&EPA
United States
Environmental Protection
Agency
Office of Solid Waste
Washington, DC 20460
EPA530-R-98-009b
May T998
Air Characteristic Study
Volume II
Technical Background Document
-------
-------
EPA530-R-98-009b
May 1998
Volume II
Air Characteristic Study
Technical Background Document
Office of Sol id 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 List of Constituents 1-2
1.2 Waste Management Units 1-2
1.3 Organization of Report 1-2
2.0 Modeling Approach and Data Sources 2-1
2.1 Overview of Modeling Approach 2-1
2.2 Conducting the Analysis 2-4
2.3 Data Sources 2-7
3.0 Waste Source Characteristics 3-1
3.1 Estimation of Missing Data 3-1
3.2 Waste Management Unit Data and Size Categories Used in
Dispersion Model 3-2
3.3 Removal of Wastepiles Containing Bevill-Excluded Wastes 3-5
3.3.1 Identification of Bevill Industrial D Facilities 3-6
3.3.2 Bevill Facilities Culled from Industrial D Data 3-7
3.4 Tanks 3-7
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 4-2
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 4-6
4.3 Development of Volatile Emissions and Waste Concentrations for
Landfills 4-9
4.4 Development of Volatile Emissions and Waste Concentrations for Land
Application Units 4-11
4.4.1 Chronic Exposure Analysis 4-11
4.4.2 Acute and Subchronic Exposure Analysis 4-14
4.5 Development of Volatile Emissions and Waste Concentrations for
Wastepiles 4-15
4.5.1 Chronic Exposure Analysis 4-15
4.5.2 Acute and Subchronic Exposure Analysis 4-17
4.6 Development of Volatile Emissions for Tanks 4-17
4.7 Development of Particulate Emissions 4-18
4.7.1 Landfills and Land Application Units 4-18
4.7.2 Wastepiles 4-21
m
-------
Table of Contents (continued)
Section
Page
5.0 Dispersion Modeling 5-1
5.1 Model Selection 5-1
5.2 Critical Parameters 5-1
5.2.1 General Assumptions 5-2
5.2.2 Meteorological Stations and Data 5-3
5.2.3 Source Release Parameters 5-6
5.2.4 Receptors 5-7
5.3 Unitized Air Concentrations 5-10
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 6-11
6.3 Acute Inhalation Health Benchmarks 6-14
7.0 Development of Risk-Specific Waste Concentration Distribution 7-1
7.1 Overview 7-1
7.2 Select Receptor Location 7-2
7.3 Obtain Unitized Air Concentrations 7-2
7.4 Calculate Air Concentration 7-3
7.5 Select Exposure Factors 7-3
7.5.1 Inhalation Rate 7-4
7.5.2 Body Weight 7-6
7.5.3 Exposure Duration 7-8
7.5.4 Exposure Frequency 7-10
7.6 Obtain Health Benchmarks 7-10
7.6.1 Carcinogens 7-10
7.6.2 Noncarcinogens 7-11
7.7 Calculate Risk or Hazard Quotient 7-12
7.8 Backcalculate Risk-Specific Waste Concentration 7-13
7.9 Adjustments for Results Not Meeting Linearity Assumption 7-14
7.9.1 Adjustments for Land-Based Units 7-14
7.9.2 Adjustments for Tanks 7-15
7.10 Methodology for Subchronic and Acute Exposure 7-16
7.10.1 Overview 7-16
7.10.2 Select Receptor Location 7-16
7.10.3 Obtain Unitized Air Concentration 7-16
7.10.4 Calculate Air Concentration 7-16
7.10.5 Select Exposure Factors 7-16
7.10.6 Obtain Health Benchmarks 7-16
7.10.7 Calculate Hazard Quotients 7-17
IV
-------
Table of Contents (continued)
Section
Page
7.10.8 Backcalculate Risk-Specific Waste Concentration 7-17
7.10.9 Adjustments for Results Not Meeting Linearity Assumptions 7-17
7.11 Modifications to Methodology for Lead 7-17
8.0 Analysis of Uncertainty 8-1
8.1 Emissions Modeling 8-2
8.2 Dispersion Modeling 8-4
8.3 Exposure Modeling/Risk Estimation 8-5
8.4 Indirect Exposures 8-6
9.0 References 9-1
Appendix A - Basic Dalenius-Hodges Procedure for Constructing Strata A-1
Appendix B - Chemical-Specific Data B-l
Appendix C - Sensitivity Analysis for Emissions Model C-l
Appendix D - Sensitivity Analysis of ISC Air Model D-l
Appendix E - Derivation of Chronic Inhalation Noncancer and Cancer Health Benchmarks .. E-l
v
-------
List of Tables
Table
Page
1-1 Constituents Modeled for All WMUs 1-3
1-2 Constituents Modeled for Tanks Only 1-5
1-3 Number of Industrial D Facilities and 1985 Waste Generation by Waste
Management Unit Type 1-6
i
3-1 Final WMU Area Strata Used for ISCST3 ModelRuns 3-5
3-2 Input Parameters for Tanks 3-8
4-1 CHEMDAT8 Land-Based Unit Model Input Requirements -... 4-5 !
4-2 CHEMDAT8 Tank Model Input Requirements 4-7
4-3 Inputs and Intermediate Values Used for Wind Erosion from Landfills and LAUs .. 4-20
4-4 Calculated Particulate Emission Rates for Landfills and LAUs 4-21 ,
4-5 Calculated Particulate Emission Rates for Wastepiles 4-23
5-1 Air Dispersion Model Capabilities 5-2
5-2 Meteorological Stations Used in the Air Characteristics Study 5-4
5-3 Area Modeled for Landfills and Land Application Units 5-9
5-4 Areas and Source Heights Modeled for Wastepile 5-9
5-5 Areas Modeled for Aerated and Storage Tanks 5-9
5-6 Maximum Annual Average Unitized Air Concentrations G"g/m3 / Aig/s-m2) for
Landfills and Land Application Units 5-11
5-7 Maximum Annual Average Unitized Air Concentrations Og/m3 / jUg/s-m2)
for Wastepiles 5-12
5-8 Maximum Annual Average Unitized Air Concentrations (//g/m3/^ig/s-m2) for :
Aerated Tanks 5-13 :
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 6-12
6-5 ATSDR Acute Inhalation MRLs 6-15
6-6 CalEPA's 1-Hour Acute Inhalation Reference Exposure Levels (RELs) 6-16
7-1 , Estimated Parameters for Inhalation Rate for Residents Assuming Lognormal
Distribution 7-5
7-2 Recommended Inhalation Rates for Workers 7-5
7-3 Body Weights for Males, All Races, Ages 18-74 Years (kg) 7-6
7-4 Body Weights for Male Children, Ages 6 Months to 18 Years (kg) 7-7
7-5 Pooled Body Weights for Children (kg) 7-8
7-6 Descriptive Statistics for Residential Occupancy Period for Males (years) 7-8
7-7 Descriptive Statistics for Population Mobility for Children (years) 7-9
7-8 Distribution of Age at Start of Exposure and Exposure Duration for Children 7-9
VI
-------
List of Tables (continued)
Table
Page
7-9 Summary of Inputs for IEUBK Model 7-18
7-10 Results of IEUBK Modeling 7-19
8-1 Summary of How Uncertainties Were Addressed in the Study 8-3
VII
-------
List of Figures
Kgure
Page
2-1 Conceptual diagram of a waste site 2-2
2-2 Model framework 2-5
2-3 Combination of results for individual WMUs into a distribution across all WMUs ... 2-8
3-1 Landfill characteristics from Industrial D Screening Survey 3-3
5-1 Meteorological station regions 5-5
5-2 Air concentration vs. size of area source 5-8
5-3 Maximum UAC by meteorological location (landfills and LAUs) 5-14
5-4 Maximum UAC by meteorological location (2-m wastepiles) 5-15
5-5 Maximum UAC by meteorological location (5-m wastepiles) 5-16
5-6 Maximum UAC by meteorological location (tanks) 5-17
6-1 Approach used to select chronic inhalation health benchmark values 6-2
6-2 Approach used to select subchronic noncancer inhalation benchmark values 6-11
6-3 Approach used to select acute noncancer inhalation health benchmark values 6-16
vm
-------
Voiume ff
Section 2.O
1.0 Introduction
The U.S. Environmental Protection Agency (EPA) has undertaken a study that evaluates
the need for developing a hazardous waste characteristic that addresses risk to human health
through,the direct inhalation pathway. This air characteristic study includes an analysis of gaps
in the hazardous waste characteristics and relevant Clean Air Act (CAA) controls, and the
resulting potential risks to human health, posed by the inhalation of gaseous and nongaseous air
emissions from wastes managed in certain waste management units (WMUs). These units are:
tanks, landfills, wastepiles, land application units, and a subset of surface impoundments. The
study covers 105 specific constituents that are volatile organic compounds (VOCs), semivolatile
chemicals, and metals.
On November 15, 1996, under a deadline negotiated with the Environmental Defense
Fund (EDF), the EPA Office of Solid Waste (OSW) completed the Hazardous Waste
Characteristic Scoping Study. 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 followup analysis on potential gaps in regulatory
coverage under the Clean Air Act and Subpart CC of the Resource Conservation and Recovery
Act (RCRA), OSW has identified air emissions from waste management units to be one of the
areas meriting further analysis to better characterize potential risks. EPA is under a consent
decree to complete the first portion (all units except surface impoundments) of the study by
May 15, 1998. The second portion of the study covers surface impoundments receiving
wastewaters that never exhibited a characteristic. The surface impoundment study is due to be
completed March 26, 2001.
This report describes a national, screening-level analysis designed to assess the potential
risk attributable to inhalation exposures when certain chemicals and metals are managed as a
waste in certain types of WMUs. Of particular interest are chemicals and metals managed as
wastes that are not regulated under RCRA as hazardous wastes. The purpose of this approach is
to determine which chemicals and waste management units are of potential national concern,
purely from a risk perspective; it is not intended to present 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 potential risk to assist the Agency in determining the need to expand regulatory
coverage in the future.
1-1
-------
Volume II
Section 1.0
1.1 List of Constituents
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. Table 1-1 lists chemicals that may be present in
nonhazardous industrial waste and that were assessed for all WMUs included in the analysis.
Table 1-2 lists chemicals that were included in the analysis of releases from tanks and surface
impoundments only.
1.2 Waste Management Units
The waste management units assessed are tanks, landfills, wastepiles, and land treatment
units. Because surface impoundments receiving wastewaters that are not hazardous by
characteristic are being addressed under a separate study, this analysis does not include surface
impoundments as a WMU.
The waste management scenario modeled in this analysis is disposal or treatment of
industrial waste streams in RCRA Subtitle D WMUs. For landfills, wastepiles, and land
application units (LAUs), the 1985 Screening Survey of Industrial Subtitle D Establishments
(Shroeder et al., 1987) was the source for data on WMU dimensions and annual waste volumes,
along with facility locations for all WMU types. This survey collected information on land-based
Industrial D waste management operations in the United States for the 17 industry groups shown
in Table 1-3. The resulting data set has been used to represent Industrial D facility locations and
WMU characteristics in a variety of RCRA regulatory initiatives, including the Hazardous Waste
Identification Rule (HWIR). Although these data are over 10 years old, they represent the largest
consistent set of data available.
Tanks were included in this analysis but were not part of the Industrial D survey. Tank
characteristics were taken from a previous EPA regulatory analysis supporting the RCRA
Subpart CC rules (U.S. EPA, 1991). The tank data used in that analysis were from a 1991 EPA
survey of hazardous waste generators and treatment, storage, and disposal facilities (Westat
Survey) (U.S. EPA, 1991).
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
distribution. 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 E.
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
1,2-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 dichloride]
cis-1,3-Dichloropropylene
trans-1,3-Dichloropropylene
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]
Ethyl benzene
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
n-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-n-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 630-20-6
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
Acrylic acid
Aniline
Benzidine
Benzo(a)pyrene
2-Chlorophenol [o-chlorophenol]
Cresols, total
7,12-Dimethylbenz[a]anthracene
N,N-Dimethyl formamide
3,4-Dimethylphenol
2,4-Dinitrotoluene
1,2-Diphenylhydrazine
Ethylene glycol
Hexachlorobenzene
Hexachloro-1,3-butadiene [hexachlorobutadiene]
Hexachlorocyclopentadiene
Isophorone
3-Methylcholanthrene
Nitrobenzene
Phenol
Phthalicanhydride
2,3,7,8-TCDD [2,3,7,8-tetrachlorodibenzo-p-dioxin]
o-Toluidine
1,2,4-Trichlorobenzene
79-06-1
79-10-7
62-53-3
92-87-5
50-32-8
95-57-8
1319-77-3
57-97-6
68-12-2
95-65-8
121-14-2
122-66-7
107-21-1
118-74-1
87-68-3
77-47-4
78-59-1
56-49-5
98-95-3
108-95-2
85-44-9
1746-01-6
95-53-4
120-82-1
1-5
-------
Table 1-3. Number of Industrial D Facilities and 1985 Waste Generation by Waste Management Unit Type
;..;..•'-•'. • . ' .v' ••'•'• ;••'•.•••• '
j • . •:• .• • • .•.',:•••..'•
j..;' ' •'••.• - "... '.' ...-.";. .....' '.: ...
IsiCCode- Industry :
2865, 2869 Organic chemicals
331 2-3321 Primary iron and steel
2873-2879 Fertilizer and agricultural
chemicals
491 1 Electric power generation
2821 Plastics and resins manufacturing
281 2-281 9 Inorganic chemicals
32 Stone, clay, glass, and concrete
26 Pulp and paper
3331-3399 Primary nonferrous metals
20 Food and kindred products
4941 Water treatment
29 Petroleum refining
30 Rubber and miscellaneous
products
37 Transportation equipment
2822, 2824, Selected chemicals and allied
2851,2891 products
22 Textile manufacturing
31 Leather and leather products
Total
Landfills
No. Facilities
17
202
31
155
32
120
1,257
259
111
194
121
61
77
63
21
28
9
2,757
0.7%
9%
1.3%
7%
1%
5%
54%
11%
5%
8%
5%
3%
3%
3%
0.9%
1%
0.4%
1985 Waste (metric
\ toils)
238,575
3,352,467
5,263,323
48,590,046
77,539
2,927,324
6,883,106
5,338,781
1,250,125
3,268,486
143,099
246,952
472,520
155,926
101,953
62,655
8,460
78,381,336
0.3%
4%
7%
62%
0%
4%
9%
7%
2%
4%
0%
0%
1%
0%
0.1%
0.08%
0.01%
Land Application Units
No. Facilities
27
76
160
43
17
24
309
139
9
3,128
147
114
16
11
17
72
0
4,309
0.6%
2%
4%
1%
0%
0.6%
7%
3%
0.2%
73%
3%
3%
0.4%
0.3%
0.4%
2%
0%
1985 Waste
(metric tons)
1,660,572
68,932
686,728
300,637
150,518
97,959
46,099
8,128,977
339,098
69,034,641
8,140,583
359,888
47,448
264
388,676
693,872
0
90,144,891
1.8%
0.08%
0.8%
0.3%
0.2%
0.1%
0.05%
9%
0.4%
77%
9%
0.4%
0.05%
0.0003%
0.4%
0.8%
0%
Waslepiles .'.. j
No. Facilities
79
464
50
110
32
98
2,528
232
312
540
48
158
123
362
41
103
54
5,335
1%
9%
0.9%
2%
0.6%
2%
47%
4%
6%
10%
0.9%
3%
2%
7%
0.8%
2%
1%
1 985 Waste (metric j
: tons)
43,771
5,571,660
4,381,910
755,800
2,743,691
37,566,042
8,348,981
1,334,766
7,966,935
417,940
8,238
71,511
53,481
644,023
6,630
16,210
10,201
69,941,790
.- ...-;....]
0.06%
8%
6%
1%
4%
54%
12%
2%
11%
0.6%
. 0.01%
0.1%
0.1%
0.9%
0.01%
0.02%
0.01%
Source: Screening Survey [Database] of Industrial Subtitle D Establishments (Shroeder et al., 1987).
-------
Volume II
Section 2.O
2.0 Modeling Approach and Data Sources
This section provides a general overview of the modeling approach and primary data
sources used and describes how the risk analysis was conducted. Further detail on the 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 assess
the risk to human health via inhalation exposure of 105
constituents managed as waste in landfills, land
application units (LAUs), wastepiles, and tanks. The
purpose of this analysis was to estimate the
concentrations of constituents managed in WMUs that
are protective for adults, children, and workers for three
different types of exposures or risk endpoints: chronic
(1 year), subchronic (1 month), and acute (1 day).
The analytical approach for this analysis was
based on three primary components:
• Emissions modeling—characterizing
emissions from a WMU
• Dispersion modeling-describing the
transport of these emissions through the
ambient environment
• 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).
Figure 2-1 is a conceptual diagram of a waste site. Constituents managed in the WMU
can be released as gases or particles or both. People who live and work at various distances and
in various directions from the unit will have different inhalation exposures depending on wind
patterns at the site. Workers at the site may also be exposed.
The Air Characteristic Study addresses:
• 105 constituents
• 4 WMU types
- landfill
- land application unit (LAU)
- wastepile (WP)
- tank
• 4 receptors
- adult resident
- child resident
- offsite worker
- onsite worker
• direct inhalation only
• volatiles and particulates
• 7 distances from the site
• 3 averaging times/risk endpoints
- chronic (1 year)
- subchronic (1 month) - LAU, WP
- acute (1 day) - LAU, WP
2-1
-------
Volume II
Section 2.0
Dispersion
Volatilization
Particulates
•' v: .•.".' Deposition
Figure 2-1. Conceptual diagram of a waste site.
Preliminary model requirements for this 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 (1 year),
subchronic (1 month), and acute (1 day) releases
• An exposure model for locating receptors proximate to the WMUs and estimating
their exposure
• A risk model that combines the exposure estimate with a dose-response
relationship
• 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 the 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, was needed (rather than a deterministic
approach, which would produce a point estimate). A probabilistic approach considers the
variability in the inputs requked to estimate the concentration nationally. For this analysis, EPA
used a Monte Carlo simulation. This is a type of probabilistic analysis in which the distribution
2-2
-------
Volume II
Section 2.0
of some or all input variables is known or estimated. A large number of iterations on 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, Q,) 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.
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 and wastepiles as elevated sources. To obtain the emission rate of
constituent sorbed to particulate matter, the emission rate of particulate matter was multiplied by
the soil concentration calculated by CHEMDAT8. Landfills and tanks have a continuous waste
loading, while LAUs and wastepiles have noncontiguous, episodic loadings. To capture potential
peaks in emissions immediately after 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). These are
dispersion coefficients based on a unit emission (i.e., 1 ^g/m2-s) for use in a backcalculation.
UACs varied depending on the averaging time, the size of the WMU, the distance and direction
of the receptor from the WMU, and the associated meteorological station.
The air concentration at any specific receptor is the product of the emission rate
(in ug/m2 -s) and appropriate UAC (in [ug/m3]/[ ug/m2 -s]). Air concentrations were estimated
for chronic, subchronic, and acute exposures, based on a combination of volatile and particulate
emissions.
The data used to identify and characterize WMUs contained no information on the
location and types of receptors near 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 living at that exact point. Since 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. 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.
For this analysis, four receptors were included: an adult resident, a child resident, an
offsite worker, and an onsite worker. The onsite worker was located at the 0-m distance or edge
2-3
-------
Volume II
Section 2.0
of the WMU. The adult and child resident and off site worker 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. For acute and subchronic exposures, receptors were modeled at 0,25, 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. Emission and dispersion modeling were performed first and the results used as
inputs to the exposure modeling/risk estimation (Shroeder et al., 1987). 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, this was done by forward calculating a risk associated with a unit waste
concentration (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 within certain restrictions (e.g., that the concentration does not exceed the saturation
concentration). This was accounted for by checking backcalculated results that fell outside the
restrictions and modifying any results that did so.
Emissions modeling was performed for all WMUs and all chemicals, assuming a unit
concentration of the chemical hi 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 21 representative WMU areas and 29
meteorological locations, assuming a unit emission rate of 1 /zg/m2-s. This produced UACs for
each area, 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, m
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).
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 790
landfill units. The database also has a sampling weight for each facility that defines how many
facilities 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
2-4
-------
Volume II
Section 2.O
Emissions Modeling
For each chemical and
WMU, estimate:
-volatile emissions
- particulate emissions
- remaining average
waste/soil concentration
1
Select constituent and WMU type
i.
Select WMU site from Industrial D Survey data
Characterize site (area and met station)
Dispersion Modeling
For each of 21 WMU
areas and 29
meteorological stations,
estimate UAC at each of
112 receptor locations.
Select receptor location for each distance
Interpolate UAC for site
(based on area and met station)
Select exposure factors for each receptor (chronic
exposure to carcinogens only)
Determine risk-specific waste concentration (Cw) 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 (1000)
Figure 2-2. Model framework.
2-5
-------
Volume II
Section 2.0
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 seven
distances from the site.
In Step 4, a UAC was interpolated for the WMU. UACs were modeled for only 21
selected WMU areas for each meteorological station and receptor location. To calculate a UAC
corresponding to the WMU's actual area, EPA 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. For a WMU with an actual area of 100 m2, the UAC was interpolated from the
UACs for 20 and 162 m2. For a WMU with an actual area of 200 m2, the UAC was interpolated
from the UACs for 162 and 486 m2.
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, 1997) 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) 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.
Steps 3 through 6, which form the core of the Monte Carlo simulation, were then repeated
1,000 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 (0,25, 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 1,000 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
2-6
-------
Volume IJ
Section 2.0
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.
These cumulative distributions are intended to encompass the variability across WMUs.
Thus, the variability in WMU characteristics and in meteorological settings are included in these
distributions.
This process was repeated from Step 1 for each chemical and WMU type analyzed in this
study.
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 EPA's Background Document
2-7
-------
Volume II
Section 2.0
WMU1
s
st
p
g-o
100
90
80
70
60
50
40
so
20
10
a (90th %lle)
(50th %IIe)
0.1 1 10 100 1,000 10,000
Cw (ppm)
WMU 2
b (SOth %lle)
(50th %!le)
100
2 so
f g 80
g'g 70
60
50
11 40
S» "S
•gCQ 20
;S 10
All WMUs
0.1 1 10 100 1,000 10,000
Cw (ppm)
WMUS
0.1
i
gO
EE
100
90
80
70
60
SO
40
30
20
10
0
(SOth %lle)
10 100 1,000
Cw (ppm)
(50 h %llo)
0.1 1 10 100 1,000 10,000
Cw (ppm)
Figure 2-3. Combination of results for individual WMUs
into a distribution across all WMUs.
supporting the RCRA Subpart CC rules on air emissions from hazardous waste treatment,
storage, and disposal facilities (U.S. EPA, 1991). 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.
2-8
-------
Volume II
Section 2.O
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 facility
to a meteorological station, EPA used a 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, the
Industrial D sites were then overlaid on this GIS coverage and meteorological station
assignments were then exported for use in the modeling exercise. Four sites in Alaska and four
in Hawaii were deleted from the analysis at this point because the 29 meteorological stations are
limited to the continental United States.
2-9
-------
-------
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, and aerated and storage tanks. The dimensions and operating characteristics of
these units 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.
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 WMU
types except tanks) on the number of each type of WMU, the 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 air 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. Tank data sources are described in Section 3.4.
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 facilities, using the DUNS number, name, street address, and
zip code, 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 Data
The Industrial D Screening Survey included all of the data needed to estimate emissions
except landfill depth and wastepile height, which had to be estimated. In the absence of data on
wastepile height, two heights, 2 m and 5 m, were chosen. These values, estimated based on the
height a frontloader might be able to reach, have been used before in the Hazardous Waste
Identification Rule (RTI, 1995). All wastepiles in the Industrial D Survey were modeled twice,
once at each height.
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
(fixed by unit type) waste bulk densities used in the previous modeling efforts:
depth = total waste capacity / (surface area x bulk density) .
3-1
-------
Volume II
Section 3.0
The bulk density used for landfills was 1.09577 g/cm3.
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 790 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.
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 depth as described above. Figure 3-1 shows the regression
plots, including the replaced ("random capacity") values, for landfills.
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).
Initially, a regular distribution of strata cutpoints (5th, 25th, 50th, 75th, and 95th
percentiles) was considered to provide points for interpolation. 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, the Dalenius-Hodges procedure was used as a
starting point. This technique is designed to break the distribution of the known variable X (in
3-2
-------
Volume II
Section 3.O
Log (Average Surface Area, acres) Line Fit Plot
Q.
8
1
0>
O)
s
I
D)
O
Actual capacity values
Best-fit line
Random capacity values
3--
-2
-1 H o 1
Log (Average Surface Area, acres)
Landfill Regression Plot
Figure 3-1. Landfill characteristics from Industrial D Screening Survey.
3-3
-------
Volume II
Section 3.0
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, EPA modified the
procedure 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 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 as shown in Table 3-1.
fl!j
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 3-1 shows the final WMU area strata, the medians used for the air dispersion
modeling runs and the cumulative percentile of the distribution each stratum represents.
3-4
-------
Volume If
Section 3.0
Table 3-1. Final WMU Area Strata Used for ISCST3 Model Runs
(modified Dalenius- Hodges procedure)
X
Strata ;
v >
""* ifow"
f ^ j t.
Median;
:m2) K /c
V, „' ' ]
, I High
-------
Volume If
Section 3.0
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.
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
2816 Inorganic Pigments 10-1099
2819 Ind. Inorganic Chem., Nee. 12-1299
2874 Phosphatic Fertilizers 13 -1399
3312 Blast Furnaces & Steel Mills 14 - 1499
3331 Primary Copper
3339 Primary Nonferrous Metals, Nee.
Description
Metal Mining
Coal Mining
Oil & Gas Extraction
Mining & Quarrying
ofNonmetallic
Minerals Except Fuels
In addition, 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.
33.1 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
6
3
2
9
4
Group Description SIC Codes
Inorganic Chemicals 2812-2819
Fertilizer & Agr. Chem. 2873 - 2879
Primary Iron & Steel 3312-3321
Primary Nonferrous Metals 3331 - 3399
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
3-6
-------
Volume II
Section 3.0
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
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 for the analysis.
3.3.2 Bevill Facilities Culled from Industrial D Data
To decide whether 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, 82
wastepiles were removed because they contained Bevill wastes, leaving 742 facilities with
wastepiles for the risk analysis.
3.4 Tanks
The Industrial D Survey did not include tanks. Therefore, two model tanks were used
from the Hazardous Waste TSDF—Background Information for Proposed RCRA Air Emissions
Standards (U.S. EPA, 1991). These two tanks were located at each of the 29 meteorological
locations selected to represent the climate regions of the United States (see Section 5.0), resulting
in 58 tanks modeled. Each tank was then modeled four times, using the following assumptions:
With aeration and biodegradation
With aeration and no biodegradation
Without aeration and with biodegradation
Without aeration or biodegradation.
