xvEPA
United States
Environmental Protection
Agency EPA 510-R-13-001
Evaluation Of Empirical Data To
Support Soil Vapor Intrusion
Screening Criteria For Petroleum
Hydrocarbon Compounds
U.S. Environmental Protection Agency
Office of Underground Storage Tanks
Washington, DC 20460
-------
[This page intentionally left blank.]
11
-------
Evaluation Of Empirical Data To Support Soil
Vapor Intrusion Screening Criteria For
Petroleum Hydrocarbon Compounds
Prepared by
Dr. Ian Hers
Golder Associates
500 - 4260 Still Creek Drive
Burnaby, British Columbia
Canada V5C 6C6
and
Robert S. Truesdale
RTI International
3040 East Cornwall! s Road
Post Office Box 12194
Research Triangle Park, NC 27709-2194
under contract to
Skeo Solutions
921 Second Street SE
Charlottesville, VA 22902
Contract No. GS-10F-0309N
for
U.S. Environmental Protection Agency
Office of Underground Storage Tanks
Washington, DC 20460
January 2013
in
-------
[This page intentionally left blank.]
IV
-------
Disclaimer
This document has been reviewed in accordance with U.S. Environmental Protection Agency
policy and approved for publication. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
-------
Acknowledgements
James B. Cowart, David J. Folkes, and Dr. Jeffrey P. Kurtz (EnviroGroup Limited); Todd A.
McAlary (Geosyntec Consultants) and Dr. Mark A. Widdowson (Virginia Polytechnic Institute
and State University) provided reviews.
Robin Davis (Utah Department of Environmental Quality) provided the May 2011 database,
which served as the basis for the database described in this report. Peter Eremita (Maine
Department of Environmental Protection) and Jackie Wright (Environmental Risk Sciences Pty
Ltd) provided significant data contributions that enhanced the Davis database.
VI
-------
Table of Contents
Disclaimer v
Acknowledgements vi
List of Acronyms xiii
Executive Summary ES-1
ES.l Purpose and Document Focus ES-1
ES.2 Methodology ES-2
ES.3 Findings and Conclusions ES-3
1. Introduction 1
1.1 Background 1
1.2 Goal and Objectives 2
1.3 Document Development and EPA Peer Review 2
1.4 Document Organization 3
2. Conceptual Site Model and Select Case Studies 3
2.1 Aerobic Biodegradation Processes in the Vadose Zone 4
2.2 Factors Influencing Biodegradation of Petroleum Hydrocarbons 5
2.3 Dissolved versus LNAPL Vapor Sources 6
2.4 Anaerobic Biodegradation and Methane Generation 7
2.5 Conditions for Increased Potential for Petroleum Vapor Intrusion 8
2.6 Case Studies Indicating Confirmed or Likely Complete Transport Pathway for
Petroleum Vapor Intrusion 9
2.6.1 Refinery Site, Perth, Australia (Patterson and Davis, 2009) 9
2.6.2 Chatterton Petrochemical Site, Vancouver, B.C., Site (Hers et al., 2000;
Hers et al., 2002) 10
2.6.3 Refinery Site, Casper, Wyoming (Luo et al., 2009) 10
2.6.4 Former Refinery Site (confidential location) (Luo et al., 2010) 11
2.6.5 Refinery Site, Hartford, Illinois (Illinois Department of Public Health,
2010) 11
2.6.6 USTSite, Stafford, New Jersey (Sanders and Hers, 2006) 11
2.6.7 UST Site, Ogden, Utah, Mini-Mart Release (McHugh et al., 2010) 12
2.6.8 UST Site, Gunnison, Utah, Top Stop Release (McHugh et al., 2010) 12
3. Summary of Modeling Studies 12
3.1 Abreu Three-Dimensional Model Simulations 13
3.1.1 Three-Dimensional Model Simulations—Below-Building
Contamination Source and Homogeneous Soil Conditions 13
3.1.2 Three-Dimensional Model Simulations—Lateral Migration Scenario and
Homogeneous Soil Conditions 16
3.1.3 Three-Dimensional Model Simulations—Surface Capping Scenario 18
vn
-------
3.1.4 Comparison of Modeled to Measured Soil Vapor Concentration Data 21
3.2 DeVaull (2007b) Study (BioVapor Model Development) 21
3.3 DeVaull (2010) Study of BioVapor Application 21
3.4 Summary of Modeling Studies 23
4. Review of Empirical Database Studies of Petroleum Hydrocarbon Vapor Attenuation 23
5. EPA PVI Database Development, Structure, and Content 24
5.1 EPA PVI Database Development and Checking 25
5.1.1 Quality Control and Data Quality Ranking 25
5.2 EPA PVI Database Structure 27
5.3 EPA PVI Database Content 28
6. EPA PVI Database Analysis Approach and Methods 29
6.1 Source Zone Identification (LNAPL versus Dissolved Indicators) 30
6.1.1 Groundwater Concentration Data 31
6.1.2 Soil Concentration Data 31
6.1.3 Proximity to Fuel Storage/Dispensing Facilities 32
6.2 Data Analysis Methods 32
6.2.1 Exploratory Data Analysis 32
6.2.2 Vertical Distance Method 32
6.2.3 Clean Soil Method 33
6.3 Soil Vapor Concentration Thresholds 36
6.3.1 Sub-slab to Indoor Air Attenuation Factors 36
6.3.2 Risk-based Concentration Thresholds 36
7. EPA PVI Database Analysis Results 37
7.1 Exploratory Data Analysis 37
7.1.1 Comparison of Groundwater and Soil Vapor Concentrations 37
7.1.2 TPH Vapor versus Oxygen Concentrations 39
7.1.3 Methane Concentrations 41
7.1.4 Comparison between Benzene and TEX Vapor Concentrations 42
7.2 Vertical Distance Method 43
7.2.1 All Data 43
7.2.2 Influence of Surface Cover 52
7.2.3 Influence of Soil Type 56
7.3 Clean Soil Method 56
8. Discussion 59
8.1 Conceptual Site Model and Mathematical Models 59
8.2 Methods and Characteristics of the Database 60
8.3 Data Analysis Results 61
Vlll
-------
8.4 Exclusion Distance Assessment Framework 61
8.5 Lateral Inclusion Distances 62
8.6 Comparison with Other Studies 63
9. Findings and Conclusions 63
10. References 66
Appendix A: Review of Exclusion/Inclusion Distances in Existing Vapor Intrusion
Guidance
Appendix B: Data Quality and Database Content
Appendix C: Analysis of Australian PVI Database
Appendix D: PVI Database Data Dictionary
Appendix E: PVI Database Entity Relationship Diagram
Appendix F: Analysis of Lead Scavengers: Ethylene Dibromide and 1,2-Dichloroethane
IX
-------
List of Tables
1. Summary of Case Study Sites with Confirmed or Likely Occurrences of
Petroleum Vapor Intrusion 9
2. Select Three-Dimensional Abreu and Johnson (2005) Model Simulation Results
from U.S. EPA (2012c) 20
3. Number of Sites by Country and States in the EPA PVI Database (November
2012) 28
4. Potential LNAPL Hydrocarbon Indicators 30
5. Risk-based Indoor Air Concentration for Primary Chemicals of Potential Concern 37
6. Summary of Results for Vertical Distance Method 49
7. Summary of Results for Clean Soil Method 57
List of Figures
1. Typical vertical concentration profile in the unsaturated zone for PHCs, carbon
dioxide, and oxygen (modified from U.S. EPA, 2012a) 5
2. Conceptual model illustrating the potential for vapor intrusion for a) free-phase
LNAPL sources, b) residual-phase LNAPL sources, and c) dissolved-phase
sources. (Source: Lahvis et al., In prep.; used with permission) 7
3. Vapor intrusion attenuation factors predicted by Abreu and Johnson (2005) three-
dimensional model for a range of source total hydrocarbon (benzene) vapor
concentrations and separation distances for a residential house scenario (adapted
from Abreu et al., 2009) 14
4. Effect of source depth on soil gas distribution and vapor intrusion attenuation
factors predicted by Abreu and Johnson (2005) three-dimensional model for a
source total hydrocarbon (benzene) vapor concentration of 100 mg/L,
biodegradation rate of 0.79 h"1, and three source depths: 3 m (top), 7 m (middle),
and 9 m (bottom) bgs. Hydrocarbon and oxygen concentrations are normalized by
source and atmospheric concentrations, and the building is a residential house
(from Abreu et al., 2009) 15
5. Vapor intrusion attenuation factors predicted by Abreu and Johnson (2005) three-
dimensional model as a function of separation distance below foundation and
first-order biodegradation rate for a residential house scenario and 10 mg/L vapor
source concentration (from Abreu et al., 2009) 16
6. Relationship between source-building lateral separation distance and normalized
indoor air concentration (a) for a NAPL source, two source depths, and three
biodegradation rates (X). The source-building lateral separation is measured from
the edge of the source zone to the center of the building with a basement; negative
values and values of less than 5 m indicate that the source is to some extent
beneath the building. The source vapor concentration is 200 mg/L. (Source: 33 in
U.S. EPA [2012c]) 17
7. Effects of various layered soil scenarios (rows A-D) on hydrocarbon and oxygen
distribution in soil gas and normalized indoor air concentration (a) for an
-------
underpressurized basement (-5 Pa, left panels) and an overpressured basement (+5
Pa, right panels). Hydrocarbon and oxygen concentration contour lines are
normalized by source and atmospheric concentrations, respectively. The source
vapor concentration is 200 mg/L located 8 m bgs. Biodegradation rate (X) = 0.18
h'1. (Source: 37 in U.S. EPA [2012c]) 19
8. Effect of source vapor concentration on hydrocarbon and oxygen distribution in
soil gas and normalized indoor air concentration (a) for scenarios with low
permeability soils at the ground surface (e.g., soil layer scenario on row D of 7).
