.30 -T
~ 0.25
i 0.20
J "&JD
§ ^ 0.15
U ft
S 0.10
» 0.05
0.00
\
ii
22. 5° Interval
10° Interval
0 2000 4000 6000 8000 10000
Distance from the edge of the unit (m)
Figure C-8d. Maximum Concentrations (5th Percentile, LAU, Little Rock, AR)
1
n
-------
£4.0
E 3.0
11 2.0
u|io
"&JD
50.0
0
22.5° Interval
10° Interval
2000 4000 6000 8000 10000
Distance from the edge of the unit (m)
O
Figure C-8e. Maximum Concentrations (50 Percentile, LAU, Little Rock, AR)
I
X'
n
-------
o
0
22.5° Interval
10° Interval
2000 4000 6000 8000 10000
Distance from the edge of the unit (m)
Figure C-8f. Maximum Concentrations (95th Percentile, LAU, Little Rock, AR)
1
n
-------
Volume II Appendix C
C.4 An Analysis on Windroses at the 29 Sites
The hourly meteorological data from the 29 meteorological stations used in the Air
Characteristic Study were used to generate windroses. A windrose consists of 16 directions, with
the angle between any two adjacent directions being 22.5°. The prevailing wind directions for
the 29 meteorological stations were counted to estimate the number of entries in each wind
directions category. The results are presented in Figure C-9.
The narrowness of the most frequent wind directions for each of the 29 meteorological
stations was examined. Based on the narrowness (or angles) of the most frequent wind
directions, four categories were used to distinguish the windroses for the 29 meteorological
stations. The four categories of windroses are:
• Narrowly distributed: most frequent wind directions no greater than 45 °
• Moderately distributed: most frequent wind directions no greater than 90°
• Evenly distributed: no obvious predominant wind directions
• Bimodally distributed: most frequent wind directions are from two opposite
directions.
The number of meteorological stations in each category is given in Table C-3. Figure C-10 gives
some examples of windroses for each category. The windroses for the 29 meteorological stations
are available and can be provided upon request.
An examination of the windroses and the maximum unitized annual average air
concentrations from the Air Characteristic Study revealed that the sites with high concentrations
are those with narrowly distributed wind directions. Simply put, persistent wind direction
consistently blows pollutants from the source to the same receptors. Therefore, the more often
the wind blows in a certain direction, the more likely high cumulative concentrations will occur
at sites in that direction.
Air concentrations from a source are inversely proportion to windspeed. Given the same
distribution of wind directions, a site with lower windspeed will have higher concentrations. The
windroses show that, in the prevailing wind direction, the percentage of light wind occurring at a
site with narrowly distributed wind directions is often higher than that at a site with evenly
distributed wind directions. Therefore, we can conclude that a site with narrowly distributed
wind directions will most likely produce the highest long-term average air concentrations.
C-3 5
-------
Volume II
Appendix C
8 n
7 -
6
"3 5 -
54-
0 3 -
o
1
0
n
2! 2! 2! W W W
g M g M
—
GO
W
i— i i— i
n
n
n
Prevailing Wind Direction
Figure C-9. Counts of Prevailing Wind Directions in Each Direction
Table C-3. No. of Met Stations with Different Shapes of Windroses
Shape of Windrose
Narrowly distributed
Moderately distributed
Evenly distributed
Bi-modally distributed
No. of Stations
10
4
6
9
C-36
-------
Volume II
Appendix C
NNW
NW
NE
WNW
W
WSW
ENE
ESE
sw
SE
Narrowly Distributed Windrose
NNW
NW
NE
WNW
W
WSW
ENE
ESE
SW
SE
SSW
SSE
Moderately Distributed Windroses
Figure C-10. Examples of Different Shapes of Windroses
C-37
-------
Volume II
NNW
NW
NE
WNW
W
WSW
ENE
ESE
SW
SE
Evenly Distributed Windrose
Appendix C
NNW
NW
WNW
W
WSW
ENE
SW
SSW
SSE
Bi-modally Distributed Windrose
Figure C-10 (Continued). Examples of different Shapes of Windroses
C-38
-------
Appendix D
Derivation of Chronic Inhalation
Noncancer and Cancer Health
Benchmark Values
-------
NONCARCINOGENS
DERIVATION OF INHALATION REFERENCE CONCENTRATIONS
-------
Volume II
Appendix D
RfC:
Basis for RfC:
Critical Study:
Critical Dose:
Critical Effect:
Species:
Route of Exposure:
Duration:
Uncertainty Factor:
Modifying Factor:
2-Chlorophenol
CAS #95-57-8
0.0014mg/m3
Route-to-route extrapolation from the RfD
Exon, J.H., and L.D. Roller. 1982. Effects of transplacental exposure to
chlorinated phenols. Environ Health Perspect 46:137-140 (as cited in
U.S. EPA, 1998).
5 mg/kg/d
[X ] NOAEL [ ] LOAEL
Increase in conception rate and number of stillbirths and decrease in
size of litters
Rat
Drinking water
10 weeks
1000:
10 for extrapolation from animals to humans
10 for protection of sensitive human subpopulations
10 for use of a subchronic study
1
Calculations:
RfC = RfD x 1/70 kg x 20 m3/d = 0.005 mg/kg/d x 1/70 kg x 20 m3/d = 0.0014 mg/m3
where:
70 kg = default adult human body weight
20 m3/d = default human daily rate of inhalation
Calculations assume 100% absorption.
Summary of Study:
The RfD is based on a NOAEL of 5 mg/kg/d with a LOAEL of 50 mg/kg/d for reproductive
effects in a subchronic drinking water study in rats (Exon and Roller, 1982, as cited in U.S. EPA,
1998). In this study, groups of 12 to 20 weanling female Sprague-Dawley rats were exposed to 0,
5, 50, or 500 ppm of 2-chlorophenol in the drinking water and bred after 10 weeks of
2-chlorophenol treatment. Treatment was continued during breeding, gestation, and weaning.
The weanling rats were evaluated for percent conception, litter size, birth weight, weaning
weight, number of stillbirths, and hematology (hematocrit, hemoglobin levels, red and white cell
counts, and mean corpuscular volume). The evaluations revealed an increase in the conception
rate and in the number of stillborns as well as a decrease in the size of the litters in the rats
exposed to 500 ppm, which can be converted to a dosage of 50 mg/kg/d-the LOAEL. No effects
D-l
-------
Volume II Appendix D
were observed at 50 ppm, which can be converted to a dosage of 5 mg/kg/d. Dividing the
NOAEL of 5 mg/kg/d by an uncertainty factor of 1,000 (10 factors each for animal to human
extrapolation, interspecies variability, and the use of subchronic data), yields the RfD of 0.005
mg/kg/d (EPA, 1998).
Rationale for Route-to-Route Extrapolation:
A first pass in the liver or respiratory tract is not expected to contribute to the toxicity of
2-chlorophenol because it has been demonstrated that the toxic action of the lower chlorinated
phenols is due to the undissociated molecule. In studies with rats, it was observed that the
toxicity of chlorophenols administered via subcutaneous and intraperitoneal routes is similar to
that which is observed in orally administered chlorophenols (Deichmann and Keplinger, 1981).
Since the dermal irritation index for 2-chlorophenol is low, no significant portal of entry effect is
expected from inhalation exposure to 2-chlorophenol (HSDB, 1998).
Consequently, route-specific difference in toxicity is not expected for 2-chlorophenol. Therefore,
in accordance with EPA guidelines (U.S. EPA, 1994), the oral toxicity data for 2-chlorophenol
are adequate for use in the calculation of an inhalation RfC for the substance.
Strengths and Uncertainties:
The strength of the RfC is that it is based on an RfD on IRIS that has undergone rigorous EPA
peer review.
The major uncertainty of the RfC is the lack of inhalation toxicity studies in humans or animals
and the use of default values in the route-to-route extrapolation.
References:
Deichmann, W.B., and M.L. Keplinger. 1981. Aromatic Hydrocarbons. In: G.D. Clayton and
FE. Clayton (eds). Patty's Industrial Hygiene and Toxicology. 3rd revised edition. Volume 2A:
Toxicology. New York: John Wiley and Sons, pp. 3325-3415.
Exon, J.H., and L.D. Roller. 1982. Effects of transplacental exposure to chlorinated phenols.
Environ Health Perspect 46:137-140 (as cited in U.S. EPA, 1998).
Hazardous Substances Databank (HSDB): 2-Chlorophenol. 1998. Online database. National
Library of Medicine, Bethesda, MD.
U.S. Environmental Protection Agency. 1994. Methods for derivation of inhalation reference
concentrations and application of inhalation dosimetry. Research Triangle Park, NC:
Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment,
Office of Research and Development, U.S. EPA. EPA/600/8-90-066F.
U.S. Environmental Protection Agency. 1998. Integrated Risk Information System (IRIS).
2-Chlorophenol. Environmental Criteria and Assessment Office, Office of Health and
Environmental Assessment, Cincinnati, OH.
D-2
-------
Volume II
Appendix D
Cobalt
CAS #7440-48-4
RfC:
Critical Study:
Critical Dose:
Critical Effect:
Species:
Route of Exposure:
Duration:
Uncertainty Factor:
Modifying Factor:
0.00001 mg/m3
National Toxicology Program (NTP). 1996. NTP Technical Report on
the toxicology and carcinogenesis studies of cobalt sulfate heptahydrate
(CAS No. 10026-24-1) in F344/N rats and B6C3Fj mice (inhalation
studies). Draft. U.S. Department of Health and Human Services, Public
Health Service, National Institutes of Health. NIH Publication no. 96-
3961.NTPTR-471.
0.11 mg/m3
[ ] NOAEL [X] LOAEL
Hyperplasia of the lateral wall of the nose and atrophy of the olfactory
epithelium
Rat
Inhalation
104 weeks
300:
10 for use of a LOAEL
10 for protection of sensitive human subpopulations
3 for extrapolation from animals to humans with the use of a LOAEL
adjusted for human equivalent concentration
1
Calculations:
RfC = LOAEL
HEC
- UF = 0.004 mg/m3 - 300 = 0.00001 mg/m3
Summary of Study:
Groups of 50 male and 50 female rats and mice were exposed to aerosols containing 0, 0.3, 1.0,
and 3.0 mg/m3 cobalt sulfate heptahydrate (0, 0.11, 0.4, and 1.1 mg cobalt/m3) 6 hours/day, 5
days per week, for 104 weeks (NTP, 1996). Survival of both rats and mice were similar to
controls. The respiratory tract was the primary site of nonneoplastic lesions and neoplasms.
Cobalt sulfate heptahydrate caused a spectrum of inflammatory, fibrotic, and proliferative lesions
in the respiratory tract. Olfactory epithelial atrophy and hyperplasia of the lateral wall of the nose
were increased significantly at all dose levels in both male and female rats, and severity of these
lesions increased with increasing exposure concentration. Olfactory epithelial atrophy was seen
at 0.4 and 1.1 mg cobalt/m3 in male and female mice. Hyperplasia of the adrenal medulla was
increased significantly in male rats at 0.11 mg cobalt/m3 and cytoplasmic vacuolization of the
bronchus was increased significantly in male and female mice at this concentration.
D-3
-------
Volume II Appendix D
A LOAEL of 0.11 mg cobalt/m3 was selected based on respiratory effects (olfactory epithelial
atrophy and hyperplasia of the lateral wall of the nose) observed at this lowest dose in rats. This
LOAEL was adjusted for continuous exposure (0.02 mg/m3). A LOAELj^c was calculated as per
EPA's inhalation dosimetry methodology (1994), using equation 4-5 (insoluble, approximately
spherical particles). An uncertainty factor of 300 was applied: 10 for use of a LOAEL, 10 for
protection of sensitive human subpopulations, and 3 for extrapolation from animals to humans
with the use of a LOAEL adjusted for human equivalent concentration.
Conversion Factors:
Dose levels are:
% cobalt in cobalt sulfate heptahydrate = 38% = 0.38.
0.3 mg CoSO4»7H2O/m3 x 0.38 = 0.11 mg Co/m3; 1.0 mg CoSO4»7H2O/m3 = 0.4 mg Co/m3;
3.0 mg CoSO4»7H2O/m3 = 1.1 mg Co/m3.
LOAEL ^ = 0.11 mg Co/m3 x (6 h/24 h) x (5/7 d) = 0.02 mg
LOAEL HEC= LOAEL ^ x RDDR,
LOAEL HEC= 0.02 mg Co/m3 x 0.209 = 0.004 mg Co/m3
Co/m3
where
LOAEL AJJJ is the adjusted LOAEL, LOAELj^c is the human equivalent concentration
LOAEL, RDDRr (regional deposited dose ratio) is a multiplicative factor used to adjust an
observed inhalation particulate concentration mass median aerodynamic diameter of an
animal to the predicted inhalation particulate exposure concentration for a human; based on
mass median aerodynamic diameter MMAD = 1.5 //m, geometric standard deviation (sigma
g) = 2.2, mean body weight of female rat = 237 g, extrathoracic respiratory effects; RDDR
calculated using EPA RDDR Program (computer disk).
Additional Information:
Human studies support the critical endpoint (the respiratory system) identified in the NTP (1996)
study. Respiratory irritation, wheezing, asthma, pneumonia, and fibrosis have been reported in
epidemiological studies at concentrations ranging from 0.007 to 0.893 mg cobalt/m3 as cobalt
dust (Antilla et al., 1986; Davison et al., 1983; Demedts et al., 1984; Hartung et al., 1982;
Shirakawa et al., 1988, 1989; Van Cutsem et al., 1987, as cited in ATSDR, 1992). Sprince et al.
(1988) reported a strong relationship between work-related wheezing and cobalt dust exposure in
workers engaged in tungsten carbide production. The relative odds for work-related wheezing
doubled when cobalt exposure exceeded 0.050 mg/m3 compared with exposure to less than 0.050
mg/m3.