The tanks modeled under the first two assumptions are referred to as aerated tanks and the tanks
modeled under the last two assumptions are referred to as storage tanks throughout this
document. Table 3-2 summarizes the characteristics of the two model tanks.
3-7
-------
Volume II
Section 3.0
Table 3-2. Input Parameters for Tanks
>, v^
W ^
Input Parameter
Plant 1
Model
Plant 2
Wind speed (m/s)
Temp (°C)
Depth/height (m)
Average surface area (m2)
Flow rate (m3/s)
Active biomass cone, (kg/m3)
Biomass in (kg/m3)
Inlet cone, (g/m3)
Inlet COD (g/m3)
Total biorate (mg/g-h)
Fraction agitated
aerated:
storage:
aerated:
storage:
Submerged airflow (m3/s)
Number of impellers
O2 transfer rate (Ib Oa/h-hp)
Power (total, hp)
Power efficiency
Impeller diam. (cm)
Impeller speed (rad/s)
Variable
Variable
4
27
0.0075
0
Oand2
0 and 0.05
1
100
19
0.518519
0
0
1
3
7.5
0.83
61
130
Variable
Variable
3.7
430
0.088
0
Oand2
0 and 0.05
1
100
19
0.581395
0
0
2
3
120
0.83
61
130
3-8
-------
Volume H
Section 4.O
i
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 paniculate 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. 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 the landfill, land application unit, and wastepile
(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 (exit
criteria) for wastes, the ideal emission model would provide as accurate emission estimates 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 conservatism) 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. And, 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 emissions 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
extensive review by both EPA and industry representatives. The CHEMDAT8 spreadsheet
-------
Volume II
Section 4.0
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 and hydrolysis for tanks, absorption for land-based
WMUs, and biodegradation for all units. 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. 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 conservative) estimates of air emissions from the
various emission sources.
The CHEMDAT8 model was used to estimate the emissions for all WMUs with some
minor modifications. For example, for the land-based WMUs (landfills, land application units,
and wastepiles), first-order biodegradation rates from the Hazardous Waste Identification Rule
(RTI, 1995) were used rather than the biodegradation rate equations in the CHEMDAT8 model.
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
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
conservatism) for each of the different emission sources under consideration, and the emission
Specifically, the CHEMDAT8 model was modified as follows:
Constituent-specific first-order biodegradation rate constants, applicable for land-based units and
converted to units of s"1 (see Appendix B).
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
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 and 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 emissions 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 results for metals other
than mercury (which do not volatilize and are therefore based solely on particulate emissions) are
somewhat conservative as a result of this overestimation.
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 the Section 4.2.1. Critical input parameters for the remaining sets of inputs are
discussed first for land-based WMU (land treatment) and then for tanks (aerated and storage
tanks). A sensitivity analysis was performed to better understand the impact of certain modeling
assumptions on the model results. Summaries of the results of the sensitivity analysis are
provided in Appendix C for the CHEMDAT8 volatilization emissions model.
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 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, log octanol-water partition coefficient, and the soil biodegradation rate constants.
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 (for adjusting vapor pressure to temperature), and biodegradation
rate constants for tanks. The biodegradation rate constants in the downloaded CHEMDAT8
4-3
-------
Volume II
Section 4.0
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 rates 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 the 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 No. 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 assuming unit concentration (concentration set to 1 mg/kg, assuming
Henry's law applies) and assuming pure component (concentration set to 1E+6 mg/kg, assuming
Raoult's law applies).
The annual waste quantity is a critical source (site-specific input) parameter. This
parameter along with assumptions regarding the frequency of contaminant addition and the
dimensions of the unit combine to influence a number of model input parameters (Input DD Nos.
Ll,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
meteorological data input that potentially impacts the emissions results for the CHEMDAT8
model for the land-based WMU.
The total porosity and air porosity values that were used in the emissions 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 1.09577
g/cm3.
For aqueous wastes, the molecular weight of the waste (Input ID No. L6) should not
impact the calculations. However, CHEMDAT8 converts Henry's law to a mole ratio using the
molecular weight of water and then converts back to mass basis using the input waste molecular
weight, resulting in an impact in the estimated volatile emission rate (see Appendix C). The
CHEMDAT8 model was modified so that the conversion back to a mass basis always uses the
molecular weight of water when the aqueous waste flag was set to 1 (alternatively, use 18 g/mol
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
L1
L2
L3
L4
L5
L6
L, Loading (g waste/cm3 soil)
Concentration in waste (ppmw)
I, Depth of tilling (or unit) (cm)
Total porosity
Air porosity (0 if unknown)
MW waste
L7 For aqueous waste, enter 1
L8 Time of calc. (days)
L9 For biodegradation, enter 1
L10 Temperature (°C)
L11 Windspeed (m/s)
L12 Area (m2)
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
for the molecular weight for aqueous waste). The molecular weight for the "pure component"
runs was set to 147 g/mol, which is the CHEMDAT8 default value for this input parameter. If
the waste were truly a pure component, then the appropriate molecular weight input is the
specific constituent's molecular weight. However, the CHEMDAT8 model does not provide for
constituent-specific molecular weights for the waste, and the pure component run is used to
backcalculate an appropriate waste concentration limit less than pure component. Therefore, a
singular waste molecular weight is appropriate. The molecular weights of the constituents
modeled were evaluated. The average molecular weight of the modeled constituents is 120
g/mol; the highest is 272 g/mol. A higher molecular weight increases the estimated volatilization
emission rates. As such, the assumed molecular weight value of 147 g/mol provides a reasonable
to conservative emission estimate on average for pure components, but may underestimate the
volatile emissions for the heaviest compound by a factor of approximately 30 percent.
As 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 first-order biodegradation rate was included for all
LAU runs and for the long-term wastepile emission runs; biodegradation losses were not
included in the short-term wastepile emission runs or any of the landfill runs, because these units
are not designed specifically for biodegradation and because of the relatively short-time period
for which the waste was assumed to be in the unit. Biodegradation was included for wastepiles
as the median retention time of waste in the wastepiles is greater than 5 years. Note that the
default CHEMDAT8 model method of calculating the biodegradation rate was not used because
the biodegradation rates used in CHEMDAT8 were derived primarily from wastewater studies
and then applied with an assumed, low-biomass concentration to model biodegradation in soils.
4-5
-------
Volume II
Section 4.0
Instead, the first-order biodegradation rate constants developed for the HWIR project based, for
the most part, on contaminant half lives in soils were used. Therefore, the HWIR biodegradation
rate constants provide a more direct link to soil-based biodegradation and are considered more
appropriate for all land-based WMUs that include biodegradation.
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 and U.S. EPA, 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 Hmited; 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 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.
Important input parameters for this calculation 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 NOAA (1992) and are summarized in
Section 4.7.2.
4.2.3 Critical Parameters for Tank Emissions Model
The input parameters for aerated tanks are presented in Table 4-2. The annual waste
quantity (flow rate) and the dimensions of the tank are critical input parameters for both aerated
and storage tanks. Unlike the other WMUs, site-specific / unit-specific data were not available
for tanks. Therefore, the flow rate and dimensions of the tanks must be estimated. Two model
tanks that have been used previously by EPA to assess emission from aerated tanks were used for
this analysis for both aerated and storage tanks (U.S. EPA, 1991).
Factors that impact the'relative surface area of turbulence and the intensity of that
turbulence are important factors in determining the fate of chemicals in aerated tanks. The
4-6
-------
Volume II
Section 4.0
Table 4-2. CHEMDATS Tank Model Input Requirements
Input Parameter
.Data Source/Assumption
For both aerated and storage tanks:
Windspeed (m/s)
Temperature (°C)
Depth (m)
Area (m2)
Flow rate (ms/s)
Active biomass (g/L)
Biomass solids in (g/L)
VO inlet cone. (mg/L)
Total organics in (mg/L)
Total biorate (mg/g bio-h)
For aerated tanks only:
Fraction agitated
Submerged air flow (m3/s)
Number impellers
Oxygen trans, rating (IbCVh-HP)
Power (total)(HP)
Power efficiency
Impeller diameter (cm)
Impeller speed (cm)
Set by location of WMU
Set by location of WMU
Input based on model unif
Input based on model unif
Input based on model unit3
With biodegradation: set to 2 for aerated tanks,
set to 0.05 for storage tanks
Without biodegradation: set to 0.
Assumed value of 0
Unit cone, assumed - set to 1
Assumed value of 100
Assumed value of 19
Input based on model unif
Set to 0 (mechanical aeration)
Input based on model unif
Input based on model unif = 3
Input based on model unif
Input based on model unif = 0.83
Input based on model unif = 61
Input based on model unif = 130
a Two model units were run at each meteorological location both with and without active biodegradation. The input
parameters for the two model tanks were developed by EPA in the development of emission standards for
hazardous wastes (U.S. EPA, 1991). Set to 0 for storage tanks.
aerated tank model has several input parameters that impact the degree and intensity of the
turbulence created by the aeration (or mixing). These input parameters include Input ID Nos.
T10, Tl 1, T12, T13, T14, T15, T17 and T18. The aerated tank model is most sensitive to the
fraction aerated; the total power, power per aerator (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 (see Appendix C). The values of most of these parameters were selected
from the model aerated tank parameters developed by EPA (U.S. EPA, 1991). Submerged air
flow was set to zero (assumed mechanical aeration only).
Factors that influence the rate of biodegradation are important in determining emissions
from both aerated and storage tanks. Unlike the biodegradation rate model that was used for the
land-based units, the biodegradation rate model used in CHEMDATS for tanks is dependent on
the amount of active biomass in the WMU. Therefore, the active biomass concentration (Input
ID No. T5) is a critical parameter for aerated and storage tanks. Because this parameter can vary
4-7
-------
Volume II
Section 4.0
widely for different types of tanks, duplicate model runs were performed at two different biomass
concentrations for both aerated and storage tanks. For both aerated and storage tanks, one model
run was made using a biomass concentration of zero (no biodegradation) and one model run was
made that included biodegradation. A zero biomass concentration is appropriate for short-term
storage tanks or for aerated tanks designed primarily for mixing, but not for enhanced
biodegradation units. Examples of tanks that might be actively mixed (aerated) but have little or
no biodegradation are equalization basins and neutralization tanks. For the second aerated tank
run, the CHEMDAT8 aerated tank default value of 2.0 g/L was used for the biomass
concentration. This biomass concentration is appropriate for aerated tanks designed for enhanced
biodegradation (e.g., activated sludge wastewater treatment tanks). For the second storage tank
run, a biomass concentration of 0.05 g/L was selected. This value is appropriate for wastewater
treatment tanks that are not specifically designed for biodegradation but where it may occur
passively.
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 ("VO
inlet") 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 model tanks were run
at a low unit concentration (i.e., 1 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 the storage
tank, the emission estimates are impacted by both temperature and windspeed. As the emissions
for the aerated tank are predominantly driven by the turbulent area and associated mass transfer
coefficients, the emissions from the aerated tank are not strongly impacted by the windspeed.
Aerated tank emissions are impacted by temperature. Note that, dependent on the residence time
of the waste in the tank, the temperature of the waste in the tank is not expected to vary
significantly with changing atmospheric temperatures, and annual average temperatures were
used to estimate the average waste temperature in the tanks. Each model tank was run for each
of the 29 meteorological regions used in the analysis for each of the model assumptions (aerated
and nonaerated, with biodegradation and without biodegradation, and high concentration and low
concentration).
The "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 CHEMDAT8 default value for solids in is zero, and this value was used for all aerated tank
model runs. The CHEMDAT8 default values for contaminant concentration ("VO inlet cone.")
and total organics concentration are 100 and 250 mg/L, respectively. Because a unit contaminant
concentration of 1 mg/L was used, it seemed appropriate to similarly reduce the total organics
4-8
-------
Volume II
Section 4.0
concentration. Therefore, the total organics input was set at 100 mg/L. The CHEMDAT8
default value of 19 mg/g-h for "total biorate" was used for the tank model runs.
4.3 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 equal cells 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 (either the cells are depleted
of the constituent or capped after 1 year).
• The waste is homogeneous with an initial concentration of 1 mg/kg.
• The waste matrix is aqueous (Henry's law partitioning applies).
• Annual average temperature is used (determined by assigned meteorological
station).
• Acute and subchronic exposures were not modeled.
Emissions based on the aqueous waste matrix assumption sometimes yielded
backcalculated concentration values that exceeded the soil saturation level for certain
compounds. Therefore, an additional model run was made for landfills using a concentration of
waste of 1E+6 mg/kg (pure component) and turning off the aqueous waste flag (i.e., employed
Raoult's law partitioning). These model runs provided a means to check the appropriateness of
backcalculated concentration, and they also provide a more environmentally conservative
backcalculated concentration estimate for constituents with high vapor pressure but low Henry's
law constants (e.g., formaldehyde).
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:
1. 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.
2. Loading = bulk density = 1.09577 g/cm3.
4-9
-------
Volume II
Section 4.0
3. Tilling depth (cm) = landfill depth, 1, calculated as follows:
j . . = 100 (cm/m) x [Capacity (Mg) x IE+6 (g/Mg)]
[Area (acre) x 4047 (m.zlacre) x 5w/fc Density (g/cm3) x (100 cm/m)3]
If the calculated depth was less than 50 cm or more than 832 cm, then the method
described in Section 3.1 was used.
4. Total landfill surface area was divided by 20 to get surface area of landfill cell.
5. The total landfill capacity was divided by 20 to get the average annual quantity of waste,
vtannual*
The landfill cell areas and depth were entered into the CHEMDAT8 input table (along
with average ambient temperature), and the emission fraction (in Column AU - "total emissions"
output 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:
Emiss.Rate (g/m2-s) =
x C^Jg/Mg) x Emiss.Fmct.
Areacell(m2) x 365.25jt24x3600(;y/;yr)
(4-2)
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:
Cwaste,t -
X eXP (-
(4-3)
At the end of 1 year, Cwaste/Cwaste>0 = 1 - emiss.fract. - biodegr.fract. Therefore, the Kliall t
term, at the time period for which the fraction loss terms were calculated, is simply:
K, ,, t = -In (\-emiss.Jract.-biodegr.fract.)
(4-4)
The concentration versus time profile (Equation 4-3) can then be integrated to calculate
the average waste concentration, Cwaste>ave, over the time period of the calculation:
4-10
-------
Volume If
Section 4.0
c = c
waste,ave waste,o
[1 - exp (-K,all t)]
K
(4-5)
\,all
Because columns AW and AX in the CHEMDAT8 spreadsheet model were not needed
for the landfill runs, Equations 4-4 and 4-5 were entered in these columns to calculate the
average output soil concentrations present in the active landfill cell.
The input parameters required for the landfill are presented in Table 4-1. The aqueous
waste flag is set to 1 (the most critical model flag), and the annual waste quantity and unit
dimensions are the critical source parameters. For landfills, the loading rate is pure waste
material so that Input ID No. LI is basically the waste density. A waste density of 1.09577 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).
4.4 Development of Volatile Emissions and Waste Concentrations for Land
Application Units
4.4.1 Chronic Exposure Analysis
Because the same basic CHEMDAT8 model was used for landfills and land application
units, the emissions estimates for land application units have some similarities to the landfill
emission estimates, but there are a number of differences. The basic modeling assumptions used
for modeling land application units are as follows:
• The land application unit operates at pseudo-steady-state (pseudo-steady-state
may be reached in 1 to 2 years or as time approaches infinity, depending on the
application and loss rates).
• Waste application occurs quarterly.
• The waste is homogeneous with an initial concentration of 1 mg/kg.
• The waste matrix is aqueous (Henry's law partitioning applies).
• Temperature is determined by assigned meteorological station; quarterly average
temperature was used.
- —
-------
Volume II
Section 4.0
• Biodegradation occurs at temperatures greater than 5 °C.
Emissions based on the aqueous waste matrix assumption sometimes yielded
backcalculated concentration values that exceeded the soil saturation level for some compounds.
Therefore, an additional model run was made for LAU using a concentration of waste of 1E+6
mg/kg (pure component) and turning off the aqueous waste flag (i.e., employed Raoult's law
partitioning). These model runs provided a means to check the appropriateness of backcalculated
concentration, and they also provide a more environmentally conservative backcalculated
concentration estimate for constituents with high vapor pressure but low Henry's law constants
(e.g., formaldehyde).
1.
2.
3.
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 land application units at
the facility to get land application unit-specific estimates.
, (Mg/yr)/Area(m2) 5 0.2 . If QannM/Area > 0.2, then
Loading rate, L, is calculated as follows:
Tilling depth (cm) = 20 cm i
depth (cm) = 100 x Q^^/Area.
L (g/cm3) =
Q
annual
X °-25
Area x (depth/lQQ)
(4-6)
4. Biodegradation flag input (Input ID No. L9) = 1 if temperature is greater than 5 °C. If the
temperature is 5 °C or lower, the biodegradation flag input = 0.
5. Time of calc. = 365.25/4 = 91.31 days.
6. Quarterly/seasonal temperature and windspeeds were calculated be averaging the monthly
temperature and windspeeds using the following groupings by month:
• Winter: December through February
• Spring: March through May
• Summer: June through August
• Fall: September through November.
The CHEMDAT8 model was run for each land application for each of the four seasons.
Equation 4-4 was used to estimate Klja]11 for a given season. However, for the land application
unit, additional waste input occurs after the modeled time step. 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. At pseudo-steady-state (approaching infinite time), the total contaminant mass lost
4-12
-------
Volume. II
Section 4.0
from the system equals the total contaminant mass added to the system. Including the
contaminant loss by burial/removal, the initial soil concentration at the pseudo-steady-state is
C x
waste>° [Bulk Dens.,
- |exp (-Klra[[ 0 x 11 -
Dens.
(4-7)
The average pseudo-steady-state concentration over the time interval of the calculation, in
this case 91.31 days, is calculated using Equation 4-5, but using Coss as calculated in
Equation 4-7 in the place of Cwaste 0. In this manner, the long-term average concentration for a
given season was calculated. Once the long-term average concentration was calculated for all
four seasons, the arithmetic average of these seasonal concentrations was calculated and output.
The emission fraction (in Column AU - "total emissions" output for the "intermediate
time" [91.31 days]) was multiplied by the pseudo-steady-state initial soil concentration (from
Equation 4-7), the bulk density, and the tilling depth and divided by the time period of the
calculation to determine the long-term average emission rate for that season as follows:
Emirate
C x Emiss.Fract. x Bulk Dens, x (/ft7/100)
°"
Time
x [24x3600(^]
(4-8)
The long-term average seasonal emission rates were calculated for all four seasons, then the
arithmetic average of these seasonal emission rates was calculated and output. The input
parameters required for the land application unit are presented in Table 4-1. The aqueous waste
flag is set to 1 (the most critical model flag). The annual waste quantity and unit dimensions are
the critical source parameter. 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 can hold the entire annual waste quantity. Quarterly waste applications were
employed (Input ID No. L8 was set to 91.3 days) so that any given quarterly waste application
occupied no more than 25 percent of the land application unit.
A waste density of 1.09577 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. The air
porosity has the greatest influence on the predicted emissions of these parameters. Nonetheless,
as these parameters typically do not vary over wide ranges, these parameters are considered
secondary parameters for the emissions estimates.
4-13
-------
Volume II
Section 4.0
Because biodegradation both lowers the emissions 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 land application unit runs; however, the default CHEMDAT8 model method of
calculating the biodegradation rate was not used. Instead, the HWIR values (RTI, 1995) for first-
order biodegradation rate constants were used. The HWIR biodegradation rate constants are
based, for the most part, on contaminant half-lives in soil. The biodegradation rates used in
CHEMDAT8 were derived primarily from wastewater studies, and then a low biomass
concentration was used in CHEMDAT8 to model biodegradation in soils. Because the HWIR
biodegradation rate constants were more directly linked to soil-based WMUs, the HWIR
biodegradation rate constants were used for all land-based WMUs that included biodegradation.
4.4.2 Acute and Sufochronic Exposure Analysis
Short-term (1-day and 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 or first 30 days immediately
after a waste application (subsequent to reaching steady state). As with the chronic exposure
emission estimates, biodegradation was included in the model.
The short-term emission estimates
were projected from the CHEMDAT8
"instantaneous" emission values by
weighting the instantaneous emission rate
with a corresponding time interval (see
box).
In the absence of biodegradation,
higher temperatures would produce higher
volatile emissions. However, when
biodegradation is modeled, it slows to zero
at temperatures below 5°C, thus increasing
volatile emissions at low temperatures. In
attempts to develop reasonable worst-case
emission estimates, each LAU was run at a
high and low temperature. The high
temperature used was the highest average
monthly temperature for the assigned meteorological location. The low temperature used v/as
either the lowest average monthly temperature or 5 °C, whichever was higher. Because the acute
(1-day) emission estimates are dependent on windspeed, and higher windspeeds provide higher
emissions, the windspeed used for the acute emission estimate was the annual average windspeed
for the assigned meteorological location multiplied by 1.5 to estimate a reasonable worst-case
windspeed. The higher of the emission rates calculated from the high and low temperature runs
was used as input into the exposure model. Average soil concentrations corresponding to the
emission estimates were also output for the exposure model.
Time at Which
Instantaneous
Emissions were
Calculated
0.25 h
Ih
4h
10 h
18 h
Period of Time
Represented by
Instantaneous
Emissions
0 - 0.5 h
0.5-2h
2-6h
6 - 14 h
14-24h
4-14
-------
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. 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, 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.
4.5 Development of Volatile Emissions and Waste Concentrations for
Wastepiles
4.5.1 Chronic Exposure Analysis
Again, 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 "seasonal" 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; fixed depths of 2 and 5 meters were
assumed; all wastepiles were modeled at both assumed depths.
• The waste is homogeneous with an initial concentration of 1 mg/kg.
• The waste matrix is aqueous (Henry's law partitioning applies).
• Quarterly average temperatures are used; temperature determined by assigned
meteorological station.
• Maximum monthly temperature was used for acute and subchronic exposures.
• Biodegradation occurs at temperatures greater than 5 °C.
• No biodegradation was assumed for acute and subchronic exposures.
Emissions based on the aqueous waste matrix assumption sometimes yielded
backcalculated concentration values that exceeded the soil saturation level for some compounds.
Therefore, an additional model run was made for wastepiles using a concentration of waste of
1E+6 mg/kg (pure component) and turning off the aqueous waste flag (i.e., employed Raoult's
law partitioning). These model runs provided a means to check the appropriateness of back-
calculated concentration, and they also provide a more environmentally conservative back-
calculated concentration estimate for constituents with high vapor pressure but low Henry's law
constants (e.g., formaldehyde).
4-15
-------
Volume II
Section 4.0
2.
3.
4.
Inputs that were calculated for the wastepile 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 wastepiles at the facility to
get wastepile unit-specific estimates.
Loading = waste bulk density = 1.09577 g/cm3.
Tilling depth (cm) = 200 and 500 (all units run at both depths).
Biodegradation flag input (Input ID No. L9) = 1 if temperature is greater than 5 °C. If the
temperature is 5 °C or lower, the biodegradation flag input = 0.
5. Time of calc. = average residence time of waste in the wastepile, tave, as follows:
Time Calc.(d) =
Area x (Jft7/100) x Bulk Dens, x 365.25(d/yr)
Q
•annual
(4-9)
6. The same quarterly/seasonal temperature and windspeeds calculated for land application
units were used for wastepiles.
The average waste concentration and emission rate for the wastepile can be calculated
using the equations presented for the landfill model (Equations 4-2 and 4-5). (Note: Because the
loading rate equals the bulk density, Equation 4-7 reduces to C0>ss = C0iWaste, and the equations
used for land application units could also be used; substitution of Equation 4-9 into 4-8 yields
Equation 4-2). Due to the extremely long residence times calculated for some of the wastepiles,
an if-statement was required when calculating (Kt>aU1) to prevent taking the logarithm of zero.
The model was found to truncate when: 1 - emiss.fract. - biodegr.fract < 10E-15.
The same input parameters required for land application units and landfills were used for
wastepiles (see Table 4-1). The aqueous waste flag is set to 1 (the most critical model flag). The
annual waste quantity and unit dimensions are still the critical source parameter. As with
landfills, the loading for wastepiles is pure waste material so that Input ID No. LI is basically the
waste density. A waste density of 1.09577 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 at two set wastepile depths. 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 continue to have some impact on emissions. The
biodegradation flag was set to 1 (biodegradation included) for wastepiles. The model remains
insensitive to molecular weight of the waste (for aqueous wastes) and windspeed (for long-term
emission estimates).
4-16
-------
Volume II
Section 4.0
4.5.2 Acute and Subchronic Exposure Analysis
Acute (1-day) and subchronic (1-month) emission estimates were made for wastepiles
following the same approach used for LAU as described in Section 4.4.2, with the following
differences:
• ' The wastepile was assumed to be completely filled with new waste at time t = 0.
• Biodegradation was not included (a lag period is generally associated with
biodegradation, especially for the assumed "all new waste" scenario).
• Since there was no biodegradation, only high-temperature runs were needed.
4.6 Development of Volatile Emissions for Tanks
The only output for tanks is the annual air emissions. Consequently, additional equations
did not have to be added to the CHEMDAT8 model to run the tank model as was needed for the
land-based WMUs. However, for the aerated tank emission estimates, a number of if-staterr~
were needed to prevent division by zero for chemicals that did not have biodegradation rate
constants, vapor pressure, Henry's law constant, or diffusivity inputs (e.g., metals). These if-
statements set the ultimate biodegradation rate or volatilization rate to zero when the
biodegradation rate or volatilization rate inputs were either zero or missing.
•statements
The basic modeling assumptions used (or inherent in CHEMDAT8) for the aerated tank
model emission estimates include:
• The WMU operates at steady state.
• The WMU is well mixed.
• Emission estimates were performed for two different influent concentrations: an
influent concentration of 1 mg/L (i.e., "VO Inlet cone." input = Cinfl = 1 mg/L = 1
g/m3) and an influent concentration equal to the constituent's solubility.
• The waste matrix is aqueous (Henry's law partitioning applies).
• Annual average temperatures are used; temperature determined by assigned
meteorological stations.
• Biodegradation rate is first order with respect to biomass concentrations.
• Biodegradation rate follows Monod kinetics with respect to contaminant
concentrations.
• Hydrolysis rate is first order with respect to contaminant concentrations.
4-17
-------
Volume II
Section 4.0
• Acute and subchronic exposures were not modeled.