Hydrocarbon and oxygen concentration contour lines are normalized by source
and atmospheric concentrations, respectively. Source located at 8 m bgs
(basement scenario). Biodegradation rate (X) = 0.18 h"1. (Source: 38 in U.S. EPA
[2012c]) 20
9. Estimates of indoor air benzene concentrations using Biovapor model for varied
effective airflow through the basement foundation. Foundation effective airflow
statistics: 5th percentile = 0.3 L/min, 50th percentile = 3 L/min, 95th percentile =
30 L/min. Key model parameters: Vapor mixing height = 2.44 m; indoor air
1 99
exchange rate = 0.25 h" ; building footprint area = 100 m (1,076 ft ) (from
DeVaull, 2010) 22
10. Example calculation of clean soil method distances 35
11. Groundwater concentrations measured near soil vapor sampling locations for
dissolved and LNAPL source zones (all refers to UST, fuel terminal, refinery, and
petrochemical sites) 38
12. Groundwater concentrations measured near soil vapor sampling locations for
dissolved and LNAPL source zones (all refers to UST, fuel terminal, refinery, and
petrochemical sites). Only detectable benzene vapor concentrations shown 39
13. TPH vapor versus oxygen concentrations for dissolved and LNAPL source zones
(all refers to UST, fuel terminal, refinery, and petrochemical sites). Data points
shown are where both TPH vapor and oxygen concentrations were above
detection limits. Shaded areas and ellipse encompass data that generally support
the aerobic mineralization paradigm 41
14. Methane concentrations versus distance and benzene vapor concentrations (all
refers to UST, fuel terminal, refinery, and petrochemical sites) 42
15. Relationship between benzene and ethylbenzene (left) and benzene and xylenes
(right) vapor concentrations. (All refers to UST, fuel terminal, refinery, and
petrochemical sites.) Data points are shown where both compounds were above
detection limits 42
16. Vertical distance method: benzene (a), oxygen (c), and xylenes (d) data for
dissolved-source sites (KM = Kaplan-Meier). Panel b shows the benzene
probability data 44
17. Vertical distance method: PHC fraction (a-c) and hexane (d) data for dissolved-
source sites 45
18. Vertical distance method: benzene (a, b), oxygen (c), and xylenes (d) data for
LNAPL sources at UST sites (KM = Kaplan-Meier) 46
19. Vertical distance method: PHC fraction (a-c) and hexane (d) data for LNAPL
sources at UST sites 47
XI
-------
20. Vertical distance method: 2,2,4-trimethylpentane (IMP), MtBE, 1,2,4-
trimethylbenzene (1MB), and naphthalene data for LNAPL sources atUST sites 48
21. Vertical distance method-benzene, xylenes, and oxygen data for LNAPL sources
at fuel terminal, refinery, and petrochemical (non-UST) sites. Red plots over blue 51
22. Vertical distance method-2,2,4-trimethylpentane (TMP), naphthalene, and 1,2,4-
trimethylbenzene (TMB) data for LNAPL sources at fuel terminal, refinery, and
petrochemical (non-UST) sites 52
23. Comparison of probability for benzene soil vapor concentrations to be less than
threshold and oxygen concentrations for different surface covers for LNAPL
sources at UST sites. Below detection limit concentrations replaced with half the
detection limit for analysis 53
24. Comparison of probability for benzene soil vapor concentrations to be less than
threshold and oxygen concentrations for different surface covers for LNAPL
sources at all sites (UST, fuel terminal, refinery, and petrochemical). Below
detection limit concentrations replaced with half the detection limit for analysis 55
25. Comparison of probability for benzene soil vapor concentrations to be less than
the threshold for different soil types (coarse and fine grained). Below detection
limit concentrations replaced with half the detection limit for analysis 56
26. Results of clean soil method for dissolved-source sites. 47 sites, N = 170 58
27. Results of clean soil method for LNAPL sources at UST Sites. 53 sites,N= 172 58
28. Results of clean soil method for LNAPL sources at fuel terminal, refinery and
petrochemical (non-UST) sites. 60 sites, N = 216 59
xn
-------
List of Acronyms
ATSDR Agency for Toxic Substances and Disease Registry
CC>2 Carbon Dioxide
CSM Conceptual Site Model
1,2-DCA 1,2-Dichloroethane
DQ Data Quality
DRO Diesel Range Organics
EPA U.S. Environmental Protection Agency
EDB Ethylene Dibromide
GRO Gasoline Range Organics
H2O Water
IARC International Agency for Research on Cancer
ILCR Incremental Lifetime Cancer Risk
IRIS Integrated Risk Information System
IUR Inhalation Unit Risk
LIF Laser Induced Fluorescence
LNAPL Light Nonaqueous Phase Liquid
MassDEP Massachusetts Department of Environmental Protection
MCL Maximum Contaminant Level
MCLG Maximum Contaminant Level Goal
MMR Massachusetts Military Reservation
MRL Minimal Risk Level
MtBE Methyl Tert-Butyl Ether
NaCl Sodium Chloride
OUST EPA's Office of Underground Storage Tanks
PAH Polycyclic Aromatic Hydrocarbon
PHC Petroleum Hydrocarbon
PID Photoionization Detector
PVI Petroleum Vapor Intrusion
QA/QC Quality Assurance/Quality Control
RBCV Risk-Based Soil Vapor Concentration
SFO Oral Cancer Slope Factor
Xlll
-------
TEX Toluene, Ethylbenzene, and Xylenes
1,2,4-TMB 1,2,4-Trimethylbenzene
2,2,4-TMP 2,2,4-Trimethylpentane
TPH Total Petroleum Hydrocarbons
UST Underground Storage Tank
UV Ultraviolet
UVIF Ultraviolet Induced Fluorescence
VI Vapor Intrusion
VOC Volatile Organic Compound
xiv
-------
Executive Summary
This report presents an evaluation of empirical data and select modeling studies of the behavior
of petroleum hydrocarbon (PHC) vapors in subsurface soils at petroleum release sites and how
these vapors can affect subsurface-to-indoor air vapor intrusion. Specifically, the report develops
an inclusion distance approach for screening petroleum release sites for vapor intrusion that
should improve the efficiency of petroleum release site investigations and help focus resources
on the sites of most concern for petroleum vapor intrusion (PVI).
ES.l Purpose and Document Focus
The purpose of this report is to support the development of a new soil vapor screening approach
for PHC compounds. Consequently, the report focuses primarily on characterizing the vapor
intrusion pathway at Solid Waste Disposal Act Subtitle I underground storage tank (UST) sites
with petroleum fuel releases. However, the report also presents and discusses PVI data from
other types of sites with PHC contamination (i.e., fuel terminals, petroleum refineries, and
petrochemical plants).
In support of its general guidance development effort for the PVI exposure pathway, the U.S.
Environmental Protection Agency (EPA) compiled an empirical database of measurements of
subsurface media (soil gas, soil, and groundwater) and supporting data from PHC sites. The
rationale for this focus on subsurface media measurements is that in the presence of oxygen,
PHC vapors can rapidly biodegrade. Compared with chlorinated hydrocarbons, PHC vapors
bioattenuate to much lower concentrations in soil gas (EPA, 2012a). Therefore, screening for
PVI using the same methodology used for chlorinated hydrocarbons is overly conservative; a
different approach is needed for PVI.
The goal of this report is to provide information on the subsurface vapor attenuation of PHCs
that would support establishing a better approach for evaluating PVI potential, with the intent of
determining when PVI may result in indoor air PHC concentrations that exceed safe levels for
human health (i.e., when the PVI exposure pathway is complete).
Because bioattenuation processes for PHCs are well documented and widespread (EPA, 2012a),
the analysis of subsurface soil gas data from actual petroleum release sites provides an
opportunity to develop improved and more realistic approaches for evaluating the potential for
PVI when PHCs are released into the subsurface. Data from real-world sites can be used to
identify an inclusion distance, defined in this report as the vertical separation distance from the
contamination source beyond which the potential for PVI is insignificant. Applying the inclusion
distance approach is potentially more efficient than current approaches for investigating PHC
release sites and can quickly focus resources on the sites where distances less than the inclusion
criteria or exceptional conditions indicate a greater potential for PVI.
This report describes the activities EPA conducted to develop and support an inclusion distance
approach:
• Assemble an empirical database from petroleum release sites where the PVI pathway
has been evaluated primarily via soil gas and groundwater measurements;
ES-1
-------
• Consolidate and independently evaluate the quality of both existing and new data;
• Analyze the data and available case studies to determine when and under what
conditions there is the potential for a complete PVI pathway at petroleum release
sites;
• Summarize published modeling studies on PHC vapor transport and intrusion as
supporting evidence of aerobic biodegradation and PHC vapor concentration
attenuation; and
• Identify an approach and criteria that can be used determine when to exclude
petroleum release sites from further PVI investigation and concern. Detailed
protocols, such as site investigation methods, to implement this approach are beyond
the scope of this report.
ES.2 Methodology
Given the importance of the contamination source type on soil vapor concentrations, the analyses
described in this report were conducted separately for light nonaqueous phase liquid (LNAPL)
and dissolved PHC sources. Therefore, the first step of the data analysis used LNAPL indicators
to determine source type (LNAPL or dissolved-phase). Once source type was identified, the data
analysis consisted of three main parts:
• Exploratory data plots to identify trends and relationships;
• Estimation of vapor concentration attenuation distance using the vertical distance
method, (Lahvis et al., In prep.); and
• Estimation of non-contaminated vertical soil thickness needed for concentration
attenuation using the clean soil method (Davis, 2009).
The vertical distance method plots soil vapor concentration versus distance above a
contamination source and estimates the probability for the soil vapor concentrations to be less
than a risk-based concentration threshold. The probabilities were calculated for two benzene
concentration thresholds, 50 and 100 ng/m3.
The clean soil method (Davis, 2009; 2010) is an analysis of the thickness of unimpacted clean
soil (i.e., soil without NAPL) required for soil vapor benzene concentrations to attenuate to
below a defined threshold, which for this study is 100 ng/m3. A clean soil thickness was not
calculated when the vertical distance between soil gas probes was greater than 10 ft (3.0 m)
because there is then insufficient resolution (i.e., spacing between probes) for meaningful
estimation.
The analysis method either replaced benzene vapor concentrations that were below the reporting
limits with half the reporting limit, a common first approximation, or used the Kaplan-Meier
method (Kaplan and Meier, 1958) to estimate the concentration distribution of the entire dataset,
including non-detects.1 The analysis used risk-based soil gas vapor concentration thresholds for a
1 The Kaplan-Meier method is a robust, non-parametric method for considering data below reporting limits,
particularly when there are multiple reporting limits (Helsel, 2005; 2006).
ES-2
-------
residential receptor scenario, continuous lifetime exposure to vapors, and a shallow soil vapor-to-
indoor air attenuation factor of 1 x 10~2.
ES.3 Findings and Conclusions
Critical factors affecting PVI include facility type (e.g., UST versus non-UST: fuel terminal,
refinery, petrochemical plant), which influences the size of the release; PHC source type
(dissolved versus LNAPL); and the vertical separation distance between the source and receptor
(or building foundation). These factors are important metrics for site screening.
Findings from analysis of dissolved sources in the PVI database include:
• For the vertical distance method, approximately 97 percent of the benzene soil vapor
concentrations are less than 100 ng/rn3 and 94 percent of the concentrations are less
than 50 ng/m3 for contamination source-building separation distances as small as 0 ft.
For other compounds evaluated, measured soil vapor concentrations are less than the
risk-based concentrations for distances greater than 3 ft (0.9 m).
• For the clean soil method, the 95th percentile vertical clean soil thickness for benzene
vapor attenuation to below 100 ng/m3 is approximately 5.4 ft (1.6 m).
• The analysis indicates there is a low probability of exceeding risk-based
concentrations even for small separation distances.
Findings from analysis of LNAPL sources at UST sites in the PVI database include:
• Approximately 95 percent of the benzene soil vapor concentrations are less than
100 |-ig/m3, and 93 percent of the concentrations are less than 50 ng/m3 at a
contamination source-building separation of approximately 15 ft (4.6 m). For other
compounds evaluated, measured soil vapor concentrations are less than the risk-based
concentrations beyond 11 ft (3.4 m).
• For the clean soil method, the 95th percentile vertical clean soil thickness for benzene
vapor attenuation is approximately 13.5 ft (4.1 m).
Findings from analysis of LNAPL sources at fuel terminal, refinery, and petrochemical (non-
UST) sites in the PVI database include:
• For the vertical distance method, approximately 90 percent of the benzene soil vapor
concentrations are less than the thresholds for a contamination source-building
separation of approximately 18 ft (5.5 m). The probability does not increase above
90 percent beyond this distance because data are limited for larger separation
distances. For other compounds evaluated, measured soil vapor concentrations are
less than the risk-based concentrations beyond 12 ft (3.6 m).
• For the clean soil method, there are insufficient data to estimate percentiles, but the
maximum vertical clean soil thickness for benzene vapor attenuation is approximately
20 ft (6.1m).
ES-3
-------
Other conclusions from this work include the following.