Animal studies have also shown a variety of respiratory effects. Exposure of rats and rabbits to
mixed cobalt oxides resulted in lesions of the alveolar region of the respiratory tract (Johanssen
et al., 1984, 1987, as cited in ATSDR, 1992). Lifetime exposure of hamsters to cobalt oxide
resulted in emphysema (Wehner et al., 1977, as cited in ATSDR, 1992). An NTP subchronic
(13-week) study reported squamous metaplasia of the larynx, histiocytic infiltrates in the lung,
and acute inflammation of the nose in rats and mice (NTP, 1991). A LOAEL of 0.3 mg cobalt
sulfate heptahydrate/m3 was identified for this endpoint. Asthma and sensitization were reported
in workers exposed to 0.007 mg Co/m3 as cobalt dust (Shirakawa et al., 1988). The LOAEL of
0.007 mg Co/m3 identified in the Shirakawa et al. (1988) occupational study is similar to the
EM
-------
Volume II _ Appendix D
c calculated in the NTP study; therefore, the use of the cobalt sulfate heptahydrate
study to derive the RfC would not be expected to be overly conservative.
Strengths and Uncertainties:
The strengths of the RfC are that it was based on a well-designed chronic study from the NTP
that involved extensive clinical and pathological examinations in two species and the critical
effect noted in the study has been observed in numerous other human and animal studies.
Respiratory effects have been reported in workers and animals exposed to a variety of cobalt
compounds, including cobalt dust. Similar respiratory effects have been reported across several
species exposed to various cobalt compounds at similar exposure levels (adjusting to mg Co/m3
and human equivalent concentrations).
The major uncertainty of the RfC is the lack of a NOAEL from this study or other studies. The
high incidence of adverse effects at the lowest dose tested presents uncertainties as to the true
NOAEL for cobalt.
References:
Agency for Toxic Substances and Disease Registry (ATSDR). 1992. Toxicological profile for
cobalt. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service.
Antilla, S., S. Sutinen, M. Paananen, et al. 1986. Hard metal disease: A clinical, histological,
ultrastructural and X-ray microanalytical study. Eur J Respir Dis 69:83-94 (as cited in ATSDR,
1992).
California Environmental Protection Agency (CalEPA). 1997. Technical support document for
the determination of noncancer chronic reference exposure levels, Draft for Public Review.
Office of Environmental Health Hazard Assessment, Air Toxicology and Epidemiology Section,
Berkeley, CA.
Davison, A.G., P.L. Haslam, B. Corrin, et al. 1983. Interstitial lung disease and asthma in hard-
metal workers: Bronchoalveolar lavage, ultrastructual and analytical findings and results of
bronchial provocation tests. Thorax 38: 1 19-128 (as cited in ATSDR, 1992).
Demedts, M., B. Gheysens, J. Nagels, et al. 1984. Cobalt lung in diamond polishers. Am Rev
Respir Dis 130:130-135 (as cited in ATSDR, 1992).
Hartung, M., K.H. Schaller, and E. Brand. 1982. On the question of the pathogenetic importance
of cobalt for hard metal fibrosis of the lung. Int Arch Occup Env Health 50:53-57 (as cited in
ATSDR, 1992).
Johanssen, A., P. Camner, C. Jarstrand, et al. 1984. Lung morphology and phospholipids after
experimental inhalation of soluble cadmium, copper, and cobalt. Environ Res 34:295-309 (as
cited in ATSDR, 1992).
D-5
-------
Volume II Appendix D
Johanssen, A., B. Robertson, and P. Camner. 1987. Nodular accumulation of type II cells and
inflammatory lesions caused by inhalation of low cobalt concentrations. Environ Res 43:227-243
(as cited in ATSDR, 1992).
National Toxicology Program (NTP). 1991. NTP report on the toxicity studies of cobalt sulfate
heptahydrate in F344/N rats and B6C3FJ mice (inhalation studies). Research Triangle Park, NC:
U.S. Department of Health and Human Services, Public Health Service, National Institutes of
Health. Nffl Publication no. 91-3124. NTP TOX-5.
National Toxicology Program (NTP). 1996. NTP Technical Report on the toxicology and
carcinogenesis studies of cobalt sulfate heptahydrate (CAS No. 10026-24-1) in F344/N rats and
B6C3FJ mice (inhalation studies). Draft. U.S. Department of Health and Human Services, Public
Health Service, National Institutes of Health. NIH Publication no. 96-3961. NTP TR-471.
Shirakawa, T., Y. Kusaka, N. Fujimura, et al. 1988. The existence of specific antibodies to
cobalt in hard metal asthma. CM Allergy 18:451-460 (as cited in ATSDR, 1992).
Shirakawa, T., Y. Kusaka, N. Fujimura, et al. 1989. Occupational asthma from cobalt sensitivity
in workers exposed to hard metal dust. Chest 95:29-37 (as cited in ATSDR, 1992).
Sprince, N.L., L.C. Oliver, E.A. Eisen, R.E. Greene, and R.I. Chamberlin. 1988. Cobalt
exposure and lung disease in tungsten carbide production: a cross-sectional study of current
workers. Am Rev Respir Dis 138:1220-1226.
U.S. Environmental Protection Agency. 1994. Methods for derivation of inhalation reference
concentrations and application of inhalation dosimetry. Research Triangle Park, NC:
Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment,
Office of Research and Development, U.S. EPA. EPA/600/8-90-066F.
Van Cutsem, E.J., J.L. Ceuppens, L.M. Lacquet, et al. 1987. Combined asthma and alveolitis
induced by cobalt in a diamond polisher. Eur JRespir Dis 70:54-61 (as cited in ATSDR, 1992).
Wehner, A.P., R.H. Busch, RJ. Olson, et al. 1977. Chronic inhalation of cobalt oxide and
cigarette smoke by hamsters. Am IndHygAssoc .738:338-346 (as cited in ATSDR, 1992).
D-6
-------
Volume II
Appendix D
Cresols
CAS #1319-77-3
RfC:
Critical Study:
Critical Dose:
Critical Effect:
Species:
Route of Exposure:
Duration:
Uncertainty Factor:
Modifying Factor:
0.0004 mg/m3
Uzhdavini, E.R., K. Astaf yeva, A.A. Mamayeva, and G.Z. Bakhtizina.
1972. [Inhalation toxicity of o-cresol]. Trudy Ufimskogo Nauchno-
Isseldovatel'skogo Institute Gigiyeny Profzabolevaniya, 7:115-9.
(Russian) [as cited in CalEPA, 1997, and U.S. EPA 1985, 1986]
9 mg/m3
[ ] NOAEL [X] LOAEL
Alterations in bone marrow cellularity
Rat
Inhalation
4 months
3000:
10 for use of a LOAEL
10 for extrapolation from animals to humans
10 for protection of sensitive human subpopulations
3 for extrapolation from subchronic to chronic exposure
1
Calculations:
RfC =
H- up = 1.3 mg/m3 - 3000 = 0.0004 mg/m3
Summary of Study:
Male and female rats were exposed to 0 or 9.0 mg/m3 o-cresol via inhalation, first for 2 months
(6 h/d, 5 d/wk) and then for 2 more months (4 h/d, 5 d/wk) (Uzhdavini et al., 1972, as cited in
CalEPA, 1997). The following endpoints were examined: elemental conditioned defensive
reflex, white blood cell levels, bone marrow elements, and liver function (as indicated indirectly
by hexobarbital narcosis). Both exposed and control animals showed some loss of the defensive
reflex, with the effect occurring in all exposed animals before the end of the second month and in
control animals at later times. White blood cell counts were elevated in male animals, peaking at
the end of the exposure period and returning to normal 1 month after cessation of exposure.
Exposed animals also showed a statistically significant change in the leukoid-to-erythroid ratio in
the bone marrow. Liver toxi city was suggested by an extension of hexobarbital narcosis duration
in treated animals. A LOAEL of 9 mg/m3 for hematological effects was identified.
The LOAEL of 9 mg/m3 was adjusted for continuous exposure (1.3 mg/m3). A LOAELj^c was
calculated as per EPA's inhalation dosimetry methodology (1994), using equation 4-48a
(category 3 - extrarespiratory effects). An uncertainty factor of 3000 was applied: 10 for use of a
D-7
-------
Volume II _ Appendix D
LOAEL, 10 for extrapolation from humans to animals, 10 for human variability, and 3 for
extrapolation from subchronic to chronic exposure.
Conversion Factors:
j = 9 mg/m3 x (5/24 h) x (5/7 d) = 1.3 mg/m3
c = LOAEL ADJ x RGDR
c = LOAEL^ x (Hb/g)A/(Hb/g)H
c =1.3 mg/m3 x 1 = 1.3 mg/m3
where
LOAEL Ajjj is the adjusted LOAEL, LOAELj^c is the human equivalent concentration
LOAEL, RGDR is the regional gas dose ratio (animal :human), and (Hb/g)A/(Hb/g)H is the ratio
of blood:gas partition coefficient; (Hb/g)A/(Hb/g)H defaults to 1 where Hb/g values are not
known.
Additional Information:
In humans, inhalation exposure is reported to cause respiratory effects, including the
development of pneumonia, pulmonary edema, and hemorrhage (Clayton and Clayton, 1981).
Irritation of the nose and throat, nasal constriction, and dryness was reported in 8 of 10
individuals briefly exposed to 6 mg/m3 (Uzhdavini et al., 1972, as cited in CalEPA 1997).
Signs of respiratory irritation (as indicated by increased paratid gland secretions) were observed
in cats exposed to 5 to 9 mg/m3 o-cresol for 30 minutes (Uzhdavini et al., 1972, as cited in
CalEPA 1997). Exposure of mice to 50 mg/m3 o-cresol for 2 h/d for 1 month did not affect
mortality; however, heart muscle degeneration and degeneration of nerve cells and glial elements
were reported (Uzhdavini et al., 1972, as cited in CalEPA, 1997, U.S. EPA, 1985).
Strengths and Uncertainties:
Major areas of uncertainty are the lack of quantitative human data, the scarcity of animal
inhalation data, and the lack of a NOAEL for this study. Also, the data presented were
incomplete, the number of animals used is not known, exposure and control conditions were not
described, statistical analyses were not provided, and the purity of the compound tested could not
be ascertained. However, the Uzhdavini et al. (1972) study was used by CalEPA in the
derivation of their chronic inhalation reference exposure level (REL) and is the only study with
quantitative data on cresols available at this time.
References:
California Environmental Protection Agency (CalEPA). 1997. Technical support document for
the determination of noncancer chronic reference exposure levels, Draft for Public Review.
Office of Environmental Health Hazard Assessment, Air Toxicology and Epidemiology Section,
Berkeley, CA.
Clayton, G.D., and F.E. Clayton (eds). 1981. Patty's Industrial Hygiene and Toxicology. 3rd
revised edition. Volume 2A: Toxicology. New York: John Wiley and Sons, pp. 2597-2601.
D-8
-------
Volume II Appendix D
U.S. Environmental Protection Agency. 1985. Health and environmental effects profile for
cresols. Cincinnati, OH: Environmental Criteria and Assessment Office, Office of Research and
Development, U.S. EPA. EPA/600/X-85-358.
U.S. Environmental Protection Agency. 1986. Health effects assessment for cresols. Cincinnati,
OH: Environmental Criteria and Assessment Office, Office of Health and Environmental
Assessment, Office of Research and Development, U.S. EPA. EPA/540/1-86-050.
U.S. Environmental Protection Agency. 1994. Methods for derivation of inhalation reference
concentrations and application of inhalation dosimetry. Research Triangle Park, NC:
Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment,
Office of Research and Development, U.S. EPA. EPA/600/8-90-066F.
Uzhdavini, E.R., K. Astaf yeva, A. A. Mamayeva, and G.Z. Bakhtizina. 1972. [Inhalation
toxicity of o-cresol]. Trudy Ufimskogo Nauchno-Isseldovatel'skogo Instituto Gigiyeny
Profzabolevaniya, 7:115-9. [as cited in CalEPA, 1997, and U.S. EPA 1985, 1986]
D-9
-------
Volume II Appendix D
3,4-Dimethylphenol
CAS # 95-65-8
RfC: Data are inadequate to support the derivation of an RfC at this time.
Supporting Data:
An RfD of 0.001 mg/kg/d is listed in IRIS (U.S. EPA, 1998), based on a subchronic feeding
study in rats. Changes in blood pressure and body weight and histopathological changes in liver,
kidney and spleen were reported (Veldre and Janes, 1979). Route-to-route extrapolation of an
RfC from the RfD is not recommended because of the potential for respiratory tract effects
following inhalation exposure and first-pass effects following ingestion exposure.
Although dimethylphenols have been detected in tobacco smoke, automobile exhausts, and
exhausts from stationary sources, they have not been detected in ambient air (U.S. EPA, 1986).
3,4-Dimethylphenol is not likely to occur at detectable concentrations in ambient air because it is
a solid at ambient temperatures and has a low vapor pressure. Consequently, inhalation
exposures are unlikely to be important for the general population. Skin absorption and ingestion,
which can be evaluated by the RfD, are likely to be the predominant exposure pathways.
Very little toxicity or metabolism data specific to 3,4-dimethylphenol are available.
Dimethylphenols and related compounds (phenol and methylphenols [cresols]) are rapidly
absorbed following ingestion, inhalation, or skin contact and are corrosive to skin, eyes, mucous
membranes, and the respiratory tract. Therefore, portal-of-entry effects are likely to be important
and cannot be addressed from route-to-route extrapolation. First-pass effects also may be
important. These compounds are metabolized predominantly to glucuronide and sulfate
conjugates and excreted in the urine (U.S. EPA, 1986). Skowronski et al. (1994) suggested that a
lack of first-pass metabolism in the liver may contribute to the toxicity of phenol following skin
absorption; therefore, differences in metabolism following ingestion and inhalation exposures
also could affect toxicity.
References:
Skowronski, G.A., A.M. Kadry, R.M. Turkall, et al. 1994. Soil decreases the dermal penetration
of phenol in male pig in vitro. J Toxicol Environ Health 41:467-479.
U.S. Environmental Protection Agency. 1986. Health and environmental effects profile for
dimethylphenols. Environmental Criteria and Assessment Office, Cincinnati, OH.
EPA/600/X-86/256.
U.S. Environmental Protection Agency. 1994. Methods for derivation of inhalation reference
concentrations and application of inhalation dosimetry. Research Triangle Park, NC:
Environmental Criteria and Assessment Office, Office of Health and Environmental Assesment,
Office of Research and Development, U.S. EPA. EPA /600//8-90-066F.
U.S. Environmental Protection Agency. 1998. Integrated Risk Information System (IRIS). 3,4-
Dimethylphenol. Environmental Criteria and Assessment Office, Office of Health and
Environmental Assessment, Cincinnati, OH.
D-10
-------
Volume II Appendix D
Veldre, I.A., and HJ. Janes. 1979. Toxicological studies of shale oils, some of their components
and commercial products. Environ Health Perspect 30:141-146 (as cited in U.S. EPA, 1998).