Emissions based on the aqueous waste matrix assumption sometimes yielded
backcalculated concentration values that exceeded the aqueous solubility level for some
compounds. Therefore, additional model runs were made using a concentration of waste of 1E+6
mg/L (pure component) and modifying the CHEMDAT8 model equation for the dimensionless
Henry's law constant to be the constituent's partial vapor pressure (i.e., to employ Raoult's law
partitioning). These model runs provided a means to check the appropriateness of backcalculated
concentration values higher than the solubility limit. It also provided a more environmentally
conservative backcalculated concentration estimate for constituents with high vapor pressure but
low Henry's law constants (e.g., formaldehyde).
For tanks, the surface area, depth, and flow rate are all directly specified by the model
units. Two model units were run: a small tank arid a large tank. The emissions from these units
were modeled at all 29 of the selected meteorological regions, both with and without biomass
(i.e., 2 units x 2 biomass conditions x 29 meteorological locations x 2 concentrations = 232
runs). The storage tank emission estimates were performed using the same model tanks that
were used for aerated tanks, but the fraction agitated (T10) was set to zero.
The CHEMDAT8 model is used to calculate the emission fractions for the model units.
The emission rate, in g/m2-s, is calculated from the fraction emitted, the flow rate, waste
concentration, and the surface area as follows:
Emiss.Rate (g/m^-s) =
Qflow(m3/s) x Cinfl(glm3) x Emiss.Fract.
Area(m2)
(4-10)
4.7 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 and 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.
4.7.1 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 (U.S. EPA, 1988).
However, this is an event-based model that requires detailed site-specific information unavailable
4-18
-------
Volume II
Section 4.O
for this analysis. Therefore, it was not used. The older Cowherd model tends to 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. Since 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
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):
[u]
Z10 = 0.036 x (1-V) x IJSL
x F(x)
(4-11)
where ^
E10 = emission rate of PMi0(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
2.06 -0.33*
F(x) = 2.6-x
2.9-1.3*
0.18(8;c3+12jc)e
0.5<*<0.8
0.82
17,
x = 0.886 x —-
(4-13)
4-19
-------
Volume II
Section 4.0
The threshold windspeed, Ut, is a function of threshold friction velocity and roughness
height, as follows:
1 I 71
U = — x In — x U*
1 <>-4 U
(4-14)
where
U, = threshold windspeed (m/s)
Z = anemometer height (cm)
ZQ = roughness height (cm)
U* = threshold friction velocity (m/s).
The inputs used for these equations (other than average annual windspeed and
anemometer height, which are location-specific) are summarized in Table 4-3.
The vegetative cover, V, is an important modifying factor that can range from 0 (bare
ground) to 1 (100 percent vegetative cover). Because the surface being modeled is either an
active landfill cell or an active LAU, it seemed reasonable to assume that no vegetation would
have the opportunity to grow on the surface of the WMU while it was active. Therefore,
vegetative cover was set to 0.
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.
Table 4-3. Inputs and Intermediate Values Used for Wind
Erosion from Landfills and LAUs
Symbol
V
ZD
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
4-20
-------
Volume II
Section 4.0
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 4-4
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.
4.7.2 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 for use with this equation.
Table 4-4. Calculated Particulate Emission Rates for Landfills and LAUs
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
c; Z(c^
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
" :iW
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
n/s)T' (m/s)
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
x
(unitless)
1.8
1.5
1.1
1.5
1.1
1.7
1.5
1.5
1.9
2.0
1.6
1.9
1.7
2.0
1.4
1.4
2.0
1.8
1.6
1.3
1.5
2.5
1.5
1.7
1.5
1.5
1.2
1.4
1.9
•** v
' (unifless)
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
x »
-------
Volume II
Section 4.0
Emissions were calculated as follows:
= 1.9 x PSM x — x 365 p x -L
1.5 235 15
(4-15)
where
E10
S
P
f
PSM
emission rate of PM10 (kg/ha-d)
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) (%)
particle size multiplier for PM10 (unitless) = 0.5.
The emission rate was converted to g/m2-h as follows:
E(g/m2-h) = E(kg/ha-d) x
lQ,QQOm2/hax24h/d
(4-16)
Data on the silt content of the wastes being modeled were not available. As a
conservative assumption, the silt content of miscellaneous fill materials from AP-42 was used.
This value is 12 percent.
The number of precipitation days and the frequency of windspeed greater than 5.4 m/s
were location specific; values were obtained from NOAA (1992) and are summarized in
Table 4-5. Also included in Table 4-5 are the calculated particulate emission rates in both kg/ha-
d and g/m2-h; these values have been rounded to two significant figures.
4-22
-------
Volume II
Section 4.O
Table 4-5. Calculated Particulate Emission Rates for Wastepiles
v* A f ^ 5?™!*- > \r*^
*••*•'><./•
Location ^ ?
Albuquerque
Atlanta
Bismarck
Boise
Casper
Charleston
Chicago
Cleveland
Denver
Fresno
Harrisburg
Hartford
Houston
Huntingdon
Las Vegas
Lincoln
Little Rock
Los Angeles
Miami
Minneapolis
Philadelphia
Phoenix
Portland, ME
Raleigh-Durham
Salem, OR
Salt Lake City
San Francisco
Seattle
Winnemucca
"; <*&>''
58
116
96
91
95
113
125
157
89
89
125
126
101
142
27
91
104
33
128
113
117
37
129
110
146
92
63
157
67
*- s
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
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
(kg/ha-d^
15
11
19
12
28
10
16
15 .
12
4.4
8.6
11
9.3
3.9
19
18
8.1
11
13
19
14
4.8
12
8.0
7.2
12
24
9.9
11
" E
(g/mP-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
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
4-23
-------
-------
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 ug/m2-s) (unitized air concentrations, or UACs) at a variety of potential receptor
locations. The following sections discuss model selection, the critical parameters of the model,
and the model results or UACs.
5.1 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 dispersion of vapors
and particulates from landfills, land application units, wastepiles, and tanks to receptors both on-
and offsite 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) onsite
and offsite 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, 1995
Industrial Source Complex - Long Term v.3 (ISCLT3) - U.S. EPA, 1995
Toxic Modeling System - Short Term (TOXST) - U.S. EPA, 1994c
Fugitive Dust Model (RDM) - 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, 1995) was selected for all
aspects of this analysis because it met all the criteria. This model, however, has considerable run
times, which limited the number of meteorological stations included in this analysis.
5.2 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 F.
5-1
-------
Volume II
Section 5.0
Table 5-1. Air Dispersion Model Capabilities
Source
Geometry
Area
Model Source
Source
Height
Elevated Ground
Receptor
Location
Onsite ' /
Air Offsite
Cone. Air Cone.
Chemical Phase
Vapor Particulate
Averaging
Period
Annual Monthly
, -
batty
ISCST3
ISCLT3
TOXST
FDM
COMPDEP
aMinimnm height of source for modeling is 0.5 meters.
5.2.1 General Assumptions
This section discusses depletion,
rural vs. urban, and terrain assumptions.
5.2.1.1 Depletion. Air
concentrations can be calculated in ISCST3
with or without wet and dry depletion.
Modeled concentrations without depletions
are higher than those with depletions. A
sensitivity analysis was conducted that
showed that the differences in the
maximum concentrations with depletion
and without depletion are small at close-to-
source receptors, increasing only slightly as
the distance from the source increases. The
sensitivity analysis also shows that the run
time for calculating concentrations using
the ISCST3 model with depletion options
is 15 to 30 times longer than the run time
without depletions for the 5th and 95th
percentile of the sizes of land application
units. (The difference is greater for larger
sources; see sensitivity analysis in Appendix
calculated without depletions in this analysis
could be modeled in the time available.
Assumptions Made for Dispersion Modeling
Dry and wet depletion options were not activated in
the dispersion modeling.
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 0,25, 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 |Ug/s-m2.
F for details.) Therefore, concentrations were
so that a greater number of meteorological locations
5.2.1.2 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 heavily metropolitan
5-2
-------
Volume II
Section 5.0
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.
5.2.1.3 Terrain. Rat 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. Complex terrain
applications are extremely site-specific; therefore, 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.
5.2.2 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. 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 this analysis.
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
• South Pacific Coastal
• Southwest
Northwest Mountains
Central Plains
Southeast
Midwest
Northern Atlantic
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
5-3
-------
Volume II
Section 5.0
Table 5-2. Meteorological Stations Used hi the Air Characteristic Study
f 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
MM
GA
ND
ID
WY
SC
IL
OH
CO
CA
PA
CT
TX
WV
NV
ME
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
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
* Longitude1
- ^ 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).
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
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 survey data, the Industrial D 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. Four
sites in Alaska and four in Hawaii were deleted from the analysis at this point because the 29
meteorological stations are limited to the continental United States. Figure 5-1 shows the final
5-4
-------
Map Legend
• Industrial Subtitle D WMU
(*) Met. Stations (29)
O Met. Regions (29)*
r~l 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
Subtropical Division
j Subtropical Regime Mountains
•> i Temperate Desert Division
$ Temperate Desert Regime Mountains
ff 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
, .*»*
*Met. station regions created from thlessen polygons
and Bailey's Ecoregion boundaries.
Created SIS/SB
File: eoo_29met_sites.api
Figure 5-1. Meteorological station regions.
-------
Volume II
Section 5.0
meteorological station boundaries used for the study along with the locations of the Industrial D
facility sites.
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 F.
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 centerline concentration.
5.2.3 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.
5.23.1 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.
5.2.3.2 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
5-6
-------
Volume II
Section 5.0
orientations is not available, a square source was chosen to minimize the errors caused by source
shapes and orientations. (See sensitivity analysis in Appendix F for details.)
5.2.3.3 Source Areas Modeled. In the modeling analysis, five types of WMUs were
considered (i.e., landfill, land application unit, wastepile, aerated tank, and storage 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 5-2, 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.
In order to address this model sensitivity, yet avoid modeling approximately 2,000
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. In addition, separate areas were
modeled for tanks, because these were based on model units, rather than the Industrial D Survey
database. (See Section 3 for further details.) Fourteen area strata were selected for landfills and
land application units, seven for wastepiles, and two for tanks. The median area size for each
stratum was used in the dispersion modeling analysis. Tables 5-3, 5-4, and 5-5 present the source
areas and heights used in the modeling analysis.
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.
5.2.4 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 F 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
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
5-7
-------
Volume II
Section 5.0
downwind distance increases. Therefore, for annual average concentrations, the receptor points
were placed on 0, 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. The first receptor square (i.e., 0 meter) is
at the edge of the unit. For monthly and daily averaging periods used in the subchronic and acute
assessment, the receptors were placed on 0, 25,50, and 75 meter receptor squares.
5.3 Unitized Air Concentrations
Unitized air concentrations (UACs) were calculated by running ISCST3 with a unit
emission rate (i.e., 1 /zg/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 inputs 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 21 areas 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 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. It
was assumed that the greatest risk of acute exposure would be closest to the site; therefore, the
receptors points were placed at 0,25, 50, and 75 meters from the edge of the WMU, with 16
equally spaced directions at each distance.
The maximum annual average UACs are presented in Tables 5-6 through 5-8 for the
different types of WMUs. 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 (i.e., 0-m receptors or onsite receptors). 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. The results in Tables 5-6 through 5-8 show that the annual
average UACs increase with the increasing area size of the sources.
Figures 5-3 through 5-6 show that 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.
5-10
-------
Table 5-6. Maximum Annual Average Unitized Air Concentrations Cug/m3 / /zg/s-m2) for Landfills and Land Application Units
1 f <
1 t\
-------
to
Table 5-7. Maximum Annual Average Unitized Air Concentrations Owg/m3 / /zg/s-m2) for Wastepiles
)••..-•• • -,. .
I .-A •..;•• • • .
; Met Station
Albuquerque, NM
Atlanta, GA
Bismarck, ND
Boise, ID
Casper, WY
Charleston, SC
Chicago, IL
Cleveland, OH
Denver, CO
Fresno, CA
Harrisburg, PA
Hartford, CT
Houston, TX
Huntington, WV
Las Vegas, NV
Lincoln, NE
Little Rock, AR
Los Angeles, CA
Miami, FL
Minneapolis, MN
Philadelphia, PA
Phoenix, AZ
Portland, ME
Raleigh-Durham, NC
Salem, OR
Salt Lake City, UT
San Francisco, CA
Seattle, WA
Winnemucca, NV
Station
No.
23050
13874
24011
24131
24089
13880
94846
14820
23062
93193
14751
14740
12960
3860
23169
14939
13963
24174
12839
14922
13739
23183
14764
13722
24232
24127
23234
24233
24128
Area (m^-m Height Wastepiles)
20
0.037
0.043
0.035
0.056
0.040
0.038
0.038
0.049
0.054
0.077
0.047
0.049
0.042
0.057
0.045
0.032
0.045
0.055
0.041
0.033
0.045
0.062
0.046
0.043
0.048
0.052
0.046
0.053
0.040
162
0.171
0.195
0.155
0.235
0.181
0.168
0.170
0.214
0.237
0.344
0.214
0.212
0.191
0.248
0.194
0.142
0.201
0.255
0.181
0.147
0.198
0.274
0.196
0.191
0.209
0.232
0.207
0.240
0.172
486
0.378
0.431
0.343
0.520
0.405
0.372
0.380
0.479
0.518
0.744
0.477
0.474
0.424
0.548
0.432
0.317
0.442
0.564
0.404
0.326
0.439
0.597
0.433
0.424
0.466
0.514
0.464
0.540
0.380
2,l
-------
Volume II
Section 5.0
Table 5-8. Maximum Annual Average Unitized Air
Concentrations C"g/m3///g/s-m2) for Aerated and Storage Tanks
Met Station S '/
Albuquerque, NM
Atlanta, GA
Bismarck, ND
Boise, ID
Casper, WY
Charleston, SC
Chicago, IL
Cleveland, OH
Denver, CO
Fresno, CA
Harrisburg, PA
Hartford, CT
Houston, TX
Huntington, WV
Las Vegas, NV
Lincoln, NE
Little Rock, AR
Los Angeles, CA
Miami, FL
Minneapolis, MN
Philadelphia, PA
Phoenix, AZ
Portland, ME
Raleigh-Durham, NC
Salem, OR
Salt Lake City, UT
San Francisco, CA
Seattle, WA
Winnemucca, NV
Station s
. "N*>r ^
23050
13874
24011
24131
24089
13880
94846
14820
23062
93193
14751
14740
12960
3860
23169
14939
13963
24174
12839
14922
13739
23183
14764
13722
24232
24127
23234
24233
24128
Area (a
27
"\ V
0.00286
0.00333
0.00245
0.00519
0.00425
0.00257
0.00248
0.00408
0.00383
0.00652
0.00378
0.00462
0.00321
0.00403
0.00265
0.00336
0.00272
0.00779
0.00328
0.00235
0.00350
0.00506
0.00317
0.00302
0.00532
0.00465
0.00543
0.00594
0.00282
i*»;
430s
0.04652
0.06414
0.04142
0.09329
0.08087
0.04466
0.04656
0.07670
0.06834
0.12357
0.06610
0.07620
0.06281
0.07845
0.04930
0.05724
0.04850
0.12923
0.05823
0.04401
0.05938
0.08872
0.05184
0.05285
0.08962
0.08360
0.09108
0.10704
0.04978
5-13
-------
45 i
»»t' * n i iB
So 13 S
s
y>
5
5 5
2
Meteorological Location
e 5-3» Maximum UAC by meteorological location (landfills and LAUs)*
o,
-------
Volume II
Section 5.0
i
n T
mm
C^ r* ^H ^H ^H r*
Qui-s/8/ui/gn)^
3
^^Hi San Francisco, CA
4
3
3 1
• -
••' t>
-™_ ^
^^jOT^^^l
^^^^PS^^a
•^
•^ra Albuquerque, MM
f[\ nmunxBp\[
I
o
i
V)
2
I
5-15
-------
Volume II
Section 5.0
D|Area«2;lOOn2
^Arn2
h a
T "«
~S
™ *
73^
( ui-s/§ / m/3n) 3Vfi ranunxep\[
£ £
ce
41
I
I
g
i
Bfi
O
I
1
i
21
5-16
-------
Ll-S
a
i
1
i
o
o
Maximum UAC (ug/m / g/s-m2)
o
3
S'
-------
-------
Volume II
Sectitm 6.O
6.0 Development of Inhalation Health
Benchmarks
Inhalation health benchmarks for chronic, subchronic, and acute exposure durations were
needed. This section describes the health benchmarks used in this study.
6.1 Chronic Inhalation Health Benchmarks Used in This Study
Chronic inhalation health benchmarks used in this study include inhalation reference
concentrations (RfCs) for noncarcinogens and inhalation unit risk factors (URFs) and inhalation
cancer slope factors (CSFs) for carcinogens. URFs 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, 1998). 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 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 no IRIS or HEAST chronic inhalation health benchmarks were 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
Naphthalene
Phenol
Pyridine
Tetrachloroethylene
1,1,1 -Trichloroethane
Xylenes.
6-1
-------
Volume I!
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.O
Table 6-1. Chronic Inhalation Health Benchmarks Used in the
Air Characteristic Analysis
r^
/•
a
"CAS*^
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
•^
vV
" -/ Name,^ ^ ^ >'
Acetaldehyde
Acetone
Acetonitrile
Acrolein
Acrylamide
Acrylic acid
Acrylonitrile
Allyl chloride
Aniline
Arsenic
Jariurn
Benzene
3enzidine
3enzo(a)pyrene
Beryllium
Bromodichloromethane
3romoform
(Tribromomethane)
Butadiene, 1,3-
Cadmium
Carbon disulfide
Carbon tetrachloride
Chloro-l,3-butadiene, 2-
(Chloroprene)
Chlorobenzene
Chlorodibromomethane
CWoroform
Chlorophenol, 2-
ChromiumVI
Cobalt
Cresols (total)
Cumene
Cyclohexanol
Noncarcinogens
RfC
^(mgfm?)
9.0E-03
3.1E+01
5.0E-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
NA
l.OE-05
4.0E-04
4.0E-01
6.0E-05
v «*- x
/ \
RfC Target Orgsffi*
Respiratory
Neurological
Jver
Respiratory
Respiratory
Respiratory
Sfeurological
Spleen
Reproductive
Respiratory
Reproductive
Respiratory
Kidney and liver
Repro/developmental
Respiratory
Hematological
Kidney and adrenal
Muscle
/
Ref*
I
A
H
I
I
I
I
I
H
I
I
H
H
D
D
D
I
FR
Carcinogens ;-
InhalURF
(Hgto?)J
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
1.7E-03
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
InhalCSF
(mg/kg/dXS
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
6.0E+00
8.4E+00
6.2E-02
3.9E-03
9.8E-01
6.3E+00
NA
5.3E-02
NA
NA
8.4E-02
8.1E-02
NA
4.2E+01
NA
NA
NA
NA
V
Ref
I
I
I
I
I
I
I
D
I
I
I
I
D
I
I
(continued)
6-3
-------
Volume II
Section 6.0
Table 6-1. (continued)
CAS#
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
Name
Dibromo-3-cbloropropane,
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-
Dimethylformainide, 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, -
" Noncarcinogens
V <. ^ * -, y M -~~ *& _,
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
/ A
•V t
BfC Target Oi^an
Reproductive
Body weight
Reproductive
Liver
Respiratory
Respiratory
Respiratory
Liver
NA
No liver, kidney, or
hemato effects
Respiratory
Respiratory
Reproductive (male),
hemato
Reproductive
Developmental
Reproductive
Respiratory
Respiratory
Respiratory
Respiratory and
neurological
SfiP
I
H
I
H
I
I
I
I
D
I
I
D
I
I
H
D
H
H
I
,•> " Carcinogens *
InhalURF
Xfig/m3)-1
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
InhalCSF^
(mg/fcg/d)-*~
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
/<•*
ReC
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)
t ^^ S ? '•
< ' * •
v" s*
>t
.CASfr
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
V "* -S'^-' > J (
• >^ '
^ ^ s*. <
' Name*- ;
isophorone
Manganese
Mercury
VIethanol
Methoxyethanol acetate,, 2-
Vlethoxyethanol, 2-
Methyl bromide
(bromomethane)
Methyl chloride
(chloromethane)
Methyl ethyl ketone
Methyl isobutyl ketone
VIethyl methacrylate
VIethyl tert-butyl ether
Methylcholanthrene, 3-
Vlethylene chloride
Naphthalene
Nickel
Nitrobenzene
Nitropropane, 2-
Nitrosodiethylamine
Nitrosodi-n-butylamine
re-Nitrosopyrrolidine
Phenol
Phthalic anhydride
Propylene oxide
Pyridine
Styrene
TCDD, 2,3,7,8-
Tetrachloroethane, 1,1,1,2-
Tetrachloroethylene
Tetrachloroethane, 1,1,2,2-
^t' s ^ Noncarcinogens^ ^
v. *&•
•*|MC
/m^M3)
4.0E-03
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:OE+00
NA
3.0E+00
l.OE-02
NA
2.0E-03
2.0E-02
NA
NA .
NA
2.0E-02
1.2E-01
3.0E-02
7.0E-03
l.OE+00
NA
NA
3.0E-01
NA
RfC Target Organ
Body weight
Neurological
STeurological
Developmental
Reproductive (male)
Reproductive
Respiratory
Developmental
Kidney and liver
Respiratory
Kidney and li ver
Liver
Respiratory
Kidney, liver,
hematological,
adrenal
Liver
No effects
Respiratory
Respiratory
Liver
Neurological
Neurological
x s
Ref
FR
I
I
D
D
I
I
I
H
I
I
H
A
H
I
FR
H
I
O
I
A
*' Carcinogens
InhalURF
Ivfrtf'
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
4.30E-02
1.60E-03
6.10E-04
NA
NA
3.7E-06
NA
NA
NA
7.4E-06
NA
5.8E-05
InhalCSF^
tragfcg/d)-1*
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
NA
2.0E-01
ReF,
H
D
I
I
'H
I
I
I
I
H
I
I
(continued)
6-5
-------
Volume II
Section 6.0
Table 6-1. (continued)
CAS# .'.
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
Name
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)
"* *** " ^6?% "^ *- f %f
- _ Noncarcinogens
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
3.0E-01
,' V
RfC Target Organ
Respkatory and
neurological
Body weight
Liver
Neurological
Kidney and
respiratory
No respiratory effects
Respkatory
Respkatory
Neurological
Re£a
I
H
H
SF
H
I
D
I
A
— '.*•> i- f j f
Carcinogens
InhalURF
(ug/rn3)''
NA
6.9E-05
NA
NA
NA
1.6E-05
1.7E-06
NA
NA
NA
NA
8.4E-05
NA
InhalCSF
(mg^gVd)-1
NA
2.4E-01
NA
NA
NA
5.6E-02
6.0E-03
NA
NA
NA
NA
3.0E-01
NA
Ref
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, 1998)
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, 1996c; U.S. EPA, n.d.)
FR = 61FR 42317-354 (U.S..EPA, 1996b)
D = Developed for this study.
O = Other source (see Sections 6.1.1 and 6.1.2).
For acetone, naphthalene, tetrachloroethylene, and total xylenes, ATSDR's chronic
inhalation MRLs were used. Naphthalene is currently undergoing review by EPA's IRIS pilot
program (future publication date not known) and a new RfC may be available soon. Provisional
RfCs were identified for cyclohexanol, isophorone, and phenol in a Federal Register notice (61
FR 42317) concerning solvents listings (U.S. EPA, 1996b). An inhalation 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
-A
N f
'CAS*
67-64-1
108-93-0
78-59-1
91-20-3
108-95-2
110-86-1
127-18-4
71-55-6
1330-20-7
^ «
X v A» N
x. 4.
Chemical Name
Acetone
(2-propanone)
Cyclohexanol
Isophorone
Naphthalene
Phenol
Pyridine
Tetrachloroethylene
1,1,1-
Trichloroethane
Xylenes (total)
^ /
Inhalation Benchmark
and Benchmark Value
x
RfC=13ppm(31
mg/m3)
Provisional RfC =
0.00006 mg/m3
Provisional RfC= 0.004
mg/m3
RfC = 0.002 ppm
(0.01 mg/m3)
Provisional RfC =
0.02 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 (0.3
mg/m3)
H Target
Organ
Neurological
Muscle
Body weight
Respiratory
No effects
Liver
Neurological
Neurological
Neurological
• ; v \ < •- ' /
!•
< ^ Source
ATSDR 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. NHS PB82-172917
61 FR 42343 (U.S. EPA, 1996b);
standard RfC methodology
61 FR 42345 (U.S. EPA, 1996b);
standard RfC methodology
ATSDR chronic inhal MRL based
on NTP TR-410 (1992); value may
change because naphthalene is
undergoing review in the IRIS pilot
program
61 FR 42336 (U.S. EPA, 1996b);
standard RfC methodology
Cited in Health and Environmental
Effects Profile (KEEP) for Pyridine
(EPA/600/X-86-168)
ATSDR chronic inhal MRL based
on Ferroni et al. 1992.
Neurobehavioral and
neuroendocrine effects of
occupational exposure to
perchloroethylene.
Neurotoxicology 12: 243-247
Superfund risk issue paper
(U.S. EPA 1996c)
ATSDR chronic inhal MRL based
on Uchida et al., 1993. Symptoms
and signs in workers exposed
predominantly to xylenes. IntArch
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
6-7
-------
Volume II
Section 6.0
1,4-Dioxane
2-Ethoxyethanol acetate
Ethylene glycol
Methanol
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 for
2-chlorophenol (U.S*. EPA, 1998). RfCs were derived for 2-ethoxyethanol acetate and
2-methoxyethanol acetate based on RfCs for 2-ethoxyethanol and 2-methoxyethanol,
respectively.
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, 1998) 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 (1997b) were used as the cancer benchmarks.
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,
1998), route-to-route extrapolation is not recommended because of the potential for respiratory
tract effects following inhalation exposure and "first-pass" effects following ingestion exposure.
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 E.
6-8
-------
Table 6-3. Chronic Inhalation Health Benchmarks Derived for This Study
<,
/
1 CA'ss '
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
*, ' -Vs"
fo 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 = 0.0014 mg/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
x RfC Target
•-. Organ
Repro/
developmental
Respiratory
Hematological
Liver, kidney,
hematological
! t > "• ^
Method o]f 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
Instituto Gigiyeny Profzabolevaniya, 7:115-9.
(Russian)
Inhal CSF and URF derived by CalEPA (1997)
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.
I
(continued)
-------
Table 6-3. (continued)
j
j 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-Methylcholanthrene
o-Toluidine
Vanadium
Inhalation Benchmark
and Benchmark Value
RfC= 0.3 mg/m3
RfC= 0.6 mg/m3
RfC = 13 mg/m3
RfC=0.03mg/m3
Inhal CSF = 7.4E+00 per mg/kg/d
Inhal URF = 2.1E-03 per ng/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
i
Method of Derivation f
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 (1997)
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)
p\
o
-------
Volume II
Section 6.0
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 included
subchronic RfCs and ATSDR intermediate inhalation MRLs. Figure 6-2 describes the approach
used to select the subchronic inhalation health benchmarks used in this analysis.