• The available data indicate benzene is the risk driver for the sites evaluated, with
exceedances of the risk-based vapor concentrations for benzene occurring at larger
contamination source-building separation distances than observed for the other PHCs
with EPA toxicity values.
• There was significantly less attenuation in vapor concentrations for the aliphatic
hydrocarbon 2,2,4-TMP compared with benzene, although data were relatively
limited. However 2,2,4-TMP does not have a toxicity benchmark and so cannot be
evaluated in the vertical distance or clean soil method.
• The data analysis indicates a poor correlation between benzene concentrations in
groundwater and deep soil vapor taken above a groundwater source. The implication
is that a screening approach for vapor intrusion based on groundwater concentrations
is not appropriate for PVI sites. However, groundwater concentrations can be used as
an approximation to identify LNAPL sources.
• The analysis of surface cover indicated:
— For LNAPL sources at UST sites, there were lower oxygen concentrations and
less benzene vapor attenuation below paved surfaces, but not below buildings,
compared to bare ground cover, and
— For fuel terminal and refinery sites, there were lower oxygen concentrations
below buildings but not below paved surfaces. The lower oxygen levels beneath
buildings may result from larger petroleum releases and consequent increased
oxygen demand at such sites, compared with typical LNAPL releases at UST
sites.
The results are variable and not conclusive as to the effect of surface cover, but they
suggest that there can be reduced oxygen availability below hard surfaces (pavement
or building foundations), for the sites evaluated.
• Because the vertical distance method evaluation includes soil vapor concentration
data from below buildings at 39 sites, the results are considered reasonably robust
with respect to the potential influence of surface cover (although further evaluation of
this factor is recommended).
The mathematical modeling studies reviewed strongly support the empirical analysis and
inclusion distances for dissolved sources. For LNAPL sources, although the modeling generally
supports the empirical analysis, further evaluation of factors potentially influencing oxygen
supply and demand is warranted. Such factors include source vapor concentration, source size,
building size, surface cover and soil layer properties, and natural soil oxygen demand.
Inclusionary criteria or conditions not analyzed in this report include non-UST facilities, organic-
rich soils (e.g., peat), large building foundations (e.g., apartment complexes, commercial or
industrial buildings), and significant subsurface preferential pathways (e.g., utilities, karst,
fractured rock). Where these conditions are present at a site, a more detailed PVI assessment may
be warranted, especially when LNAPL is present.
ES-4
-------
Releases of certain ethanol blends of gasoline may also warrant additional consideration for
inclusion and PVI assessments, although further research is required to determine the
significance of ethanol content with respect to inclusion distances. Biodegradation of ethanol
may generate methane at a greater rate than gasoline alone, consuming oxygen that would
otherwise be available for biodegradation of PHCs and thus increasing the potential for PVI.In
addition, inclusion criteria may not apply at sites where there is significant methane generation
because of the potential for safety hazards, advective soil gas transport, and reduced
biodegradation of other PHCs (due to oxygen demand represented by methane). High methane
generation potential has been documented at large diesel and gasoline spills at non-UST sites.
ES-5
-------
[This page intentionally left blank.]
ES-6
-------
1. Introduction
This report describes an evaluation of empirical data and select modeling studies of the behavior
of petroleum hydrocarbon (PHC) vapors in subsurface soils at petroleum release sites and how
these vapors can affect subsurface-to-indoor air petroleum vapor intrusion (PVI). The purpose of
this report is to support the development of a soil vapor screening methodology for PHC
compounds that can be used for characterizing PVI at Solid Waste Disposal Act Subtitle I2 UST
sites with petroleum fuel releases. However, PVI data from other types of petroleum release sites
(fuel terminals, petroleum refineries) are also presented and discussed.
1.1 Background
In support of its general guidance development effort for the PVI exposure pathway, EPA
compiled an empirical database of measurements of subsurface media (soil gas, soil and
groundwater) and supporting data at PHC sites. The rationale for this focus on subsurface media
measurements is that in the presence of oxygen, PHC vapors can rapidly biodegrade. Compared
to chlorinated hydrocarbons, PHC vapors attenuate to much lower concentrations in soil gas
(U.S. EPA, 2012a). Therefore, screening for PHCs using the same methodology as chlorinated
hydrocarbons is overly conservative.
Because bioattenuation processes are well documented and widespread (U.S. EPA, 2012a), the
analysis of subsurface soil gas data from sites provides an opportunity for developing improved
and more realistic screening evaluation methods for PHC compounds based on the observed
attenuation. These data can be used to identify an inclusion distance3, defined in this report as the
contamination source-separation distance beyond which the potential for PVI may be
insignificant. An inclusion distance approach is potentially more efficient than current
approaches for investigating PVI sites. It also focuses resources on sites within the inclusion
zone that may have significant potential for PVI issues.
Davis (201 la) compiled a large quantity of the data in the EPA PVI database. The May 2011
version of the Davis database was used as the starting point for the EPA PVI database. A
significant quantity of data from other sources was added to the EPA PVI database for this effort,
including data from Maine (Eremita, 2011), Canada, and Australia (Wright 2011, 2012).
However, for purposes of evaluation of inclusion distances, the North American data (primarily
sites from the U.S.) and Australian data were analyzed separately, given the differences in site
conditions in these two countries.
Several similar complementary efforts using somewhat different datasets are in progress in the
U.S. (Lahvis et al., In prep.; Peargin and Kolhatkar, 2011). Section 8 (below) compares these
complementary studies and their results with those in this report.
2 Subtitle I of the Solid Waste Disposal Act
3 An exclusion distance concept and compilation of a PVI database to support estimation of exclusion distances was
first developed by Davis (2009, 2010, 201 la, and 201 Ib). The exclusion and inclusion distance concepts are similar,
although each has a slightly different focus. The inclusion concept establishes criteria for identifying sites that are
screened in for further assessment, whereas the exclusion concept establishes criteria that, when met, indicate low
potential for PVI and thus a basis for screening sites out the PVI assessment process.
-------
Several states, including New Jersey, California, and Wisconsin, are in the process of developing
or have recently developed guidance for PHCs based on a pathway inclusion or exclusion
distance approach. Appendix A provides a review of existing state vapor intrusion guidance
focusing on approaches for PHCs.
1.2 Goal and Objectives
The goal of this report is to provide information on the behavior of PHCs with respect to
subsurface vapor attenuation that would support establishing an assessment framework for
evaluating potential petroleum vapor migration from subsurface to indoor air. The main intent of
the evaluation is to determine if the vapor migration pathway is complete (or incomplete) to
indoor air, which in this context is the potential to exceed human health-based concentration
criteria in indoor air due to PVI.
This report provides data and analyses in support of a key part of this framework: the evaluation
of PHC vapor attenuation and identification and justification of inclusion distances between
contamination and receptors that can be used to quickly assess whether the PVI pathway is likely
to be complete. However, the scope of this report does not include development of the detailed
protocols that will be needed to apply these inclusion distances, which are described in separate
guidance on PVI being prepared by EPA (Wilson et al., In press).
To develop and support the inclusion distances, the objectives of work described in this report
were to:
• Assemble an empirical database from petroleum release sites where the PVI pathway
has been evaluated via primarily soil gas and groundwater measurements;
• Consolidate and independently evaluate the quality of selected existing databases and
newly available data;
• Analyze the database and available case studies to determine when and under what
conditions there is the potential for a complete PVI pathway at petroleum release
sites;
• Summarize published modeling studies on PHC vapor transport and intrusion as
supporting evidence of aerobic biodegradation and vapor concentration attenuation;
and
• Identify an approach and criteria that can be used to screen out certain petroleum
release sites from further PVI investigation and concern so that resources may be
focused on sites within the inclusion zone where the PVI pathway may potentially be
complete.
1.3 Document Development and EPA Peer Review
The draft document was subjected to EPA's external peer review process from May to June
2012. The peer review contractor independently selected five experts not affiliated with EPA.
The experts were James B. Cowart, David J. Folkes, and Dr. Jeffrey P. Kurtz of EnviroGroup
Ltd.; Todd McAlary of Geosyntec Consultants; and Dr. Mark A. Widdowson of Virginia
Polytechnic Institute and State University. The expertise of the peer review panel includes:
-------
• Practical and theoretical understanding of the petroleum vapor intrusion pathway,
including how volatile organic contaminants move and distribute in the subsurface
(soil gas), indoor air, and outdoor air from dissolved and nonaqueous phase liquid
sources;
• Experience in planning and conducting site-specific vapor intrusion studies, including
developing and refining conceptual site models of the migration and distribution of
volatile contaminants; and
• Expertise in collecting and statistically analyzing vapor intrusion pathway data,
applying and calibrating models using site-specific data, and interpreting results to
make decisions at vapor intrusion sites.
The peer reviewers were tasked to review the draft report and provide opinion and perspective
regarding:
• The scientific appropriateness of the database for EPA's purposes;
• Whether the reported analyses are based on sound scientific principles, methods, and
practices; and
• Whether the reported conclusions are adequately supported by the data and analyses.
1.4 Document Organization
This report is organized as follows:
• Section 2 describes the conceptual site model (CSM) for aerobic biodegradation of
PHC vapors and select case studies where PVI has been documented.
• Section 3 provides a summary of select modeling studies of the biodegradation of
PHCs in the subsurface.
• Section 4 provides a review of empirical database studies of PHC vapor attenuation.
• Section 5 describes the EPA PVI database development, structure, and content.
• Section 6 describes the EPA PVI database analysis approach and methods.
• Section 7 describes the EPA PVI database analysis results.
• Section 8 provides a discussion of the results and comparisons with other studies.
• Section 9 provides findings and conclusions of this report.
The key sections of this report supporting the inclusion distance approach are the empirical
database methods and analysis in Sections 6 and 7. The CSM discussion and summary of model
studies are intended as supporting evidence for the PVI inclusion criteria.
2. Conceptual Site Model and Select Case Studies
The CSM for PVI described in this section builds on the general vapor intrusion CSM described
in U.S. EPA (2012a), with additional emphasis on the difference in PHC vapor concentrations
for light nonaqueous phase liquid (LNAPL) and dissolved groundwater contamination sources.
-------
The scope of this CSM discussion is on specific aspects relating to PHC fate and transport and
does not address general aspects of the general vapor intrusion CSM, which are covered in other
documents (e.g., U.S. EPA, 2012a,b,c). In addition, select case studies indicating a confirmed or
likely complete transport pathway for PVI are described.
2.1 Aerobic Biodegradation Processes in the Vadose Zone
Petroleum liquids (e.g., gasoline, diesel) are moderately soluble in water and often form separate
phase liquids commonly referred to as LNAPLs when released into the environment. When an
LNAPL reaches the water table, it tends to accumulate and spread laterally and vertically (as the
water table rises and falls), to form a smear zone where residual LNAPL partially occupies soil
pore spaces across the water table. A dissolved hydrocarbon groundwater plume that extends
beyond the LNAPL source zone is formed as PHC compounds dissolve from the LNAPL into
groundwater or as water percolates through residual LNAPL in the unsaturated soil (vadose)
zone.