D-ll
-------
Volume II
Appendix D
1,4-Dioxane
CAS # 123-91-1
RfC:
Critical Study:
Critical Dose:
Critical Effect:
Species:
Route of Exposure:
Duration:
Uncertainty Factor:
Modifying Factor:
0.8 mg/m3
Torkelson, T.R., B.K.J. Leong, RJ. Kociba, et al. 1974. 1,4-Dioxane.
II. Results of a 2-year inhalation study in rats. ToxicolAppl
Pharmacol 30:287-298.
400 mg/m3
[X] NOAEL [ ] LOAEL
No effect on liver, kidney, or hematological endpoints
Rat
Inhalation
2 years
100:
10 for extrapolation from animals to humans
10 for protection of sensitive human subpopulations
1
Calculations:
RfC =
- UF = 83.3 mg/m3 - 100 = 0.8 mg/m3 (0.2 ppm)
Summary of Study:
Groups of Wistar rats were exposed to 0 or 1 1 1 ppm (0 or 400 mg/m3) 1,4-dioxane 7 h/d, 5 d/wk
for 2 years (Torkelson et al., 1974). Animals were observed for signs of toxicity, including
behavioral changes, eye and nasal irritation, respiratory distress, and skin condition. Body
weight was measured weekly. Hematological measurements were made at 16 and 23 months and
included serum glutamic-pyruvic transaminase (SGPT) activity, blood urea nitrogen (BUN),
alkaline phosphatase (AP) activity, and total protein determinations. At sacrifice, gross necropsy
of all animals was performed, and organs were examined for tumors. Histological examination
of tissues was conducted.
No significant differences in survival, body weight, general appearance, or behavior were
reported. Packed cell volume (PCV), red blood cells, and hemoglobin were slightly, but
significantly (p<0.05), increased and white blood cells were significantly decreased in exposed
males; however, the study authors note that these differences were within normal physiological
levels and not considered of toxic importance. Slightly decreased BUN and AP values observed
in exposed males were not considered to be biologically significant by the investigators based on
the fact that an increase, not a decrease, in these parameters would indicate kidney or liver
damage. Increased total protein in exposed males was also reported but not considered to be
biologically significant. No significant differences in liver, kidney, or spleen weights, or gross or
D-12
-------
Volume II _ Appendix D
microscopic alterations were observed. Tumor incidence (including hepatic and nasal) was not
significantly different in any of the organs examined.
The NOAEL of 400 mg/m3 was adjusted for continuous exposure (83.3 mg/m3). A
was calculated as per EPA's inhalation dosimetry methodology (1994), using equation 4-48a
(category 3 - extrarespiratory effects). An uncertainty factor of 100 was applied: 10 for
extrapolation from humans to animals and 10 for human variability.
Conversion Factors:
0.4 mg/L x 1,000 L/m3 = 400 mg/m3
= 400 mg/m3 x (7/24 hr) x (5/7 d) = 83.3 mg/m3
= NOAEL^j x RGDR
= NOAEL^ x (Hb/g)A/(Hb/g)H
= 83.3 mg/m3 x 1 = 83.3 mg/m3
where
NOAELAQj is the adjusted NOAEL, NOAELj^c is the human equivalent concentration
NOAEL, RGDR is the regional gas dose ratio (animal: human), and (Hb/g)A/(Hb/g)H is the
ratio of blood:gas partition coefficient; (Hb/g)A/(Hb/g)H defaults to 1 where Hb/g values are not
known.
Additional Information:
The major metabolite of 1,4-dioxane in rats is beta-hydroxyethoxyacetic acid (HEAA), which is
excreted in the urine (Braun and Young, 1977). Results from a study by Young et al. (1978)
show that the fate of 1,4-dioxane in rats is markedly dose-dependent due to a limited capacity to
metabolize dioxane to HEAA. Exposure to 1,4-dioxane by ingestion results in saturation of
metabolism above a single dose of 100 mg/kg, or as low as 10 mg/kg when administered in
multiple doses. When rats were exposed to 50 ppm for 6 hours, nearly all the inhaled
1,4-dioxane was also metabolized to HEAA (99%); the plasma half-life was 1 . 1 hours (Young et
al., 1978). The correlation of the dose-dependent fate of 1,4-dioxane with the results of
toxicological studies in rats supports the conclusion that there is an apparent threshold for the
toxic effects of dioxane that coincides with saturation of the metabolic pathway for its
detoxification (Young et al., 1978). 1,4-Dioxane and HEAA were also found in the urine of
dioxane plant workers exposed to an average concentration of 1.6 ppm (TWA) for 7.5 hours
(Young etal., 1976, 1977).
In a study by Kociba et al. (1974), Sherman rats were exposed to 0, 0.01, 0.1, or 1.0%
1,4-dioxane in drinking water for up to 2 years. No hematologic changes were reported.
Histopathologic examination revealed hepatocellular and renal tubular degenerative changes,
accompanied by regenerative activity, in rats exposed to the two highest dose levels, but not at
the low dose (Kociba et al., 1974). The lack of hematological effects observed in the ingestion
study suggests that the toxicity of 1,4-dioxane may be route-specific. Studies suggest that the
inhalation of 1,4-dioxane may lead to adverse effects, but good dose-response data are not
available. The toxicity of 1,4-dioxane may be a function of the saturation of the mechanism of
metabolism (Young et al., 1978).
D-13
-------
Volume II Appendix D
Strengths and Uncertainties:
The strengths of the RfC are that it is based on a lifetime study, with a large number of toxic
endpoints examined and a large sample size (n=192-288). The weaknesses of the inhalation
benchmark value include the use of a free-standing NOAEL, that only one exposure level was
used in the Torkelson et al. (1974) study, the limited human data, the limited inhalation data in
animals, and the lack of developmental and reproductive studies.
References:
Braun, W.H., and J.D. Young. 1977. Identification of beta-hydoxyethoxyacetic acid the major
urinary metabolite of 1,4-dioxane in the rat. ToxicolApplPharmacol 39:33-38.
California Environmental Protection Agency (CalEPA). 1997. Technical support document for
the determination of noncancer chronic reference exposure levels, Draft for Public Review.
Office of Environmental Health Hazard Assessment, Air Toxicology and Epidemiology Section,
Berkeley, CA.
Kociba, R.J., S.B. McCollister, C. Park, et al. 1974. 1,4-Dioxane. I. Results of a 2-year
ingestion study in rats. Toxicol Appl Pharmacol 30:275-286.
Torkelson, T.R.,B.KJ. Leong, RJ. Kociba, etal. 1974. 1,4-Dioxane. II. Results of a 2-year
inhalation study in rats. Toxicol Appl Pharmacol 30:287-298.
U.S. Environmental Protection Agency. 1994. Methods for derivation of inhalation reference
concentrations and application of inhalation dosimetry. Research Triangle Park, NC:
Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment,
Office of Research and Development, U.S. EPA. EPA/600/8-90-066F.
Young, J.D., W.H. Braun, and PJ. Gehring. 1978. Dose-dependent fate of 1,4-dioxane in rats.
J Toxicol Environ Health 4:709-726.
Young, J.D., W.H. Braun, PJ. Gehring, et al. 1976. 1,4-Dioxane and beta-hydroxyethoxyacetic
acid excretion in urine of humans exposed to dioxane vapors. Toxicol Appl Pharmacol 38:643-
646.
Young, J.D., W.H. Braun, L.W. Rampy, et al. 1977. Pharmacokinetics of 1,4-dioxane in
humans. J Toxicol Environ Health 3:507-520.
D-14
-------
Volume II
Appendix D
RfC:
Basis for RfC:
Critical Study:
Critical Dose:
Critical Effect:
Species:
Route of Exposure:
Duration:
Uncertainty Factor:
Modifying Factor:
2-Ethoxyethanol Acetate
CAS #111-15-9
0.3 mg/m3
Calculated from RfC for 2-ethoxyethanol
Barbee, S.J., J.B. Terrill, DJ. DeSousa, and C.C. Conaway. 1984.
Subchronic inhalation toxicology of ethylene glycol monoethyl ether in
the rat and rabbit. Environ Health Perspect 57: 157-163.
68 mg/m3 (2-ethoxyethanol)
[X]NOAEL []LOAEL
Decreased testis weight, seminiferous tubule degeneration, and
decreased hemoglobin (2-ethoxyethanol)
Rabbit
Inhalation
13 weeks
300:
3 for extrapolation from animals to humans
10 for intraspecies extrapolation
10 for use of a subchronic study
1
Summary of Study:
Groups of New Zealand rabbits (10/sex/group) were exposed to 0, 25, 103 or 403 ppm (0, 92,
380, or 1,485 mg/m3) 2-ethoxyethanol 6 h/d, 5 d/wk for 13 weeks (Barbee et al., 1984). The
duration-adjusted exposure concentrations were 0, 16, 68, and 265 mg/m3, respectively. Groups
of Sprague-Dawley rats were similarly exposed to 2-ethoxyethanol (0, 25, 103, or 403 ppm).
Physical examination and body weight measurements were conducted weekly. Hematological,
chemical, and histopathological changes (including bone marrow of the sternum) were also
assessed. No respiratory effects (nasal turbinates, tracheas, and lungs were examined) were
observed in either species.
In both male and female rabbits, body weight was significantly depressed at 403 ppm.
Significantly decreased testes weight was observed in rabbits exposed to 403 ppm. Pathological
changes in the testes were characterized as minimal to slight focal degeneration of the
seminiferous tubules with loss of epithelium in 3 of 10 rabbits. Spermatogenic activity in the
affected males was judged by overall organ morphology and deemed normal. Additionally, both
sexes exhibited significantly decreased hemoglobin, hematocrit, and erythrocyte count at 403
ppm.
D-15
-------
Volume II _ Appendix D
Hematological, reproductive, and developmental effects were also observed in pregnant Wistar
rats (Doe, 1984). Neurobehavioral and neurochemical effects were reported in the offspring of
female Sprague-Dawley rats, as well as in the dams (Nelson et al., 1981, 1982a, 1982b).
Based on the observed testicular and hematopoietic effects at 403 ppm (Barbee et al., 1984), the
NOAELjjEc and LOAELj^c for 2-ethoxyethanol in the rabbit are identified as 68 and 265 mg/m3,
respectively. The NOAEL was adjusted for intermittent exposure. A NOAELj^ was calculated
as per EPA's inhalation dosimetry methodology (1994), using equation 4-48a (category 3 -
extrarespiratory effects). An uncertainty factor of 300 was applied (10 for intraspecies
extrapolation, 10 for use of a subchronic study, and 3 to account for interspecies extrapolation).
An RfC of 0.2 mg/m3 was calculated for 2-ethoxyethanol (U.S. EPA, 1998).
Calculations and Conversion Factors:
Dose levels are:
25 ppm 2-ethoxyethanol x (90.12/24.45) = 92 mg/m3 2-ethoxyethanol; 103 ppm = 380 mg/m3;
403 ppm = 1,485 mg/m3.
= 380 mg/m3 x (6/24 h) x (5/7 d) = 68 mg/m3
= NOAEL^j x RGDR
= NOAEL^ x (Hb/g)A/(Hb/g)H
= 68 mg/m3 x 1 = 68 mg/m3
where
NOAELAQj is the adjusted NOAEL, NOAELj^c is the human equivalent concentration
NOAEL, RGDR is the regional gas dose ratio (animal: human), and (Hb/g)A/(Hb/g)H is the ratio
of blood:gas partition coefficients; (Hb/g)A/(Hb/g)H defaults to 1 where Hb/g values are not
known.
RfC for 2-ethoxyethanol = NOAHE^ + UF = 68 mg/m3 + 300 = 0.2 mg/m3
RfC (2-ethoxyethanol) x [MW (2-ethoxyethanol acetate) - MW (2-ethoxyethanol)] =
RfC (2-ethoxyethanol acetate)
0.2 mg/m3 x (132.16/90.12) = 0.3 mg/m3 = RfC (2-ethoxyethanol acetate)
Additional Information:
The RfC of 0.2 mg/m3 for 2-ethoxyethanol (U.S. EPA, 1998) is based on hematological and male
reproductive effects in New Zealand rabbits. It is assumed that 2-ethoxyethanol acetate would be
equitoxic to 2-ethoxyethanol on a molar basis, since both compounds are metabolized to
ethoxyacetic acid by the alcohol dehydrogenase pathway (DOL-OSHA, 1993). The RfC for
2-ethoxyethanol acetate was calculated from the RfC for 2-ethoxyethanol of 0.2 mg/m3 (U.S.
EPA, 1998), resulting in an RfC of 0.3 mg/m3 for 2-ethoxyethanol acetate.
Strengths and Uncertainties:
The strength of the RfC is that it is based on an RfC in IRIS that has undergone Agency review.
The major uncertainty of the RfC for 2-ethoxyethanol acetate is that it is based on an RfC for
2-ethoxyethanol and converted using the molecular weight ratio. However, it is assumed that
2-ethoxyethanol acetate would be equitoxic to 2-ethoxyethanol on a molar basis.
D-16
-------
Volume II Appendix D
References:
Barbee, S.J., J.B. Terrill, DJ. DeSousa, and C.C. Conaway. 1984. Subchronic inhalation
toxicology of ethylene glycol monoethyl ether in the rat and rabbit. Environ Health Perspect 57:
157-163.
Doe, I.E. 1984. Ethylene glycol monoethyl ether and ethylene glycol monoethyl ether acetate
teratology studies. Environ Health Perspect 57:33-41.
Department of Labor, Occupational Safety and Health Administration (DOL-OSHA). 1993.
Federal Register: Occupational exposure to 2-methoxyethanol, 2-ethoxyethanol and their
acetates (glycol ethers); Proposed rule. 58FR15526. March 23, 1993.
Nelson, B.K., W.S. Brightwell, J.V. Setzer, et al. 1981. Ethoxyethanol behavioral teratology in
rats. Neurotoxicology 2:231-249.
Nelson, B.K., W.S. Brightwell, and J.V. Setzer. 1982a. Prenatal interactions between ethanol
and the industrial solvent 2-ethoxyethanol in rats: Maternal and behavioral teratogenic effects.
Neurobehav Toxicol Teratol 4(3):387-394.