ATSDR intermediate inhalation MRLs are for use with exposure durations of 15 to 364
days and are derived from subchronic lexicological or epidemiological studies. Subchronic
inhalation RfCs were identified in HEAST (U.S. EPA, 1997'a) or were derived from existing
RfCs or chronic inhalation MRLs. A number of chronic RfCs cited in 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. Five RfCs cited in IRIS were based on chronic
studies; for these constituents, a modifying factor (MF) of 10 was applied to extrapolate from
chronic to subchronic duration to derive subchronic RfCs. Chronic MRLs were available for
1,2-dichloroethane and nickel (0.2 ppm and 2.0E-4 mg/m3, respectively); a modifying factor was
used to extrapolate from chronic to subchronic duration, resulting in interim subchronic RfCs
(8.1 mg/m3 and 2.0E-3 mg/m3, respectively). RfCs derived for 1,4-dioxane, naphthalene, and
tetrachloroethylene were based on chronic inhalation studies; a modifying 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,
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 MF
Subchronic
health
benchmark
value
for use
in
modeling
if no chronic RfC available
Used ATSDR intermediate
inhalation MRL
Figure 6-2. Approach used to select subchronic noncancer
inhalation benchmark values.
6-11
-------
Volume II
Section 6.0
Table 6-4. Subchronic Inhalation Health Benchmarks
fcs •• .'
;",;; 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
10061-02-6
Name
Acetaldehyde
Acetone
Acetonitrile
Acrolein
Acrylic acid
Acrylonitrile
Allyl chloride
Aniline
Barium
Benzene
Carbon disulfide
Carbon tetrachloride
Chloro-l,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-
Dichloropropene, trans-1,3-
Subchronic
RfC(mg/m3)
9.0E-02
3.1E+01
5.0E-01
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
2.0E-05
3.0E-05
l.OE-04
4.0E+00
6.0E-04
2.0E-03
2.0E+00
2.5E+00
2.0E+00
8.1E+00
7.9E-02
1.3E-02
2.0E-02
2.0E-02
Target Organ-
Respkatory
Neurological
Liver
Respiratory
Respkatory
Respiratory
Neurological
Spleen
Developmental
Neurological
Neurological
Liver
Respkatory
Liver, kidney
Liver
Respkatory
Respkatory
Hematological
Kidney, adrenal
Muscle
Reproductive
Body wt
Liver
Liver
Liver
Liver
Respkatory
Respkatory
Respkatory
" .' : /,-v
- ' f „ > ' ' '*
Source
RfC based on Subchronic study -
removed UF
ATSDR intermediate MRL
HEAST
RfC based on subchronic study -
removed UF
HEAST
RfC based on chronic study -
adjusted w/ MF
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 MRL avail. - adjusted w/ MF
ATSDR intermediate MRL
HEAST
HEAST
HEAST
(continued)
6-12
-------
Volume II
Section 6.0
Table 6-4. (continued)
r *p.
?- *
-------
Volume II
Section 6.0
CAS#
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
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-
TricWorobenzene, 1,2,4-
Trichloroethane, 1,1,1-
Trichloroethylene
Trichlorofluoromethane
Triethylamine
Vanadium
Vinyl acetate
Vinyl chloride
Xylenes (total)
Subchronic
RfCimg/m3)
2.0E-03
2.0E-02
2.0E-02
2.0E-01
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-04
2.0E-01
7.7E-02
3.0E+00
>
s Target Organ
Respiratory
Hemato, adrenal,
kidney, liver
Liver
No effects
Respiratory
Respiratory
Neurological
Liver
Neurological
Respiratory,
neurological
Body wt
Liver
Neurological
Neurological
Kidney,
respiratory
No effects
Respiratory
Respiratory
Liver
Developmental
<-_ > * s V
Source » /
chronic MRL avail. - adjusted w/ MF
HEAST
HEAST
RfC based on subchronic study -
removed UF
HEAST
HEAST
HEAST
ATSDR intermediate MRL
RTI-derived RfC based on chronic
study - adjusted w/ MF
RfC based on chronic study -
adjusted w/ MF
HEAST
HEAST
ATSDR intermediate MRL
ATSDR intermediate MRL
HEAST
RfC based on subchronic study -
removed UF
RfC based on subchronic study -
removed UF
HEAST
ATSDR intermediate MRL
ATSDR intermediate MRL
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 in 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
inhalation MRLs, EPA acute exposure guideline levels (AEGLs), and CalEPA (1995) 1-h acute
6-14
-------
Volume If
Section 6.O
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 lexicological or epidemiological studies (see Table 6-5). AEGLs have been
derived for aniline and ethylene oxide. CalEPA (1995) 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
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
* ' •> Name ^ „ - ' \
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 tert-butyl ether
Methylene chloride
Tetrachloroethylene
Toluene
Trichloroethane, 1,1,1-
Trichloroethylene
Vanadium
Vinyl chloride
Xylenes (total)
TSDR acute inhi
MRL(mg/m3)
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
1.4E+00
1.4E+00
1.1E+01
1.1E+01
1.1E+01
2.0E-04
1.3E+00
4.3E+00
- ~ Target Organ
Neurological
Ocular
Neurological
Immunological
Liver
Liver
Developmental
Immunological
Respiratory
Kidney
Respiratory
Neurological
Neurological
Neurological
Neurological
Neurological
Neurological
Neurological
Neurological
Neurological
Respiratory
Developmental
Neurological
6-15
-------
Volume II
Section 6.0
Table
• 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
67-56-1
109-86-4
78-93-3
7440-02-0
108-95-2
75-56-9
100-42-5
6-6. CalEPA's 1-Hour Acute Inhalation
* CafiEPX 1-h KEL
Name (mg/m3)
Acrylic acid
Arsenic
Carbon disulfide
Dioxane, 1,4-
Epichlorohydrin
Ethoxyethanol acetate, 2-
Ethoxyethanol, 2-
Mercury
Methanol
Methoxyethanol, 2-
Methyl ethyl ketone
Nickel
Phenol
Propylene oxide
Styrene
Used ATSDR acute
inhalation MRLs
6.0E+00
3.9E-04
3.1E+00
1.8E+00
8.0E-01
3.3E-01
8.8E-02
1.8E-03
2.8E+00
2.3E-02
1.5E-01
1.6E-03
1.5E-01
9.3E-01
2.2E+01
Reference Exposure Levels (RELs)
•V,!. 1, ^ **
EffectLevel -Effect
I
n
n
i
i
n
n
n
i
i
i
i
i
i
i
Respiratory irritation
Repro/developmental
Repro/developmental
Eye irritation
Eye & respir irritation
Repro/developmental
Hemato,
Repro/developmental
Repro/developmental
Mild neuro
Hemato
Respiratory irritation
Immuno
Respiratory irritation
Eye & respir irritation
Eye & respir irritation
if not available
\ '
Used EPA
AEGL values
if not available
\ r
UsedCaIEPA1-h
acute inhalation RELs
V
Acute
health
benchmark
value
for use
in
modeling
A
Figure 6-3. Approach used to select acute noncancer
inhalation health benchmark values.
6-16
-------
VoZume II
Section 6.0
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:
• LevellandAEGL-l: discomfort or mild effect level
• Level n and AEGL-2: disability or serious effect level
• Level ICE 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 4 were for
semivolatiles not modeled in LAUs and wastepiles.
6-17
-------
-------
Volume II
Section 7.0
7.0 Development of Risk-Specific Waste
Concentration Distribution
This section describes the Monte Carlo model that combines 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 exposures, and the model used to calculate risk-specific waste concentrations for
subchronic and acute exposures. In addition, it describes modifications made to the methodology
for lead.
7.1 Overview
A Monte Carlo analysis was performed in which the location of the receptor 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). 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.
A thousand 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 95"
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
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 facilities in the United
States 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.
Each iteration of the Monte Carlo simulation required the following steps:
• Select a receptor location
rth
7-1
-------
Volume II
Section 7.0
• Obtain the appropriate unitized air concentration
• Calculate air concentration
• Select exposure factors from distributions
• Obtain health benchmarks
• Calculate risk or hazard quotient
• Backcalculate risk-specific waste concentration (Cw).
These steps are discussed in more detail in the following sections.
7.2 Select Receptor Location
Dispersion coefficients were modeled for each of 7 distances and 16 directions as
described in Section 5.0. The 7 distances were 0,25,50,75,150,500, 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.
Four receptors were modeled: an adult resident, a child resident, an offsite worker, and an
onsite worker. The onsite worker was always located at 0 meters from the WMU. The adult and
child residents and the offsite worker were assessed at all of the six offsite distances (any
distance except 0 meters). 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, both on- and offsite, for a particular iteration.
7.3 Obtain Unitized Air Concentrations
After on- and offsite receptor locations had been selected, UACs for those two locations
were calculated. For each location, the UAC was estimated by interpolating between the UACs
developed for areas immediately above and below the actual area of the unit, as follows:
( A-At]
UAC=\ - x (UACj-UAC}} + UAC{
where
UAC
A
unitized air concentration for specific WMU ([ng/m3]/[|j.g/m2-s])
area of specific WMU (m2)
area modeled in dispersion modeling immediate below area of specific WMU
(m2)
area modeled in dispersion modeling immediate above area of specific WMU
(m2)
7-2
-------
Volume II
Section 7.0
UAQ = unitized air concentration developed for area i ([ug/m3]/[ug/m2-s])
UACj = unitized air concentration developed for area j ([ug/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, A, and UACj
were set to the values for the smallest area modeled, and Aj 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.
7.4 Calculate Air Concentration
Two air concentrations, one for the onsite worker and one for the offsite receptors, were
calculated from the WMU emission rate (which remains constant through all iterations of the
Monte Carlo simulation for a particular WMU) and the on- and offsite dispersion coefficients.
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:
Cair = (Evapor x 106 ug/g + CM x Eparticl 3600 slh) x UAC
(7-2)
where
Cai,. = air concentration associated with a unit waste concentration ([|ig/m3]/[mg/kg]
or [ug/m3/[mg/L])
Evapor = emission rate of constituent in vapor phase ([g/m2-s]/[mg/kg] or
[g/m2-s]/mg/L])
Csoa = average annual soil concentration in unit ([mg/kg]/[mg/kg] = [ug/g]/[mg/kg] or
= emission rate of particulates (g/m2-h)
UAC = unitized air concentrations ([ug/m3]/[ng/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
Exposure factor distributions are described in the following subsections. Values for each
exposure factor (inhalation rate, body weight, and exposure duration) were selected
7-3
-------
Volume II
Section 7.0
independently for adults, children, and workers. The same values were used for a particular
iteration for both on- and offsite workers. Exposure frequency remains constant for all iterations,
though it varies by receptor. All data in this section are from the draft Exposure Factors
Handbook (U.S. EPA, 1997; hereafter, the draft 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 used for consistency with the data on inhalation rate in
the draft EFH.
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 being 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.
Single values for inhalation rates are presented as recommended values for long-term
dose assessments in the draft EFH. Values of 11.3 mVd 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. No distributional data are recommended for this parameter.
Although the draft 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 draft EFH. As part of this effort, Myers et al. (1997) present two fitted
distributions, one lognormal and one gamma, for inhalation rate for male and female residents in
6 age ranges (0-3 years, 4-10 years, 11-18 years, 19-30 years, 31-60 years, and >60 years). Myers
et al. find that the difference between the two distributions is negligible and recommend using
the lognormal distribution.
7-4
-------
Volume II
Section 7.0
Table 7-1. Estimated Parameters for Inhalation Rate for
Residents Assuming Lognormal Distribution
Age5(yr) "
0-3
0-3
4-10
4-10
11-18
11-18
19-30
19-30
31-60
31-60
>60
>60
Male
Female
Male
Female
Male
Female
Male
Female
Male
Female
Male
Female
v^s
7.52
5.75
9.30
8.65
14.58
10.76
16.75
11.14
16.32
10.95
12.69
10.44
(%-*)
73
71
30
31
36
31
31
30
32
29
34
29
' Stdlfev
"
-------
Volume II
Section 7.0
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 (to be used for both
adult residents and workers) and male children ages 0-3,4-10, and 11-18 years were needed.
The draft EFH has summary data on body weight for adults broken down by gender and
race. The data for males of all races ages 18 to 74 years were used. These are summarized in
Table 7-3, Body weight is typically assumed to be normally distributed, and the percentiles
shown in Table 7-3 agree well with a normal distribution. Therefore, a normal distribution was
defined using the mean and standard deviation for males 18 to 74 years of. age (these values are
shown in bold in the table). Unlike a lognormal distribution, which cannot contain negative
values, a normal distribution can take on negative values. Given the mean and standard
deviation used, this is extremely unlikely; however, the distribution was explicitly truncated at
zero to ensure that negative values did not occur.
For children, the draft EFH contains mean and standard deviations of body weights for
1-year age intervals (e.g., 1 year, 2 years). These values, summarized in Table 7-4, had to be
pooled into the age ranges used for inhalation rates (0-3,4-10, and 11-18 years).
If X denotes a random variable of interest (in this case, body weight), then the variance
(or standard deviation squared) of X satisfies the following:
o2(X) =
- E(X) x E(X)
(7-3)
Table 7-3. Body Weights for Males, All Races, Ages 18-74 Years (kg)
Percentiles
A *«. vr i;
Age Sample
•(years) Size
18-74 5,916
18-24 988
25-34 1,067
35-44 745
45-54 690
55-64 1,227
65-74 1,199
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
5*
58.6
56.8
59.5
59.7
50.8
59.9
54.4
10*
62.3
60.4
62.9
65.1
65.2
63.8
58.5
15th'
64.9
61.9
65.4
67.7
67.2
66.4
61.2
25*
68.7
64.8
69.3
72.1
71.7
70.2
66.1
n /
76.9
72.0
77.5
79.9
79.0
77.7
74.2
~ 75*
85.6
80.3
85.6
88.1
89.4
85.6
82.7
85*
91.3
85.1
91.1
94.8
94.5
90.5
87.9
90*
95.7
90.4
95.1
98.8
99.5
94.7
91.2
95* '
102.7
99.5
102.7
104.3
105.3
102.3
96.6
Note: Bolded values were used in this analysis.
Source: U.S. EPA, 1997, Table 7-3.
7-6
-------
Volume II
Section 7.0
Table 7-4. Body Weights for Male Children, Ages 6 Months to 18 Years (kg)
S f VS
Age (years) ~*
6-11 months
1
2
3
4
5
6
7
8
9
Mean
9.4
11.8
13.6
15.7
17.8
19.8
23.0
25.1
28.2
31.1
St^Der*
1.3
1.9
1.7
2.0
2.5
3.0
4.0
3.9
6.2
6.3
Age (years)
10
11
12
13
14
15
16
17
18
" / Mean
36.4
40.3
44.2
49.9
57.1
61.0
67.1
66.7
71.1
* StdDev
7.7
10.1
10.1
12.3
11.0
11.0
12.4
11.5
12.7
Source: U.S. EPA, 1997, Table 7-2.
where
a2(X) = variance of X (the standard deviation, a, squared)
E(X2) = mean of X2
E(X) = mean of X.
In general, if there are L finite populations to be pooled, of sizes nl5 n2,.... nL, and the mean, E(X),
and standard deviation, a, are known for each population, then the mean and standard deviation
of the pooled population may be found as follows. The ith population accounts for a proportion
Pi = njA^-H^+.-.+nL) of the total. The mean of X2, E(X2), may be calculated for each population
by solving Equation 7-3 for E(X2). Then the pooled values of E(X) and E(X2) are found by
averaging these quantities of the L populations, using the proportions p{ as probabilities in
calculating the average:
n
n
E(X) = £ p. x E(X) or E(X2) = pt x E(X}
(7-4)
Once these values are obtained, the pooled variance (and standard deviation) can be calculated
from Equation 7-3.
For body weights for children, it was assumed that each population to be pooled (i.e.,
each 1-year age range) was of equal size N. The pj values were therefore equal to N/N*L, or 1/L
(where L was the number of populations to be pooled). Table 7-5 summarizes the resulting mean
and standard deviations of body weight for each age group. These were then used to define a
normal distribution, which, like the one for adults, was truncated at zero to prevent nonsensical
results.
-------
Volume II
Section 7.0
Table 7-5. Pooled Body Weights for Children (kg)
Age Group (years)
Mean
StdDev
0-3
4-10
11-18
12.63
25.90
57.18
2.90
7.91
15.61
7.5.3 Exposure Duration
Data on exposure duration in the draft EFH were obtained from the distributions
presented for population mobility (Chapter 14.3). Because data on males were used for
inhalation rate and body weight, data on males were used for exposure duration for consistency.
For adult residents, regression methods described in Myers et al. (1997) were used to fit
gamma, lognormal, Weibull, and generalized gamma distributions to the tabulated percentiles
from Table 14-158 in the draft EFH. The gamma model fit best of the two-parameter models and
fit adequately relative to the generalized gamma model (goodness-of-fit p-values of .96, .07, and
.70 for females, males, and males/females combined). Therefore, the gamma distribution was
used. Table 7-6 summarizes the data from Table 14-158 of the draft EFH and the parameters of
the fitted gamma distribution.
Table 14-159 of the draft EFH provides descriptive statistics on population mobility for
children by age. These data are summarized here in Table 7-7. Because children were divided
into three age groups, exposure duration had to be approached in a somewhat different manner
for children. Instead of using a distribution for exposure duration for children in the Monte Carlo
analysis, a distribution on the age of the child at the start of exposure was used, and the 90th
percentile of residential occupancy period for that age was fixed as the exposure duration.
Table 7-6. Descriptive Statistics for Residential Occupancy Period for Males (years)
•f, • ' ' ' \
Statistic or Parameter
Alpha (shape)
Beta (scale)
Mean
5th percentile
10th percentile
25* percentile
50* percentile
75* percentile
90* percentile
95* percentile
98* percentile
99* percentile
NA s Not applicable.
Source: Data column: U.S. EPA, 1997, Table 14-158.
, %% „ .
NA
NA
11.1
2
2
4
8
15
24
31
39
44
Fitted Distribution
1.32
8.37
11
1
2
4
8
15
24
30
39
45
7-8
-------
Volume fl
Section 7.O
Table 7-7. Descriptive Statistics for Population Mobility for Children (years)
m **
Percentile
Current- Age (years)
3
6
9
12
15
18
2S
3
4
5
5
5
4
50
5
7
8
9
8
7
7?-,
8
10
12
13
12
11
90
13
15
16
16
16
16
95
17
18
18
18
18
19
99
22
22
22
23
23
23
Source: U.S. EPA, 1997, Table 14-159.
Exposure was constrained to begin at the beginning of one of the age ranges (so at 0,4, or
11 years). Each of the three ages had a probability assigned to it reflecting the relative number of
years in that age range. Those probabilities and the 90th percentile residential occupancy period
used are summarized in Table 7-8. When an exposure duration was longer than the age range,
exposure was continued in the next age range until the exposure duration or age 18 was reached.
For example, if exposure began at 0 years, then 3 years was modeled as age 0 to 3,7 years was
modeled as age 4 to 10, and 3 years was modeled as 11 to 18, for a total of 13 years.
For workers, the typical default exposure values used in the past were an 8-h shift,
240 d/wk, for 40 years. The draft 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
draft 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 draft 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.
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 draft EFH to develop a
distribution for exposure duration. Therefore, the average of 6.6 years was used.
Table 7-8. Distribution of Age at Start of Exposure and Exposure Duration for Children
.
Age RangeXyears)
. -"f 4,X
- - -?s. !•"* •
Years in Range
' " _r< Exposure Duration
Probability^ > " ~^ (years)
0-3
4-10
11-18
3
7
8
0.17
0.39
0.44
13
15
16
Source: U.S. EPA, 1997, Table 14-159.
7-9
-------
Volume II
Section 7.0
7.5.4 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 240 d/yr was used for workers.
These are based, respectively, on 7 d/wk and 5 d/wk for 50 wk/yr and account for the receptor
being elsewhere on vacation for 2 wk/yr.
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). The draft
EFH (U.S. EPA, 1997) cautions that care must be taken that the assumptions about population
parameters in the exposure analysis are consistent with the population parameters used in the
dose-response analysis used to develop these health benchmarks. If the exposure scenario is not
consistent with the standard factors applied in developing the dose-response relationships, then
adjustment must be made for the dose-response relationship to reflect the exposure being
modeled.
7.6.1 Carcinogens
In the IRIS derivation of cancer dose-response (D-R) assessments, a standard exposure
scenario is assumed in calculating the slope factor (i.e., human cancer risk per unit dose) on the
basis of either animal bioassay data or human data. This standard scenario has traditionally been
assumed to be typical of the U.S. population and includes a body weight of 70 kg, an inhalation
rate of 20 mVd, and a lifetime of 70 years. The use of these specific values has depended on
whether the slope factor was derived from animal or human epidemiological data.
For D-R studies based on animal data, the animal dose is scaled to human equivalent
doses using a human body weight assumption of 70 kg. No explicit lifetime adjustment is
necessary because the assumption is made that events occurring in the lifetime animal bioassay
will occur with equal probability in a human lifetime, whatever that might happen to be.
(For D-R studies based on human studies (either occupational or general population), the
Agency has usually made no explicit assumption of body weight or human lifetime. For both of
these parameters, there is an implicit assumption that the population usually of interest has the
same descriptive parameters as the population analyzed by the Agency. In the rare situation
where this assumption is known to be wrong, the Agency has made appropriate corrections so
that the D-R parameters represent the national average population.
The draft EFH provides a table of correction factors for the D-R values tabulated in the
IRIS database for carcinogens. Because risks were characterized for three distinct subpopulations
(i.e., adults, children and workers), each of which has different body weights and inhalation rates,
adjusted cancer slope factors were calculated for each subpopulation using the following
equation from the draft EFH:
7-10
-------
Volume II
Section 7.0
= CSFstd x
(7-5)
where
CSFadj = cancer slope factor adjusted to population of interest (per mg/kg-d)
CSFstd = standard cancer slope factor from IRIS (per mg/kg-d)
BW = body weight for the subpopulation of interest (kg).
Befcause this adjustment depends on body weight, it was performed anew for each iteration of the
Monte Carlo simulation.
7.6.2 Noncarcinogens
For noncarcinogens, the inhalation reference concentrations were applied without
adjustment. RfCs are derived using a NOAEL (no observed adverse effect level) or LOAEL
(lowest observed adverse effect level) that has been identified from an epidemiological or animal
study. To adjust for interspecies differences, the NOAEL or LOAEL must be converted into a
human equivalent concentration (HEC) when using animal data. To do so, interspecies
differences such as lung tidal volume, breathing frequency, fractional deposition in respiratory
tract regions, surface area of respiratory tract region of interest, body weight, and blood:gas (air)
partition coefficient must be accounted for. Once a human equivalent concentration for the
identified NOAEL or LOAEL has been determined, uncertainty factors are applied. The
uncertainty factors account for the use of a LOAEL, interspecies variability, and human
variability (e.g., for the protection of sensitive subpopulations, such as children, the elderly, the
ill, or those who have been exposed previously).
RfC dosimetry adjustments are based on adult input parameter values. However, an
uncertainty factor (usually a factor of 10 or 3) for human variability is intended to account for the
uncertainty and variability that might be due to pharmacokinetic and pharmacodynamic
considerations, including those related to age differences. For example, airway architecture,
breathing patterns and rates, metabolic rates, and sensitivity differ between adults and children,
but the application of this uncertainty factor is expected, on a basic level, to include this
uncertainty. Therefore, Agency guidance does not recommend additional adjustment of the RfC
using ratios of body weights and inhalation rates of children and adults. It would also not be
appropriate to convert the RfC to an oral RfD and then modify this value using ratios of body
weights and inhalation rates of children and adults. Mechanisms of lexicological action are
different for oral exposures, and the same health effect may not occur from the inhalation
pathway.
7-11
-------
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 for
each iteration based on the calculated air concentration and the exposure factors selected for the
iteration.
Risk for carcinogens was calculated as follows:
Risk
C. x l
-------
Volume II
Section 7.O
Cajj. = air concentration associated with a unit waste concentration
([Hg/m3]/[mg/kg])
RfC = reference concentration (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 and an onsite worker (which
differ only in location) were modeled for noncarcinogens. As discussed in Section 7.6.2, the
uncertainty factors used in establishing a reference concentration for noncarcinogens should be
considered to account for differences between different human receptors.
A single risk or hazard quotient was calculated for adult receptors (adult resident, offsite
worker, and onsite worker). However, because the exposure factors for children can vary greatly
as the child ages from infancy to adulthood, the child receptor was broken into three age ranges
for the purposes of calculating risk. A separate risk or hazard quotient was calculated for each
age range for each iteration. For carcinogens, the overall risk to a child resident was the sum of
the risks for the three age ranges. For noncarcinogens, it is not appropriate to sum hazard
quotients; therefore, the maximum of the three age ranges was used as the overall hazard
quotient. Not all the age ranges were used for each iteration. The age at which exposure is
assumed to begin was varied for each iteration, and a fixed exposure duration associated with
each starting age was then used. Some starting age/exposure duration combinations result in
exposure only during a single age range, while others result in exposure over two of the age
ranges. None result in exposure over all three age ranges. The values used for starting age and
exposure duration for children are detailed in Section 7.5.3.
7.8 Backcalculate Risk-Specific Waste Concentration
The final step in each iteration 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, in the models used in this analysis, this may be done
by a simple ratio technique:
Risk
w ~ ~Risk
'crit
or
crit
'calc'd
HQ,
(7-8)
calc'd
where
Cw
Riskcrit
Riskcalc,d
risk-specific waste concentration (mg/kg)
risk criterion (unitless)
risk associated with unit waste concentration (per mg/kg)
hazard quotient criterion (unitless)
hazard quotient associated with unit waste concentration (per mg/kg).
7-13
-------
Volume II
Section 7.0
When a particular constituent had both carcinogenic and noncarcinogenic effects, the
carcinogenic risk was used to continue the calculations, because it is generally more
conservative.
7.9 Adjustments for Results Not Meeting Linearity Assumption
As mentioned above, 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 within certain restrictions (e.g., that the
concentration does not exceed the saturation concentration). Therefore, it was necessary to check
backcalculated results that did not meet those restrictions and modify them. The restrictions on
the emissions model, and therefore the checks performed, differ for land-based units and tanks.
These are discussed separately in the following sections.