Volatilization of PHCs occur from both LNAPL and dissolved (groundwater) hydrocarbon
sources. In addition to PHCs, fuel oxygenates, such as ethanol and methyl tert-butyl ether
(MtBE), and fuel additives, such as ethylene dibromide (EDB) and 1,2-dichloroethane (DCA),
can also be present in the vapor phase in the unsaturated zone proximate to LNAPL source
zones. The biodegradation of PHC vapors is relatively rapid when oxygen is present; therefore,
aerobic biodegradation can typically limit the concentration and subsurface migration of PHC
vapors in unsaturated soils and in groundwater. Modeling studies (Abreu and Johnson, 2006;
DeVaull, 2007a; Abreu et al., 2009) and field studies (Ririe et al., 2002; Hers et al., 2000;
Roggemans et al., 2001; Fitzpatrick and Fitzgerald, 2002) indicate that the potential for PVI is
greatly reduced (orders-of-magnitude concentration attenuation) when aerobic biodegradation
processes occur in soils between the hydrocarbon source and receptor (building foundation). A
typical vertical concentration profile in the unsaturated zone for PHCs, carbon dioxide, and
oxygen is shown in Figure 1.
The aerobic biodegradation processes between the hydrocarbon source and receptor may be
conceptualized with respect to fluxes where the oxygen availability must exceed microbial
metabolically driven oxygen demand associated with the hydrocarbon source (Lahvis et al.,
In prep.). The PHC and oxygen fluxes are primarily due to diffusion, which is influenced by soil
moisture and porosity. PHC biodegradation rates are rapid (e.g., half-lives on the order of hours
to days; DeVaull, 2007b, 2011; Davis et al., 2009) and much faster than the rate of hydrocarbon
transport by diffusion within the unsaturated zone. For this reason, there are typically sharp
reaction fronts where the PHC vapor concentrations attenuate by orders of magnitude over short
distances (e.g., 1 to 5 ft [0.3 to 1.5 m]) and where there is a corresponding decrease in the oxygen
concentrations, as observed in several field studies (Lahvis and Baehr, 1999; Hers et al., 2000;
Sanders and Hers, 2006; Davis et al., 2009; Luo et al., 2009). DeVaull et al. (2007b) report that
the lower threshold oxygen concentrations required to support aerobic biodegradation of PHC
vapors range from 1 to 4 percent.
-------
Land Surface
/
Oxygenated
Soil
Impacted
Soil
Oxygen
Flux
PHC + CH4
Flux
Increasing Concentration
Figure 1. Typical vertical concentration profile in the unsaturated zone for PHCs, carbon
dioxide, and oxygen (modified from U.S. EPA, 2012a).
2.2 Factors Influencing Biodegradation of Petroleum Hydrocarbons
Biodegradation of PHCs has been documented in publications for more than 70 years (Zobell,
1946; Atlas, 1981; Leahy and Colwell, 1990). These papers describe biodegradation of many
types of PHC: liquids and gases; straight, branched, and ring-structure compounds; and
compounds with single and multiple carbon bonds. Many microbial species, including more than
30 genera of bacteria, more than 25 genera of fungi, and several algae degrade PHCs, although
not every microbial species degrades every chemical. PHC biodegradation has been documented
in marine, freshwater, sediment, and soil environments and by direct metabolism and co-
metabolism (co-oxidation). Microbial degradation of petroleum produces biomass, intermediate
products (e.g., alcohols, aldehydes, organic acids), and the ultimate mineralization products
carbon dioxide (CO2) and water (H2O).
In general, relatively fast acclimation times are observed, absent other limits, by population
enrichment (fast biomass growth) and/or plasmid transfer. Acclimation times tend to be shorter
with prior chemical exposure. Environmental conditions under which petroleum biodegradation
has been observed range from 0° to 70°C, salinity up to 25 parts per thousand sodium chloride
(NaCl), and pH from 6 to 10, although optimal conditions can be narrower. Aerobic
biodegradation is the primary mechanism in the unsaturated zone, but anaerobic biodegradation
near source zones may also occur in the presence of other strict or facilitative electron acceptors
(e.g., nitrate, sulfate) or under fermentative or methanogenic conditions (DeVaull et al., 1997;
Madigan et al., 2010). There have been extensive compilations of rates of aerobic degradation
specific to vadose zone aerobic soils (e.g., Leeson and Hinchee, 1996; DeVaull et al., 1997; Hers
et al., 2000; Ririe et al., 2002; Davis et al., 2009; DeVaull, 2011).
-------
The U.S. Air Force Bioventing Initiative study of 125 sites indicated that environmental factors,
such as soil moisture, nutrients, and pH, did not significantly influence biodegradation activity
and respiration rates, except for one site located in the Mohave Desert (California, USA) with
very dry soils (moisture content of 2 percent), although some biological activity did still occur
there (Leeson and Hinchee, 1996). Biological activity is limited when the moisture content is at
or below the permanent wilting point (Zwick et al., 1995; Holden et al., 1997), a condition
expected to be rare in most geological and climatic environments. A literature search for soil
moisture effects did not indicate other studies specifically addressing vadose zone attenuation of
PHC vapors under dry conditions. However, the empirical database assembled for this study
includes sites in relatively dry climates (e.g., Utah, Australia) that can be considered fairly direct
indicators for the influence of soil moisture.
Aerobic biodegradation of PHCs is a robust process that has been demonstrated under a wide
range of environmental conditions. Important factors influencing aerobic biodegradation of PHC
vapors include:
• Vapor source hydrocarbon concentration, flux, and composition (including methane);
• Minimal oxygen concentration required to support aerobic biodegradation;
• Oxygen demand (i.e., the oxygen required to biodegrade the available hydrocarbons)
and supply (i.e., flux balance);
• Distance between the vapor source and the building;
• Soil type and properties (e.g., soil porosity and moisture); and
• Size and characteristics of the building and adjacent land surface.
2.3 Dissolved versus LNAPL Vapor Sources
The PHC vapor source concentration is highly dependent on whether partitioning occurs from
compounds present as a dissolved phase in groundwater or directly from LNAPLs present above
the capillary fringe4 (either as mobile LNAPL or trapped as residual LNAPL in the smear zone
in soils above the water table). The vapor mass flux from LNAPLs present in the unsaturated
zone soils will be higher than for a dissolved groundwater source. In the case of a dissolved
source, chemicals must diffuse through water in the capillary fringe before reaching continuous
gas-filled soil pores, and hydrocarbons may also be attenuated through biodegradation and
sorption within the capillary fringe. The vapor mass flux for LNAPL source zones will also tend
to be sustained for longer periods of time because of the larger contaminant mass compared with
dissolved sources. In addition, the vapor composition will differ depending on whether the vapor
source is LNAPL or the dissolved phase. For LNAPL sources, there will tend to be a higher
proportion of relatively insoluble PHC compounds, including aliphatic hydrocarbons and
polycyclic aromatic hydrocarbons (PAHs), such as naphthalene. For dissolved sources, there will
tend to be higher concentrations of the more soluble chemicals, including single-ring aromatic
hydrocarbons, such as benzene, toluene, ethylbenzene, and xylenes (Lahvis et al., In prep.).
4 The capillary fringe is the tension-saturated zone in soils just above the water table. It will vary in height
depending on soil permeability with greater heights with lower permeability soils.
-------
The distribution of LNAPL sources and presence of residual LNAPL within the unsaturated soil
zone (i.e., in release zones) compared with LNAPL at the water table can influence the
volatilization potential. For LNAPL sources at the water table, the position of the water table
relative to the smear zone can be important, and seasonally higher volatilization rates can occur
when the water table is low and below a portion of the smear zone. The differences in the PHC
vapor concentrations and fluxes for LNAPL and dissolved vapor sources are an important
distinction for defining exclusion distances (Figure 2). A free-phase and residual-phase LNAPL
source will tend to act the same with respect to a vapor source. A residual-phase LNAPL source
will not yield separate-phase LNAPL to the monitoring well and therefore may look the same as
a monitoring well that intercepts a dissolved source with limited concentration attenuation.
Conceptually, the source type (dissolved or LNAPL) will affect the position of the aerobic
reaction front in the unsaturated zone relative to the oxygen source. For dissolved sources, the
reaction front will be located close to the hydrocarbon source (Roggemans et al., 2001; Golder
Associates, 2006; Abreu et al., 2009), while for LNAPL sources, the reaction front position is
more variable but typically is located at greater distances from the source compared with
dissolved sources because of the greater PHC flux (Roggemans et al., 2001; Golder Associates,
2006; Abreu et al., 2009). For dissolved vapor sources, case studies and database evaluations
reported in the literature indicate no confirmed cases of PVI for a wide range of site conditions
(Davis, 2009; McHugh et al., 2010).
ACT THE SAME
LOOKTHESAME
MW
MW
MW
UNSATURATED
ZONE
SATURATED
ZONE
UNSATURATED
ZONE
1 (CAPILLARY ZONE(
SATURATED
ZONE
—
UNSATURATED
ZONE
j (CAPILLARY ZONE
SATURATED
ZONE
A
<
a) free-phase
LNAPL source
MW= Monitoring Well
b) residual-phase
LNAPL source
c) dissolved-phase
source
Figure 2. Conceptual model illustrating the potential for vapor intrusion for a) free-phase LNAPL
sources, b) residual-phase LNAPL sources, and c) dissolved-phase sources. (Source:
Lahvis et al., In prep.; used with permission)
2.4 Anaerobic Biodegradation and Methane Generation
Where anaerobic conditions exist, methane may be produced through the breakdown of PHC
compounds and ethanol, if present, by microbes through the process of methanogenesis. The
terminal electron acceptor in methanogenesis is carbon, and either carbon dioxide or acetic acid
can act as terminal electron acceptors (Ririe and Sweeney, 1995; Wiedemeier et al., 1996). The
-------
formation of the biogases methane and carbon dioxide is of potential concern with respect to
explosion hazard and asphyxiation. Concentrations of methane above the lower explosive limit
can present a flammability and explosion risk. Methane generation also poses potential issues for
PVI, as discussed in Section 2.5.
2.5 Conditions for Increased Potential for Petroleum Vapor Intrusion
As described in U.S. EPA (2012a), site conditions that may result in increased potential for PVI
include:
• Direct contact between contamination (either dissolved or LNAPL) and a
building foundation. Most documented cases of PVI are for this condition (McHugh
etal.,2010).
• Insufficient separation distance. For biodegradation to limit the potential for PVI,
there needs to be a sufficiently thick layer of clean, oxygenated soil between the
building foundation and the contamination. (Clean soil is defined as unimpacted by
residual LNAPL.) The required thickness will depend on hydrocarbon source
concentration and oxygen supply and demand.
• Preferential transport pathways. If a preferential pathway, such as coarse-grained
utility backfill, fractured rock, or karst, connects a contamination source to a building,
the chemical transport can be faster and extend farther than transport through the
surrounding soils.
• Surface capping effect. Building foundations, paved surfaces, and surficial soils with
low effective diffusivity and soil-air permeability (e.g., moist clay layers) can act as a
surface cap to reduce oxygen transfer to the subsurface. The importance of this effect
is not well understood, although significant diffusive oxygen transport through intact
concrete can occur, as indicated by measured rates (typically between 1x10" and
5x10~4 cm2/s) reported in the literature (Branco and de Brito, 2004; Kobayashi and
Shuttoh, 1991; Tittarelli, 2009). Advective transport of atmospheric air to the
subsurface also can occur through openings (e.g., cracks, drains, sumps) in the
building foundation during periods when the building is positively pressurized. These
mechanisms can limit the potential for low oxygen conditions beneath a building.