Nelson, B.K., W.S. Brightwell, J.V. Setzer, and T.L. O'Donohue. 1982b. Prenatal interactions
between ethanol and the industrial solvent 2-ethoxyethanol in rats: Neurochemical effects in the
offspring. Neurobehav Toxicol 7erato/4(3):395-401.
Research Triangle Institute (RTI). 1996. Assessment of Risks from the Management of Used
Solvents. U.S. Environmental Protection Agency, Office of Solid Waste, Washington, DC. July
1996. [In support of 61FR42317]
U.S. Environmental Protection Agency. 1994. Methods for derivation of inhalation reference
concentrations and application of inhalation dosimetry. Research Triangle Park, NC:
Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment,
Office of Research and Development, U.S. EPA. EPA/600/8-90-066F.
U.S. Environmental Protection Agency. 1996. Federal Register: Hazardous Waste Management
System; Identification and Listing of Hazardous Waste; Solvents; CERCLA Hazardous
Substance Designation and Reportable Quantities; Proposed Rule. 61FR42317-354. August 14,
1996.
U.S. Environmental Protection Agency. 1998. Integrated Risk Information System (IRIS) file
for 2-ethoxyethanol.
D-17
-------
Volume II
Appendix D
Ethylene glycol
CAS # 107-21-1
RfC:
Critical Study:
Critical Dose:
Critical Effect:
Species:
Route of Exposure:
Duration:
Uncertainty Factor:
Modifying Factor:
0.6 mg/m3
Wills, J.H., F. Coulston, E.S.Harris, et al. 1974. Inhalation of
aerosolized ethylene glycol by man. din Toxicol 7:463-476.
67 mg/m3
[X ] NOAEL [ ] LOAEL
Throat and upper respiratory tract irritation
Humans
Inhalation
30 days
100:
10 for protection of sensitive human subpopulations
10 for use of a subchronic study
1
Calculations:
RfC =
- UF = 55.8 mg/m3 - 100 = 0.6 mg/m3
Summary of Study:
Twenty volunteer male prisoners were exposed to ethylene glycol in mean daily concentrations
between 3 and 67 mg/m3 for 30 days, 20 h/d, without effect (Wills et al., 1974). Irritation was
noted after 15 minutes at an exposure concentration of 188 mg/m3 and was judged intolerable at
244 mg/m3. No effects were observed in clinical serum enzyme levels for liver and kidney
toxicity, hematotoxicity, or psychological responses. The irritation resolved soon after exposure
with no effects noted after a 6-week followup period.
A NOAEL of 67 mg/m3 was selected and adjusted for continuous exposure (55.8 mg/m3). An
uncertainty factor of 100 was applied: 10 for use of a subchronic study (30 day-duration) and 10
for protection of sensitive human subpopulations.
Conversion Factors:
NOAEL^^ 67 mg/m3 x 20/24 h = 55.8 mg/m3
Additional Information:.
Animal studies are inconclusive regarding the respiratory effects of ethylene glycol. Suber et al.
(1989, as cited in ATSDR, 1997) report thickened respiratory epithelium with enlarged goblet
cells in rats that inhaled ethylene glycol over 90 days. Another study in rhesus monkeys and rats
showed no respiratory effects from continuous exposure to propylene glycol for 13 to 18 months
(Robertson et al., 1947, as cited in ATSDR, 1997). Developmental effects have been seen in
D-18
-------
Volume II Appendix D
animal studies. Tyl et al. (1995a, 1995b, as cited in CalEPA, 1997) reported reduced ossification
in humerus, zygmotatic arch, and the metatarsals in fetuses of rats and mice exposed to ethylene
glycol on days 6 through 15 of gestation.
Strengths and Uncertainties:
The major strength of the RfC is that it was based on human data with controlled inhalation
exposures and the observation of a NOAEL. The major uncertainty to the RfC is the lack of
chronic inhalation studies in humans and confirming studies in animals.
References:
Agency for Toxic Substances and Disease Registry. 1997. Toxicological profile for ethylene
glycol and propylene glycol. Atlanta, GA: U.S. Department of Health and Human Services,
Public Health Service.
California Environmental Protection Agency (CalEPA). 1997. Technical support document for
the determination of noncancer chronic exposure levels, Draft for Public Review. Office of
Environmental Health Hazard Assessment, Air Toxicology and Epidemiology Section, Berkeley,
CA.
Robertson, O.K., C.G. Loosli, and T.T. Puck. 1947. Test for chronic toxicity of propylene
glycol and triethylene glycol on monkeys and rats by vapor inhalation and oral administration. J
PharmacolExper Therap 91:52-76 (as cited in ATSDR 1997).
Suber, R.L., R.D. Deskin, I. Nikiforov, et al. 1989. Subchronic nose-only inhalation study of
propylene glycol in Sprague-Dawley rats. FoodChem Toxicol 27(9):573-5&4 (as cited in
ATSDR, 1997).
Tyl, R.W., B. Ballantyne, L.C. Fisher, et al. 1995a. Evaluation of the developmental toxicity of
ethylene glycol aerosol in CD-I mice by nose-only exposure. Fundam Appl Toxicol 27:49-62 (as
cited in CalEPA, 1997).
Tyl, R.W., B. Ballantyne, L.C. Fisher, et al. 1995b. Evaluation of the developmental toxicity of
ethylene glycol aerosol in CD rat and CD-I mouse by whole-body exposure. Fundam Appl
Toxicol 24:51'-75 (as cited in CalEPA, 1997).
U.S. Environmental Protection Agency. 1994. Methods for derivation of inhalation reference
concentrations and application of inhalation dosimetry. Research Triangle Park, NC:
Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment,
Office of Research and Development, U.S. EPA. EPA/600/8-90-066F.
Wills, J.H., F. Coulston, E.S. Harris, et al. 1974. Inhalation of aerosolized ethylene glycol by
man. din Toxicol 7:463-476.
D-19
-------
Volume II
Appendix D
Methanol
CAS # 67-56-1
RfC:
Critical Study:
Critical Dose:
Critical Effects:
Species:
Route of Exposure:
Duration:
Uncertainty Factor:
Modifying Factor:
13 mg/m3
Rogers, J.M., M.L. Mole, N. Chernoff, et al. 1993. The developmental
toxicity of inhaled methanol in the CD-I mouse, with quantitative
dose-response modeling for estimation of benchmark doses.
Teratology 47(3): 175-188.
1,310 mg/m3
[X] NOAEL [ ] LOAEL
Developmental malformations (increased cervical ribs, exencephaly,
and cleft palate)
Mouse
Inhalation
Gd6-15
100:
10 for extrapolation from animals to humans
10 for protection of sensitive human subpopulations
1
Calculations:
RfC =
+ UF = 13 10 mg/m3 - 100 = 13 mg/m3 (10 ppm)
Summary of Study:
Groups of pregnant CD-I mice were exposed to 1,000, 2,000, 5,000, 7,500, 10,000, or 15,000
ppm methanol (1,310, 2,620, 6,552, 9,828, 13,104, or 19,656 mg/m3) for 7 h/d on days 6 through
15 of gestation (Rogers et al., 1993). Three groups of controls were used. Sham-exposed
controls were exposed to filtered air. Additional control groups remained in their cages and
received food and water ad libitum or were food-deprived for 7 h/d (to match the food
deprivation experienced by the exposed mice). Dams were observed twice daily and weighed on
alternate days during the exposure period. Blood methanol concentrations were determined in
three mice per exposure level on gestation days 6, 10, and 15. On day 17, the remaining mice
were weighed and sacrificed and the gravid uteri removed. Implantation sites, live and dead
fetuses, and resorptions were counted, and fetuses were examined externally and weighed as a
litter. Half of each litter were examined for skeletal morphology and the other half of each litter
were examined for internal soft tissue anomalies.
One dam died in each of the three highest exposure groups, but no dose-response relationship
was evident for maternal death. The sham-exposed and food-deprived controls, as well as all
methanol-exposed dams, gained less weight than did unexposed dams fed ad libitum, but
methanol did not exacerbate this effect. Significant increases in the incidence of exencephaly
D-20
-------
Volume II Appendix D
and cleft palate were observed at 6,552 mg/m3 and above, increased embryo/fetal death at 9,828
mg/m3 and above (including an increasing incidence of full-litter resorptions), and reduced fetal
weight at 13,104 mg/m3 and above. A dose-related increase in cervical ribs (small ossification
sites lateral to the seventh cervical vertebra) was significant at 2,620 mg/m3 and above.
Therefore, a NOAEL of 1,310 mg/m3 for developmental toxicity in mice was identified in this
study.
Because this is a developmental study, the NOAEL of 1,310 mg/m3 was not adjusted for
continuous exposure. A NOAELj^ was calculated as per EPA's inhalation dosimetry
methodology (1994), using equation 4-48a (category 3 - extrarespiratory effects). An uncertainty
factor of 100 was applied: 10 for extrapolation from humans to animals and 10 for human
variability.
Conversion Factors:
Dose levels are:
(1,000 ppm x 32.04)/ 24.45 = 1,310 mg/m3; 2,000 ppm = 2,620 mg/m3; 5,000 ppm = 6,552
mg/m3; 7,500 ppm = 9,828 mg/m3; 10,000 ppm = 13,104 mg/m3; 15,000 ppm = 19,656 mg/m3
= NOAEL x RGDR
= NOAEL x (Hb/g)A/(Hb/g)H
=1310 mg/m3 x 1 = 1310 mg/m3
where
NOAELjjEc is the human equivalent concentration NOAEL, RGDR is the regional gas dose
ratio (animal:human) and (Hb/g)A/(Hb/g)H is the ratio of blood:gas partition coefficient;
(Hb/g)A/(Hb/g)H defaults to 1 where Hb/g values are not known.
Additional Information:
Developmental effects were also reported in a study by Nelson et al. (1985). Pregnant
Sprague-Dawley rats were exposed to methanol at concentrations of 0, 5,000, 10,000, and 20,000
ppm (0, 6,552, 13,104, and 26,208 mg/m3) 7 h/d on days 1 through 19 of gestation (high dose rats
were exposed on Gd 7-15 only). Dams were sacrificed on Day 20. Half of the fetuses were
examined for visceral defects, and the other half were examined for skeletal defects. No effect
on the numbers of corpora lutea or implantations or the percentage of dead or resorbed fetuses
was observed. At the two highest concentrations, a dose-related decrease in fetal weights was
reported. The highest concentration of methanol produced slight maternal toxicity and a high
incidence of congenital malformations (p<0.001), predominantly extra or rudimentary cervical
ribs and urinary or cardiovascular defects. Similar malformations were seen in the 10,000 ppm
group, but the incidence was not significantly different from controls. No adverse effects were
noted in the 6552 mg/m3 group (Nelson et al., 1985).
Strengths and Uncertainties:
The major strengths of the Rogers et al. (1993) study are the identification of a NOAEL and the
demonstration of a dose-response relationship. The study was well performed, large numbers of
animals were used (n=20-44 per group), and effects at six exposure concentrations were
examined. The results are also supported by an additional developmental study (Nelson et al.,
1985).
D-21
-------
Volume II Appendix D
The major uncertainties of the RfC are the lack of human data for chronic inhalation exposure
and the lack of comprehensive, long-term muliple dose studies.
References:
California Environmental Protection Agency (CalEPA). 1997. Technical support document for
the determination of noncancer chronic reference exposure levels, Draft for Public Review.
Office of Environmental Health Hazard Assessment, Air Toxicology and Epidemiology Section,
Berkeley, CA.
Nelson, B.K., W.S. Brightwell, D.R. MacKenzie, et al. 1985. Teratological assessment of
methanol and ethanol at high inhalation levels in rats. Fundam Appl Toxicol 5:727-736.
Rogers, J.M., M.L. Mole, N. Chernoff, et al. 1993. The developmental toxicity of inhaled
methanol in the CD-I mouse, with quantitative dose-response modeling for estimation of
benchmark doses. Teratology 47(3): 175-188.
U.S. Environmental Protection Agency. 1994. Methods for derivation of inhalation reference
concentrations and application of inhalation dosimetry. Research Triangle Park, NC:
Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment,
Office of Research and Development, U.S. EPA. EPA/600/8-90-066F.
D-22
-------
Volume II
Appendix D
RfC:
Basis for RfC:
Critical Study:
Critical Dose:
Critical Effect:
Species:
Route of Exposure:
Duration:
Uncertainty Factor:
Modifying Factor:
2-Methoxyethanol Acetate
CAS #110-49-6
0.03 mg/m3
Calculated from RfC for 2-methoxyethanol
Miller, R.R., J.A. Ayres, J.T. Young, and MJ. McKenna. 1983.
Ethylene glycol monomethyl ether. I. Subchronic vapor inhalation
study with rats and rabbits. FundAppl Toxicol 3(l):49-54.
17 mg/m3 (2-methoxyethanol)
[X]NOAEL []LOAEL
Testicular effects (2-methoxyethanol)
Rabbit and rat
Inhalation
13 weeks
1,000:
10 for protection of sensitive humans
10 for use of a subchronic study
10 to account for extrapolation from animals to humans and for
database deficiencies
1
Summary of Study:
Groups of New Zealand white rabbits (5/sex/dose) and Sprague-Dawley rats (10/sex/dose) were
exposed to 0, 30, 100, or 300 ppm 2-methoxyethanol (0, 93, 311, or 934 mg/m3) 6 h/d, 5 d/wk for
13 weeks (Miller et al., 1983). The duration-adjusted exposure concentrations were 0, 17, 56,
and 167 mg/m3, respectively. Toxicity was assessed by clinical observations, body and organ
weights, hematology, clinical chemistry, urinalysis (rats only), and gross and histopathological
examination of major organs including respiratory tract, heart, liver, kidney, bone marrow, testes,
uterus, and ovaries.
Effects reported in both sexes of rabbits exposed to 300 ppm included reduced body weight,
hematological changes (pancytopenia), lymphoid tissue atrophy (thymus), and a significant
decrease in testicular weight with small flaccid testes in the males. A slight to moderate decrease
in testes size was also reported in 2/5 and 4/5 male rabbits exposed to 30 and 100 ppm,
respectively. Microscopic lesions included degenerative changes in the germinal epithelium of
the testes in 3/3, 3/5, and 1/5 male rabbits exposed to 300, 100, and 30 ppm, respectively. The
decrease in testes weight was considered to be concentration-dependent in the male rabbits. No
effects on the reproductive organs of the female rabbits were found. Thymus weights were
significantly decreased in both sexes exposed to 300 ppm. Based upon the testicular effects in
D-23
-------
Volume II Appendix D
rabbits, a NOAEL of 30 ppm (NOAELADJ=17 mg/m3) and a LOAEL of 100 ppm (LCAFL^^Se
mg/m3) were identified.