7.9.1 Adjustments for Land-Based Units
For this analysis, wastes were initially assumed to be aqueous phase (i.e., dilute wastes
that partition primarily to water within the soil). However, aqueous phase wastes can only occur
up to the soil saturation limit. At concentrations above the soil saturation limit, wastes can only
occur in organic phase. The soil saturation limit is calculated as follows:
where
C =
'-'sat
Q __
Pb =
H'
a
soil saturation limit (mg/kg)
solubility limit (mg/L)
bulk density of soil / waste matrix (kg/L)
soil-water partition coefficient (L/kg)
water-filled soil porosity (unitless)
dimensionless Henry's law constant (unitless, = H/RT)
air-filled soil porosity (unitless).
Wastes can also occur in the organic phase at concentrations below the soil saturation
limit, but for most chemicals, the aqueous phase produces greater emissions than the organic
phase for the same concentration. A few chemicals (most notably formaldehyde) have greater
emissions from the organic phase than the aqueous phase. If the aqueous phase emission rate
results in a backcalculated waste concentration that exceeds the soil saturation limit, then the
chemical poses no risk in the aqueous phase. However, it may still pose risks in the organic
phase at higher concentrations.
In order to address both chemicals with greater emissions from the organic phase and
chemicals that pose no risk in the aqueous phase, emissions were also modeled in the organic
7-14
-------
Volume ff
Section 7.0
phase using a concentration of 1,000,000 mg/kg, or pure component; these were then normalized
to a unit concentration by dividing by 1,000,000. When a backcalculated waste concentration
based on the aqueous phase assumption exceeded the soil saturation limit, it was recalculated
using the normalized organic phase emission rate. Similarly, if the normalized organic phase
emission rate was greater than the aqueous phase emission rate, it was used instead. If the
backcalculated waste concentration based on the organic phase emission rate exceeds
1,000,000 mg/kg, then the chemical poses no risk in the organic phase.
All results were flagged to indicate the emission rate on which they were based (aqueous
or organic). If results were based on the organic phase emission rate, the flag also indicates
whether they produced higher emissions and therefore higher risk, or whether there was no risk
from the aqueous phase. When there was no risk from either phase, the waste concentration was
set to 1,000,000 ppm and flagged as no risk. These flags are very important to note when
comparing WMUs and receptors for a particular chemical. As a consequence of the differences
in emission rate between the aqueous phase and the organic phase, there can be a significant drop
in emission rate (and therefore in risk) right around the soil saturation limit, causing a significant
increase in the backcalculated waste concentration from one WMU to another or one receptor to
another.
7.9.2 Adjustments for Tanks
For tanks, the concentration limit for the aqueous phase is the solubility of the chemical
in water. This was addressed in a manner similar to that described in Section 7.9.1 for land-
based units, except that, if the aqueous phase posed no risk and the organic phase did, the waste
concentration was capped at the solubility (essentially limiting concentrations to the range that
can occur in the aqueous phase.) This was done because it was not considered plausible that
pure, organic phase waste would be disposed of directly in tanks.
In addition, for tanks with biodegradation, the biodegradation rates are not linear. At low
concentrations, biodegradation 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 the solubility. This emission rate then was normalized to a unit
concentration by dividing by the solubility. When the backcalculated waste concentration
exceeded the half-saturation constant, suggesting that biodegradtion would be zero order, it was
recalculated based on the normalized solubility limit emission rate.
All results were flagged to indicate whether they were based on the aqueous emission rate
or capped at the solubility and, for tanks with biodegradation, whether emissions were based on
first-order or zero-order kinetics. When there was no risk from either aqueous phase or pure
component, the waste concentration was set to 1,000,000 mg/L and flagged as no risk. These
flags are very important to note when comparing WMUs and receptors for a particular chemical.
As a consequence of the differences in emission rate between the aqueous phase and the organic
phase, and the difference between first-order and zero-order biodegradation kinetics, there can
be a significant drop in emission rate (and therefore in risk) right around the solubility limit or
the half-saturation limit, causing a significant increase in the backcalculated waste concentration
from one WMU to another or one receptor to another.
7-15
-------
Volume II
Section 7.0
7.10 Methodology for Subchronic and Acute Exposure
Exposure and risk modeling for subchronic and acute exposures differed somewhat from
the modeling for chronic exposures in several respects. This section describes any differences in
the methodology for subchronic and acute exposures for each step of the modeling described in
Sections 7.1 through 7.9 for chronic exposures.
7.10.1 Overview
No Monte Carlo analysis was performed for subchronic and acute exposures. Instead, a
point estimate of Cw was calculated for each WMU in the Industrial D database. This point
estimate is most comparable to the 100th percentile of the distribution generated by the Monte
Carlo model for chronic exposures and 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 what is
presented in the results for subchronic and acute exposures.
7.10.2 Select Receptor Location
For subchronic and acute exposures, the most exposed of the 16 receptors at each
distance from the site was used. The direction of this receptor varied from WMU to WMU
depending on prevailing wind directions.
7.10.3 Obtain Unitized Air Concentration
Unitized air concentrations were interpolated as described in Section 7.3 for chronic
exposures, using the maximum UACs at each distance for subchronic and acute averaging times.
7.10.4 Calculate Air Concentration
Air concentration was calculated as described in Section 7.4 (Equation 7-2), using
emission rates and residual soil concentrations developed for subchronic and acute averaging
times.
7.10.5 Select Exposure Factors
The health benchmarks used for subchronic and acute exposure are analogous to the
chronic, noncarcinogen benchmarks used in the chronic exposure analysis. They are expressed
as ambient air concentrations, and are compared directly to the modeled air concentration,
without the use of exposure factors.
7.10.6 Obtain Health Benchmarks
Health benchmarks for subchronic exposures were obtained for 64 chemicals, and acute
health benchmarks were obtained for 35 chemicals, as described in Section 6.
7-16
-------
Volume II
Section 7.0
7.10.7 Calculate Hazard Quotient
Hazard quotient was calculated as described in Section 7.7 (Equation 7-7).
7.10.8 Backcalculate Risk-Specific Waste Concentration
Risk-specific waste concentration was backcalculated as described in Section 7.8
(Equation 7-8).
7.10.9 Adjustments for Results Not Meeting Linearity Assumptions
The adjustments described in Section 7.9 were also made for subchronic and acute
exposures.
7.11 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.
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 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
7-17
-------
Volume II
Section 7.0
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 7-9. 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 1EUBK 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. Ihgestion 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 hours per day 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.
Table 7-9. Summary of Inputs for IEUBK Model
Input
Age(yr)
Value
Source
Exposure Factors
Inhalation rate (mVd) 0-3
3-7
Body weight (kg) 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 (ng/g)
Indoor dust (ng/g)
Water (fag/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
7-18
-------
Volume II
Section 7.O
Model outputs were produced for three target blood lead levels (10, 8, and 5 |ig/dL) and
two background soil concentrations (75 and 200 |Jg/g); these are shown in Table 7-10.
Depending on the target PbB and the background soil concentration, acceptable air lead
concentrations ranged from background to 3.3 ug/m3. For comparison, the default background
air lead concentration used by the IEUBK is 0.1 ug/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 ug/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 ug/g;
therefore, that concentration was used. This resulted in air concentrations of 0.5 and 1.8 ug/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.
Table 7-10. Results of IEUBK Modeling
-. > - ^*° <*< ^ ,
Target Blood Level Soil/Dust Concentration
(pg/dL), > ^ * HagfeT
10
8
5
10
8
5
200
200
200
75
75
75
?"£ , ^ <£ Air Concentration3 t^g/rn3) ^
X0- to3-year-old
0.5
Ob
Ob
2.4
1.25
Ob
/ *• ,
3- to 7-year-old
1.8
0.6
Ob
3.3
2.1
0.35
1 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.
7-19
-------
-------
Volume II
Section 8.0
8.0 Analysis of Uncertainty
In any national assessment of this type there are numerous potential uncertainties and a
wide range of variability in many of the input parameters. 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 sensitivity analyses. However, much of the uncertainty could
only be addressed qualitatively. The 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. The variability in
input parameters that was addressed in this analysis, as well as some significant sources of
uncertainty that were not, are discussed in this section according to four broad categories:
emissions modeling, dispersion modeling, exposure modeling/risk estimation, and decision rule.
This study does not attempt to address modeling error in this analysis directly. Details of the
derivation of parameter values based on sensitivity analyses are presented in Appendixes A
through D.
Uncertainty introduced
into the analysis includes model
uncertainty, parameter uncertainty,
and parameter variability. The
first two, model uncertainty and
parameter uncertainty, are
generally recognized by risk
assessors as major sources of
uncertainty.
Parameter uncertainty
occurs when parameters in
equations 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.
Main Sources of Uncertainty in Risk Assessment
General Type
Model uncertainty
Parameter uncertainty
Parameter variability
Specific Source of Uncertainty
Surrogate variables
Excluded variables
Abnormal conditions
Incorrect model form
Measurement errors
Random errors
Systematic errors
Heterogeneity
8-1
-------
Volume II
Section 8.0
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, as well as the 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 that 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.
Variability, the third source of uncertainty, is often used interchangeably with the term
"uncertainty," but this is not strictly correct. Variability may be tied to variations in physical and
biological processes and cannot be reduced with additional research or information, although it
may be known with greater certainty (e.g., age distribution of a population may be known and
represented by the mean age and its standard deviation). "Uncertainty" is a description of the
imperfection in knowledge of the true value of a particular parameter or its real variability in an
individual or group. In general, uncertainty is reducible by additional information gathering or
analysis activities (better data, better models), whereas real variability will not change (although
it may be more accurately known) as a result of better or more extensive measurements.
Table 8-1 presents the major categories of uncertainty and how they have been addressed
in this study. The columns in the table show model uncertainty, parameter uncertainty, and
parameter variability. The rows present the four main model components in the analysis:
emissions model, dispersion model, exposure model, and risk model.
8.1 Emissions Modeling
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 particulate emission model. As discussed in Section 3.1, there
are many features of this model that meet the needs of this analysis. However, the model was
developed to address only volatile emissions from these waste management units. Competing
mechanisms such as runoff and erosion and leaching are not included in the model. In so much
as these competing processes actually occur, the model would tend to slightly overestimate the
volatile emissions and waste/soil concentrations in the waste management unit. 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 waste/soil concentration that is the basis for estimating particulate emissions of a
contaminant.
Among the many parameters needed as inputs to the emissions model, four are critical to
the emission estimates: (1) area and depth (or capacity) of the WMU, (2) temperature,
(3) volatility, and (4) biodegradation. Both volatility and biodegradation are correlated with
8-2
-------
Volume If
Section 8.0
Table 8-1. Summary of How Uncertainties Were Addressed in the Study
** *i "* •*"•*•< Y™
Model Uncertainty' 'Parameter Uncertainty Parameter,Variability
Emissions Model
Instantaneous release
model used for acute
and subchronic peak
releases
Dependencies of
biodegradation, volatility,
and temperature addressed
through sensitivity analysis
and use of seasonal
temperature variations
Facility-specific locations,
dimensions and waste
volumes used to address
variability in WMU
parameters
Dispersion Model
Model error increased
by 2%-10% by not
using plume depletion
option
Acute, subchronic
and chronic averages
used
Sensitivity analyses 29
conducted on a number of
parameters including shape
and orientation of WMU, 14
meteorologic data, and
receptor grid
meteorologic stations
used to represent
climate regions
surface areas used to
represent distribution
of surface area for
landfills, LAUs
surface areas and 2
heights used for
wastepiles
model tanks
Exposure Model
Acute, subchronic
and chronic
exposures estimated
Sensitivity analysis
conducted for receptor grid
16 receptor locations at
each distance used at each
WMU in Monte Carlo
analysis
Distributions developed
for exposure factors
(inhalation rate, body
weight, and exposure
duration) and used in
Monte Carlo analysis
Risk Model
Acute, subchronic
and chronic health
benchmarks used
Not addressed
Not addressed
temperature and therefore do not vary as independent variables. With regard to area and depth,
the data used were derived from facility-specific data contained in the Industrial D database
(Shroeder et al., 1987). When more than one WMU of the same type existed at a facility, the
combined area of those WMUs was reported in the database. This analysis used the average area
of WMUs of a single type when more than one unit existed at the same facility.
In order to address the variability in the surface area and capacity of a waste management
unit, unit-specific data were used in the emissions model. The variability in these parameter
values spans several orders of magnitude for all WMU types except tanks. Since these are very
8-3
-------
Volume II
Section 8.0
sensitive inputs to the emissions model, it was deemed important to capture this variability in the
analysis.
The uncertainty associated with the information contained in the Industrial D database is
unknown. There are several sources of this uncertainty including (1) missing data on waste
volumes or capacity, (2) multiple WMUs of the same type associated with a combined surface
area and waste volume, and (3) the accuracy of the reported data (i.e., measurement error). For
multiple units of the same type, an average value for the facility was used to model a single
WMU. 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.
A sensitivity analysis was conducted to determine the impact of seasonal variations in
temperature on both volatility and biodegradation. Section 4 provides a detailed discussion of
the assumptions made in modeling emissions from each WMU. For land application units and
wastepiles, biodegradation is an important parameter that lowers both emission rates and the
average waste/soil concentration. Biodegradation was assumed to occur at temperatures greater
than 5°C. In order to turn this on and off, quarterly seasonal temperatures were calculated for
each site based on regional-specific meteorologic 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 these land-based WMUs. The landfill scenario assumed no biodegradation
occurred and temperature was therefore a less important parameter for this WMU than for the
land treatment unit or wastepile. 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. However, for a few chemicals (see Appendix B) there was a very
large impact for certain 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.
Seasonal temperatures were not used to adjust volatility or biodegradation rates for tanks.
The assumption in this case is that biologically active tanks must maintain conditions suitable for
the biodegradation to occur throughout the year. In addition, the tanks are assumed to have a
steady flow-through, and the tank temperature is more a function of the temperature of the
entering wastewaters than the ambient temperature.
8.2 Dispersion Modeling
The EJCST3 model was used to calculate the dispersion of particle and gas emissions
from a WMU. This model has many capabilities need for this assessment such as the ability to
model area, volume, or point sources for chronic, subchronic, and acute averages, and the ability
to provide onsite concentrations that were used to evaluate exposure to onsite workers. For
dispersion modeling of this type, it is considered a fairly good 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 the plume depletion option. As discussed in Section 5, this option dramatically increases
8-4
-------
Volume II
Section 8.0
run time and would have required a much longer schedule to complete. The sensitivity analyses
presented in Appendix D show that model error is increased by 2 to 10 percent when plume
depletion is turned off.
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 wind-
speed 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., maximum 15 day) and acute (i.e., maximum 1 day) UACs as
discussed in Section 5. 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. These datasets were combined with 14
surface areas representing the distribution for landfills and land treatment units. Combined, this
provides 406 different sets of UACs (a set consists of all UACs at various distances and
directions at one location) to use with emissions data to estimate air concentrations. Similarly,
wastepiles used 406 different sets of UACs based on 7 surface areas, 2 heights, and 29
meteorological stations. Tanks used the same 29 meteorological stations and two different
characterizations of tanks, providing 58 sets of UACs. To minimize error associated with the use
of discrete surface areas, an interpolation routine was used to estimate the UAC for a specific
surface area, using the UACs for the areas immediately above and below the WMU used.
Obviously, 29 meteorological stations do not represent every site-specific condition that
could exist in the continental United States.
However, based on the EQM report (1993), it
is believed that these stations provide a
representation of the variability in wind rose
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.
However, a detailed analysis of the wind rose
for the 29 met stations is presented in
Appendix B.
Shape of Wind Rose for
29 Meteorological Stations
Shape of Wind Rose
Narrowly distributed
Moderately distributed
Evenly distributed
Bimodally distributed
No. of Stations
10
6
8.3 Exposure Modeling/Risk Estimation
The potential location of receptors was the main variability addressed in the exposure
modeling. The data used to identify and characterize WMUs contained no information on the
location and types of receptors near 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 living at that exact point. Since 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
8-5
-------
Volume II
Section 8.0
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 being located in any of 16 directions at each
distance.
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. 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
uncertainty 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 noncarcinogens incorporates
uncertainty only about receptor location since exposure factors are not used in the calculation.
Risk estimation is a straightforward calculation using health benchmarks. There is
recognized uncertainty in the health benchmarks; however, this uncertainty has not been
explicitly addressed in this analysis.
The overall output of the analysis includes specific consideration of the variability in site-
specific WMU information, regional-specific meteorological conditions, the location of
receptors, and the 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 IE.
8.4 Indirect Exposures
This risk analysis addresses only inhalation exposure to humans. This is considered 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 the project was to be completed in 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
8-6
-------
Volume II
Section 8.O
potential concern for humans or domestic animals such as beef or dairy cows that might graze in
fields adjacent to a WMU.
In summary, EPA looked at a waste stream in which polychlorinated aromatic
hydrocarbons (PAHs) were the constituents of concern from a risk perspective. Releases from an
LAU 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 thought to have a high level of uncertainty.
Studies are currently underway 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 group. At this time, it is not known which scenario
would be the most appropriate to use for the air characteristic analysis or 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, fti a sense, this sensitivity analysis should be
considered a bounding analysis since it is based on a set of inputs that is biased toward
determining if the air-to-plant pathway could be a concern for any of the constituents.
8-7
-------
-------
Volume II
Section P. O
9.0 References
Agency for Toxic Substances and Disease Registry (ATSDR). Minimal Risk Levels (MRLs) for
Hazardous Substances. http://atsdrl.atsdr.cdc.gov:8080/mrls.hrml
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/
CalEPA (California Environmental Protection Agency). 1995. Technical Support Document for
the Determination of Acute Toxicity Exposure Levels for Airborne Toxicants. Draft for Public
Comment. Office of Environmental Health Hazard Assessment, Berkeley, CA.
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.
Coburn, J., C. Allen, D. Green, and K. Leese. 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.
Environmental Quality Management, Inc., and E.H. Pechan & Associates. 1993. Evaluation of
Dispersion Equations in Risk Assessment Guidance for Superfund (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.
Myers, L., J. Lashley, and R. 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.
9-1
-------
Volume II
Section 9.0
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). 1994. NIOSH Pocket 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.
OSHA (Occupational Safety and Health Administration). 1997. Occupational Safety and Health
Standards: Subpart Z - Toxic and Hazardous Substances. 29 CFR 1910. 7/1/97 edition of Code
of Federal Regulations.
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.
U.S. 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. Environmental Protection Agency. 1985b. Rapid Assessment of Exposures to Particulate
Emissions from Surface Contamination Sites. Office of Health and Environmental Assessment,
Washington, DC.
U.S. 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. 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. 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. 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. Pp. C-19
through C-30.
9-2
-------
Volume II
Section 9.0
U.S. 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. 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. 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. 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.
U.S. Environmental Protection Agency. 1995. 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. Environmental Protection Agency. 1996a. EPACMTP Background Document. Office of
Solid Waste. Washington, DC.
U.S. EPA (Environmental Protection Agency). 1996b. Hazardous Waste Management System;
Identification and Listing of Hazardous Waste; Solvents; CERCLA Hazardous Substance
Designation and Reportable Quantities; Proposed Rule. 61 FR 42317-354. August 14.
U.S. EPA (Environmental Protection Agency). 1996c. 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. 62 FR 58839-58851.
October 30.
U.S. Environmental Protection Agency. 1997c. Exposure Factors Handbook. Draft. Office of
Research and Development, National Center for Environmental Assessment.
9-3
-------
Volume II
Section 9.0
U.S. EPA (Environmental Protection Agency). 1998. 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.
9-4
-------
Appendix A
Basic Dalenius-Hodges Procedure for
Constructing Strata
-------
-------
Volume IT
Appendix A
A. Basic Dalenius-Hodges Procedure for Constructing Strata
(applies to census data)
Select an available variable X correlated with the variable Y of interest. (Both X and Y
are assumed to be continuous.)
Form a relative frequency histogram of X with respect to M prespecified intervals with
breakpoints a,,, al9 a^ ...,aM. Let fj denote the frequency count for the i"1 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=l,2,...M), compute Z. = ^L/i where L{ = a.-a.^ is the length
of the i* interval.
4. Compute CUMZi =
j=l
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 1th 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).
1 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 IB
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, log octanol-water partition coefficient, and the soil biodegradation rate constants.
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.
Kollig, H.P., J.J. Ellington, S.W. Karickhoff, B.E. Kitchens, J.M. Long, EJ. Weber, and N.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-1. Chemical Specific Input Parameters
w
CAS# COMPOUND NAME
50000 Formaldehyde
50328 Benzo(a)pyrene
55185 N-Nltrosodlethylamine
56235 Carbon tetrachloride
56495 3-Methylcholanthrene
57976 7,1 2-Dlmethylbenz[a]anthracene
62533 Aniline
67561 Methanol
67641 Acetone
67663 Chloroform
67721 Hexachloroethane
68122 N.N-Dimethyl formamide
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
7521 8 Ethylene oxide
75252 Tribromomethane
75274 Bromodichloromethane
75354 1,1-DichloroethyIene
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
79016 Trichloroethylene
79061 Acrylamide
79107 Acrylic acid
Mol.
Wt. Density
fo/mol) fo/cc)
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
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
log
VAP. HLaw Dlffuslvity Diffuslvity Antolnes1 Vapor Oct Hydrol.
Press. Const. In Water In Air Pressure Coefficients Water Kmax K1 Rate
(atm- Part. mgVO/g-
frrimHa) m3/mol) (cm2/sec) (cm2/sec) A B C Coeff. hr. L/a-hr. sec-1
5240 3.4E-07
5.5E-09 1.1E-06
0.86 3.6E-06
115 0.0304
7.7E-09 9.4E-07
5.6E-09 3.1E-08
0.49 1.9E-06
126 4.6E-06
230 3.9E-05
197 0.00367
0.21 0.00389
4 1.9E-07
95 0.00558
124 0.0172
1620 0.00624
4300 0.00882
2980 0.027
91.1 3.5E-05
902 7.9E-05
433 0.00219
359 0.03022
1094 0.00012
5.51 0.00054
50 0.0016
600 0.0261
532.1 8.5E-05
803 0.097
4850 0.343
332 0.4815
0.0596 0.027
0.438 6.6E-06
52 0.0028
95.3 5.6E-05
23.3 0.00091
73.5 0.0103
0.007 1E-09
4 1.2E-07
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
1.00E-05
6.80E-06
1.92E-05
9.80E-06
8.80E-06
1.21E-05
6.50E-06
1.23E-05
1.66E-05
1.41E-05
1.17E-05
1.00E-05
1.45E-05
1.03E-05
1.06E-05
1.04E-05
1.00E-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
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
1.01E-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
7.195 971
9.246 3724
6.934 1242
8.164 3364
6.955 2163
6.950 1467
7.897 1474
7.117 1211
6.493 929
7.228 1348
6.928 1401
6.905 1211
6.827 1147
7.566 1301
7.093 949
6.991 969
7.119 1314
8.005 1600
6.968 1074
6.942 1169
7.128 1055
7.988 2159
7.966 1847
6.972 1099
7.067 1133
6.884 1043
7.590 1329
8.784 1894
8.415 2835
7.963 2481
6.980 1380
7.112 1305
7.192 1480
6.518 1019
11.293 3940
5.652 649
244 -0.05
273 6.11
273 0.48
230 2.73
273 6.42
171 6.62
177 0.98.
229 -0.71
230 -0.24
196 1.92
133 4
196 -1.01
221 2.13
219 2.48
273 1.19
249 0.91
251 1.5
230 -0.34
292 1.25
223 1.25
242 2
238 -0.3
273 2.35
273 2.1
237 2.13
236 0.03
237 2.53
273 2.16
273 3.16
273 5.39
273 1.7
223 1.97
229 0.28
229 2.05
193 2.71
273 -0.96
155 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
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
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 Solubility
sec-1 mo/L
6.08E-10 5.50E+05
4.61 E-08 2.50E-02
1.56E-08 9.30E+04
3.13E-08 7.93E+02
1.22E-07 3.23E-03
2.43E-09 2.50E-02
6.95E-10 3.61 E+04
6.08E-10 1.00E+06
6.08E-10 1.00E+06
2.43E-09 7.92E+03
1.56E-08 5.00E+01
1.00E-20 1.00E+06
1.39E-09 1.75E+03
2.37E-08 1.33E+03
2.43E-09 1.52E+04
2.43E-09 5.33E+03
1.56E-08 2.76E+03
2.43E-09 1.00E+06
1.00E-20 1.00E+06
2.43E-09 1.30E+04
1.00E-20 1.19E+03
1.00E-20 3.83E+05
1.56E-08 3.10E+03
1.00E-20 6.74E+03
1.56E-08 2.25E+03
1.00E-20 4.76E+05
3.13E-08 1.10E+03
1.56E-08 2.80E+02
1.00E-20 1.70E+02
2.43E-09 1.80E+00
2.43E-09 1.20E+04
1.12E-07 2.80E+03
6.08E-10 2.23E+05
3.17E-08 4.42E+03
3.13E-08 1.10E+03
6.52E-11 6.40E+05
1.00E-20 1.00E+06
-------
TABLE B-l. (continued)
W
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
961 28 1 ,2-Dibromo-3-chloropropane
98011 Furfural
98828 Cumene
98953 Nitrobenzene
100414 Ethylbenzene
100425 Styrene
106467 p-Dichlorobenzene
106887 1,2-Epoxybutane
106898 Epichlorohydrin
106934 Ethylene dibromide
106990 1,3-Butadiene
107028 Acrolein
107051 Allyl chloride
107062 1,2-Dichloroethane
107131 Acrylonitrile
107211 Ethylene glycol
108054 Vinyl acetate
108101 Methyl isobutyl ketone
108883 Toluene
108907 Chlorobenzene
108930 Cyclohexanol
108952 Phenol
109864 2-Methoxyethanol
110496 2-Methoxyethanol acetate
110543 n-Hexane
110805 2-Ethoxyethanol
110861 Pyridine
Mol.
Wt.
(a/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
fq/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
log
VAP. H Law Diffusivity Diffusivity Antoines' Vapor Oct
Press. Const, in Water in Air Pressure Coefficients Water Kmax K1
(atm- Part. mgVO/g-
(mmHcrt mS/moH (cm2/sec) fcm2/seri ABC Coeff. hr. L/a-hr.