• Production of methane. Methane may be produced through microbial breakdown of
PHC compounds in anaerobic source zones. The presence of ethanol in a source zone
may increase the methane generation rate compared with a gasoline LNAPL-only
source zone (Nelson et al., 2010; Spalding et al., 2011; Jourabchi et al., 2012). Note
that some sites with releases of fuel containing 10 percent ethanol (E10) are probably
present in the EPA PVI database (see Section 5.3) and in other data compilations
(e.g., Lahvis et al., In prep.). Evaluation of methane generation from ethanol fuel
blends is an area of active research (Jewell and Wilson, 2011). Methane production
can result in soil gas pressures and flow toward receptors and may deplete oxygen
that otherwise could be used for biodegradation of the PHC vapors (Jourabchi et al.,
2012). Elevated methane therefore could increase the potential for PVI at a PHC
release site. However, no published cases were identified where pressure build-up at a
UST site caused soil gas advection to be an issue.
-------
• High organic matter content. Oxygen that would otherwise be used to degrade PHC
vapors may be consumed in settings with high soil organic matter content (e.g., peat
bogs).
• Atmospheric pressure changes. Atmospheric pressure changes could result in
transient advective soil gas flow at sites with deep water tables and coarse-grained
soils. However, such processes are not expected to result in longer-term conditions
where there would be significant differences in the aerobic biodegradation profile,
compared with a diffusion-only transport paradigm.
2.6 Case Studies Indicating Confirmed or Likely Complete Transport Pathway for
Petroleum Vapor Intrusion
Confirmed occurrences of subsurface PVI to indoor air or elevated sub-slab hydrocarbon vapor
concentrations at petroleum sites are rare in the literature5 but important for defining the
inclusion criteria—that is, sites that warrant PVI evaluation because PVI is likely to be found.
Eight case study sites were identified in the literature where PVI was confirmed or likely
(Table 1, following page). Five sites were refinery or petrochemical sites, and three were UST
sites. Common site conditions involved the following: large fuel releases, extensive LNAPL
contamination at the water table, and shallow depth to LNAPL contamination, although at two
sites the separation distances between the building and LNAPL source were approximately 25 to
30 ft (7.6 to 9.1 m). Factors that appeared to contribute to PVI at the two sites with deeper
contamination were a possible geological capping effect at a former refinery site (unknown
location) and heavy rain and a sharp water table rise at a site with a very large petroleum fuel
release (Hartford, Illinois).
2.6.1 Refinery Site, Perth, Australia (Patterson and Davis, 2009)
Monitoring at a former refinery site near Perth, Australia, with a kerosene LNAPL source
indicated elevated (up to 20 mg/L) PHC vapor concentrations and depleted oxygen (<1 percent)
below the interior of a building but much lower hydrocarbon and near-atmospheric oxygen
concentrations near the edge and beside the building. The slab-on-grade building footprint area
was 2,700 ft (251m) with a 30-ft (9. l-m)-wide concrete apron on three sides of the building
and uncovered open ground on the other side. The building is underlain by sand with a LNAPL
zone across the water table at approximately 10 ft (3.0 m) below ground surface (bgs). The
effective diffusion coefficient for chemical transport through concrete measured at the site was
relatively low compared with published data, indicating the concrete slab is not overly porous.
The relatively low diffusivity of the concrete may have reduced oxygen transport to the
subsurface under the building. In addition, because of the wide concrete aprons, the effective
area of the building with respect to oxygen transport restrictions may be larger than its footprint.
5 There are also anecdotal accounts of PVI occurrances at buildings typically with shallow fuel sources and
preferential pathways (e.g., sewer lines, drains) connecting the fuel source to the building.
-------
Table 1. Summary of Case Study Sites with Confirmed or Likely Occurrences
of Petroleum Vapor Intrusion
Site
Perth, Australia
(Patterson and
Davis, 2009)
Chatterton,
Vancouver, B.C.
(Hers etal., 2000)
Casper, Wyoming
(Luo etal., 2009)
Unknown (Luo et
al., 2010)
Hartford, Illinois
(Illinois DPH,
2010)
Stafford, New
Jersey (Sanders
and Hers, 2006)
Ogden, Utah,
Mini-Mart
(McHugh et al.,
2010)
Gunnison, Utah,
Top-Stop
(McHugh et al.,
2010)
Distance
LNAPL to
Building
10 ft (3.0m)
5 ft (1.5m)
1-5 ft
(0.3-1. 5m)
25-30 ft
(7.6-9.1 m)
Depth to
groundwater
~ 23-33 ft
(-7-1 Om)
5.25ft
(1.6m)
3.3ft
(1.0m)
Depth to
groundwater
-13-16 ft
(-4-5 m)
Building Source Vapor
Size Concentrations
2,700ft2 TPH=20mg/L
610ft2 TPH=40mg/L
2,300ft2 TPH = 100mg/L
2,100ft2 TPH~60-160mg/L
N/A N/A (gasoline
source)
700 ft2 Benzene = 0.66
mg/L; 2,2,4-TMP =
2.1 mg/L; MtBE =
5.9 mg/L
N/A N/A (gasoline
source)
N/A N/A (gasoline
source
Facility
Refinery
Petro-
chemical
Refinery
Refinery
Refinery
UST
UST
UST
Comments
30-ft (9.1-m) building apron
on 3 sides of building
PVI only when building
depressurization was - 10
Pa
Shallow LNAPL source
Capping effect from
geology observed
Very large spill, episodic
PVI events when heavy
rain or sharp rise in water
table
PVI observed for MtBE,
2,2,4-TMP and
cyclohexane but not for
BTEX
Large release, odors
detected in building
Sudden 20,000-gallon
(75,708 L) release, odors
detected in buildings up to
500 ft (152m)
downgradient of source
TPH = total petroleum hydrocarbons; N/A = not available
2,2,4-TMP = 2,2,4-trimethyl pentane; MtBE = methyl tert-butyl ether; BTEX = benzene,
toluene, ethylbenzene, and xylenes2.6.2 Chatterton Petrochemical Site,
Vancouver, B.C., Site (Hers et al., 2000; Hers et al., 2002)
At the former Chatterton petrochemical site near Vancouver, B.C., a greenhouse was constructed
above a residual LNAPL source comprising benzene, toluene, and xylene. Monitoring indicated
depletion of oxygen (<1 percent) and a complete PVI pathway when the building was continually
depressurized (to approximately 10 Pa), but only partial oxygen depletion and no complete
pathway under natural (near-neutral) pressure conditions. The slab-on-grade building footprint
was 610 ft2 (57 m2), the building was underlain by sand, and the depth to the LNAPL smear zone
was 5 ft (1.5 m) below the building foundation slab.
2.6.3 Refinery Site, Casper, Wyoming (Luo et al., 2009)
Monitoring at a site in Casper, Wyoming, indicated a complete PVI pathway at a refinery site
with a light distillate (gasoline-range) LNAPL source. LNAPL contamination was present in the
10
-------
unsaturated zone at depths between about 1 and 5 ft (0.3 and 1.5m) below a small warehouse-
type building with a slab-on-grade foundation. Monitoring of sub-slab soil gas indicated oxygen
was depleted (<1 percent) below the interior regions of the building (except near the saw-cut
expansion joints during times of positive building pressures) but not below the edges of the
building.
2.6.4 Former Refinery Site (confidential location) (Luo et al., 2010)
Monitoring at a former refinery site with a building overlying a light distillate (gasoline-range)
LNAPL source indicated relatively uniform and elevated (60 to 160 mg/L) PHC vapor
concentrations and depleted oxygen beneath and around the building foundation. The building
footprint area is 2,100 ft2 (195 m2) with a basement depth of 5 ft (1.5 m) bgs. The ground surface
is predominantly grass covered, except on one side of the building where there is an
asphalt/concrete parking lot. Soils with LNAPLs were first encountered at about 30 to 35 ft
(9.1 to 10.7 m) bgs; therefore, the separation distance between the building foundation and
contamination (LNAPL) is approximately 25 to 30 ft (7.6 to 9.1 m). Detailed soil respiration and
soil-air permeability test results suggest two possible reasons for the observed behavior and
elevated shallow PHC concentrations: 1) significant background oxygen uptake in surface soils
or 2) physically limited oxygen transport from the atmosphere. Soil oxygen uptake rates in
shallow soil ranged from 2 to 25 mg-oxygen/kg-soil/day. There were silt and clay layers between
2 to 5 ft (0.7 to 1.5 m) and 7 to 8 ft (2.1 to 2.4 m) bgs, both with soil-air permeabilities of less
than 1 x io~14 m2. The results from Luo et al. (2010) suggest both of these reasons are plausible
for the observed soil vapor behavior.
2.6.5 Refinery Site, Hartford, Illinois (Illinois Department of Public Health, 2010)
Soil gas monitoring at a Hartford, Illinois refinery site with a very large petroleum fuel spill
(several million gallons) indicated episodic PVI into buildings when there were heavy rains and a
sharp rise in the water table. (Note that there are sites where the opposite effect is observed: soil
vapor concentrations rise when the water table falls below LNAPL source zones.) The vadose
zone soils consisted of coarse sand overlain by fine sediments, and the depth to groundwater
ranged from 23 to 33 ft (7 to 10 m) bgs.
2.6.6 UST Site, Stafford, New Jersey (Sanders and Hers, 2006)
Monitoring of a house with a basement above a residual gasoline LNAPL source at a site with
sandy soils indicated PVI of methyl tert-butyl ether (MtBE), 2,2,4-trimethylpentane
(2,2,4-TMP), and cyclohexane but not benzene, toluene, and xylene compounds. The depth to
the LNAPL source was 10.75 ft (3.27 m), which was 5.25 ft (1.60 m) below the basement
foundation. The source soil vapor concentrations of benzene, 2,2,4-TMP, and MtBE were
0.66 mg/L, 2.1 mg/L, and 5.9 mg/L, respectively. It was inferred that, compared with benzene,
MtBE attenuated to a lesser degree because of its lower degradation rate. Also, 2,2,4-TMP
attenuated to a lesser degree than benzene because of its lower solubility (biodegradation occurs
in the water phase). No PVI was detected at a nearby slab-at-grade building above residual
LNAPL or other buildings above a dissolved groundwater source.
11
-------
2.6.7 UST Site, Ogden, Utah, Mini-Mart Release (McHugh et al., 2010)
A large release at a UST site in Ogden, Utah resulted in free product gasoline floating on shallow
groundwater less than 3.3 ft (1 m) below the bottom of the slab. Petroleum odors reported in a
building were mitigated by installing a positive pressure HVAC system.
2.6.8 UST Site, Gunnison, Utah, Top Stop Release (McHugh et al., 2010)
A large, sudden gasoline release (20,000 gallons [75,700 L]) occurred at a UST site in Gunnison,
Utah. The vadose zone soils consisted of silty sand and gravel overlain by sandy silt, and the
depth to groundwater was 13 to 16 ft (4 to 5 m) bgs. The soil headspace photoionization detector
(PID) readings in the LNAPL source zone were in the range of 100's to 1,000's ppmv (parts per
million by volume). In the first several months after the release occurred, people complained of
gasoline odors, and elevated PID readings were reported in several buildings up to 1,640 ft
(500 m) from the release site in the direction of groundwater flow.