The authors reported a significant decrease in body weight in the male rats exposed to 300 ppm
and in the females exposed to 100 ppm or more. Effects reported in both sexes of rats exposed to
300 ppm included hematological changes (pancytopenia), lymphoid tissue atrophy, a decrease in
liver weight, and changes in clinical chemistry parameters. In the 300 ppm group (male and
female rats), the mean values for total serum protein, albumin, and globulins were lower than the
control values. A significant decrease in testicular weight and small flaccid testes were also
reported in the male rats exposed to 300 ppm. Microscopic examination showed moderate to
severe degeneration of the germinal epithelium in the seminiferous tubules at the highest
exposure. There were no microscopic changes in the testes in the animals exposed to 100 or 30
ppm. The authors found no effects in the reproductive organs of the female rats.
Based on the observed testicular effects at 100 ppm (Miller et al., 1983), a NOAEL of 30 ppm for
2-methoxyethanol in the rabbit was identified (NOAELj^c =17 mg/m3). The NOAEL was
adjusted for intermittent exposure. A NOAELj^c was calculated as per EPA's inhalation
dosimetry methodology (1994), using equation 4-48a (category 3 - extrarespiratory effects). An
uncertainty factor of 1000 was applied (10 for protection of sensitive humans, 10 for use of a
subchronic study, and 10 to account for extrapolation from animals to humans and for database
deficiencies). An RfC of 0.02 mg/m3 was calculated for 2-methoxyethanol (U.S. EPA, 1998).
Calculations and Conversion Factors:
Dose levels are:
30 ppm 2-methoxyethanol x (76.09/24.45) = 93 mg/m3 2-methoxyethanol; 100 ppm =311
mg/m3; 300 ppm = 934 mg/m3.
= 93 mg/m3 x (6/24 hours) x (5/7 days) = 17 mg/m3
= NOAEL^j x RGDR
= NOAEL^ x (Hb/g)A/(Hb/g)H
= 17 mg/m3 x 1 = 17 mg/m3
where
NOAELAQj is the adjusted NOAEL, NOAELj^c is the human equivalent concentration
NOAEL, RGDR is the regional gas dose ratio (animal:human), and (Hb/g)A/(Hb/g)H is the ratio
of blood:gas partition coefficients; (Hb/g)A/(Hb/g)H defaults to 1 where Hb/g values are not
known.
RfC for 2-methoxyethanol = NOAHE^ + UF = 17 mg/m3 + 1,000 = 0.02 mg/m3
RfC (2-methoxyethanol) x [MW (2-methoxyethanol acetate) + MW (2-methoxyethanol)] =
RfC (2-methoxyethanol acetate)
0.02 mg/m3 x (118.13/76.09) = 0.03 mg/m3 = RfC (2-methoxyethanol acetate)
Additional Information:
The RfC of 0.02 mg/m3 for 2-methoxyethanol (U.S. EPA, 1998) is based on male reproductive
effects in New Zealand rabbits. It is assumed that 2-methoxyethanol acetate would be equitoxic
to 2-methoxyethanol on a molar basis, since both compounds are metabolized to methoxyacetic
D-24
-------
Volume II Appendix D
acid by the alcohol dehydrogenase pathway (DOL-OSHA, 1993). The RfC for 2-methoxyethanol
acetate was calculated from the RfC for 2-methoxyethanol of 0.02 mg/m3 (U.S. EPA, 1998),
resulting in an RfC of 0.03 mg/m3 for 2-methoxyethanol acetate.
Strengths and Uncertainties:
The strength of the RfC is that it is based on an RfC in IRIS that has undergone Agency review.
The major uncertainty of the RfC for 2-methoxyethanol acetate is that it is based on an RfC for
2-methoxyethanol and converted using the molecular weight ratio. However, it is assumed that
2-methoxyethanol acetate would be equitoxic to 2-methoxyethanol on a molar basis.
References:
Department of Labor, Occupational Safety and Health Administration (DOL-OSHA). 1993.
Federal Register: Occupational exposure to 2-methoxyethanol, 2-ethoxyethanol and their
acetates (glycol ethers); Proposed rule. 58FR15526. March 23, 1993.
Miller, R.R., J.A. Ayres, J.T. Young, and MJ. McKenna. 1983. Ethylene glycol monomethyl
ether. I. Subchronic vapor inhalation study with rats and rabbits. FundApplToxicol3(l)A9-54.
Research Triangle Institute (RTI). 1996. Assessment of Risks from the Management of Used
Solvents. U.S. Environmental Protection Agency, Office of Solid Waste, Washington, DC. July
1996. [In support of 61FR42317]
U.S. Environmental Protection Agency. 1994. Methods for derivation of inhalation reference
concentrations and application of inhalation dosimetry. Research Triangle Park, NC:
Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment,
Office of Research and Development, U.S. EPA. EPA/600/8-90-066F.
U.S. Environmental Protection Agency. 1996. Federal Register: Hazardous Waste Management
System; Identification and Listing of Hazardous Waste; Solvents; CERCLA Hazardous
Substance Designation and Reportable Quantities; Proposed Rule. 61FR42317-354. August 14,
1996.
U.S. Environmental Protection Agency. 1998. Integrated Risk Information System (IRIS) file
for 2-methoxyethanol.
D-25
-------
Volume II
Appendix D
RfC:
Critical Study:
Critical Dose:
Critical Effect:
Species:
Route of Exposure:
Duration:
Uncertainty Factor:
Modifying Factor:
Vanadium
CAS # 7440-62-2
0.00007 mg/m3
Zenz, C., and B.A. Berg. 1967. Human responses to controlled
vanadium pentoxide exposure. Arch Environ Health 14:709-712.
0.06 mg vanadium/m3
[ ] NOAEL [X] LOAEL
Upper respiratory symptoms (coughing)
Human
Inhalation
8 hours
300:
3 for use of a minimal LOAEL in a human study
10 for use of an acute study
10 for protection of sensitive human subpopulations
1
Calculations:
RfC =
- UF = 0.02 mg/m3 - 300 = 0.00007 mg/m3
Summary of Study:
Nine human volunteers were exposed to vanadium pentoxide at levels of 0.1, 0.25, and 1.0
mg/m3 (0.06, 0.1, and 0.6 mg vanadium/m3) for 8 hours (Zenz and Berg, 1967). Subjects
exposed to 0.06 mg vanadium/m3 experienced no symptoms during or immediately after
exposure. Within 24 hours, considerable mucus formed, which resulted in slight coughing that
increased after 48 hours, subsided within 72 hours, and disappeared completely after 4 days. All
individuals exposed to 0. 1 mg vanadium/m3 developed a loose cough the following morning,
while those individuals exposed to 0.6 mg vanadium/m3 developed frequent coughing that
persisted for 8 days. There were no other signs of irritation, fever, or increased pulse rate.
A LOAEL of 0.06 mg/m3 was selected, since the subjects experienced slight coughing at this
dose. This LOAEL was adjusted for exposure duration (0.02 mg/m3). An uncertainty factor of
300 was applied: 3 for use of a minimal LOAEL in a human study, 10 for use of an acute study
(8 hours' duration), and 10 for protection of sensitive human subpopulations, since this study
involved only nine volunteers who cannot be assumed to be the most sensitive individuals to this
type of effect.
D-26
-------
Volume II Appendix D
Conversion Factors:
Dose levels are:
% vanadium in vanadium pentoxide = 56.02% = 60% = 0.6. 0.1 mg V2O5/m3x 0.6 = 0.06 mg
vanadium/m3; 0.25 mg V2O5/m3 = 0.15 mg vanadium/m3; 1 mg V2O5/m3 = 0.6 mg vanadium/m3.
LOAEL A^ 0.06 mg/m3 x 8/24 h = 0.02 mg/m3
Additional Information:
Other human studies support the critical effect (upper respiratory symptoms, coughing) reported
in the Zenz and Berg (1967) study. Workers exposed to vanadium pentoxide for durations
ranging from 1 day to 6 or more years showed mild respiratory distress, such as cough, wheezing,
chest pain, runny nose, or sore throat (Levy et al., 1984; Musk and Tees, 1982; Orris et al., 1983;
Sjoeberg, 1956; Thomas and Stiebris, 1956, as cited in ATSDR, 1992); quantitative data were
not available for these occupational studies. Animal studies also support the human findings;
monkeys, rats, and rabbits experienced respiratory effects from inhalation exposure to vanadium
pentoxide (Knecht et al., 1985; Lee and Gillies, 1986; Sjoeberg, 1950; as cited in ATSDR, 1992).
Strengths and Uncertainties:
The major strength of the RfC is that it was based on human data with a critical effect that has
been reported in numerous other studies. ATSDR's acute inhalation MRL (0.0002 mg/m3) is
also based on the Zenz and Berg (1967) study. The RfC developed for the Air Characteristic
Study (0.00007 mg/m3) is also supported by the Sjoeberg (1950) study. Dyspnea was reported in
rabbits exposed to 0.8 mg V/m3 (as vanadium pentoxide) 1 h/d for 8 months. Using the Sjoeberg
(1950) study, adjusting for intermittent exposure (LOAEL^j = 0.03 mg V/m3), and applying an
uncertainty factor of 3,000 for animal-to-human extrapolation, use of a LOAEL, human
variability, and extrapolation of subchronic to chronic duration would result in an RfC (0.00001
mg/m3) similar to the RfC based on the Zenz and Berg (1967) study. Although the Zenz and
Berg (1967) study was of brief duration (8 hours), the lowest LOAEL among available data was
identified in this study and quantitative human data are generally preferred over animal data
(when available). Therefore, the RfC of 0.00007 mg/m3 based on the Zenz and Berg (1967) data
is used in the Air Characteristic Study.
The major uncertainty to the RfC is the short duration (8 hours) and the limited number of
individuals (9) in the Zenz and Berg (1967) study. In addition, a NOAEL was not identified in
the study.
References:
Agency for Toxic Substances and Disease Registry (ATSDR). 1992. Toxicological profile for
vanadium. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service.
Knecht, E.A., WJ. Moorman, J.C. Clark, et al. 1985. Pulmonary effects of acute vanadium
pentoxide inhalation in monkeys. Am Rev Respir Dis 132:1181-1185 (as cited in ATSDR, 1992).
Lee, K.P., and PJ. Gillies. 1986. Pulmonary response and intrapulmonary lipids in rats exposed
to bismuth orthovanadate dust by inhalation. Environ Res 40:115-135 (as cited in ATSDR,
1992).
D-27
-------
Volume II Appendix D
Levy, B.S., L. Hoffman, and S. Gottsegen. 1984. Boilmakers'bronchitis. J OccupMed26:
567-570 (as cited in ATSDR, 1992).
Musk, A.W., and J.G. Tees. 1982. Asthma caused by occupational exposure to vanadium
compounds. MedJAust 1:183-184 (as cited in ATSDR, 1992).
Orris, P., J. Cone, and S. McQuilkin. 1983. Health hazard evaluation report HETA 80-096-
1359, Bloomington, IL. Washington, DC: U.S. Department of Health and Human Services,
National Institute of Occupational Safety and Health. NTIS-PB85-163574 (as cited in ATSDR,
1992).
Sjoeberg, S.G. 1950. Vanadium pentoxide dust - a clinical and experimental investigation on its
effects after inhalation. ActaMedScandSupp. 238:1-188 (as cited in ATSDR, 1992).
Sjoeberg, S.G. 1956. Vanadium dust, chronic bronchitis, and possible risk of emphysema. Ada
MedScand 154:381-386 (as cited in ATSDR, 1992).
Thomas, D.L., and K. Stiebris. 1956. Vanadium poisoning in industry. The MedicalJournal of
Australia 1:607-609 (as cited in ATSDR, 1992).
U.S. Environmental Protection Agency. 1994. Methods for derivation of inhalation reference
concentrations and application of inhalation dosimetry. Research Triangle Park, NC:
Environmental Criteria and Assessment Office, Office of Health and Environmental
Assessment, Office of Research and Development, U.S. EPA. EPA/600/8-90-066F.
Zenz, C., and B.A. Berg. 1967. Human responses to controlled vanadium pentoxide exposure.
Arch Environ Health 14:709-712.
D-28
-------
Volume II Appendix D
CARCINOGENS
DERIVATION OF INHALATION UNIT RISK FACTORS
AND CANCER SLOPE FACTORS
D-29
-------
Volume II Appendix D
D-30
-------
Volume II Appendix D
Bromodichloromethane
CAS #75-27-4
Inhalation Unit Risk Factor: 1.8E-05 (jig/m3)'1
Slope Factor: 6.2E-02 (mg/kg/d)4
Critical Effects: Tubular cell adenoma and tubular cell adenocarcinoma
Species: Mice
Route of Exposure: Gavage, corn oil
Duration: 2 years
Basis for Toxicity Values:
EPA has not developed an inhalation reference concentration (RfC) for bromodichloromethane.
An oral reference dose (RfD) value of 0.02 mg/kg/d, based on a chronic gavage study in mice for
renal cytomegaly is available on IRIS for bromodichloromethane (U.S. EPA, 1998).
Based on inadequate human data and sufficient evidence of carcinogenicity in animals, EPA
considers bromodichloromethane a probable human carcinogen (Class B2) by the oral route and
has calculated an oral cancer slope factor (CSF) of 0.062 (mg/kg/d)"1 for the substance. In a
National Toxicology Program (NTP) study, 2-year gavage administration of bromodichloro-
methane to both sexes of F344/N rats and B6C3F1 resulted in compound-related statistically
significant increases in tumors of the kidney in male mice, the liver in female mice, and the
kidney and large intestine in male and female rats (NTP, 1987, as cited in U.S. EPA, 1998).
In male mice, the incidences of tubular cell adenomas and the combined incidence of tubular cell
adenomas and adenocarcinomas of the kidneys were significantly increased in the high-dose
animals. In female mice, there were significant increases of hepatocellular adenomas and
hepatocellular carcinomas. The combined incidence of hepatocellular adenomas or carcinomas
in vehicle control, low-dose, and high-dose groups were 3/50, 18/48, and 29/50, respectively.