4.62 0.00035
18 0.00012
38.4 0.00034
0.00052 1.6E-08
0.221 0.00815
0.085 0.00048
8E-09 3.9E-11
1.36 0.0019
0.32 2.7E-06
2.34 0.00039
0.05836 2.3E-07
0.58 0.00015
2.21 4E-06
4.5 1.16
0.245 2.4E-05
9.6 0.00788
6.12 0.00275
1 0.0024
207.912 0.00046
16.4 3E-05
13.3 0.00074
2110 0.0736
274 0.00012
368 0.011
78.9 0.00098
109 0.0001
0.092 6E-08
90.2 0.00051
19.9 0.00014
28.4 0.00664
12 0.0037
1.22 4.5E-06
0.276 4E-07
2.55697 2.6E-07
9.28503 1.6E-06
151 0.0143
5.31 3.5E-07
20.8 8.9E-06
7.90E-06
1.01E-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
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 192
1531 229
1052 188
2869 273
1956 215
1968 223
2626 163
1538 205
1683 191
1472 193
1940 197
2436 273
1199 163
1461 208
1747 202
1424 213
1437 208
1690 218
1141 228
2087 273
1675 245
1145 269
1297 247
1494 273
1293 225
1336 238
2089 204
1296 227
1168 192
1345 219
1431 218
913 109
1517 175
273
273
1171 224
1844 234
1374 215
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
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 Solubility
sec-1 ma/L_
3.82E-09 2.97E+03
1.56E-08 1.70E+04
2.43E-09 1.50E+04
1.00E-20 6.20E+03
1.56E-08 3.23E+00
4.17E-09 3.10E+01
6.95E-10 5.00E+02
1.56E-08 1.56E+02
6.08E-10 1.66E-f04
1.00E-20 2.20E+04
1.00E-20 4.00E+04
1.56E-08 1.23E+03
1.00E-20 1.10E+05
6.95E-10 6.13E+01
1.71E-08 2.09E+03
8.69E-10 1.69E+02
2.43E-09 3.10E+02
1.00E-20 7.38E+01
1.00E-20 4.28E+04
2.43E-09 6.59E+04
1.56E-08 4.18E+03
1.00E-20 7.35E+02
2.43E-09 2.13E+05
1.21E-09 3.37E+03
1.56E-08 8.52E+03
2.00E-09 7.40E+04
1.00E-20 1.00E+06
1.00E-20 2.00E+04
6.08E-10 1.90E+04
1.91E-09 5.26E402
1.30E-08 4.72E+02
1.00E-20 3.60E+04
8.69E-10 8.28E+04
1.00E-20 1.00E406
1.00E-20 1.00E+06
1.00E-20 1.24E401
2.43E-09 1.00E+OS
6.08E-10 1.00E+OS
-------
TABLE B-1. (continued)
CAS # COMPOUND NAME
111159 2-Ethoxyethanol acetate
118741 Hexachlorobenzene
120821 1,2,4-Trichlorobenzene
1211422,4-Dlnitrotoluene
121448 Triethylamine
122667 1,2-Dlphenylhydrazine
123911 1,4-Dioxane
124481 Chlorodlbromomethane
126998 Chloroprene
127184 Tetrachloroethylene
630206 1,1,1,2-Tetrachloroethane
924163 N-Nitrosodi-n-butylamine
930552 N-Nitrosopyrrolidine
131 9773 Cresols (total)
1330207 Xylenes
1634044 Methyl tert-butyl ether
1746016 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
1 0061 015 cis-1 ,3-Dichloropropylene
1 0061 026 trans-1 ,3-Dichloropropylene
16065831 Chromium (III)
18540299 Chromium (VI)
Mol.
WL
(ci/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
Density
fa/ccl
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.
fmmHo)
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
HLaw DlffusIvUy Diffusely
Const. In Water In Air
(alm-
m3/mon (cm2/sec) (cm2/sec}
2.2E-06 8.00E-06 8.00E-02
0.00132 5.91E-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
ONA NA
0
0.0092 6.30E-06 3.07E-02
ONA NA
ONA NA
ONA NA
ONA NA
ONA NA
ONA NA
ONA 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
log
Antolnes' Vapor Oct Hydro).
Pressure Coefficients Water Kmax K1 Rate
Part. mgVO/g-
A 8 C Coeff. hr. L/a-hr. sec-1
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 0
203 5.89
253 4.01
280 2.01
223 1.45
273 2.94
238 -0.39
273 2.17
180 2.08
218 2.67
192 2.63
273 2.41
273 -0.19
273 0
273 3.17
223 1.901
159 6.64
273 5.447
273
273 4.978
273 1.914
273 1.462
273 2.724
273 1.845
273 2.21
273 1.255
273
273 1.699
230 2
230 2
273 1.255
273 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
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 Solubility
sec-1 rno/L
1.00E-20 1.00E+06
1.82E-07 6.20E+00
1.56E-08 3.00E+02
1.56E-08 2.70E+02
1.00E-20 5.50E+04
1.00E-20 6.80E+01
1.56E-08 1.00E+06
1.56E-08 2.60E+03
1.56E-08 1.74E+03
3.13E-08 2.00E+02
5.81 E-09 1.10E+03
1.00E-20 1.27E+03
1.56E-08 1.00E+06
1.00E-20 2.20E+04
2.43E-09 1.86E+02
1.00E-20 3.88E+04
1.00E-20 1.90E-05
1.00E-20
1.00E-20
1.00E-20 5.62E-02
1.00E-20
1.00E-20
1.00E-20
1.00E-20
1.00E-20
1.00E-20
1.00E-20
1.00E-20
9.81 E-10 2.72E+03
9.81 E-10 2.72E+03
1.00E-20
1.00E-20
-------
Appendix C
Sensitivity Analysis for Emissions Model
-------
-------
Volume II
Appendix C
C. Sensitivity Analysis for Emissions Model
C.1 General Approach
A sensitivity analysis was performed on the CHEMDAT8 emission models used to
estimate the emissions from the various waste management units (WMUs). As CHEMDAT8
employs exactly the same emission model equations for land application units, open landfills,
and wastepiles, only a single sensitivity analysis was needed to encompass the variations for
these units. However, three scenarios were modeled in the land application / open landfill /
wastepile (or simply the LAU) emissions model sensitivity analysis to cover the range of
assumptions that might be used in modeling these WMUs.
For a given WMU emission model, a base condition was defined and the annual
emissions were estimated. Then, each of the emissions model input parameters were altered one
at a time using a step function (either increasing or decreasing the input parameter by a factor of
2), and the annual emissions were compared with the base case annual emissions. For certain
pairs of parameters, both input parameters were simultaneously altered to better illustrate the
impact of indirect model parameters such as process residence time.
In addition to the emissions model input parameter, the emissions predictions are also
expected to be a function of the contarninant's chemical and physical properties. Because
constituent volatility is expected to have the greatest influence on the emissions estimates, the
sensitivity analysis was performed for a number of constituents that range the spectrum of
volatility.
C.2 Selection of Constituents for Sensitivity Analysis
Ten compounds were selected from the complete list of constituents to be modeled in the
analysis based on a ranking of the constituent's Henry's law constant. The constituents were
ranked using Henry's law constant because the most conservative volatile emissions estimates
would result from aqueous wastes (rather than organic wastes, where constituent vapor pressure
is the key parameter). The constituent properties database developed for the HWIR project was
sorted and the constituents with the 5th, 15th, 25th, etc., to the 95th percentile Henry's law constants
were selected for the analysis. The constituent properties database developed for the HWIR
project was used as the primary source of the constituent-dependent input data. As the
CHEMDAT8 model uses Monod kinetics for estimating the biodegradation rates in the tank
model, the CHEMDAT8 model constituent database was used to supplement the HWIR
constituent properties database (primarily for biodegradation rate constants, but also for
hydrolysis and photolysis rates) prior to performing the sensitivity analysis. Table C-l
summarizes the constituent properties input data for the 10 constituents selected for the analysis.
Except for the use of these 10 constituents in the sensitivity analysis, no other separate analysis
was performed to assess the various models' sensitivity to a specific constituent input parameter.
C-l
-------
Volume II
Appendix C
There was one unusual compound in the selected group of constituents modeled. Phthalic
anhydride reacts with water (undergoes hydrolysis) to form phthalic acid. Hydrolysis is only a
loss mechanism in the surface impoundment and aerated tank models. The selection of a
hydrolyzing compound was more by coincidence than by design, but it is interesting to note the
impact (or the lack of impact) the model input parameters have for the predicted air emissions.
C.3 Sensitivity Analysis for Land Application Unit, Landfill, and Wastepile Emissions
Three potential modeling assumptions were investigated for the LAU emissions model
sensitivity analysis as potential base cases. First, the LAU model was assumed to have
biodegradation, and the waste (or soil waste mixture) was assumed to be an organic waste
matrix. The base case LAU input parameters are provided in Table C-2. As seen in Table C-2,
the LAU model has an input flag for biodegradation and aqueous waste matrix. For the second
series of LAU emissions estimates, the biodegradation flag was set to zero (no biodegradation in
an organic waste matrix). In the third series of LAU emissions estimates, the aqueous waste flag
and the biodegradation flag were both set to 1 (biodegradation with an aqueous waste matrix).
The results of the sensitivity model runs are summarized in Table C-3.
As stated previously, most of the parameters were altered by a factor of 2 in the step
function analysis. For a few parameters, a factor of 2 change was not employed. Specifically, in
reducing the air porosity, a reduced air porosity value of 0.2 was used (rather than 0.51 4- 2).
Additionally, the total porosity could only be lowered when the air porosity was lowered, and a
value of 0.4 was used (rather than 0.6 4- 2). Finally, an air temperature of 37 °F was used in the
step function for the high-end air temperature (rather than 25 x 2). Note: For the air temperature
step function increase, a value of 30 °F may have been more appropriate since the CHEMDAT8
model temperature adjustment uses a temperature of 20 °F as a reference temperature.
In reviewing the data in Table C-3, note that a value of -50 percent indicates a factor of 2
reduction in the annual emissions and a value of 100 percent indicates a factor of 2 increase in
the annual emissions (i.e., direct proportionality provided a factor of 2 change in the input
parameter) from the base case for that section of the table. The CHEMDAT8 LAU model is
insensitive to windspeed for long-term emission estimates. Also, because the CHEMDAT3
model uses an application rate per volume of soil, simply increasing the surface area or the
waste/till depth also increases the total quantity of waste applied. That is why the "low depth"
emission runs do not show an increase in annual emissions (the total amount of contaminant
applied is half that of the base case). Rerunning the "low depth" with a simultaneous increase in
loading indicates a 50 percent increase in annual emissions for the organic waste matrix and the
low volatile compounds, and base level of emissions for the volatile compounds in the aqueous
waste matrix (i.e., nearly 100 percent volatilization).
In comparing the emissions results for the base case for the aqueous versus the organic
waste matrix (Tables C-3a and C-3c), the emissions are an order of magnitude or several orders
of magnitude higher for aqueous wastes than for organic wastes. Only acetone, which has a high
volatility and is miscible in water, has comparable emission rates for the different waste matrices.
The impact of including biodegradation can be seen by comparing the base emissions in
C-2
-------
Volume II
Appendix C
Tables C-3a and C-3b. The inclusion of biodegradation for the organic waste matrix reduced the
annual emission rate (by a factor of approximately 4) for all compounds.
C.4 Sensitivity Analysis for Aerated Tank Emissions
Two basic modeling approaches were investigated for the aerated tank emission model
sensitivity analysis: with biodegradation and without biodegradation. The base case aerated tank
input parameters are provided in Table C-4. For the model runs without biodegradation, the
active biomass input parameter was set to zero. Note: The CHEMDAT8 aerated tank emission
model provides two different types of aeration (mechanical or submerged air) and a whole host
of input parameters that affect aeration, especially mechanical aeration. Thus, a preliminary
analysis was performed solely on the influence of altering the aeration input parameters when no
biodegradation is present.
Table C-5 summarizes the sensitivity model runs for the aeration input parameters. For
this analysis, mechanical aeration was primarily investigated. Submerged aeration has only one
input parameter: submerged air flow rate. When submerged air flow was included, the fraction
agitation .was set to zero (no mechanical aeration). Due to the differences in the mass transfer
mechanisms, some compounds may increase in emissions with submerged aeration (heptachlor
epoxide at submerged air flow = 0.4 m3/s), while others remain the same or have lower
emissions.
For the aerated tank aeration parameter sensitivity analysis (Table C-5), the only input
parameter that was not varied by a factor of 2 was the aerator power efficiency; an efficiency of
0.5 was used (rather than 0.83 -r 2; a typical variation for this parameter is from 0.80 to 0.85). As
seen in Table C-3, the aerated tank model is most sensitive to the fraction aerated, with the total
power, power per aerator, and impeller diameter having some impact on the emission results.
The other aeration input parameters (oxygen transfer rate, power efficiency, and impeller speed)
had very little impact on the emission results.
Table C-6 summarizes the other input parameters for mechanically aerated tanks with' and
without active biomass. The only input parameter that was not varied by a factor of 2 was the
influent solids content. The base case assumes no influent solids; the influent solids case
assumed an influent solids concentration of 1 g/L. The emissions for the higher volatility
compounds (ethyl ether and the dichloroethanes) increased by a factor of 2 when biodegradation
was turned off; the emissions for the less volatile compounds, except for phthalic anhydride,
increased by a factor of 4 to 100 when biodegradation was turned off. As phthalic anhydride is
primarily removed by hydrolysis, the inclusion of biodegradation did not significantly alter its
predicted emissions.
The impact of other model parameters can be seen by comparing the results within either
Table C-6a or C-6b. However, it is interesting to see the difference in the impacts of some of the
variables between the two table sections. For example, increasing the depth, which increases the
residence time in the impoundment, reduces the emissions when biodegradation is present, but
has no impact when biodegradation is not present (except for the hydrolyzing compound). On
the other hand, increasing the surface area, which also increases the residence time, has little
. , —
-------
Volume II
Appendix C
impact on the emissions when biodegradation is present, but increases the emissions when
biodegradation is absent for those compounds that do not already have emissions fractions over
80 or 90 percent. Note: At a fixed residence time, increasing the surface area (with an equal
decrease in depth) increases the emissions regardless of the presence or absence of biodegra-
dation. Increasing the flow rate also favors air emissions over biodegradation because it reduces
the retention time at a given surface area.
Windspeed and influent solids have limited to no impact on the predicted emissions rates.
Temperature, on the other hand, can have a very significant impact on the predicted emissions
rates. Temperature impacts the air diffusivity raised to the 1.75 power. There is also a
temperature-dependent correction factor used in CHEMDAT8 for the biodegradation rate. That
is why the increased temperature reduced the air emissions for a few contaminants (those nearing
100 percent emissions) when biodegradation was present.
C-4
-------
Table C-1. Constituent Input Parameters for the Emission Model Sensitivity Analysis
COMPOUND NAME
ETHYL ETHER
DICHLOROETHANE(U)
DICHLOROETHANE 1,2
ACENAPHTHALENE
ACETONE
HEPTACHLOR EPOXIDE
TOLUIDENE(o-)
DINITROPHENOL2.4
PHTHALIC ANHYDRIDE
PHENYLENEDIAMINE (m-)
DENSITY VAP.PRESS
HLAW
CONST
M.W. (g/cc) (mmHg) (atm-m3/mol)
74.12 0.71 7.07e-01
98.96 1.17 2.99e-01
98.96 1.26 1.04e-01
154.21 1.02 3.29e-06
58.08 0.79 3.036-01
389.32 1.57 2.57e-08
107.16 0.989 4.21e-04
184.11 1.68 6.716-06
148.12 1.33 6.80e-07
108.14 1.14 3.58e-06
3.30e-02
5.62e-03
9.79e-04
1.55e-04
3.88e-05
9.50e-06
2.72e-06
4.43e-07
1.63e-08
1.35e-10
DIFF.
WATER
(cm2/sec)
9.30e-06
1.05e-05
9.90e-06
7.69e-06
1.14e-05
4.68e-06
9.12e-06
9.06e-06
9.60e-06
9.88e-06
Table C-2. Land Application Model Input
TNPT IT PAR AMF.TPR
L,Loading (g oil/cc soil)
Concentration in oil(ppmw)
l,Depth of tilling (cm)
Total porosity
Air Porosity(0 if unknown)
MWoil
For aqueous waste, enter 1
Time of calc. (days)
For biodegradation,enter 1
Temperature (Deg. C)
Wind Speed (m/s)
Area (m2)
LOG
DIFF. AIR OCT/ Kmax
WATER
(cm2/sec) PART. mgVO/
CORF a-hr
7.40e-02 0.83 17.56
7.42e-02 1.79 10.76
1.04e-01 1.47 2.1
4.21e-02 3.92 31.1
1.246-01 -0.24 1.3
1.22e-02 5.00 10.76
7.14e-02 1.34 31.1
2.73e-02 1.55 8.0
7.10e-02 -0.62 17.56
6.63e-02 0.05 9.7
Parameters
VAT .TIP.
0.036
2000
20
0.61
0.5
282
0
365.25
1
25
4.47
25000
Kl HYDROL.
L/g-hr. (I/sec)
0.57 0
2.3 0
0.98 0
2.7 0
1.15 0
11.2 0
0.86 0
0.62 0
0.078 5.55e-04
0.28 0
-------
Table C-3. Summary of Sensitivity Analysis for Land Application Model Emissons
C-3a. LAND APPLICATION // OPEN LANDFILL // WASTE PILE - WITH BIODEGRADATION - ORGANIC MATRIX
COMPOUNDS
ETHYL ETHER
DICHLOROETHANE(U)
DICHLOROETHANE 1,2
ACENAPHTHALENE
ACETONE
HEPTACHLOR EPOXIDE
TOLUIDINECo-)
DINLTROPHENOL^
PHTHALIC ANHYDRIDE
PHENYLENEDIAMINE (m-)
BASE
Annual
Em.
(Mg/yr)
5.14e-02
4.27e-02
5.21e-02
6.29e-05
1.19e-01
5.08e-06
9.27e-04
1.42e-04
4.94e-05
1.47e-04
Percent Increase (or decrease) from BASE emissions
lower
loading
-29%
-29%
-27%
-29%
-23%
-29%
-29%
-29%
-29%
-29%
semi-
annual
41%
39%
20%
41%
41%
39%
41%
36%
41%
38%
half
cone.
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
lower air
porosity
-86%
-86%
-87%
-86%
-89%
-86%
-86%
-86%
-86%
-86%
lower air
& total
porosity
-72%
-72%
-73%
-72%
-78%
-72%
-72%
-72%
-72%
-72%
lower
MWtoil
-29%
-29%
-31%
-29%
-39%
-29%
-29%
-29%
-29%
-29%
higher
temp
-4%
-1%
22%
22%
26%
35%
24%
38%
19%
19%
higher
wind
speed
0%
0%
0%
0% '
0%
0%
0%
0%
0%
0%
low till
depth
high area
100%
100%
87%
100%
105%
100%
100%
100%
100%
100%
high till
depth
low area
-50%
-50%
-51%
-50%
-59%
-50%
-50%
-50%
-50%
-50%
low
area
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
low
depth
0%
0%
-6%
0%
2%
0%
0%
0%
0%
0%
n
6\
C-3b. LAND APPLICATION // OPEN LANDFILL // WASTE PILE - WITHOUT BIODEGRADATION - ORGANIC MATRIX
COMPOUNDS
ETHYL ETHER
DICHLOROETHANE(U)
DICHLOROETHANE 1,2
ACENAPHTHALENE
ACETONE
HEPTACHLOR EPOXIDE
TOLUIDINECo-)
DINITROPHENOL2,4
PHTHALIC ANHYDRIDE
PHENYLENEDIAMINE fm-)
BASE
Annual
Em.
(Mg/yr) ,
1.95e-01
1.28e-01
8.90e-02
3.19e-04
1.66e-01
1.52e-05
4.70e-03
3.67e-04
1.88e-04
4.17e-04
Percent Increase (or decrease) from BASE emissions
lower
loading
-32%
-29%
-29%
-29%
-30%
-29%
-29%
-29%
-29%
-29%
semi-
annual
23%
32%
38%
50%
27%
50%
49%
50%
50%
50%
half
cone.
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
lower air
porosity
-86%
-86%
-86%
-86%
-86%
-86%
-86%
-86%
-86%
-86%
lower air
& total
porosity
-72%
-72%
-72%
-72%
-72%
-72%
-72%
-72%
-72%
-72%
lower
MWtoil
-29%
-29%
-29%
-29%
-29%
-29%
-29%
-29%
-29%
-29%
higher
temp
24%
29%
33%
60%
29%
77%
63%
80%
56%
55%
higher
wind
speed
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
low till
depth
high area
69%
95%
100%
100%
83%
100%
100%
100%
100%
100%
high till
depth
low area
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
low
area
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
low
depth
-15%
-2%
0%
0%
-9%
0%
0%
0%
0%
0%
-------
Table C-3. (Continued)
1
J L
1
1
J L
J L
C-3c. LAND APPLICATION // OPEN LANDFILL // WASTE PILE - WITH BIODEGRADATION - AQUEOUS MATRIX
COMPOUNDS
ETHYL ETHER
DICHLOROETHANE(U)
DICHLOROETHANE 1,2
ACENAPHTHALENE
ACETONE
HEPTACHLOR EPOXIDE
TOLUIDINE (o-)
DINITROPHENOL2.4
PHTHALIC ANHYDRIDE
PHENYLENEDIAMINE (m-)
BASE
Annual
Em.
(Mg/yr)
3.59e-01
3.59e-01
3.59e-01
3.57e-01
3.59e-01
3.26e-01
3.06e-01
2.19e-01
4.96e-02
5.87e-03
Percent Increase (or decrease) from BASE emissions
lower
loading
-50%
-50%
-50%
-50%
-50%
-48%
-46%
-39%
-29%
-29%
semi-
annual
0%
0%
0%
0%
0%
5%
8%
22%
41%
38%
half
cone.
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
lower air
porosity
0%
0%
0%
-21%
-2%
-73%
-78%
-85%
-86%
-86%
lower air
& total
porosity
0%
0%
0%
-10%
-1%
-57%
-67%
-77%
-79%
-79%
lower
MWtoil
0%
0%
0%
-1%
0%
-8%
-12%
-22%
-29%
-29%
higher
temp
0%
0%
0%
0%
0%
4%
5%
20%
19%
19%
higher
wind
speed
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
low till
depth
high area
0%
0%
0%
0%
0%
7%
12%
39%
100%
100%
high till
depth
low area
0%
0%
0%
-2%
0%
-20%
-28%
-41%
-50%
-50%
low
area
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
-50%
low
depth
-50%
-50%
-50%
-50%
-50%
-46%
-44%
-30%
0%
0%
2
-------
Table C-4. Aerated Tank Input Parameters for Base Case with Biodegradation
WINDSPEED
DEPTH
AREA
FLOW
ACTIVE BIOMASS
BIOMASS SOLIDS IN
VO INLET CONC.
TOTAL ORGANICS IN
TOTAL BIORATE
FRACT. AGITATED
SUBMERGED ATR FLOW
4.47
4
27
0.0075
2
0
10
250
19
0.7
0
m/s
m
m2
m3/s
g/1
gfl
mg/1
mg/1
mg/g bio-hr
m3/s
Number impellers
Oxygen trans, rat.
POWR (total)
Power efficiency
Temperature
impeller dia
impeller speed
3
7.5
0.83
25
61
126
Ib02/h-hp
HP
degC
cm
rad/s
Table C-5. Aerated Tank Without Biodegradation- Sensitivity Analysis of Aeration Parameters
COMPOUNDS
ETHYL ETHER
DICHLOROETHANE(U)
DICHLOROETHANE 1,2
ACENAPHTHALENE
ACETONE
HEPTACHLOR EPOXIDE
TOLUIDINE(o-)
DINTrROPHENOL2,4
PHTHALIC ANHYDRIDE
PHENYLENEDIAMINE ( m-)
BASE
Annual
Em.
(Mg/yr)
2.26e-01
2.19e-01
1.89e-01
7.63e-02
4.12e-02
3.88e-03
2.76e-03
2.80e-04
1.87e-06
1.35e-07
Percent increase (or decrease) from BASE emissions
2 Aerators
0%
-1%
-5%
-17%
-20%
-23%
-22%
-23%
-22%
-22%
higher O2
xfer rate
2%
2%
2%
1%
0%
0%
0%
0%
0%
0%
lower
power
efficiency
-3%
-2%
-2%
-1%
0%
0%
0%
0%
0%
0%
lower fract
agitated
-1%
-4%
-14%
-38%
-41%
-41%
-38%
-39%
-37%
-37%
higher tot.
power
2%
3%
6%
20%
25%
30%
29%
30%
29%
29%
higher
impeller
diameter
0%
-1%
-5%
-16%
-19%
-21%
-21%
-21%
-21%
-21%
higher
impeller
speed
0%
-1%
-2%
-9%
-11%
-12%
-12%
-12%
-12%
-12%
submerged
air @ 0.2
m3/sec
2%
-7%
-35%
-51%
-67%
-21%
-51%
-37%
-49%
-48%
submerged
air @ 0.4
m3/sec
3%
0%
-15%
-19%
-46%
40%
-26%
4%
-24%
-22%
-------
Table C-6. Summary of Sensitivity Analysis Model Runs for Aerated Tank Emissions Model
r.-fia AERATED TANK- WITH BIODEGRADATION
COMPOUNDS
ETHYL ETHER
DICHLOROETHANE(U)
DICHLOROETHANE 1,2
ACENAPHTHALENE
ACETONE
HEPTACHLOR EPOXIDE
TOLUIDINE(o-)
DINITROPHENOL2,4
PHTHALIC ANHYDRIDE
PHENYLENEDIAMINE (m-)
BASE
Annual
Emissions
(Mg/yr)
1.86e-01
9.14e-02
7.54e-02
4.88e-03
5.17e-03
4.38e-05
3.55e-04
4.75e-05
1.75e-06
4.15e-08
Percent Increase (or decrease) from BASE emissions
higher
biomass
-15%
-37%
-38%
-48%
-49%
-50%
47%
-46%
-6%
-41%
higher tot.
biorate
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
higher
concentr.
100%
101%
104%
101%
117%
102%
101%
102%
100%
101%
larger
depth
-15%
-37%
-38%
-48%
-49%
-50%
-47%
-46%
-47%
-41%
larger
area
-13%
-21%
-7%
0%
1%
0%
6%
8%
5%
18%
larger area
half depth
2%
19%
40%
92%
94%
99%
99%
99%
99%
99%
influent
solids
0%
0%
-1%
-5%
-1%
-6%
-1%
-2%
-1%
-3%
higher
temp
-3%
-6%
2%
47%
-1%
80%
61%
101%
132%
61%
higher
flow rate
93%
95%
88%
92%
92%
100%
78%
72%
81%
53%
C-6b AERATED TANK - WITHOUT BIODEGRADATION (no biomass)
COMPOUNDS
ETHYL ETHER
DICHLOROETHANE(U)
DICHLOROETHANE 1,2
ACENAPHTHALENE
ACETONE
HEPTACHLOR EPOXIDE
TOLUIDINE(O-)
DINITROPHENOL2.4
PHTHALIC ANHYDRIDE
BASE
Annual
Em.
(Mg/yr)
2.26e-01
2.19e-01
1.89e-01
7.63e-02
4.12e-02
3.88e-03
2.76e-03
2.80e-04
1.87e-06
1.35e-07
Percent Increase (or decrease) from BASE emissions
higher
concentr.