3. Summary of Modeling Studies
Numerous modeling studies of aerobic biodegradation have been conducted to evaluate
biodegradation processes, identify factors influencing biodegradation, compare modeled to
measured hydrocarbon vapor attenuation, and estimate first-order biodegradation rates (e.g., Jury
et al., 1983; Lahvis and Baehr, 1999; Hers et al., 2000; Ririe et al., 2002; Grathwohl and Maier,
2002; Robinson and Tursczynowisz, 2005; Abreu and Johnson, 2005; Abreu and Johnson, 2006;
DeVaull, 2007b; Abreu et al., 2009; Davis et al., 2009; DeVaull, 2011; Hers et al., In prep.;
U.S. EPA, 2012c). Modeling studies using representative first-order biodegradation rates indicate
that aerobic biodegradation is a rapid and, in some cases, essentially instantaneous process and
that attenuation of petroleum hydrocarbon vapor concentrations occurs over relatively short
distances (a few feet), consistent with the observed field data (e.g., Hers et al., 2000; Davis et al.,
2009). An important input to modeling studies is the first-order biodegradation rate; a
comprehensive compilation of such rates is provided in DeVaull (2011).
The biodegradation of aliphatic hydrocarbon compounds is less well studied, but available data
suggest bioattenuation distances may be greater for aliphatic hydrocarbons than for aromatic
hydrocarbon compounds. For example, greater concentration attenuation between deep and
shallow soil vapor was observed for benzene (and aromatics in general) than for 2,2,4-TMP (and
aliphatics in general). Examples from two sites illustrate this behavior:
• At the Stafford site (discussed previously in Section 2.6.6), the ratio between deep
and shallow soil vapor concentrations was 220 times lower for benzene than for
2,2,4-TMP (Sanders and Hers, 2006); and
• At a site in North Battleford, Saskatchewan, this ratio was 40 times lower for benzene
than for 2,2,4-TMP (Hers et al., In prep.).
The modeling studies reviewed below were selected to provide insight on the vertical and lateral
attenuation of PHC vapors and, where possible, the influence of factors such as source vapor
concentrations and layered soil deposits on PHC vapor migration and attenuation. By examining
12
-------
model results that predict distances over which PHC vapor concentration attenuation occurs for
varying site conditions, this review offers evidence that can help inform the development of
inclusion distances for sites where PVI is being assessed.
3.1 Abreu Three-Dimensional Model Simulations
Abreu and Johnson (2005) present the theoretical basis for a three-dimensional model for
predicting soil vapor-to-indoor air attenuation factors incorporating subsurface processes of
diffusion, gas-phase advection through building depressurization, oxygen-limited first-order
biodecay, and uniform mixing of vapors entering a building. Three-dimensional modeling
scenarios of interest are summarized below.
3.1.1 Three-Dimensional Model Simulations—Below-Building Contamination Source and
Homogeneous Soil Conditions
Abreu et al. (2009) presents model simulation results for oxygen-limited aerobic biodegradation
for a scenario where the building parameters assumed were representative of a residential house
above a homogeneous sand unsaturated zone. The hydrocarbon modeled was assumed to have
the same fate and transport properties as benzene. Oxygen-limited decay was simulated in the
aerobic portion of the unsaturated zone (i.e., when oxygen concentrations exceeded 1 percent). A
first-order biodegradation rate of 0.79 h"1 was assumed for the hydrocarbon, which is consistent
with published rates for benzene (e.g., DeVaull, 2011).
Potentially conservative attributes of the Abreu et al. (2009) model simulations include the
following:
• oxygen transport occurs only through cracks in the foundation and not through intact
concrete;
• the building is continuously depressurized; thus, for cases where pressure cycling is a
relevant condition, no atmospheric air moves downward into the soil at times when
the building is pressurized; and
• there is no oxygen recharge through pressure effects caused by wind and/or
atmospheric pressure changes.
Spatially variable soil properties (e.g., moisture, porosity, permeability) were not considered, a
potentially non-conservative modeling assumption. Conceptually, there are scenarios where
layered systems consisting of a fine-grained, wet surface soil layer underlain by a coarser-
grained, drier soil layer could increase the potential for oxygen limitations below buildings.
The Abreu et al. (2009) model results are summarized in Figure 3. For context, the EPA PVI
database indicates representative total hydrocarbon vapor concentrations (excluding methane)
between 100 and 200 mg/L above gasoline LNAPL distributed above the capillary fringe, and
the approximate lower end of this range likely indicates weathered gasoline sources. The total
hydrocarbon vapor concentration range simulated by Abreu et al. (2009) is representative of a
gasoline source but, because of the additional oxygen demand represented by the oxidation of
methane, less representative of a source that also includes elevated methane concentrations. The
13
-------
empirical analysis indicates methane concentrations at most sites with methane data were low
(Section 7.1.3). For a dissolved vapor source, the database indicates that the maximum total
petroleum hydrocarbon (TPH) vapor concentration measured in vapor was 10 mg/L, with more
than 99 percent of the data indicating TPH vapor concentrations less than 1 mg/L.
l.E-02
l.E-03
l.E-04
l.E-05
Dissolved phase
NAPL
l.E-06
TO
3
C
2 l.E-07
l.E-08
l.E-09
l.E-10
1 10 100
Source Vapor Concentration (mg/L)
(biodegradation rate = 0.79 Fr1)
•L = lm Ji 1 = 2 m D L = 3 m
•L = 5m A L = 10 m No Biodeg, L= 1 m
No Biodeg, L= 10 m
1000
Figure 3. Vapor intrusion attenuation factors predicted by Abreu and Johnson (2005) three-
dimensional model for a range of source total hydrocarbon (benzene) vapor
concentrations and separation distances for a residential house scenario (adapted
from Abreu et al., 2009).
The model-predicted vapor attenuation factors presented in Figure 3 are highly sensitive to
source hydrocarbon concentrations above 10 mg/L. Below 10 mg/L, the attenuation factors are
relatively constant for a given separation distance. For a source vapor hydrocarbon concentration
representative of weathered gasoline (100 mg/L), the model predictions in Figure 4 for source-
building distances of 3.3 ft (1 m) and 16.4 ft (5 m) predict oxygen concentrations less than
1 percent below the building. It is not until the source-building distance is increased to 23 ft
(7 m), shown in the lower panel of Figure 4, that an aerobic reaction front and corresponding
orders-of-magnitude reduction in hydrocarbon vapor concentrations below the building is
observed. As previously discussed, a potentially conservative aspect of the model predictions in
Figure 4 is that they assume no oxygen transport through the building foundation.
14
-------
Hydrocarbon
Oxygen
Figure 4. Effect of source depth on soil gas distribution and vapor intrusion attenuation factors
predicted by Abreu and Johnson (2005) three-dimensional model fora source total
hydrocarbon (benzene) vapor concentration of 100 mg/L, biodegradation rate of
0.79 h"1, and three source depths: 3 m (top), 7 m (middle), and 9 m (bottom) bgs.
Hydrocarbon and oxygen concentrations are normalized by source and atmospheric
concentrations, and the building is a residential house (from Abreu et al., 2009).
Abreu et al. (2009) also present a chart that provides representative attenuation factors (a) that
apply to all source hydrocarbon vapor concentrations below 10 mg/L and for a range of
biodegradation rates and source depths beneath the foundation (Figure 5). The attenuation
factors are much lower when the aerobic biodegradation process is included compared to the
non-biodegradation case. For example, there is an approximate three order-of-magnitude
reduction in the attenuation factor for a source-foundation separation distance of 5 ft (1.5 m) and
first-order biodegradation rate of 0.79 h"1.
15
-------
1.E-02 i
1.E-03 i
1.E-Q4 =
o 1.E-05 =
o
^ 1.E-06 =
O :
1 1.E-07 =
C :
g 1.E-08 i
1.E-09 =
1.E-10 :
1.E-11
No Biodegradation
}. = 2 h'1
0
23456789 10
Vapor Source Depth below Foundation {m)
11
Figure 5. Vapor intrusion attenuation factors predicted by Abreu and Johnson (2005) three-
dimensional model as a function of separation distance below foundation and first-
order biodegradation rate fora residential house scenario and 10 mg/L vapor source
concentration (from Abreu et al., 2009).
3.1.2
Three-Dimensional Model Simulations—Lateral Migration Scenario and
Homogeneous Soil Conditions
U.S. EPA (2012c) presents a modeling study of conceptual model scenarios for the vapor
intrusion pathway where the Abreu and Johnson (2005) model was used for a range of
simulation scenarios, including oxygen-limited aerobic biodegradation of PHC compound vapors
(using benzene as a surrogate for TPH). The building assumptions in U.S. EPA (2012c) are
similar to those described in Abreu et al. (2009). One of the scenarios evaluated was the
influence of PHC source and building lateral separation distance on the predicted vapor
attenuation factor (Figure 6). The simulations were conducted for a TPH vapor concentration of
200 mg/L, a 2-m (6.6 ft) deep basement, two contamination source depths (3 m and 8 m [9.8 ft
and 26 ft] bgs), and a range of first-order biodegradation rates (0.018, 0.18, and 1.8 h"1). The
different biodegradation rates compared to the earlier Abreu study reflect the progression in
studies of rates over time. The different rates do not materially change the findings described
herein. The predicted vapor attenuation factors decrease rapidly as the lateral distance increases.
16
-------
For example, the vapor attenuation factor for a shallow LNAPL source that is offset
approximately 15 ft (4.6 m) from the edge of the building is 1 x 10~6 for a biodegradation rate of
•vlO.
0.018 h"1 and less than 1 x IQ"1 for a biodegradation rate of 0.18 h" . The vapor attenuation factor
for a shallow below-building LNAPL source and the same biodegradation rate is approximately
1 x io~3. A key point is that there is greater attenuation and hence, lower vapor attenuation
factors, for lateral building-contamination source separation scenarios compared with vertical
ones.
1.E-02
1.E-03T
I
j source zone no Ic
I beneath building
>nge
01 1.E-06:
1.E-08 :
1.E-09 :
1.E-10
-15 -10 -50 5 10 15
Source Edge - Building Center Separation (m)
20
25
•Source 3 m bgs, no biodegradation
-Source 3 m bgs , A = 0.018 (1/h)
-Source 3 m bgs , A =0.18 (1/h)
-Source 3 m bgs ,A=1.8(1/h)
Source 8m bgs, no biodegradation
- -A- - Source 8m bgs , A= 0.018 (1/h)
- O- -Source 8m bgs ,A= 0.18 (1/h)
- -Q- - Source 8m bgs , A= 1.8 (1/h)
Figure 6. Relationship between source-building lateral separation distance and normalized
indoor air concentration (a) for a NAPL source, two source depths, and three
biodegradation rates (A). The source-building lateral separation is measured from the
edge of the source zone to the center of the building with a basement; negative values
and values of less than 5 m indicate that the source is to some extent beneath the
building. The source vapor concentration is 200 mg/L. (Source: Figure 33 in U.S. EPA
[2012c])
17
-------
3.1.3 Three-Dimensional Model Simulations—Surface Capping Scenario
U.S. EPA (2012c) also provided aerobic biodegradation modeling simulations showing the
influence of variable soil moisture and layered soil deposits on subsurface hydrocarbon
concentrations and vapor intrusion. The simulations included various multilayer configurations
involving up to three 1-m (3.3-ft)-thick soil layers with higher moisture content (60 percent
saturation) and lower permeability within a general soil profile with a lower moisture content
(20 percent saturation) and higher permeability (Figures 7 and 8). The hydrocarbon source for
these simulations was located at 8 m (26 ft) below ground surface, beneath a house with a 2-m
(6.6-ft)-deep basement and both positive (5 Pa) and negative (-5 Pa) building pressurizations.