In male and female rats, the incidences of tubular cell adenomas, adenocarcinomas, and the
combined incidence of adenomas and adenocarcinomas of the kidneys were statistically
significantly increased only in the high-dose groups. The combined incidence of tubular cell
adenomas or adenocarcinomas in vehicle control, low-dose, and high-dose groups were 0/50,
1/49, and 13/50 for males and 0/50, 1/50, and 15/50 for females, respectively.
Tumors of the large intestines, namely adenocarcinomas and adenomatous polyps, were
significantly increased in male rats in a dose-dependent manner. These large intestinal tumors,
however, were observed only in high-dose female rats (adenocarcinomas 0/46, 0/50, 6/47;
adenomatous polyps 0/46, 0/50, 7/47 in the vehicle control, low-dose and high-dose groups,
respectively). The combined incidence of large intestine adenocarcinomas and/or adenomatous
polyps in vehicle control, low-dose, and high-dose groups were 0/50, 13/49, and 45/50 for males
and 0/46, 0/50, and 12/47 for females. The combined tumor incidences in the large intestine and
kidney in male and female rats at control, low dose, and high dose were 0/50, 13/49, 46/50 and
D-31
-------
Volume II Appendix D
0/46, 1/50, 24/48, respectively. Under the conditions of this bioassay, the NTP concluded there
was clear evidence of carcinogenicity of bromodichloromethane in male and female F344/N rats
and B6C3F1 mice (U.S. EPA, 1998).
The mechanism for the carcinogenicity of bromodichloromethane appears to be genotoxic
carcinogenesis, independent of liver activation and, hence, route-independent. In one
genotoxicity assay, bromodichloromethane was mutagenic in Salmonella typhimurium strain
TA100 in the absence of liver homogenate in a vapor phase test performed in a desiccator.
Positive results for mutagenicity were reported for bromodichloromethane in other
S. typhimurium assays in which the TA100 and TA1537 strains were used without rat liver
homogenate activation. Bromodichloromethane also induced weak mutagenic effects in
Saccharomyces cerevisiae strains D7 and XV185-14C in the absence of liver homogenate (U.S.
EPA, 1998; HSDB, 1998).
Thus, inhalation exposure to bromodichloromethane is likely to lead to carcinogenic
consequences not dissimilar from that from oral exposure. Therefore, in accordance with current
EPA guidelines, it is considered appropriate to calculate an inhalation unit risk factor for
bromodichloromethane from the oral CSF listed for that substance in IRIS (U.S. EPA, 1994,
1996).
Calculations:
URF = CSF x 1 mg/1,000 //g x 1/70 kg x 20 m3/day =
0.062 (mg/kg/d)'1 x 1 mg/1,000 //g x 1/70 kg x 20 m3/d = LSE-OS^g/m3)'1
where
70 kg = default adult human body weight
20 m3 = default adult human daily rate of inhalation
Calculations assume 100% absorption.
Additional Information:
Inhalation CSFs are often derived from oral data. Of the 51 chemicals currently listed in IRIS
(U.S. EPA, 1998) and HEAST (U.S. EPA, 1997) that have both an oral and inhalation CSF,
about 60% of the inhalation CSFs were derived from oral studies and are identical or essentially
identical to the oral CSF (see Table D-l, Figure D-l). In at least one case (benzene), the oral
CSF was based on inhalation data resulting in identical values for both routes of exposure. In
most cases (>75%) where an inhalation CSF was derived from an inhalation study, the inhalation
CSF was lower than the corresponding oral CSF. Therefore, use of an oral CSF as an interim
inhalation CSF appears reasonable and is unlikely to result in underestimating risk.
References:
Hazardous Substances Databank (HSDB): Bromodichloromethane. 1998. Online database.
National Library of Medicine, Bethesda, MD.
National Toxicology Program (NTP). 1987. NTP Technical Report on the Toxicology and
Carcinogenesis Studies of Bromodichloromethane (CAS no. 75-27-4) in F344/N Rats and
B6C3F1 Mice (gavage studies). NTP Tech. Report Series No.321. U.S. Dept. Health and Human
Services, Public Health Service, National Institute of Health (as cited in U.S. EPA, 1998).
D-32
-------
Volume II Appendix D
U.S. Environmental Protection Agency. 1994. Provisional Guidance for the Qualitative Risk
Assessment of Poly cyclic Aromatic Hydrocarbons. Prepared by the Environmental Criteria and
Assessment Office, Office of Health and Environmental Assessment, Cincinnati, OH, for the
Office of Research and Development, Cincinnati, OH. EPA/600/R-93.
U.S. Environmental Protection Agency. 1996. Proposed Guidelines for Carcinogen Risk
Assessment. Office of Research and Development. Washington, DC. EPA/600/P-92/003C.
U.S. Environmental Protection Agency. 1997. Health Effects Assessment Summary Tables
(HEAST), FY 1997 Update. Office of Emergency and Remedial Response, Washington, DC.
EPA-540-R-97-036.
U.S. Environmental Protection Agency. 1998. Integrated Risk Information System (IRIS).
Bromodichloromethane. Environmental Criteria and Assessment Office, Office of Health and
Environmental Assessment, Cincinnati, OH.
D-33
-------
Volume II Appendix D
Chlorodibromomethane
CAS #124-48-1
Inhalation Unit Risk Factor: 2.4E-05 (|ig/m3)4
Slope Factor: 8.4E-02 (mg/kg/d)'1
Critical Effects: Hepatocellular adenoma or carcinoma
Species: Mice
Route of Exposure: Gavage
Duration: 2 years
Basis for Toxicity Values:
EPA has not developed an inhalation reference concentration (RfC) for Chlorodibromomethane.
An oral reference dose (RfD) value of 0.02 mg/kg/d, based on a subchronic gavage study in rats
for hepatic lesions is available on IRIS for Chlorodibromomethane (U.S. EPA, 1998)
Based on inadequate human data and limited evidence of carcinogenicity in animals, EPA
considers Chlorodibromomethane a possible human carcinogen (Class C) by the oral route and
has calculated an oral cancer slope factor (CSF) of 0.084 (mg/kg/d)"1 for the substance. In the
study, 2-year gavage administration of Chlorodibromomethane to both sexes of B6C3F1 mice
caused increased incidence of adenomas and carcinomas in female mice and a significantly
increased incidence of hepatocellular carcinomas in high-dose male mice (NTP, 1985, as cited in
U.S. EPA, 1998). Drinking water administration of Chlorodibromomethane to both sexes of
CBAxC57Bl/6 mice also resulted in significantly increased incidence of tumors (U.S. EPA, 1998).
The mechanism for the carcinogenicity of Chlorodibromomethane appears to be genotoxic
carcinogenesis, independent of liver activation and, hence, route-independent. In one
genotoxicity assay, Chlorodibromomethane produced reverse mutations in Salmonella
typhimurium strain TA100 in a vapor-phase test performed in a desiccator. Positive results for
gene conversion in Saccharomyces cerevisiae strain D4 without, but not with, hepatic
homogenates, and negative results for mutation in strain XV185-14C both with and without
hepatic homogenates have been reported for Chlorodibromomethane. In others tests,
Chlorodibromomethane produced sister chromatid exchange in cultured human lymphocytes and
in bone marrow cells of mice treated orally (U.S. EPA, 1998; HSDB, 1998).
Thus, inhalation exposure to Chlorodibromomethane is likely to lead to carcinogenic
consequences not dissimilar from that from oral exposure. Therefore, in accordance with current
EPA guidelines, it is considered appropriate to calculate an inhalation unit risk factor for
Chlorodibromomethane from the oral CSF listed for that substance in IRIS (U.S. EPA, 1994,
1996).
D-34
-------
Volume II Appendix D
Calculations:
URF = CSF x 1 mg/1,000 //g x 1/70 kg x 20 m3/d =
0.084 (mg/kg/d)'1 x 1 mg/1,000 //g x 1/70 kg x 20 m3/d = 2.4E-05(yag/m3)-1
where
70 kg = default adult human body weight
20 m3 = default adult human daily rate of inhalation
Calculations assume 100% absorption.
Additional Information:
Inhalation CSFs are often derived from oral data. Of the 51 chemicals currently listed in IRIS
(EPA, 1998) and HEAST (U.S. EPA, 1997) that have both an oral and inhalation CSF, about
60% of the inhalation CSFs were derived from oral studies and are identical or essentially
identical to the oral CSF (see Table D-l, Figure D-l). In at least one case (benzene), the oral
CSF was based on inhalation data resulting in identical values for both routes of exposure. In
most cases (>75%) where an inhalation CSF was derived from an inhalation study, the inhalation
CSF was lower than the corresponding oral CSF. Therefore, use of an oral CSF as an interim
inhalation CSF appears reasonable and is unlikely to result in underestimating risk.
References:
Hazardous Substances Databank (HSDB): Chlorodibromomethane. 1998. Online database.
National Library of Medicine, Bethesda, MD.
National Toxicology Program (NTP). 1985. Toxicology and Carcinogenesis Studies of
Chlorodibromomethane in F344/N Rats and B6C3F1 Mice (gavage studies). NTP TR282 (as
cited in U.S. EPA, 1998).
U.S. Environmental Protection Agency. 1994. Provisional Guidance for the Qualitative Risk
Assessment of Poly cyclic Aromatic Hydrocarbons. Prepared by the Environmental Criteria and
Assessment Office, Office of Health and Environmental Assessment, Cincinnati, OH, for the
Office of Research and Development, Cincinnati, OH.
U.S. Environmental Protection Agency. 1996. Proposed Guidelines for Carcinogen Risk
Assessment. Office of Research and Development. Washington, DC. EPA/600/P-92/003C.
U.S. Environmental Protection Agency. 1997. Health Effects Assessment Summary Tables
(HEAST), FY 1997 Update. Office of Emergency and Remedial Response, Washington, DC.
EPA-540-R-97-036.
U.S. Environmental Protection Agency. 1998. Integrated Risk Information System (IRIS).
Chlorodibromomethane. Environmental Criteria and Assessment Office, Office of Health and
Environmental Assessment, Cincinnati, OH.
D-35
-------
Volume II Appendix D
7,12-Dimethy Ibenz [a] anthracene
CAS # 57-97-6
Unit Risk Factor: 2.4E-02 (jig/m3)'1
Slope Factor: 8.4E+01 (mg/kg/d)'1
Critical Effects: Malignant angioendothelioma of the mesenteric intestine
Species: Mouse (albino)
Route of Exposure: Diet
Duration: 60 weeks
Basis for Toxicity Values:
There are no human data available that may be used to address the carcinogenicity of
7,12-dimethylbenz[a]anthracene (DMBA). However, DMBA belongs to a class of chemicals
known as polycyclic aromatic hydrocarbons (PAHs), which are components of coal tar and
incomplete combustion. Many of the PAHs have been demonstrated to be carcinogenic to rats
and mice following oral exposure, skin painting, intrapulmonary injection, inhalation,
subcutaneous injection, and intraperitoneal injection; however, most of these studies are not
considered suitable for quantitative risk assessment. Nevertheless, the data do indicate that the
carcinogenic potencies vary and that DMBA is considered one of the most potent PAHs (Pitot
andDragan, 1996).
DMBA is not listed in EPA's IRIS (U.S. EPA, 1998) or HEAST (U.S. EPA, 1997) databases and
was not included in EPA's (1993) Provisional Guidance for Quantitative Risk Assessment of
PAHs. However, the California Environmental Protection Agency (CalEPA) has developed a
unit risk factor (URF) and cancer slope factor (CSF) for DMBA in support of the Air Toxics Hot
Spots Program (CalEPA, 1994a, 1994b, 1997). The CalEPA URF and inhalation CSF are listed
above and are recommended as interim values.
The CalEPA developed an "expedited" approach for deriving cancer potency values in order to
implement Proposition 65 (Hoover et al., 1995). The expedited approach was used for DMBA.
Under the expedited approach, instead of conducting a comprehensive literature review, cancer
dose response data are taken from the Carcinogenic Potency Database (CPDB) (Gold and Zeiger,
1997). The linearized multistage model is automatically used to derive cancer potency estimates
for low-dose exposures, and pharmacokinetic adjustments are not made.
Only one study was listed in the CPDB (Chouroulinkov et al., 1967). Female albino mice were
fed DMBA for 60 weeks at a dose rate of 0.39 mg/kg/d. No tumors were reported in 40 control
mice. Malignant angioendotheliomas of the intestine were reported in 49 of 75 test animals.
Twenty test animals also had nonmalignant forestomach papillomas.
D-36
-------
Volume II Appendix D
Additional Information:
The CPDB summarizes the results of 5,152 cancer tests on 1,298 chemicals. Carcinogenic
potency estimates are presented as TD50 values. TD50 is defined as that dose-rate in mg/kg body
wt/d which, if administered chronically for the standard lifespan of the species, will halve the
probability of remaining tumorless throughout that period (Gold and Zeiger, 1997). The TD50 is
analogous to the dose that is lethal to 50 percent of test animals (LD50). A low TD50 indicates
high potency, just as a low LD50 indicates high acute toxicity.
Some studies have reported high correlations between various measures of cancer potency and
the maximum tolerated dose or maximum dose tested in the carcinogenicity studies (Gaylor,
1989; Krewski et al., 1993). The correlation of TD50 values as reported in the CPDB and
inhalation CSFs derived from IRIS or HEAST was evaluated as a possible means to estimate the
CSF from the TD50. Forty-five chemicals were identified that had both a TD50 and an inhalation
CSF (see Table D-2, Figure D-2). The correlation coefficient for the regression is 0.95. The
TD50 reported for DMBA is 0.084 mg/kg/d (Gold and Zeiger, 1997). Based on a linear
regression of log TD50 as the independent variable and log (1/CSF) as the dependent variable, an
inhalation CSF of 55 (mg/kg/d)"1 and a URF of 1.6E-02 (jig/m3)"1 are predicted. These values are
in close agreement with the CalEPA values of 84 (mg/kg/d)"1 and 2.4E-02(//g/m3)"1, respectively.
References:
California Environmental Protection Agency (CalEPA). 1994a. Benzo[a]pyrene as a Toxic Air
Contaminant. Executive Summary. California Air Resources Board, Office of Environmental
Health Hazard Assessment, Berkeley, CA.
California Environmental Protection Agency (CalEPA). 1994b. Benzo[a]pyrene as a Toxic Air
Contaminant. Part B Health Effects of Benzo(a)pyrene. California Air Resources Board, Office
of Environmental Health Hazard Assessment, Berkeley, CA.