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
larger
depth
0%
0%
0%
0%
0%
0%
0%
0%
-47%
0%
larger
area
0%
2%
9%
50%
69%
96%
97%
99%
6%
99%
larger area
half depth
0%
2%
9%
50%
69%
96%
97%
99%
99%
99%
higher
windspeed
0%
0%
0%
2%
5%
5%
6%
5%
6%
6%
influent
solids
-1%
-1%
-2%
-44%
-8%
-85%
-10%
-11%
-1%
-9%
higher
temp
1%
3%
9%
69%
50%
198%
157%
221%
143%
141%
higher
flow rate
91%
86%
67%
19%
10%
1%
1%
0%
80%
0%
with
biodegr.
-18%
-58%
-60%
-94%
-87%
-99%
-87%
-83%
-7%
-69%
o
-------
-------
Appendix D
Sensitivity Analysis of ISC Air Model
-------
-------
Volume II
Appendix D
D. Sensitivity Analysis of ISC Air Model
This appendix describes sensitivity analysis on depletion options, source shape and
orientation, and receptor location and spacing.
D.I Options With and Without Depletions
A sensitivity analysis was conducted using the ISCST3 model to determine whether dry
and wet depletion options should be used in the risk analysis for five types of waste management
units. A discussion of the analysis follows.
The depletion options (dry depletion and wet depletion) may be used with concentrations
and depositions in the ISCST3 model runs. The model concentrations/depositions without
depletion are higher than those with depletion. Because it takes much longer to run the ISCST3
model with depletions than without depletions, a sensitivity analysis was performed to
investigate the differences of model outputs with and without selecting depletion options.
In this investigation, the 5th and the 95th percentile of sizes of LAUs were used to
determine the relationship between concentrations with depletions and sizes of units.
For dry depletion, two meteorological stations (Little Rock, Arkansas, and Winnemucca,
Nevada) were selected for the sensitivity analysis. The average particle sizes used in the
sensitivity analysis are 20 um and 5 um with corresponding mass fraction of 50 percent each.
The roughness length at application site was assumed as 0.4 meters.
For wet depletion, two meteorological stations were selected for the sensitivity analysis:
Atlanta, Georgia, with 49.8 inches precipitation per year (4th highest annual precipitation rate
among the 29 meteorological stations to be modeled), and Winnemucca, Nevada, with 8.1 inches
precipitation per year (3rd lowest annual precipitation rate). The reason for selecting a wet site
and a dry site was to examine (1) whether wet depletion has a more significant impact for a wet
site than a dry site; and (2) the differences of ambient concentrations that a very wet site can
make with and without selecting wet depletion.
Five-year average concentrations with and without dry depletion were calculated using
meteorological data from Little Rock and Winnemucca for the 5th and the 95th percentile of
sizes of LAUs. The results show that the differences of the maximum concentrations with and
without dry depletion are very small at close-to-source receptors. As the distance from the
source increases, the differences between the dry depletion option and without dry depletion
increase only slightly. The differences of concentrations are about 10 percent of the
concentrations for the 95th percentile and are less than 2 percent of the concentrations for the 5th
percentile at 50 meters from the edge of the LAU. The larger the area source, the larger the
differences of the maximum concentrations. The results are shown in Figures D-la through
D-ld.
D-l
-------
Volume II
Appendix D
Five-year average concentrations with and without wet depletion also were calculated
using meteorological data from Atlanta and Winnemucca for the 5th and 95th percentile of sizes
of LAUs. The results show that the differences of the maximum concentrations with and without
wet depletion are small for both Atlanta and Winnemucca sites. However, the differences in the
maximum concentrations between the wet depletion option and without wet depletion are about 5
to 10 times greater for the Atlanta site than the Winnemucca site. Tables D-la and D-lb show
that for the 95th percentile unit size, at 50 meters from the edge of the unit, the differences in the
maximum concentrations are only 0.03% and 0.37% for Winnemucca and Atlanta, respectively.
This means that model concentrations with and without wet depletion are about the same.
D-2
-------
8
1
0
w/o dry depletion
w/ dry depletion
Distance (km)
Figure D-la. Air Concentrations of Particles
(LAU, 5th Percentile, Little Rock, AR)
-------
o
~;
& •
•
1 0 -
^ U
§ -1 ^
e '
WD
3.2 -
° -3
W)
2 -A -
H^ -«+
.
\24681
i
X^_
'"''7T7:7:7:::::^^rnrrrr^^
*•••-»
Distance (km)
Figure D-lb. Air Concentrations of Particle
(LAU, 5th Percentile, Winnemucca, NV)
0 w/o dry depletion
. , , . .
w/ dry depletion
S
-------
2n
.U
<— N ^ c
^ 1-5 -
£ -1 /V
' •§> 1'° '
g
^ n c
S ^-^
v3 A fi
^ U.U
o
§ n*(
5 -U.o i
M)
S H n
iJ -I .U
1 ^
1
\T!\
\>^
) 2 ""-••4T" 6 _ 8 1
w/o dry depletion
w/ ory oepiecion
0
-1 .0
Distance (km)
Figure D-lc. Air Concentrations of Particles
(LAU, 95th Percentile, Little Rock, AR)
-------
d
ON
- w/o dry depletion
w/ dry depletion
Distance (km)
Figure D-ld. Air Concentrations of Particles
(LAU, 95th Percentile, Winnemucca, NV)
-------
Table D-la. Differences of Air Concentrations for Vapors Between Wet Depletion Option and Without Wet Depletion
(Atlanta, GA Site)
5th Pcrcentile
w/o wet depletion
Distance Concentrations
(m) (ug/m3 / g/m2-s)
19.3 (1)
47.3"'
75.2"'
100
103.2 ("
187.0 (l)
200
300
400
500
600
800
1000
1500
2000
3000
4000
5000
10000
7.40752
0.93175
0.38178
0.25129
0.21003
0.06886
0.07091
0.03390
0.02026
0.01359
0.00981
0.00590
0.00400
0.00205
0.00128 •
0.00068
0.00044 .
0.00031
0.0001 1
w/ wet depletion
Concentrations Difference
(ug/m3 / g/m2-s) (ug/m3 / g/m2-s)
7.40716
0.93159
0.38168
0.25121
0.20996
0.06882
0.07086
0.03387
0.02024
0.01357
0.00979
0.00589
0.00399
0.00205
0.00128
0.00067
0.00043
0.00031
0.00011
0.00036
0.00016
0.00010
0.00008
0.00007
0.00004
0.00005
0.00003
0.00002
0.00002
0.00002
0.00001
0.00001
0.00000
0.00000
0.00001
0.00001
0.00000
0.00000
Difference in
Percentage
0.005%
0.017%
0.026%
0.032%
0.033%
0.058%
0.071%
0.088%
0.099%
0.147%
0.204%
0.169%
0.250%
0.000%
0.000%
1.471%
2.273%
0.000%
0.000%
95th Pcrcentile
w/o wet depletion w/ wet depletion
Distance Concentrations Concentrations
(m) (ug/m3 / g/m2-s) (ug/m3 / g/m2-s)
651.9 ">
676.9 (I)
701.9(1)
726.9 (l)
80I.9(I)'
1000
1100
1200
1300
1400
1500
1600
1800
2000
3000
4000
5000
10000
0.00614
0.00574
0.00539
0.00507
0.00427
0.00400
0.00342
0.00296
0.00260
0.00230
0.00205
0.00185
0.00152
0.00128
0.00068
0.00044
0.00031
0.0001 1
0.00612
0.00573
0.00537
0.00505
0.00426
0.00399
0.00341
0.00295
0.00259
0.00229
0.00205
0.00184
0.00152
0.00128
0.00067
0.00043
0.00031
0.00011
Difference
(ug/m3/g/m2-s)
0.00002
0.00001
0.00002
0.00002
0.00001
0.00001
0.00001
0.00001
0.00001
0.00001
0.00000
0.00001
0.00000
0.00000
0.00001
0.00001
0.00000
0.00000
Difference in
Percentage
0.33%
0.17%
0.37%
0.39%
0.23%
0.25%
0.29%
0.34%
0.38%
0.43%
0.00%
0.54%
0.00%
0.00%
~ 1. 41%
2.27%
0.00%
0.00%
'" These refer to the distances from the center of emission source to the maximum concentration points along 0, 25,50,75, and 150 meter receptor squares, respectively.
-------
Table D-lb. Differences of Air Concentrations for Vapors Between Wet Depletion Option and Without Wet Depletion
(Winnemucca, NV Site)
O
oo
din rercentile
Distance
(m)
17.3 <"
it\
42.3 ("
n\
67.3 (l)
/it
92.3 '"
100
167.3 (l)
200
300
400
500
600
800
1 000
1500
2000
3000
4000
5000
1 0000
w/o wet depletion w/ wet depletion
Concentrations Concentrations Difference
(ug/m3/g/m2.s) (ug/m3/g/m2.s) (ug/m3 / R/m2-s)
7.79132
1.08468
0.48369
0.27965
0.24315
0.09949
0.07296
0.03600
0.02181
0.01475
0.01070
0.00649
0.00443
0.00229
0.00144
0.00077
0.00050
0.00036
0.00013
7.79125
1.08464
0.48367
0.27963
0.24313
0.09948
0.07295
0.03599
0.02180
0.01474
0.01070
0.00648
0.00443
0.00229
0.00144
0.00077
0.00050
0.00036
0.00013
0.00007
0.00004
0.00002
0.00002
0.00002
0.00001
0.0000 1
0.00001
0.0000.1
0.00001
0.00000
0.00001
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.00000
Difference in
Percentage
0.001%
0.004%
0.004%
0.007%
0.008%
0.010%
0.014%
0.028%
0.046%
0.068%
0.000%
0.154%
0.000%
0.000%
0.000%
0.000%
0.000%
0.000%
0.000%
95th Percentile
Distance
(m)
651.9(1)
676.9 (l)
701.9(l>
726.9 (l)
801.9(l)
1000
1100
1200
1300
1400
1500
1600
1800
2000
3000
4000
5000
10000
w/o wet depletion
Concentrations
(ug/m3/g/mz-s)
23.14326
13.86979
11.62889
10.25373
7.84900
5.85241
4.69239
3.98357
3.43255
2.99083
2.63019
2.33211
1.93762
1.65686
0.91889
0.61 160
0.45013
0.17843
w/ wet depletion
Concentrations
(ug/m3/g/m2-s)
23.13885
13.86551
11.62486
10.24985
7.84548
5.84988
4.68991
3.98130
3.43045
2.98887
2.62837
2.33042
1.93554
1.65487
0.91727
0.61020
0.44890
0.17767
Difference
(ug/m3/g/m2-s)
0.00441
0.00428
0.00403
0.00388
0.00352
0.00253
0.00248
0.00227
0.00210
0.00196
0.00182
0.00169
0.00208
0.00199
0.00162
0.00140
0.00123
0.00076
Difference in
Percentage
0.02%
0.03%
0.03%
0.04%
0.04%
0.04%
0.05%
0.06%
0.06%
0.07%
0.07%
0.07%
0.11%
0.12%
0 18%
\ff 1 0 /O
0.23%
0.27%
043%
'" These refer to the distances from the center of emission source to the maximum concentration points along 0,25,50,75, and 150 meter receptor squares, respectively.
-------
Volume II
Appendix. D
D.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 D-2).
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 D-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 D-4a through D-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 D-2a and D-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.
_
D-9
-------
30
'<1X>
10
0
Ik)
20
30
0 -5
-
*
50 -5
!0 -1
20 -1
0
Src1
0
0
I
Src3
*
1
0
1
1
3 2
Sro2
*
0 2
3 3
-
0 3
0
30
20
10
0
- 10
-20
-30
9
(meters)
Figure D-2. Source Shapes and Orientations
D-10
-------
Los Angeles, California
Little Rock, Arkansas
_
Figure D-3. Wind Roses
D-ll
-------
-1
100
50
0 '
-50
100
-1
30 -50 0 50 16
1 r I
* » «
*
* . ~
m *
Me
% * * * *
, *
*>
" Sro1 "
«,,,»,,*« -
. * * »
0 * * ' * *
*
— * ' —
. *
* ' «
1 1 _J ,-
00 -50 0 50 16
)0
100
50
'0
-50
-100
30
(meters)
Figure D-4a. Receptor Locations (Source No. 1)
D-12
-------
-100
100
50
-50
•100
-100
-50
50
-50
50
100
100
50
-50
-100
100
(meters)
Figure D-4b. Receptor Locations (Source No. 2)
D-13
-------
-100
100
-50
•100
-100
-50
50
—T~
-50
50
100
100
50
0
-50
100
-100
(meters)
Figure D-4c. Receptor Locations (Source No. 3)
D-14
-------
Table D-2a. Comparisons of Unitized Air Concentrations (ug/m3 / ug/s-m2) for Different Source Shapes and
(Littte Rock, Arkansas)
Orientations
1 (20m x 20m)
Source No. 2 (40m x
10m)
Source No. 3 (10m x 40m)
X(m)
19
38
35
71
46
92
50
too
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
Yhn)
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 in UACs
Sources No.
Dlff. 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
1 and No. 2
Differences in UACs
Sources No.
%ofDiff. Diff.InUAC
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
1 and No. 3
Differences in UACs
Sources No. 2 and No. 3
%ofDiff. Diff.InUAC
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%
0.018 9%
0.012
0.000
0.035
0.002
-0.105
-0.010
0.020
6.003
-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
%ofDiff.
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%
-------
Table D-2a. Comparisons of Unitized Air Concentrations (ug/m3 / ug/s-m2) for Different Source Shapes and Orientations
(Littte Rock, Arkansas)
Source No.
1 (20m x 20m)
Source No. 2 (40m x 10m) (Source No. 3 (10m x 40m)
Carteslon Receptor Grid
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
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
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
V(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
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
Y(ra)
-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
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 in UACs
Sources No. 1 and No. 2
Dlff.InUAC
-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
Differences in UACs
Sources No. 1 and No. 3
% of DOT. Dlff.InUAC
-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
Differences in UACs
Sources No. 2 and No. 3
%ofDlff. Dlff.InUAC
-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
% of OUT.
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%
-------
Table D-2b. Comparisons of Unitized Air Concentrations (ug/m3 / ug/s-m2) for Different Source Shapes and Orientations
(Los Angeles, California)
Source No. 1 (20m x 20m)
Source No. 2 (40m x 10m)
Source No. 3 (10m x 40m)
Polar Receptor Grid
X(m) Yftn) UAC X(m) Y(m) UAC X(m) Yfcn) UAC
19 46 0.059
38 92 0.016
35 35 0.188
71 71 0.046
46 19 0.582
92 38 0.172
50 0 0.278
100 0 0.068
46 -19 0.061
92 -38 0.015
35 -35 0.062
71 -71 0.016
19 -46 0.080
38 -92 0.023
0 -50 0.086
0 -100 0.023
-19 -46 0.099
-38 -92 0.028
-35 -35 0.122
-71 -71 0.033
-46 -19 0.218
-92 -38 0.060
-50 0 0.320
-100 0 0.093
-46 19 0.264
-92 38 0.074
-35 35 0.137
-71 71 0.037
-19 46 0.063
-38 92 0.017
0 50 0.067
0 100 0.020
19 46 0.065
38 92 0.016
35 35 0.168
71 71 0.045
46 19 0.607
92 38 0.174
50 0 0.293
100 0 0.067
46 -19 0.062
92 -38 0.015
35 -35 0.068
71 -71 0.017
19 -46 0.076
38 -92 0.022
0 -50 0.084
0 -100 0.024
-19 -46 0.092
-38 -92 0.027
-35 -35 0.119
-71 -71 0.032
-46 -19 0.223
-92 -38 0.061
-50 0 . 0.378
-100 0 0.098
-46 19 0.273
-92 38 0.075
-35 35 0.123
-71 71 0.035
-19 46 0.066
-38 92 0.017
0 50 0.058
0 100 0.018
19 46 . 0.069
38 92 0.016
35 35 0.284
71 71 0.052
46 19 0.461
92 38 0.161
50 0 0.293
100 0 0.074
46 -19 0.087
92 -38 0.016
35 -35 0.062
71 -71 0.017
19 -46 0.087
38 -92 0.024
0 -50 0.096
0 -100 0.024
-19 -46 0.108
-38 -92 0.028
-35 -35 0.143
-71 -71 0.034
-46 -19 0.226
-92 -38 0.061
-50 0 0.278
-100 0 0.087
-46 19 0.260
-92 38 0.073
-35 35 0.164
-71 71 0.039
-19 46 0.073
-38 92 0.018
0 50 0.080
0 100 0.021
Standard Deviation:
Differences in UACs
Sources No. 1 and No. 2
Differences in UACs
Sources No. 1 and No. 3
Differences in UACs
Sources No. 2 and No. 3
Diff.InUAC %ofDlff. Diff.InUAC % of Diff. Diff.InUAC % of Dlff.
0.006 9%
0.000 -1%
-0.020 -11%
-0.001 -3%
0.025 4%
0.003 2%
0.014 5%
-0.001 -2%
0.002 3%
0.000 0%
0.006 10%
0.001 4%
-0.004 -4%
-0.001 -5%
-0.003 -3%
0.000 1%
-0.006 -7%
-0.001 -2%
-0.003 -2%
0.000 -1%
0.005 2%
0.001 1%
0.057 18%
0.005 6%
0.009 3%
0.001 1%
-0.014 -10%
-0.002 -5%
0.003 . 4%
0.000 -2%
-0.008 -12%
-0.002 -9%
0.013 6%
0.010 17%
0.000 3%
0.096 51%
0.006 13%
-0.121 -21%
-0.01 1 -6%
0.015 5%
0.005 8%
0.026 43%
0.002 10%
0.000 0%
0.001 3%
0.007 9%
0.001 3%
0.009 11%
0.001 3%
0.009 9%
0.000 1%
0.021 18%
0.001 4%
0.008 4%
0.001 1%
-0.042 -13%
-0.006 -6%
-0.005 -2%
-0.001 -2%
0.027 20%
0.002 4%
0.010 15%
0.001 ' 3%
0.014 21%
0.001 6%
0.030 14%
0.005 7%
0.001 4%
0.116 69%
0.007 16%
-0.146 -24%
-0.014 -8%
0.001 0%
0.007 10%
0.025 40%
0.002 11%
-0.006 -9%
0.000 -1%
0.011 14%
0.002 8%
0.012 15%
0.000 2%
0.016 17%
0.001 3%
0.024 20%
0.002 5%
0.003 2%
0.000 0%
-0.099 -26%
-0.011 -11%
-0.013 -5%
-0.002 -2%
0.041 33%
0.003 9%
0.007 11%
0.001 5%
0.022 37%
0.003 15%
0.040 18%
-------
Table D-2b. Comparisons of Unitized Air Concentrations (ug/m3 / ug/s-m2) for Different Source Shapes and Orientations
(Los Angeles, California)
Source No. 1 (20m x 20m)
Source No. 2 (40m x 10m)
Source No. 3 (10m x 40m)
Cartesian Receptor Grid
X(m) Yfrn) UAC X(m) Y(m) UAC X(m) V(m) UAC
-10 -10 3.225
-5 -10 4.025
0 -10 3.952
5 -10 3.431
10 -10 1.683
10 -5 5.931
10 0 6.636
10 5 6.640
10 10 5.600
5 10 6.893
0 10 6.860
-5 10 6.031
-10 10 3.393
-10 5 5.649
-10 0 5.944
-10 -5 5.663
-35 -35 0.124
-17.5 -35 0.158
0 -35 0.172
17.5 -35 0.123
35 -35 0.064
35 -17.5 0.095
35 0 0.592
35 17.5 0.829
35 35 0.192
17.5 35 0.109
0 35 0.125
-17.5 35 0.113
-35 35 0.139
-35 17.5 0.387
-35 0 0.603
-35 -17.5 0.318
-20 -5 3.241
-10 -5 4.333
0 -5 4.297
10 -5 3.871
20 -5 1.592
20 -2.5 4.787
20 0 5.882
20 2.5 6.294
20 5 5.866
10 5 8.126
0 5 8.285
-10 5 7.442
-20 5 3.497
-20 2.5 5.102
-20 0 ' 5.373
-20 -2.5 5.028
-45 -30 0.139
-22.5 -30 0.183
0 -30 0.199
22.5 -30 0.124
45 -30 0.053
45 -15 0.076
45 0 0.377
45 15 0.739
45 30 0.304
22.5 30 0.195
0 30 0.144
-22.5 30 0.160
-45 30 0.166
-45 15 0.335
-45 0 0.472
-45 -15 0.275
-5 -20 2.674
-2.5 -20 3.119
0 -20 3.050
2.5 -20 2.564
5 -20 1.511
5 -10 5.570
5 0 5.644
5 10 5.524
5 20 4.325
2.5 20 4.939
0 20 4.913
-2.5 20 4.156
-5 20 2.702
-5 10 5.015
-5 0 5.167 •
-5 -10 5.104
-30 -45 0.095
-15 -45 0.123
0 -45 0.121
15 -45 0.100
30 -45 0.063
30 -22.5 0.119
30 0 0.696
30 22.5 0.683
30 45 0.101
15 45 0.072
0 45 0.100
-15 45 0.077
-30 45 0.089
-30 22.5 0.370
-30 0 0.603
-30 -22.5 0.316
Standard Deviation:
Differences in UACs
Sources No. 1 and No. 2
Differences in UACs
Sources No. 1 and No. 3
Differences in UACs
Sources No. 2 and No. 3
Dlff.InUAC %ofDiff. Dlff.InUAC %ofDifT. DOT. In UAC % ofDlff.
0.016 1%
0.308 8%
0.345 9%
0.440 13%
-0.091 -5%
-1.143 -19%
-0.754 -11%
-0.346 -5%
0.266 5%
1.232 18%
1.424 21%
1.411 23%
0.103 3%
-0.547 -10%
-0.572 -10%
-0.635 -11%
0.014 11%
0.025 . 16%
0.028 16%
0.001 0%
-O.pll -17%
-0.019 -20%
-0.215 -36%
-0.090 -11%
0.112 58%
0.086 78%
0.019 15%
0.047 42%
0.026 19%
-0.053 -14%
-0.131 -22%
-0.043 -13%
0.542 24%
-0.551 -17%
-0.906 -23%
-0.902 -23%
-0.867 -25%
-0.172 -10%
-0.360 -6%
-0.992 -15%
-1.116 -17%
-1.275 -23%
-1.955 -28%
-1.947 -28%
-1.875 -31%
-0.691 -20%
-0.634 -11%
-0.777 -13%
-0.559 -10%
-0.029 -23%
-0.035 -22%
-0.050 -29%
-0.024 -19%
-0.001 -2%
0.024 25%
0.104 18%
-0.146 -18%
-0.091 -47%
-0.037 -34%
-0.025 -20%
-0.035 -31%
-0.050 -36%
-0.017 -4%
0.000 0%
-0.002 -1%
0.614 15%
•0.567 -17%
-1.214 -28%
-1.247 -29%
-1.307 -34%
-0.081 -5%
0.783 16%
-0.238 -4%
-0.770 -12%
-1.541 -26%
-3.187 -39%
-3.371 -41%
-3.286 -44%
-0.794 -23%
-0.088 -2%
-0.205 -4%
0.076 2%
-0.043 -31%
-0.060 -33%
-0.078 -39%
-0.024 -20%
0.010 19%
0.043 57%
0.319 85%
-0.055 -7%
-0.203 -67%
-0.122 -63%
-0.044 -31%
-0.082 -52%
-0.077 -46%
0.036 11%
0.131 28%
0.04! 15%
1.026 33%
o
1—I
00
-------
Volume II
AppendixD
D.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 D-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
0,25,50,75, 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 0,25,50,
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 D-5a
through D-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.ef, 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 D-6a and D-6b for
polar receptor locations (5th percentile).
-------
Volume II
Appendix D
The results (Figures D-7a through D-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 D-7a through D-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 D-8a through D-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.
D-20
-------
(meters)
-400 -300 -200 -100 0 100 200 300 48
400
300
200
100
' 0
-100
-200
-300
-400
1 I I I I i i
•
** *.«...* *
*
» * *
» » *
* *
* *
^
* » .
_ * m * m
* * *
*
* *
* ,
- *
* * *
* »
m
* * »
* * *
* *
* _
* v
* *
m
„ * *
»
, * *
. » « *
» * *
*
, * * *
* .
» *
' * "~
* * »
. *
i - * *
.
* ************ *
* * * *
*
'
10
400
300
200
100
0
-100
-200
-300
-ARfl
-400 -300 • -200 -100
0
Figure D-5a. Cartesian Receptor Grid (50th percentUe, 64 receptors each square)
D-21
-------
(meters)
-400
400
300
200
100
0
-100
-200
-300
-300 -200
-400
-100
1
0
T
100
200 300
« a "
* * i
*
• *
* *
* » i
* * »
* *
• •
• '
400
400
300
200
100
0
-100
-200
-300
-400 -300 -200 -100
0
100 200 300 400
-400
Figure D-5b. Cartesian Receptor Grid (50th percentile, 32 receptors each square)
D-22
-------
(meters)
-400
400
-300 -200 -100
300
200
100
0
-100
-200
-300
-400
0
~r
100
—T—
200 300
J L
J L
400
400
300
200
100
0
-100
-200
-300
-400 -300 -200 -100 0 100 200
300
-400
400
Figure D-5c. Cartesian Receptor Grid (50th percentiJe, 16 receptors each square)
D-23
-------
(meters)
-750 -500
1000
"•" w <£=M 500 750 1000
1 ' ' . ' ' '
' *
1 *
750
500
-
' "
» « a
.
I • »
1 * .
250 h " *..« ^ •
*•••* +
1 - • • »
0
-250
-500
-750
1000 L
-1®
» * * *
h •*••••••"* n *".*.