For a relatively high source hydrocarbon vapor concentration (200 mg/L), low permeability
(high moisture) layers beneath the building tend to limit the diffusion of hydrocarbon vapors
from the source, while a low permeability layer at the surface (i.e., a surface cap) limits oxygen
diffusion and biodegradation in the subsurface. The individual and combined effects of low
permeability surface and subsurface layers on source-to-building attenuation factors can be seen
in Figure 7, which shows the effects of four configurations of such layers (rows A-D) on an
under pressurized (left panels) and over pressurized (right panels) residence. Figure 8 shows that
the attenuation factor predicted for a surface capping scenario with a 200 mg/L source was an
order of magnitude greater than the homogeneous (one-layer) soil scenario (Table 2).
For the lower source vapor concentration (2 mg/L) shown in Figure 8 and Table 2, the vapor
attenuation factor for the two-layer scenario was 7.1 x 10"15. Although a single-layer simulation
was not performed for the lower source strength, this two-layer scenario attenuation factor is
very low and indicates essentially complete biodegradation, with the cap having little or no effect
on oxygen levels or hydrocarbon vapor bioattenuation below the building.
It is acknowledged that available modeling addresses a limited number of capping scenarios, and
additional work in this area would be valuable. The capping scenario inputs are considered
reasonably representative. For clay soils, saturations could be greater than 60 percent over short
time periods (i.e., weeks), but the fine-grained layer modeled is relatively thick (1 m).
18
-------
•c ™
a) 0-
00
s.
in
•o
ffl
SI
c
•5
i
O 0) o 0)
ijfl
S. £ i
= -
a) 3
E °
& ^
a) c
w o
=
C
N re O
i_ i_ o
3¥ •-
t/> §
£ c
Q. O
i_ O
"2 <"
!?
re °
o
Q.
re
0)
3 "
w ^
O C
re
- re >
•-
i.0.
-B
Q.
* 2 g 5T~
w
o
'C
re
c
8
C .
8 w-
i_ a)
re re ?
0)
(w)s6qi||daa
0)
3
O)
19
-------
Hydrocarbon
Oxygen
Lithology
-2
-6-
O)
(D
-8-
g = 7.1E-15
1E-7
1E-5
1E-3
0.1
0.9
Hydrocarbon Source Vapor Cone = 2,000 mg/m3
2 4 6 8 10 120 2 4 6 8 10 12
Hydrocarbon Source Vapor Cone = 200,000 mg/m3
x (m) x (m)
Figure 8. Effect of source vapor concentration on hydrocarbon and oxygen distribution in soil
gas and normalized indoor air concentration (a) for scenarios with low permeability
soils at the ground surface (e.g., soil layer scenario on row D of Figure 7).
Hydrocarbon and oxygen concentration contour lines are normalized by source and
atmospheric concentrations, respectively. Source located at 8 m bgs (basement
scenario). Biodegradation rate (A) = 0.18 h"1. (Source: Figure 38 in U.S. EPA [2012c])
Table 2. Select Three-Dimensional Abreu and Johnson (2005) Model
Simulation Results from U.S. EPA (2012c)
Source
Hydrocarbon
Vapor
Concentration
(mg/L)
200
200
2
Building
pressurization
(Pa)
5
-5
-5
Vapor Intrusion Attenuation Factor
Single Soil Layer
(no surface cap)
9.8 x 10'21
6.7 x 10'5
N/A
Two Soil
Layers
(surface cap)
4.9 x 1Q-11
6.8 x 10'4
7.1 x 10'15
Five Soil Layers
(no surface cap)
4.0 x 10'24
8.1 x 10'13
N/A
Six Soil Layers
(surface cap)
2.1 x 10'15
5.0 x 10'7
N/A
Notes: Residential house with 2-m (6.6-ft)-deep basement, depth to hydrocarbon vapor source = 8 m (25.3 ft), first-
order biodegradation constant equal to 0.18 h"1. Lower attenuation factors indicate higher attenuation. N/A = not
available
20
-------
3.1.4 Comparison of Modeled to Measured Soil Vapor Concentration Data
The three-dimensional model predictions of vertical profiles of hydrocarbon vapor and oxygen
concentrations showed good agreement between measured and modeled data for three sites
evaluated by Abreu and Johnson (2006). The estimated first-order biodegradation rate for these
studies ranged between 0.18 and 0.4 h"1. The three-dimensional model was also applied to
compare measured and modeled hydrocarbon vapor concentrations for the former refinery site
(confidential location), described in Section 2 of this report (Luo et al., 2010). A good
comparison was obtained when the model incorporated site-specific conditions (a surface soil
layer of low diffusivity and low soil-air permeability), but when generic (homogeneous) soil
conditions were assumed, the model was not conservative and underpredicted the measured
concentrations by a factor of approximately 100. The first-order rate incorporated in the Luo
et al. (2010) model simulations was 0.18 h"1.
3.2 DeVaull (2007b) Study (BioVapor Model Development)
DeVaull (2007b) presents the theory and one-dimensional model simulation results for a
subsurface soil vapor-to-indoor air chemical PVI model that includes oxygen-limited
biodegradation. (The model described is the basis for the BioVapor model [American Petroleum
Institute (API), 2012].) The processes simulated by the algebraic model are one-dimensional
upward diffusion and aerobic biodegradation of chemicals in a homogeneous subsurface soil
layer and mixing of vapors within a building enclosure. The soil is divided into a shallow aerobic
layer where first-order decay is assumed to occur and a deeper anaerobic layer in which
biodegradation does not occur because of oxygen limitations. The boundary between the aerobic
and anaerobic zones is determined iteratively to match oxygen demand to supply.
The model results indicate that vapor intrusion of PHCs can be orders of magnitude less than
indicated by estimates that neglect biodegradation. A model sensitivity analysis using specified
ranges of scenario parameters showed a high degree of sensitivity to oxygen availability, soil
properties, and biodegradation rates. The attenuation factor varied by more than nine orders of
magnitude about a specified attenuation factor of 1 x 10"8; however, the corresponding variation
in contamination source-to-foundation separation distance was within only a factor of
approximately three (the attenuation factors for the non-biodegradation scenario ranged from
approximately 1 x 10"3 to 1 x 10"5). DeVaull (2007b) concludes that identifying a distance where
PVI is unlikely to occur is a more robust screening tool than an attenuation factor for PHC
compounds.
Favorable comparison of the one-dimensional model to the three-dimensional results of Abreu
and Johnson (2005) is shown in DeVaull (2007b). With matched model parameters, the models
show similar estimates of indoor air-to-subsurface source vapor concentrations and similar
sensitivities of both attenuation factor and exclusion distance to changes in model parameters.
3.3 DeVaull (2010) Study of BioVapor Application
DeVaull (2010) presents BioVapor model simulations in which the sensitivity of the model
predictions was evaluated for a scenario characterized by a building with a basement separated
by a distance of 5 ft (1.5 m) from a dissolved hydrocarbon vapor source (Figure 9). The model
21
-------
simulations were designed, in part, to provide insight on the applicability of the dissolved-source
exclusion distance of 5 ft (1.5 m) proposed by Davis et al. (2009). The source groundwater
concentrations were 1 mg/L benzene and 3 mg/L for each of toluene, ethylbenzene, and xylenes,
for a total benzene, toluene, ethylbenzene, and xylenes (BTEX) source concentration of 10 mg/L.
The source vapor concentrations were estimated from the Henry's Law constant and a
groundwater to deep soil vapor attenuation factor of 0.1, resulting in source benzene and BTEX
vapor concentrations of 12 mg/m3 and 120 mg/m3, respectively.
Effective Foundation Airflow (L/min)
1
I
o
O
CD
CD
N
c
CD
00
o
o
-a
100
10 •
1-
0.1 •
0.01 •
0.001 •
100
0.0001
Below nominal
indoor air
background
Biodegradation
included
H 1 1 I I Mil
Soil Layer
Entirely Aerobic
cs
o
^;
-------
oxygen mass transfer rate was converted to an effective foundation airflow rate by considering
the density and oxygen content of air. The effective foundation airflow rate is conceptually easier
to understand than the oxygen mass transfer rate because of the common usage of the soil gas
advection rate (Qsoii) parameter in modeling studies. The model predicts low indoor air benzene
concentrations (less than 1 ng/m3) for the range of effective foundation airflow rates considered.
It is also instructive to note that the model results for recalcitrant non-degrading chemicals show
an opposite trend: the indoor air VOC concentration increases as the effective foundation
airflow rate increases.
3.4 Summary of Modeling Studies
Modeling study results, particularly for LNAPL sources, cannot be easily and precisely
correlated to distances (or inclusion distances) beyond which PVI is unlikely to occur. This is, in
part, because of the sensitivity of the model predictions to key site-specific variables such as
source vapor concentration, separation distance, biodegradation rates, and oxygen diffusion
through building foundations.
For LNAPL vapor sources, the Abreu et al. (2009) three-dimensional model simulations for a
residential house scenario and homogeneous soil conditions predict that a vertical separation
distance of 23 ft (7.0 m) or more is required for aerobic reaction front development within the
unsaturated zone. The modeling results for smaller separation distances and an LNAPL source
indicate the attenuation factor calculated by the model is sensitive to a surface capping effect.
For dissolved vapor sources, the model simulations indicate very low attenuation factors and
negligible potential for a complete PVI pathway, including a modeling scenario where a surface
capping effect was simulated. The model simulations strongly support the inclusion distances for
dissolved sites subsequently described in this report.
The modeling results indicate further evaluation of factors potentially influencing oxygen supply
and demand is warranted for the LNAPL source scenario. These factors include source vapor
concentration strength, building size, surface foundation and soil layer properties, and natural
soil oxygen demand.
4. Review of Empirical Database Studies of Petroleum Hydrocarbon
Vapor Attenuation
Three published studies that analyze empirical data on PHC vapor attenuation are summarized
below.
Davis (2009) obtained soil gas data from 53 geographical locations in the United States and
Canada and from an analysis of 259 benzene and 210 TPH samples. For dissolved-phase sources,
the analysis indicates 5 ft (1.5 m) is sufficient to significantly attenuate benzene and TPH vapors.
Dissolved-phase sites were defined as sites where benzene concentrations in groundwater were
less than 1,000 ng/L. Analysis of a dataset that includes LNAPL sources of all types (e.g., USTs,
refineries, fuel terminals), indicates that a separation distance of 30 ft (9.1 m) is required for
benzene vapor attenuation. The data obtained by Davis (2009) are incorporated into the database
and further analysis of these data is described in this report.
23
-------
Peargin and Kolhatkar (2011) evaluated 218 pairs of benzene soil vapor and groundwater
concentration data from 25 sites. A rigorous quality assurance/quality control (QA/QC) program
was followed for data collection, including installation of properly sealed permanent probes and
leak tracer tests. Data were categorized in bins based on 10~2 to 10~6 excess cancer risk and
assuming a soil vapor-to-indoor air attenuation factor of 0.01 (U.S. EPA, 2002). No benzene
soil vapor concentrations exceeding 300 ug/m3 (air concentration for 10"5 cancer risk multiplied
by 0.01) were observed at vertical separation distances greater than 15 ft (4.6 m). Benzene soil
vapor concentrations exceeding 300 ug/m3 were only observed above groundwater sources
where benzene concentrations exceeded 1,000 ug/L. The authors concluded that the data support
a CSM where benzene vapor transport at concentrations exceeding target screening values can
only occur where groundwater source benzene concentrations are high, defined for the current
study as greater than 1,000 ng/L. Some of the data analyzed by Peargin and Kolhatkar (2011)
that were provided to Davis (2009) are incorporated in the current EPA PVI database. The
remaining Peargin and Kolhatkar (2011) data were not readily accessible during the data analysis
performed for this report.