California Environmental Protection Agency (CalEPA). 1997. Air Toxics Hot Spots Program
Risk Assessment Guidelines: Technical Support Document for Determining Cancer Potency
Factors. Draft for Public Comment. Office of Environmental Health Hazard Assessment.
Chouroulinkov, I, A. Gentil, and M. Guerin. 1967. Etude de 1'activite carcinogene du 9,10-
dimethyl-benzanthracene et du 3,4-benzopyrene administres par voie digestive. Bull Cancer
54:67-78 (as cited in Gold and Zeiger, 1997).
Gaylor, D.W. 1989. Preliminary estimates of the virtually safe dose for tumors obtained from
the maximum tolerated dose. Regulatory Toxicology and Pharmacology 9:1-18.
Gold, L.S., and E. Zeiger (eds). 1997. Handbook of Carcinogenic Potency and Genotoxicity
Databases. Boca Raton, FL: CRC Press, 754 pp.
Hoover, S.M., L. Zeise, W.S. Pease, et al. 1995. Improving the regulation of carcinogens by
expediting cancer potency estimation. Risk Analysis 15(2):267-280.
D-37
-------
Volume II Appendix D
Krewski, D., D.W. Gaylor, A.P. Soms, and M. Szyszkowicz. 1993. An overview of the report:
Correlation between carcinogenic potency and the maximum tolerated dose: Implications for risk
assessment. Risk Analysis 13(4):383-398.
Pitot, H.C., III, and Y.P. Dragan. 1996. Chemical Carcinogenesis. In: Casarett & Doull's
Toxicology the Basic Science of Poisons. 5th edition. C.D. Klaassen (ed). New York: McGraw-
Hill, pp. 202-203.
U.S. Environmental Protection Agency. 1993. Provisional Guidance for Quantitative Risk
Assessment of Poly cyclic Aromatic Hydrocarbons. Environmental Criteria and Assessment
Office, Office of Health and Environmental Assessment, Cincinnati, OH. EPA/600/R-93/089.
U.S. Environmental Protection Agency. 1997. Health Effects Assessment Summary Tables
(HEAST), FY 1997 Update. Office of Emergency and Remedial Response, Washington, DC.
EPA-540-R-97-036.
U.S. Environmental Protection Agency. 1998. Integrated Risk Information System (IRIS).
Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment,
Cincinnati, OH.
D-38
-------
Volume II Appendix D
2,4-Dinitrotoluene
CAS #121-14-2
Unit Risk Factor: 1.9E-04 (jig/m3)'1
Slope Factor: 6.8E-01 (mg/kg/d)4
Critical Effects: Hepatocellular carcinoma, liver neoplastic nodules, benign and
malignant mammary gland tumors.
Species: Female Sprague-Dawley rats
Route of exposure: Diet
Duration: 2 years
Basis for Toxicity Values:
There are no human data available that may be used to address the carcinogenicity of
2,4-dinitrotoluene. 2,4-Dinitrotoluene is not listed in EPA's IRIS (U.S. EPA, 1998) or HEAST
(U.S. EPA, 1997) databases. However, an oral CSF of 0.68 (mg/kg/d)4 is available in IRIS for a
mixture of 2,4-and 2,6-dinitrotoluene. The mixture was 98% 2,4-dinitrotoluene and 2%
2,6-dinitrotoluene. The oral CSF for the mixture is proposed as an interim value for the
inhalation CSF for 2,4-dinitrotoluene.
Inhalation CSFs are often derived from oral data. Of the 51 chemicals currently listed in IRIS
and HEAST that have both an oral and inhalation CSF, about 60% of the inhalation CSFs were
derived from oral studies and are identical or essentially identical to the oral CSF (see Table D-l,
Figure D-l). In at least one case (benzene), the oral CSF was based on inhalation data resulting
in identical values for both routes of exposure. In most cases (>75%) where an inhalation CSF
was derived from an inhalation study, the inhalation CSF was lower than the corresponding oral
CSF. Therefore, use of an oral CSF as an interim inhalation CSF appears reasonable and is
unlikely to result in underestimating risk.
Dose-Response Data:
The oral CSF listed in HEAST was based on a study by Ellis et al. (1979). Sprague-Dawley rats
were fed dietary concentrations of 0, 15, 100, and 700 ppm and Swiss mice were fed 0, 100, 700,
and 5,000 ppm for 2 years. Mortality was high in all treatment groups. A statistically significant
increase in liver tumors was observed in both male and female rats and a statistically significant
increase in benign mammary gland tumors was observed in female rats. In addition, an increased
incidence of kidney tumors was observed in the mid-dose male mice. Data used to derive the
CSF were based on liver and mammary tumors in female rats and are presented below as
reported in IRIS.
D-39
-------
Volume II
Appendix D
Administered Dose
(ppm)
0
15
100
700
Human Equivalent Dose
(mg/kg/d)
0
0.129
0.927
7.557
Tumor Incidence
11/23
12/35
17/27
34/35
Calculations:
URF = CSF x 1 mg/1,000 |ig x 1/70 kg x 20 m3/d =
0.68 (mg/kg/d)'1 x 1 mg/1,000 |ig x 1/70 kg x 20 m3/d = 1.9E-04(|ig/m3)-1
where
70 kg = default adult human body weight
20 m3 = default adult human daily rate of inhalation
Calculations assume 100% absorption.
Additional Information:
The California Environmental Protection Agency (CalEPA) adopted a URF of 8.9E-05 (jig/m3)4
and an inhalation CSF of 3.1E-01 (mg/kg/d)'1 for practical grade 2,4-dinitrotoluene based on a
potency factor derived by EPA (U.S. EPA, 1987) (CalEPA, 1997). These values were based on a
feeding study using Sprague-Dawley rats (Lee et al., 1978). Liver and mammary tumors in
female rats were used to develop the CSF and results were very similar to the Ellis et al. (1979)
study discussed above.
The Carcinogenic Potency Database (CPDB) summarizes the results of 5,152 cancer tests on
1,298 chemicals (Gold and Zeiger, 1997). Carcinogenic potency estimates are presented as TD50
values. TD50 values are defined as that dose-rate in mg/kg body wt/day which, if administered
chronically for the standard life span of the species, will halve the probability of remaining
tumorless throughout that period (Gold and Zeiger, 1997). The TD50 is analogous to the dose
that is lethal to 50% of test animals (LD50). A low TD50 indicates high potency, just as a low
LD50 indicates high acute toxicity.
Some studies have reported high correlations between various measures of cancer potency and
the maximum tolerated dose or maximum dose tested in the carcinogenicity studies (Gaylor,
1989; Krewski et al., 1993). The correlation of TD50s as reported in the CPDB and inhalation
CSFs derived from IRIS or HEAST was evaluated as a possible means to estimate the CSF from
the TD50. Forty-five chemicals were identified that had both a TD50 and an inhalation CSF (see
Table D-2, Figure D-2). The correlation coefficient for the regression is 0.95. The reported TD50
is 9.35 mg/kg/d (Gold and Zeiger, 1997). Based on a linear regression of log TD50 as the
independent variable and log (1/CSF) as the dependent variable, an inhalation CSF of 0.53
(mg/kg/d)'1 and a URF of 1.5E-04 (jig/m3)'1 are predicted. These values are in close agreement
with the oral CSF listed in IRIS for a mixture of 2,4- and 2,6-dinitrotoluene and the CalEPA
values.
D-40
-------
Volume II Appendix D
References:
California Environmental Protection Agency (CalEPA). 1997. Air Toxics Hot Spots Program
Risk Assessment Guidelines: Technical Support Document for Determining Cancer Potency
Factors. Draft for Public Comment. Office of Environmental Health Hazard Assessment.
Ellis, H.V., III, J.H. Hagensen, J.R. Hodgson, et al. 1979. Mammalian toxicity of munitions
compounds. Phase III: Effects of life-time exposure. Part I: 2,4-dinitrotoluene. Final report No.
7. U.S. Army Medical Bioengineering Research and Development Laboratory. Midwest
Research Institute. Report Order No. AD-A077692.
Gaylor, D.W. 1989. Preliminary estimates of the virtually safe dose for tumors obtained from the
maximum tolerated dose. Regulatory Toxicology and Pharmacology 9:1-18.
Gold, L.S., and E. Zeiger (eds). 1997. Handbook of Carcinogenic Potency and Genotoxicity
Databases. Boca Raton, FL: CRC Press, 754 pp.
Krewski, D., D.W. Gaylor, A.P. Soms, and M. Szyszkowicz. 1993. An overview of the report:
Correlation between carcinogenic potency and the maximum tolerated dose: Implications for risk
assessment. Risk Analysis 13(4):383-398.
Lee, C.C., H.V. Ellis, JJ. Kowalski, et al. 1978. Mammalian toxicity of munition compounds.
Phase II. Effects of multiple doses and Phase III. Effects of lifetime exposure. Part II.
2,4-Dinitrotoluene. U.S. Army Medical Bioengineering Research and Development Laboratory.
Midwest Research Institute, Kansas City, MO. NTIS ADA 061715.
U.S. Environmental Protection Agency. 1987. Health Effects Assessment for 2,4- and 2,6-
Dinitrotoluene. Office of Health and Environmental Assessment, Cincinnati, OH. EPA/600/8-
88/032.
U.S. Environmental Protection Agency. 1997. Health Effects Assessment Summary Tables
(HEAST), FY 1997 Update. Office of Emergency and Remedial Response, Washington, DC.
EPA-540-R-97-036.
U.S. Environmental Protection Agency. 1998. Integrated Risk Information System (IRIS).
Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment,
Cincinnati, OH.
D-41
-------
Volume II Appendix D
3-Methylcholanthrene
CAS # 56-49-5
Unit Risk Factor: 2.1E-03 (jig/m3)'1
Slope Factor: 7.4E+00 (mg/kg/day)4
Critical Effects: Mammary gland adenocarcinomas
Species: Wistar rats
Route of Exposure: Gavage
Duration: 26 to 52 weeks
Basis for Toxicity Values:
There are no human data available that may be used to address the carcinogenicity of
3-methylcholanthrene (3-MC). However, 3-MC belongs to a class of chemicals known as
polycyclic aromatic hydrocarbons (PAHs), which are components of coal tar and incomplete
combustion. Many of the PAHs have been demonstrated to be carcinogenic to rats and mice
following oral exposure, skin painting, intrapulmonary injection, inhalation, subcutaneous
injection, and intraperitoneal injection; however, most of these studies are not considered suitable
for quantitative risk assessment. Nevertheless, the data do indicate that the carcinogenic
potencies vary and that 3-MC is considered one of the most potent PAHs (Pitot and Dragan,
1996).
3-MC is not listed in EPA's IRIS (U.S. EPA, 1998) or HEAST (U.S. EPA, 1997) databases and
was not included in EPA's (1993) Provisional Guidance for Quantitative Risk Assessment of
PAHs. However, the California Environmental Protection Agency (CalEPA) has developed a
unit risk factor (URF) and cancer slope factor (CSF) for 3-MC in support of the Air Toxics Hot
Spots Program (CalEPA, 1994a, 1994b, 1997). The CalEPA URF and inhalation CSF are listed
above and recommended as interim values.
The CalEPA developed an "expedited" approach for deriving cancer potency values in order to
implement Proposition 65 (Hoover et al., 1995). The expedited approach was used for 3-MC.
Under the expedited approach, instead of conducting a comprehensive literature review, cancer
dose response data are taken from the Carcinogenic Potency Database (CPDB) (Gold and Zeiger,
1997). The linearized multistage model is automatically used to derive cancer potency estimates
for low-dose exposures, and pharmacokinetic adjustments are not made.
Fifteen studies (4 diet and 11 gavage) were listed in the CPDB (Gold and Zeiger, 1997). All of
the studies included a control group and one treatment group. No tumors were reported in any of
the dietary studies; however, a significant increase in tumors was reported in all of the gavage
studies. Doses for the gavage studies ranged from 2.46 mg/kg/d to 12.2 mg/kg/d.
Adenocarcinomas of the mammary gland were reported in nine studies and two studies identified
unspecified mammary tissue tumors. Tumor incidence ranged from 67% to 100%.
D-42
-------
Volume II Appendix D
Additional Information:
The CPDB summarizes the results of 5,152 cancer tests on 1,298 chemicals. Carcinogenic
potency estimates are presented as TD50s. TD50s are defined as that dose-rate in mg/kg body
wt/day which, if administered chronically for the standard lifespan of the species, will halve the
probability of remaining tumorless throughout that period (Gold and Zeiger, 1997). The TD50 is
analogous to the dose that is lethal to 50% of test animals (LD50). A low TD50 indicates high
potency, just as a low LD50 indicates high acute toxicity.
Some studies have reported high correlations between various measures of cancer potency and
the maximum tolerated dose or maximum dose tested in the carcinogenicity studies (Gaylor,
1989; Krewski et al., 1993). The correlation of TD50s as reported in the CPDB and inhalation
CSFs derived from IRIS or HEAST was evaluated as a possible means to estimate the CSF from
the TD50. Forty-five chemicals were identified that had both a TD50 and an inhalation CSF (see
Table D-2, Figure D-2). The correlation coefficient for the regression is 0.95. The TD50 reported
for 3-MC is 0.491 mg/kg/d (Gold and Zeiger, 1997). Based on a linear regression of log TD50 as
the independent variable and log (1/CSF) as the dependent variable, an inhalation CSF of 9.6
(mg/kg/d)"1 and a URF of 2.7E-03 (jig/m3)"1 are predicted. These values are in close agreement
with the CalEPA values of 7.4 (mg/kg/d)"1 and 2.1E-03 (//g/m3)"1, respectively.
References:
California Environmental Protection Agency (CalEPA). 1994a. Benzo[a]pyrene as a Toxic Air
Contaminant. Executive Summary. California Air Resources Board, Office of Environmental
Health Hazard Assessment, Berkeley, CA.
California Environmental Protection Agency (CalEPA). 1994b. Benzo[a]pyrene as a Toxic Air
Contaminant. Part B Health Effects of Benzo(a)pyrene. California Air Resources Board, Office
of Environmental Health Hazard Assessment, Berkeley, CA.
California Environmental Protection Agency (CalEPA). 1997. Air Toxics Hot Spots Program
Risk Assessment Guidelines: Technical Support Document for Determining Cancer Potency
Factors. Draft for Public Comment. Environmental Criteria and Assessment Office, Office of
Environmental Health Hazard Assessment.