* r
» * «
. * ' " * * *
»* * " • " * *
» * • » " • *
* ,
. * «
* » «
*
1000
750
500
250
'0
-250
-500
'
*
'
*
* *
" * • *
30-750-500-250 0 2S0 500 TeL lfta
-750
-1000
£t
Figure D-6a. Polar Receptor Grid (22.5 degree, 5th perceiatile)
D-24
-------
(meters)
-1000 -750 -500
1000
-250 0 250 500 750
750
500
250
-250
-500 -
-750 -
-1000
* *
* *
* *
0 - « .«».*. » * D
» *
« * *
1000
1000
750
500
250
0
-250
-500
-750
-1000 -750 -500 -250
0
250 500 750
1000
-1000
Figure D-6b. Polar Receptor Grid (10 degree, 5th percentile)
D-25
-------
d
tb
ON
0
— 64 receptors
— 32 receptors
16 receptors
20 40 60 80 100 120
Distance from the edge of the unit (m)
140
160
Figure D-7a. Maximum Concentrations
(5th Percentile, LAU, Los Angeles, CA)
-------
0
64 receptors
32 receptors
• 16 receptors
50 100 150
Distance from the edge of the unit (m)
Figure D-7b. Maximum Concentrations
(50th Percentile, LAU, Los Angeles, CA)
200
-------
9
00
0
0
— 64 receptors
— 32 receptors
16 receptors
50 100 150
Distance from the edge of the unit (m)
Figure D-7c. Maximum Concentrations
(95th Percentile, LAU, Los Angeles, CA)
200
-------
to
vo
0
64 receptors
32 receptors
16 receptors
50 100 150
Distance from the edge of the unit (m)
200
Figure D-7d. Maximum Concentrations
(5th Percentile, LAU, Little Rock, AR)
-------
o
o
20 -,
18-
16-
f 14-
| 12 -
J WD
§3 10-
1 8 "
§ 6-
4-
2 -
0 \
\
•
\
\
\
\
\
\
\
\
\.^
^^^-^^^^
^s>v*>s»_
^~~-~ — _________
' — - ,„
\ \ \
— 64 receptors
— 32 receptors
- 16 receptors
0 50 100 150 200
Distance from the edge of the unit (m)
Figure D-7e. Maximum Concentrations
(50th Percentile, LAU, Little Rock, AR)
-------
o
0
50 100 150
Distance from the edge of the unit (m)
— 64 receptors
-— 32 receptors
—16 receptors
200
Figure D-7f. Maximum Concentrations
(95th Percentile, LAU, Little Rock, AR)
-------
D
u>
to
0.6
0.5 -
§^0.3
I
^0.2
0.1
0.0
0
— 22.5° Interval
—10° Interval
2000 4000 6000 8000 10000 •
Distance from the edge of the unit (m)
Figure D-8a. Maximum Concentrations
(5th Percentile. LAU. Los Angeles. CA)
-------
Cfl
§
u
9
OJ
1
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
0
22.5° Interval
10° Interval
2000 4000 6000 8000 10000
Distance from the edge of the unit (m)
Figure D-8b. Maximum Concentrations
(50th Percentile, LAU, Los Angeles, CA)
Sheetl Chart 3
-------
10
8
A 4
0
22.5° Interval
10° Interval
2000 4000 6000 8000
Distance from the edge of the unit (m)
10000
Figure D-8C. Maximum Concentrations
(95th Percentile, LAU, Los Angeles, CA)
-------
Ul
§
0.3
0.3
/••••.
^02
1 U^
"S
1 0.1
N*^
^_^-
0.1
0.0
0
— 22.5° Interval
—10° Interval
2000 4000 6000 8000 10000
Distance from the edge of the unit (m)
Figure D-8d. Maximum Concentrations
(5th Percentile, LAU, Little Rock, AR)
-------
o
o\
0
— 22.5° Interval
— 10° Interval
2000 4000 6000 8000 10000
Distance from the edge of the unit (m)
Figure D-8e. Maximum Concentrations
(50th PercentiSe, LAU? Little Rock, AM)
-------
0
22.5° Interval
10° Interval
2000 4000 6000 8000
Distance from the edge of the unit (m)
10000
Figure D-8f. Maximum Concentrations
(95th Percentile, LAU, Little Rock, AR)
-------
Volume II
Appendix D
D.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 D-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 D-3. Figure D-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.
D-38
-------
Volume II Appendix D
8 -I
| 6-
3 4-
I 2"
Figure D-9. Counts of Prevailing Wind Directions in Each Direction
• • j • I • I • I • • I •
Prevailing Wind Direction
Table D-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
D-39
-------
Figure D-10. Examples of Different Shapes of Windroses
Narrowly Distributed Windrose
Moderately Distributed Windroses
D-40
-------
Figure D-10 (Continued). Examples of Different Shapes of Windroses
Evenly Distributed Windrose
Bi-modally Distributed Windroses
D-41
-------
-------
Appendix E
Derivation of Chronic Inhalation
Noncancer and Cancer Health
Benchmark Values
_
-------
-------
Volume II
AppendixE
NONCARCINOGENS
DERIVATION OF INHALATION REFERENCE CONCENTRATIONS
E-l
-------
-------
Volume II
Appendix E
2-Chlorophenol
CAS #95-57-8
0.0014 mg/m3
Route-to-route extrapolation from'the RfD
Exon, J.H., and L.D. Koller. 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
RfC:
Basis for RfC:
Critical Study:
Critical Dose:
Critical Effect:
Species:
Route of Exposure:
Duration:
Uncertainty Factor:
Modifying Factor:
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 Koller, 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 stillboms as well as a decrease in the size of the litters in the rats
E-3
-------
Volume II
Appendix E
exposed to 500 ppm, which can be converted to a dosage of 50 mg/kg/d-the LOAEL. No effects
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., andM.L. Keplinger. 1981. Aromatic Hydrocarbons. In: G.D. Clayton and
RE. 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. Koller. 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.
E-4
-------
Volume II
Appendix E
RfC:
Critical Study:
Critical Dose:
Critical Effect:
Species:
Route of Exposure:
Duration:
Uncertainty Factor:
Modifying Factor:
Cobalt
CAS #7440-48-4
0.00001 mg/m3
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. Nffl Publication no. 96-3961. NTP TR-
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
-r UF = 0.004 mg/m3 4- 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 0.1 1 mg cobalt/m3 and cytoplasmic vacuolization of the
bronchus was increased significantly in male and female mice at this concentration.
E-5
-------
Volume II
Appendix E
A LOAEL of 0. 1 1 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 LOAE!,^ 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 = 0.1 1 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.
= 0. 1 1 mg/m3 x (6 hr/24 hr) x (5/7 d) = 0.02 mg/m3
LOAEL ADJ x RDDR,
LOAEL
LOAEL
LOAEL „£<-= 0.02 mg/m3 x 0.209 = 0.004 mg/m3
where
j is the adjusted LOAEL, RDDR,. is a multiplicative factor used to adjust an
observed inhalation particulate concentration of an animal to the predicted inhalation
particulate exposure concentration for a human; based on MMAD = 1.5/tm, 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 (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 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.
E-6
-------
Volume II
Appendix E
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.
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,
ultrastractural and X-ray microanalytical study. Eur JRespir 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 38:119-128 (as cited in ATSDR, 1992).
Demedts, M., B. Gheysens, J. Nagels, et al. 1984. Cobalt lung in diamond polishers. Am Rev
RespirDis 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., B. Robertson, and P. Camner. 1987. Nodular accumulation of type E cells and
inflammatory lesions caused by inhalation of low cobalt concentrations. Environ Res 43:227-243
(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).
NTP. 1991. NTP report on the toxicity studies of cobalt sulfate heptahydrate in F344/N rats and
B6C3F! 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. NTPTOX-5.
E-7
-------
Volume II
Appendix E
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 B6C3F! 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.
Shirakawa, T., Y. Kusaka, N. Fujimura, et al. 1988. The existence of specific antibodies to
cobalt in hard metal asthma. Clin 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 RevRespirDis 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. EurJRespir 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 Ind Hyg Assoc J 38:338-346 (as cited in ATSDR, 1992).
E-8
-------
Volume II
Appendix E
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 Instituto 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 = LOAELnEc ± UF = 1.3 mg/m3 -r 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 toxicity 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 LOAfi^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
E-9
-------
Volume II
Appendix E
LOAEL, 10 for extrapolation from humans to animals, 10 for human variability, and 3 for
extrapolation from subchronic to chronic exposure.
Conversion Factors:
= 9 mg/m3 x (5/24 h) x (5/7 d) = 1.3 mg/m3
= LOAELADj x RGDR
= LOAEL^ x (Hb/g)A/(Hb/g)H
= 1.3 mg/m3 x 1 = 1.3 mg/m3
LO
where
j is the adjusted LOAEL, RGDR is the regional gas dose ratio (animal:human), and
(Hb/g)A/(Hb/g)H is the ratio of bloodrgas partition coefficient; (Hb/g)A/(Hb/g)H defaults to 1
where HWg 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 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.
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 RE. Clayton (eds). 1981. Patty's Industrial Hygiene and Toxicology. 3rd
revised edition. Volume 2A: Toxicology. New York: John Wiley and Sons, pp. 2597-2601.
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.
E-10
-------
Volume II
Appendix E
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:
Envkonmental 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]
_
E-ll
-------
Volume II
Appendix E
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 41:461-479.
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.
E-12
-------
Volume II
Appendix _E
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.
Veldre, LA., and HJ. Janes. 1979. lexicological studies of shale oils, some of their components
and commercial products. Environ Health Perspect 30:141-146 (as cited in U.S. EPA, 1998).
E-13
-------
Volume II
Appendix E
1,4-Dioxane
CAS # 123-91-1
Torkelson, T.R., B.KJ. Leong, R.J. Kociba, et al. 1974. 1,4-Dioxane.
n. Results of a 2-year inhalation study in rats. Toxicol Appl
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
RfC: 0.8 mg/m3
Critical Study:
Critical Dose:
Critical Effect:
Species:
Route of Exposure:
Duration:
Uncertainty Factor:
Modifying Factor:
Calculations:
RfC = NOAELnEc 4- UF = 83.3 mg/m3 -f 100 = 0.8 mg/m3 (0.2 ppm)
Summary of Study:
Groups of Wistar rats were exposed to 0 or 111 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
E-14
-------
Volume II
Appendix E
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).
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
j x RGDR
j x (Hb/g)A/(Hb/g)H
= 83.3 mg/m3 x 1 = 83.3 mg/m3
where
NOAELAD, is the adjusted 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 et al., 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).
E-15
-------
Volume II
Appendix E
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. Toxicol Appl Pharmacol 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.K.J. Leong, RJ. Kociba, et al. 1974. 1,4-Dioxane. E. 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.
E-16
-------
Volume II
Appendix E
2-Ethoxyethanol Acetate
CAS # 111-15-9
0.3 mg/m3
Calculated from RfC for 2-ethoxyethanol
Barbee, S.J., J.B. TerriU, D.J. 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
RfC:
Basis for RfC:
Critical Study:
Critical Dose:
Critical Effect:
Species:
Route of Exposure:
Duration:
Uncertainty Factor:
Modifying Factor:
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.
E-17
-------
Volume II
Appendix E
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 LOAELnEc for 2-ethoxyethanol in the rabbit are identified as 68 and 265 mg/m3,
respectively. The NOAEL was adjusted for intermittent exposure. A NOAE!^, 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
J x RGDR
x (H^A/(HWH
= 68 mg/m3 x 1 = 68 mg/m3
where
j is the adjusted NOAEL, RGDR is the regional gas dose ratio (animal:human),
and (HWg)A/(HWg)H is the ratio of bloodrgas partition coefficients; (Hb/g)A/(Hb/g)H defaults to 1
where HWg values are not known.
RfC for 2-ethoxyethanol = NOAE!^ -r UF = 68 mg/m3 -f 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.
E-18
-------
Volume II
Appendix E
References:
Barbee, S.J., J.B. Terrill, D.J. 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, J.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 Teratol 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.
E-19
-------
Volume II
Appendix E
Ethylene glycol
CAS # 107-21-1
0.6 mg/m3
Wills, J.H., F. Coulston, E.S. Harris, et al. 1974. Inhalation of
aerosolized ethylene glycol by man. Clin 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
, -f UF = 55.8 mg/m3 -r 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:
NOAELAu, = 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
RfC:
Critical Study:
Critical Dose:
Critical Effect:
Species:
Route of Exposure:
Duration:
Uncertainty Factor:
Modifying Factor:
Calculations:
RfC =
E-20
-------
Volume If
Appendix E
(Robertson et al., 1947, as cited in ATSDR, 1997). Developmental effects have been seen in
animal studies. Tyl et al. (1995a, 1995b, as cited in CalEPA, 1997) reported reduced ossification
in humeras, 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.H., 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
Pharmacol Exper 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. Food Chem Toxicol 27(9):573-584 (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. FundamAppl 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 24:57-75 (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. Clin Toxicol 7:463-476.
E-21
-------
Volume II
Appendix E
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
Gd 6-15
100:
10 for extrapolation from animals to humans
10 for protection of sensitive human subpopulations
1
Calculations:
RfC = NOAELnEc + 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 etal., 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
-------
Volume II
Appendix E
methanol did not exacerbate this effect. Significant increases in the incidence of exencephaly
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 NOAHI^c 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
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
E-23
-------
Volume II
Appendix E
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).
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.
E-24
-------
Volume II
AppendixE
2-MethoxyethanoI 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 ULOAEL
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
RfC:
Basis for RfC:
Critical Study:
Critical Dose:
Critical Effect:
Species:
Route of Exposure:
Duration:
Uncertainty Factor:
Modifying Factor:
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
E-25
-------
Volume II
Appendix E
rabbits, a NOAEL of 30 ppm (NOAHL^^!? mg/m3) and a LOAEL of 100 ppm
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 (NOAE!^ = 17 mg/m3). The NOAEL was
adjusted for intermittent exposure. A NOAHI^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
= NOAELADj x RGDR
= NOAELAD, x (Hb/g)A/(Hb/g)H
= 17 mg/m3 x 1 = 17 mg/m3
where
j is the adjusted NOAEL, RGDR is the regional gas dose ratio (animalrhuman),
(Hb/g)A/(Hb/g)H is the ratio of blood:gas partition coefficients; (Hb/g)A/(Hb/g)H defaults to 1
where HWg values are not known.
RfC for 2-methoxyethanol = NOAE!^ + UF = 17 mg/m3 -f 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
E-26 ' " ' " ~~
-------
Volume If
Appendix E
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.
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. FundAppl Toxicol 3(l):49-54.
Research Triangle Institute (RTT). 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 61FR423 17]
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.
US 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. 6 1FR423 17-354. August 14,
1996.
U.S. Environmental Protection Agency. 1998. Integrated Risk Information System (IRIS) file
for 2-methoxyethanol.
E-27
-------
Volume II
Appendix E
Vanadium
CAS #7440-62-2
RfC:
Critical Study:
Critical Dose:
Critical Effect:
Species:
Route of Exposure:
Duration:
Uncertainty Factor:
Modifying Factor:
Calculations:
RfC =
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:
10 for use of a LOAEL in a human study
10 for use of an acute study
3 for protection of sensitive human subpopulations
1
j -r 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: 10 for use of a LOAEL in a human study, 10 for use of an acute study (8 hours'
duration), and 3 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.
E-28
-------
Volume II
Appendix E
Conversion Factors:
Dose levels are:
% vanadium in vanadium pentoxide = 56.02% = 60% = 0.6. 0.1 mg V2O5/m3 x °-6 = °-6
vanadium/m3; 0.25 mg V2Os/m3 = °-15 mg vanadium/m3; 1 mg V2O5/m3 = 0.06 mg vanadium/m .
LOAELADJ= 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). 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. 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 RevRespirDis 132:1181-1185 (as cited in ATSDR, 1992).
Lee, K.P., and P.J. 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).
Levy, B.S., L. Hoffman, and S. Gottsegen. 1984. Boilmakers' bronchitis. J Occup Med 26:
567-570 (as cited in ATSDR, 1992).
Musk, A.W., and J.G. Tees. 1982. Asthma caused by occupational exposure to vanadium
compounds. 'Med J Aust 1:183-184 (as cited in ATSDR, 1992).
Orris, P., J. Cone, and S. McQuilkin. 1983. Health hazard evaluation report HETA 80-096-
1359' Bl'oomington, EL. 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).
E-29
-------
Volume II
Appendix E
Sjoeberg, S.G. 1950. Vanadium pentoxide dust - a clinical and experimental investigation on its
effects after inhalation. Acta Med Scand Supp. 238:1-188 (as cited in ATSDR, 1992).
Sjoeberg, S.G. 1956. Vanadium dust, chronic bronchitis, and possible risk of emphysema. Acta
Med Scand 154:381-386 (as cited in ATSDR, 1992).
Thomas, D.L., and K. Stiebris. 1956. Vanadium poisoning in industry. The Medical Journal 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.
E-30
-------
Volume IT
Appendix E
CARCINOGENS
DERIVATION OF INHALATION UNIT RISK FACTORS
AND CANCER SLOPE FACTORS
E-31
-------
-------
Volume If
Appendix E
Inhalation Unit Risk Factor:
Slope Factor:
Critical Effects:
Species:
Route of Exposure:
Duration:
Bromodichloromethane
CAS #75-27-4
1.8E-05 (ug/m3)-1
6.2E-02 (mg/kg/d)-1
Tubular cell adenoma and tubular cell adenocarcinoma
Mice
Gavage, corn oil
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
E-33
-------
Volume II
Appendix E
kidney in male and female rats at control, low dose, and high dose were 0/50,13/49,46/50 and
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
andB6C3Fl 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
5. 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 pg x 1/70 kg x 20 mVday =
0.062 (mg/kg/d)-1 x 1 mg/1,000 pg x 1/70 kg x 20 m3/d = 1.8E-05Gug/m3)
where
70 kg = default adult human body weight
20 m3 = default adult human daily rate of inhalation
Calculations assume 100% absorption.
.3\-l
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 E-l, Figure E-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.
E-34
-------
Volume II
Appendix E
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).
U.S. Environmental Protection Agency. 1994. Provisional Guidance for the Qualitative Risk
Assessment of Polycyclic 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.
E-35
-------
Volume II
Appendix E
Chlorodibromomethane
CAS #124-48-1
Inhalation Unit Risk Factor: 2.4E-05 fag/m3)'1
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). '
E-36
-------
Volume II
Appendix E
Calculations:
URF = CSF x 1 mg/1,000 fjvg x 1/70 kg x 2b m3/d =
0.084 (mg/kg/d)-1 x 1 mg/1,000 ^ x 1/70 kg x 20 m3/d = 2.4E-05(//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: . mTO
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 E-l, Figure E-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).
US Environmental Protection Agency. 1994. Provisional Guidance for the Qualitative Risk
Assessment of Polycyclic 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.
US 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.
E-37
-------
Volume II
Appendix E
Unit Risk Factor:
Slope Factor:
Critical Effects:
Species:
7,12-Dimethylbenz[a]anthracene
CAS #57-97-6
2.4E-02 (ug/m3)-1
8.4E+01 (mg/kg/d)'1
Malignant angioendothelioma of the mesenteric intestine
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.
E-38
-------
Volume II
AppendixE
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 carcinogeniciry 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 E-2, Figure E-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 (ug/m3)-1 are predicted. These values are in close
agreement with the CalEPA values of 84 (mg/kg/d)-1 and 2.4E-02Gug/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[ajpyrene 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-
dime"thyl-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.
E-39
-------
Volume II
Appendix E
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. RiskAnalysis 13(4):383-398.
Pitot, H.C., IE, 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 Polycyclic 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.
E-40
-------
Volume II
Appendix E
Unit Risk Factor:
Slope Factor:
Critical Effects:
Species:
Route of exposure:
Duration:
2,4-Dinitrotoluene
CAS #121-14-2
1.9E-04(^g/m3)-1
6.8E-01 (mg/kg/d)-1
Hepatocellular carcinoma, liver neoplastic nodules, benign and
malignant mammary gland tumors.
Female Sprague-Dawley rats
Diet
2 years
Basis for Toxicity Values: • • f
There are no human data available that may be used to address the carcmogenicity 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)-1 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 E-l,
Figure E-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.
E-41
-------
Volume II
Appendix E
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 jig x 1/70 kg x 20 m3/d =
0.68 (mg/kg/d)-1 x 1 mg/1,000 ug x 1/70 kg x 20 m3/d = 1.9E-04(jig/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
and an inhalation CSF of 3.1E-01 (nag/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 rumors 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 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 (LDSO). A low TDSO 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 IMS 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 E-2, Figure E-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 (ug/m3)-1 are predicted. These values are in close agreement
E-42
-------
Volume ff
Appendix E
with the oral CSF listed in IRIS for a mixture of 2,4- and 2,6-dinitrotoluene and the CalEPA
values.
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 m, J.H. Hagensen, J.R. Hodgson, et al. 1979. Mammalian toxicity of munitions
compounds. Phase IE: 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, J.J. Kowalski, et al. 1978. Mammalian toxicity of munition compounds.
Phase n Effects of multiple doses and Phase ffl. Effects of lifetime exposure. Part E.
2,4-Dinitrotoluene. U.S. Army Medical Bioengineering Research and Development Laboratory.
Midwest Research Institute, Kansas City, MO. NITS ADA 061715.
US. Environmental Protection Agency. 1987. Health Effects Assessment for 2,4- and 26-
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.
E-43
-------
Volume II
Appendix E
Unit Risk Factor:
Slope Factor:
Critical Effects:
Species:
Route Of Exposure:
Duration:
3-Methylcholanthrene
CAS #56-49-5
2.1E-03 (ug/m3)-1
7.4E+00 (mg/kg/day)'1
Mammary gland adenocarcinomas
Wistar rats
Gavage
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
polycycKc 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,
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%.
E-44
-------
Volume II
Appendix E
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 TD^ and an inhalation CSF (see
Table E-2, Figure E-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 TDSO 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 (ng/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 Gag/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.
E-45
-------
Volume II
Appendix E
Pilot, H.C., HI, and Y.P. Dragan. 1996. Chemical Carcinogenesis. In: Casarett & Doutt'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 Polycyclic 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.
E-46
-------
Volume II
Appendix E
o-Tohiidine (2-Methylaniline)
CAS #95-53-4
6.9E-05 (ng/rn3)'1
2.4E-01 (mg/kg/d)'1
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
Unit Risk Factor:
Slope Factor:
Critical Effects:
Species:
Route of Exposure:
Duration:
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 E-1,
Figure E-1). 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
inU.S.EPA(1987).
E-47
-------
Volume II
Appendix E
' Experimental Dose
o-ToIuidine-HCI
(mg/rat/d)
0
62
Transformed Dose
o-Toluidine
(mg/kg/d)
0
80
,_ ^
Incidence :
1/27
25/30
Calculations:
URF = CSFxlmg/l,OOO^gxl/70kgx20m3/d= .
0.24 (mg/kg/d)-1 x 1 mg/1,000 //g x 1/70 kg x 20 m3/d =
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.
E-48
-------
Volume II
Appendix E
Table E-1. Correlation of Oral and Inhalation Cancer Slope Factors
Reported in IRIS and HEAST
% <•> /
<.
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
-0.61 98
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
E-49
-------
Volume II
Appendix E
Table E-l. (continued)
iCAS #
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)
AANitrosodi-n-butylamine
A/-Nitrosodiethylamine
MNitrosodimethylamine
AA-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 Inlj."
CSF
(continued)
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
E-50
-------
Volume II
Appendix E
Figure E-l. Correlation of Oral and Inhalation Cancer Slope Factors.
E-51
-------
Volume II
Appendix E
Table E.2. Correlation of TDSOs Reported in the Cancer Potency Database and Inhalation
Cancer Slope Factors Reported in IRIS and HEAST.
Chemical
Carbon tetrachloride
DDT
Aldrin
Bis(2-chloroethyl)ether
1,3-Butadlene
Chlordane
Chlorobenzilate
Chloroform
1 ,1-Dichloroethylene
1 .3-Dichloropropene
Dieldrin
HCH alpha
HCH beta
HCH tech.
Heptachlor
Hexachloroethane
Melhylene chloride
N-Nitrosodi-n-butylamine
1,1.1 ,2,-Tetrachloroethane
1 ,1 .2,2,-Tetrachloroethane
Toxaphene
1 ,1 ,2-Trichloroethane
Acetaldehyde
Acrylamide
AcryJon'rtrile
Aramite
Azobenzene
Bis(Chloromethyl)ether
Bromoethene
Bromoform
1 ,2-Dibromoethane
1 ,2-Dichloroethane
1 ,2-Diphenylhydrazine
Ethylene oxide
Hexachlorobenzene
Hexachlorobutadiene
Hydrazine
4,4'-MethyIenebis(2-chIoroaniline)
N-Nitrosodiethylamine
N-Nitrosodimethylamine
N-Nrtrosopyrrolidine
Propylene oxide
2,3,7,8-TCDD
2,4,6,-Trichlorophenol
Vinyl chloride
CAS#
56-23-5
50-29-3
309-00-2
111-44-4
106-99-0
57-74-9
510-15-6
67-66-3
75-35-4
542-75-6
60-57-1
319-84-6
319-85-7
608-73-1
76-44-8
67-72-1
75-09-2
924-16-3
630-20-6
79-34-5
8001-35-2
79-00-5
75-07-0
79-06-1
107-13-1
140-57-8
103-33-3
542-88-1
593-60-2
75-25-2
106-93-4
107-06-2
122-66-7
75-21-8
118-74-1
87-68-3
302-01-2
101-14-4
55-18-5
62-75-9
930-55-2
75-56-9
1746-01-6
88-06-2
75-01-4
TD50a
Rat
2.29
84.7
-
-
261
-
-
262
-
94
-
11.2
-
-
-
55.4
724
0.691
-
-
-
-
153
6.15
16.9
96.7
24.1
0.004
18.5
648
1.52
8.04
5.59
21.3
3.51
65.8
0.309
19.3
0.024
0.124
0.799
74.4
2E-05
405
19.1
TDsoa
Mouse
150
12.3
1.27
11.7
13.9
2.99
93.9
90.3
34.6
49.6
0.912
6.62
27.8
14.8
1.21
338
918
1.09
182
38.3
5.57
55
-
-
-
158
-
0.182
-
-
7.45
101
26
63.7
65.1
-
2.93
-
-
0.189
0.679
912
2E-04
1070
20.9
TDSO
Geo
Meanb
18.53
32.28
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
«
*
*
*
*
*
*
*
*
*
*
*
*
it
*
*
*
*
*
*
Inh CSF°
0.053
0.34
17
1.16
0.98
1.3
0.27
0.08
0.175
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
0.11
217
0.11
0.004
0.77
0.091
0.77
0.35
1.6
0.077
17.1
0.13
151
49
2.13
0.013
150000
0.011
0.3
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
r
r
r
•I/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
0.179
38.4615
5
0.909
17.857
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
0.000
90.909
3.333
log TDsoe
(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
1.170
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
-4.629
2.607
1.281
n
r2
slope
intercept
log 1/CSF
(Y)
1.276
0.469
-1.230
-0.064
0.009
-0.114
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
-5.176
1.959
0.523
45
0.949
0.983
-0.679
a Gold and Zeiger, 1997
b Geometric mean of the TDSO reported for mice and rats (only used when the CSF was derived from both species).
C IRIS. 1 998 or HEAST, 1 997.
d Test species reported as the basis for the CSF derivation ("b" is both rats and mice, "m" is mice, and V is rats).
e Selected to correspond with the CSF test species.
n Number of chemicals with a TDSO and inhalation CSF.
r2 Correlation coefficient.
No data available.
Not calculated because the CSF was based on a single species.
E-52
-------
Volume II
Appendix E
Figure E-2. Correlation of TD50 and inhalation cancer slope factors.
E-53
-------
-------
|