Gaylor, D.W. 1989. Preliminary estimates of the virtually safe dose for tumors obtained from the
maximum tolerated dose. Regulatory Toxicology and Pharmacology 9:1-18.
Gold, L.S., and E. Zeiger (eds). 1997. Handbook of Carcinogenic Potency and Genotoxicity
Databases. Boca Raton, FL: CRC Press, 754pp.
Hoover, S.M., L. Zeise, W.S. Pease, et al. 1995. Improving the regulation of carcinogens by
expediting cancer potency estimation. Risk Analysis 15(2):267-280.
Krewski, D., D.W. Gaylor, A.P. Soms, and M. Szyszkowicz. 1993. An overview of the report:
Correlation between carcinogenic potency and the maximum tolerated dose: Implications for risk
assessment. Risk Analysis 13(4):383-398.
D-43
-------
Volume II Appendix D
Pitot, H.C., III, and Y.P. Dragan. 1996. Chemical Carcinogenesis. In: Casarett & Doull's
Toxicology the Basic Science of Poisons. 5th edition. C.D. Klaassen (ed). New York:
McGraw-Hill, pp. 202-203.
U.S. Environmental Protection Agency. 1993. Provisional Guidance for Quantitative Risk
Assessment of Poly cyclic Aromatic Hydrocarbons. Environmental Criteria and Assessment
Office, Office of Health and Environmental Assessment, Cincinnati, OH. EPA/600/R-93/089.
U.S. Environmental Protection Agency. 1997. Health Effects Assessment Summary Tables, FY
1997 Update. Office of Emergency and Remedial Response, Washington, DC. EPA-540-R-97-
036.
U.S. Environmental Protection Agency. 1998. Integrated Risk Information System (IRIS).
Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment,
Cincinnati, OH.
D-44
-------
Volume II
Appendix D
o-Toluidine (2-Methylaniline)
CAS # 95-53-4
Unit Risk Factor:
Slope Factor:
Critical Effects:
Species:
Route of Exposure:
Duration:
6.9E-05
2.4E-01 (mg/kg/d)4
Skin fibromas - also increased incidence of other tumor types
including sarcomas, mesotheliomas, carcinomas, hemangiosarcomas,
and hepatocellular carcinomas of various tissues.
F-344 rats and B6C3F1 mice
Diet
2 years
Basis for Toxicity Values:
There is limited evidence that o-toluidine is carcinogenic in humans; however, data are
inadequate for a quantitative risk assessment (U.S. EPA, 1987). o-Toluidine is not listed in
EPA's IRIS (U.S. EPA, 1998) but an oral CSF is included in HEAST (U.S. EPA, 1997). The
oral CSF of 2.4E-01 mg/kg/d is proposed as an interim value for the inhalation CSF.
Inhalation CSFs are often derived from oral data. Of the 51 chemicals currently listed in IRIS
and HEAST that have both an oral and inhalation CSF, about 60% of the inhalation CSFs were
derived from oral studies and are identical or essentially identical to the oral CSF (see Table D-l,
Figure D-l). In at least one case (benzene), the oral CSF was based on inhalation data resulting
in identical values for both routes of exposure. In most cases (>75%) where an inhalation CSF
was derived from an inhalation study, the inhalation CSF was lower than the corresponding oral
CSF. Therefore, use of an oral CSF as an interim inhalation CSF appears reasonable and is
unlikely to result in underestimating risk.
Dose-Response Data:
The oral CSF listed in HEAST was based on a study by Hecht et al. (1982). Groups of 30 male
F344 rats were fed dietary concentrations of 0 or 4,000 ppm o-toluidine hydrochloride for 73
weeks followed by 20 weeks of observation. An increased incidence of skin fibromas, mammary
fibroadenomas, spleen fibromas, and peritoneal sarcomas was reported. Skin fibromas gave the
greatest response and were used to derive the CSF. The data are summarized below as reported
in U.S. EPA (1987).
Experimental Dose
o-Toluidine»HCl
(mg/rat/d)
0
62
Transformed Dose
o-Toluidine
(mg/kg/d)
0
80
Incidence
1/27
25/30
D-45
-------
Volume II Appendix D
Calculations:
URF = CSF x 1 mg/1,000 //g x 1/70 kg x 20 m3/d =
0.24 (mg/kg/d)'1 x 1 mg/1,000 //g x 1/70 kg x 20 m3/d = e.QE-OS^g/m3)-1
where
70 kg = default adult human body weight
20 m3 = default adult human daily rate of inhalation
Calculations assume 100% absorption.
Additional Information:
The National Cancer Institute (NCI) also has conducted a cancer bioassay of o-toluidine
hydrochloride (NCI, 1979). F344 rats were fed diets containing 0, 3,000, and 6,000 ppm and
B6C3FJ mice were fed diets containing 0, 1,000, and 3,000 ppm for 2 years. Multiple site
sarcomas, subcutaneous fibromas, and multiple site mesotheliomas were observed in male rats.
Female rats had multiple site sarcomas, mammary fibroadenomas, splenic sarcomas, and urinary
bladder carcinomas. Multiple site hemangiosarcomas were seen in male mice and hepatocellular
carcinomas and adenomas were seen in female mice. U.S. EPA (1987) reported that the Hecht et
al. (1982) study was selected over the NCI (1979) study because the former resulted in a higher
cancer potency estimate.
References:
Hecht, S.S., K. El-Bayoumy, A. Rivenson, and E. Fiala. 1982. Comparative carcinogenicity of
o-toluidine hydrochloride and o-nitrosotoluene in F-344 Rats. Cancer Letters 16:103-108.
National Cancer Institute (NCI). 1979. Bioassay of o-Toluidine Hydrochloride for Possible
Carcinogenicity. TR-153. Bethesda, MD.
U.S. Environmental Protection Agency. 1987. Health and Environmental Effects Profile for
2-Methylaniline and 2-Methylaniline Hydrochloride. Office of Health and Envrionmental
Assessment, Cincinnati, OH. EPA/600/X-87/092.
U.S. Environmental Protection Agency. 1997. Health Effects Assessment Summary Tables
(HEAST), FY 1997 Update. Office of Emergency and Remedial Response, Washington, DC.
EPA-540-R-97-036.
U.S. Environmental Protection Agency. 1998. Integrated Risk Information System (IRIS).
Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment,
Cincinnati, OH.
D-46
-------
Volume II
Appendix D
Table D-l. Correlation of Oral and Inhalation Cancer Slope Factors
Reported in IRIS and HEAST
CAS#
79-06-1
107-13-1
309-00-2
140-57-8
7440-38-2
103-33-3
71-43-2
92-87-5
7440-41-7
111-44-4
542-88-1
108-60-1
75-25-2
56-23-5
57-74-9
510-15-6
67-66-3
74-87-3
50-29-3
96-12-8
106-93-4
107-06-2
75-35-4
542-75-6
60-57-1
122-66-7
106-89-8
75-21-8
319-84-6
319-85-7
608-73-1
76-44-8
1024-57-3
118-74-1
87-68-3
67-72-1
Chemical
Acrylamide
Acrylonitrile
Aldrin
Aramite
Arsenic
Azobenzene
Benzene
Benzidine
Beryllium
Bis(2-chloroethyl)ether
Bis(chloromethyl)ether
Bis(2-chloro-1-
methylethyl)ether
Bromoform
Carbon tetrachloride
Chlordane
Chlorobenzilate
Chloroform
Chloromethane
DDT
1,2-Dibromo-3-
chloropropane
1,2-Dibromoethane
1,2-Dichloroethane
1,1-Dichloroethylene
1,3-Dichloropropene
Dieldrin
1 ,2-Diphenylhydrazine
Epichlorohydrin
Ethylene oxide
HCH alpha
HCH beta
HCH tech.
Heptachlor
Heptachlor epoxide
Hexachloro benzene
Hexachlorobutadiene
Hexachloroethane
Oral
CSF
4.5
0.54
17
0.025
1.5
0.11
0.029
230
4.3
1.1
220
0.07
0.0079
0.13
1.3
0.27
0.0061
0.013
0.34
1.4
85
0.091
0.6
0.18
16
0.8
0.0099
1.02
6.3
1.8
1.8
4.5
9.1
1.6
0.078
0.014
Inh
CSF
4.5
0.24
17
0.025
15
0.11
0.029
235
8.4
1.16
217
0.035
0.0039
0.053
1.3
0.27
0.08
0.0063
0.34
0.0024
0.77
0.091
0.175
0.13
16
0.77
0.0042
0.35
6.3
1.8
1.8
4.5
9.1
1.6
0.077
0.014
log Oral
CSF
0.6532
-0.2676
1.2304
-1.6021
0.1761
-0.9586
-1.5376
2.3617
0.6335
0.0414
2.3424
-1.1549
-2.1024
-0.8861
0.1139
-0.5686
-2.2147
-1.8861
-0.4685
0.1461
1.9294
-1.0410
-0.2218
-0.7447
1.2041
-0.0969
-2.0044
0.0086
0.7993
0.2553
0.2553
0.6532
0.9590
0.2041
-1.1079
-1.8539
log Inh.
CSF
0.6532
-0.6198
1.2304
-1.6021
1.1761
-0.9586
-1.5376
2.3711
0.9243
0.0645
2.3365
-1.4559
-2.4145
-1.2757
0.1139
-0.5686
-1.0969
-2.2007
-0.4685
-2.6162
-0.1135
-1.0410
-0.7570
-0.8861
1.2041
-0.1135
-2.3768
-0.4559
0.7993
0.2553
0.2553
0.6532
0.9590
0.2041
-1.1135
-1.8539
(continued)
D-47
-------
Volume II
Appendix D
Table D-l. (continued)
CAS#
302-01-2
75-09-2
101-14-4
924-16-3
55-18-5
62-75-9
930-55-2
1336-36-3
75-56-9
630-20-6
79-34-5
8001-35-2
79-00-5
88-06-2
75-01-4
Chemical
Hydrazine
Methylene chloride
4,4'-Methylenebis(2-
chloroaniline)
/V-Nitrosodi-n-butylamine
/V-Nitrosodiethylamine
/V-Nitrosodimethylamine
/V-Nitrosopyrrolidine
PCBs
Propylene oxide
1,1 ,1 ,2,-Tetrachloroethane
1,1 ,2,2,-Tetrachloroethane
Toxaphene
1,1,2-Trichloroethane
2,4,6,-Trichlorophenol
Vinyl chloride
Oral
CSF
3
0.0075
0.13
5.4
150
51
2.1
2
0.24
0.026
0.2
1.1
0.057
0.011
1.9
Inh
CSF
17.1
0.0016
0.13
5.6
151
49
2.13
0.4
0.013
0.026
0.2
1.1
0.056
0.011
0.3
log Oral
CSF
0.4771
-2.1249
-0.8861
0.7324
2.1761
1.7076
0.3222
0.3010
-0.6198
-1.5850
-0.6990
0.0414
-1.2441
-1.9586
0.2788
log Inh.
CSF
1.2330
-2.7959
-0.8861
0.7482
2.1790
1.6902
0.3284
-0.3979
-1.8861
-1.5850
-0.6990
0.0414
-1.2518
-1.9586
-0.5229
OT ~6'
0
3.000 -
2.000 -
1.000 -
300 -5.000 -4.000 -3.000 -2.000 -1.000 (ft
-1.000 -
• «
-2.000 -
* *
-3.000 -
-4.000 -
# -5.000 -
finnn
• * **
* *» »**
^» * *
^ A .
00 1.000* * 2.000 3.000 4.(
*
00
log TD50
Figure D-l. Correlation of Oral and Inhalation Cancer Slope Factors.
D-48
-------
Volume II
Appendix D
Table D-2. Correlation of TD50s Reported in the Cancer Potency Database and Inhalation
Cancer Slope Factors Reported in IRIS and HEAST.
TD50
Geo
Mean"
18.53
32.28
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
Inh CSF°
0.053
0.34
1 7
1 .16
0.98
1.3
0.27
0.08
0.1 75
0.13
16
6.3
1.8
1 .8
4.5
0.014
0.0016
5.6
0.026
0.2
1.1
0.056
0.0077
4.5
0.24
0.025
0.1 1
217
0.1 1
0.004
0.77
0.091
0.77
0.35
1.6
0.077
17.1
0.13
151
49
2.13
0.013
CSF
test
speciesd
b
b
m
m
m
m
m
m
m
m
m
m
m
m
m
m
m
m
m
m
m
m
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
1/CSF
18.87
2.94
0.059
0.86
1 .020
0.77
3.70
12.50
5.71
7.692
0.0625
0.159
0.556
0.556
0.222
71 .429
625.000
0.1 79
38.4615
5
0.909
1 7.857
129.870
0.222
4.167
40
9.09
0.00
9.091
250.00
1 .299
10.99
1 .298701
2.857
0.625
12.987
0.058
7.692
0.007
0.020
0.469
76.9231
log TD50e
(X)
1 .268
1 .509
0.104
1 .068
1 .143
0.476
1 .973
1 .956
1 .539
1 .695
-0.040
0.821
1 .444
1 .1 70
0.083
2.529
2.963
0.037
2.260
1 .583
0.746
1 .740
2.185
0.789
1 .228
1 .985
1 .382
-2.447
1 .267
2.812
0.182
0.905
0.747
1 .328
0.545
1 .818
-0.510
1 .286
-1.625
-0.907
-0.097
1 .872
log 1/CSF
(Y)
1 .276
0.469
-1 .230
-0.064
0.009
-0.1 14
0.569
1 .097
0.757
0.886
-1.204
-0.799
-0.255
-0.255
-0.653
1 .854
2.796
-0.748
1 .585
0.699
-0.041
1 .252
2.114
-0.653
0.620
1 .602
0.959
-2.336
0.959
2.398
0.114
1 .041
0.114
0.456
-0.204
1.114
-1.233
0.886
-2.179
-1 .690
-0.328
1 .886
D-49
-------
Volume II
Appendix D
ft -6'
o
3.000 -
2.000 -
1.000 -
500 -5.000 -4.000 -3.000 -2.000 -1.000 (ft
-1.000-
-2.000 -
• *
-3.000 -
-4.000 -
! -5.000 -
.
. ' «' *
00 ' ^.000* * 2.000 3.000 4.(
logTDSO
00
Figure D-2. Correlation of TD50 and inhalation cancer slope factors.
D-50
-------
|