Wright (2011) presents data from 124 sites in Australia. There are 1,080 pairs of benzene soil
vapor and groundwater concentration data; 41 percent of the data were obtained at sites with
fractured rock aquifer systems and 12 percent represent data obtained below building
foundations (i.e., sub-slab). After removing the data from fractured rock sites and from sites or
sets of probes that did not have clean soil between the source and the soil gas sampling port, the
analysis resulted in vertical exclusion distances of 5 to 10 ft (1.5 to 3.0 m) for relatively low-
strength dissolved-phase sources (benzene < 1 mg/L and TPH < 10 mg/L) and -30 ft (-10 m) for
LNAPL and poorly characterized dissolved-phase sources (including sites with large building
slabs). The lower threshold benzene and TPH soil vapor concentration for estimating the
exclusion distances was based on 5 percent of the lowest Australian health screening levels
(Friebel and Nadebaum, 2011).6
The Australian data analyzed by Wright (2011) were incorporated into the EPA PVI database but
in this report were analyzed separately from the North American sites; Lahvis et al. (In prep.) has
analyzed the two datasets together. The analysis of the Australian database presented in
Appendix C suggests that the conclusions of the empirical analysis would not change if the
Australian data were included, although site conditions differ between the United States and
Australia.
5. EPA PVI Database Development, Structure, and Content
The database compiled by Davis (2009, 201 la) was the starting point for the EPA PVI database.
The Davis database contained data on PHC vapor behavior from over 50 sites and included
information on groundwater and soil vapor chemistry, soil properties, and other site data. The
May 2011 version of the Davis database was imported and used in this analysis.
6 The soil vapor concentrations corresponding to 5 percent of the Australian Health screening levels are as follows:
benzene = 50 |J.g/m3, toluene = 65,000 |J.g/m3, ethylbenzene = 16,500 |J.g/m3, xylenes = 11,000 |J.g/m3, and TPH(6-
16) = 15,500 |ag/m3.
24
-------
5.1 EPA PVI Database Development and Checking
The Davis database was imported into Microsoft Access and then exported into a working
Microsoft Excel spreadsheet to enable data checking, addition, and analysis. The original Davis
database was expanded by adding new data fields to describe the data types needed to meet the
objectives of this report. The data in the spreadsheet were checked against the original data when
reports, journal articles, and other references were available, and all available references are
included as electronic files linked to the EPA PVI database. As additional sets of data were
imported into the database, each was examined and corrected for internal consistency.
Once the checks and additions were complete, the Excel spreadsheets were re-imported into the
Access database and checked for consistency and accuracy of import using queries and manual
checks. The EPA PVI database is available in Microsoft Access format as a companion to this
report, along with Microsoft Excel spreadsheet outputs of the basic data used in the data analysis.
Abridged summaries of these data are provided in Appendix B.
5.1.1 Quality Control and Data Quality Ranking
For the two-part data verification process, data were screened to establish minimal acceptable
data quality for inclusion in the database and use in the analysis, and data quality indicators were
developed to ensure the included data are of known, acceptable, and documented quality. Data
that were determined to be of unacceptable quality were either not added to the final dataset or
flagged to facilitate their separation during data analysis. This effort focused on identifying data
of questionable quality for use in the subsequent analysis. Data that were considered
questionable included:
• Analytical data obtained by unacceptable methods, or data with no reported
methodology or evidence of QA/QC processes;
• Soil gas data from fractured rock systems (because of the potential for preferential
soil gas flow);
• Data where the benzene concentrations in groundwater were below the detection level
(i.e., no contamination source exists); and
• Data where lateral spacing between a groundwater monitoring well and a soil gas
probe (for paired data) was greater than about 30 ft (9 m). (Note that this information
was not available for all data).
Hydrocarbon analytical methods considered acceptable for individual air-phase compounds
included EPA Method TO-15, EPA Method TO-3, and Modified EPA 8260; for TPH, the
Massachusetts Air Phase Hydrocarbons method or the equivalent was preferred. For fixed gases
(oxygen, carbon dioxide, methane), the following analytical methods were considered
acceptable: ASTM D1946 and EPA Method 3C. Note that data quality checks were not repeated
for all the data imported from the Davis database because they had been previously checked
(e.g., Davis 2009), but some spot checks were made as the EPA PVI database was assembled.
25
-------
For the Australian data, analytical methods were generally consistent and equivalent for benzene,
but the methods used for TPH and fixed gases (field or laboratory) were not specified or
available in time for inclusion in this report. As a result, the Australian fixed gas and TPH data
were not evaluated or used in the conclusions of this report (but they are included in the EPA
PVI database). Also, the Australian data for fractured settings were included in the database but
were flagged and not used in the analysis. The U.S. and Canadian data do not include fractured
rock settings.
As part of initial data screening, consideration was given to whether to screen out data for which
there were no probe leak tracer test results. Given that a significant proportion of the data were
from older studies where leak tracer tests were either not conducted or not reported, the decision
was made not to adopt this criterion for data screening. However, it is noted that much of the
recent data include leak tracer tests or tests for consistent fixed gas data.
For the second part of the data verification process, each site's data quality was considered to
decide whether to use the data in the analysis. Sites were ranked according to several criteria:
• Availability of information to evaluate key data quality indicators such as leak
checks, probe purging, analytical method, or sampling probe installation;
• Adequacy of CSM development for locating and sampling soil gas probes;
• Use of appropriate soil gas sampling protocols (i.e., leak checks, purging, permanent
probes);
• Appropriateness of analytical methods;
• Consistency of fixed gas and volatile organic compound (VOC) soil gas results;
• Publication in a journal, EPA report, or other peer-reviewed source, and
• Oversight by a federal, state, or local UST regulatory program.
Each site was scored considering these factors on CSM robustness using a three-point ranking
(3 highest, 1 lowest) and data quality using a five-point ranking (5 highest, 1 lowest). CSM
robustness rankings were defined as follows:
• CSM-3: Well-developed CSM, appropriately located soil gas probes, vertical soil gas
profiles, well-characterized contamination source (LNAPL and dissolved), and
available ancillary data (e.g., soil properties);
• CSM-2: Less well-developed CSM with well-located probes but more limited soil
gas locations (e.g., single location) and a reasonably well-characterized contamination
source; and
• CSM-1: Limited data to develop CSM and evaluate appropriateness of soil gas probe
locations and results, generally a single soil gas location, limited or no CSM-related
information, and/or inadequate data to perform clean soil thickness analyses.
Data quality (DQ) rankings were defined as follows:
• DQ-5: Very high-quality data with fully documented QA/QC, permanent probes, leak
tracer and/or pneumatic testing, and fixed gas data consistent with hydrocarbon vapor
26
-------
concentrations. In some cases, a site's data were given a DQ-5 ranking when not all
of these aspects were met but there was a well-developed CSM and the research was
peer reviewed;
• DQ-4: High-quality data, with mostly documented QA/QC, generally permanent
probes and leak tracer testing, and fixed gas data that were consistent with
hydrocarbon vapor concentrations;
• DQ-3: Moderate-quality data, with some QA/QC documentation and fixed gas data
that may be limited in quantity or inconsistent with hydrocarbon vapor concentration
data;
• DQ-2: Low-to-moderate quality data, with limited QA/QC documentation (but
typically still collected under state program oversight), limited data documentation,
and no fixed gas results (minimum data quality for inclusion in database analysis);
and
• DQ-1: Low-quality data, unacceptable data-quality indicators or methods (data
excluded from all analyses).
Table B-l in Appendix B shows these data quality measures for the U.S. and Canadian sites in
the database. A detailed table can also be found in the PVI database that accompanies this report.
Note that the Australian data were developed using a similar but not identical CSM and data-
quality scoring system; these scores are included in the Australian data analyzed in Appendix C
and included in the EPA PVI database.
5.2 EPA PVI Database Structure
The tables and fields in the EPA PVI database and a comprehensive data dictionary and entity-
relations diagrams are provided in Appendices D and E, respectively. An Excel spreadsheet was
designed to facilitate evaluation, analysis, and presentation of data relations in the EPA PVI
database and used to perform the analyses described in this report. Filters were added for most
data fields, enabling screening of data based on site conditions and other applicable attributes.
The different data types are summarized as follows:
• Background data: Site location, geologic setting, contamination type, and generic
soil description;
• Facility type: UST, fuel terminal, petroleum refinery, and petrochemical plant;
• Site conditions: Soil type, water-filled and total porosity, and surface cover at soil
vapor probe (bare ground, asphaltic pavement, building);
• Sampling data: For each probe, vertical depth from ground surface to water table, to
top of contamination, and to media sampling locations. Lateral distance between soil
gas probe and groundwater monitoring well and between soil gas probe and UST
facility infrastructure (e.g., tanks, fuel dispensers) and buildings;
• Analytical data: Sampling date, analytical method, quality control data, and
chemistry data for soil, groundwater, and soil vapor. Analyte fields in the database are
fixed gases (oxygen, carbon dioxide, methane); benzene, toluene, and xylene; TPH;
27
-------
naphthalene; MtBE; 1,3-butadiene; hexane; heptane; 2,2,4-TMP,
1,2,4-trimethylbenzene (1MB); and aromatic and hydrocarbon fractions according to
analytical methods prescribed by the State of Massachusetts. Not all analyte data are
available for every site; and
• Building data: Building use (e.g., residential, commercial, institutional), foundation
type, and building size.
As described in Section 5.1.1, the data verification process included data quality indicators that
were developed and reviewed to ensure that data of known and acceptable quality were used in
the analysis.
5.3 EPA PVI Database Content
The number of sites in the EPA PVI database and their locations are listed in Table 3. The
contents of the database for key fields are summarized in Tables B-2 and B-3 in Appendix B.
The majority of the sites were UST release sites, although the database also includes data from
fuel terminals, petroleum refineries, and petrochemical sites. Most sites were affected by
gasoline releases, although a small number had other types of PHC contamination (e.g., diesel,
kerosene). The gasoline composition was unknown and assumed to be variable with respect to
fuel oxygenate composition, given the relatively broad time span for data collection (1995 to
2011). Gasoline containing ethanol (10 percent vol/vol) was generally introduced into the United
States in 2000, with a large increase in use in 2006 (U.S. EPA, 2009), so some sites in the EPA
PVI database where recent releases occurred probably had gasoline containing ethanol. (From
the available information, it is not possible to quantify the number of sites with ethanol.)
Table 3. Number of Sites by Country and States in the EPA PVI Database (November 2012)
Location
Sites
Location
Sites
United States
California
Maine
Maryland
Minnesota
New Jersey
North Dakota
7
13
1
22
3
1
Ohio
Oklahoma
South Carolina
Utah
U.S. unknown
4
1
1
15
1
Other Countries
Canada
4
Australia
1
Total Sites = 74
Sub-slab vapor samples were obtained at the 39 with buildings out of the 74 sites total in the
database. Almost all buildings in the EPA PVI database were residential houses or smaller
commercial buildings. (Table B-2 in Appendix B includes the building footprint area when
available.) Thus, the applicability of the database to large buildings may be limited. At a few
sites, soil gas samples were obtained from below and beside a building.
28
-------
The EPA PVI database is populated to varying degrees depending on the data type. (All statistics
with respect to number of sites with data are provided in Table B-3 in Appendix B). The
database contains data for most sites and records on facility type, vertical distances, surface
cover, soil type, and benzene and TPH vapor concentrations. Groundwater benzene and TPH
data are also available for many sites, and the database includes fixed gas data for a majority of
the sites. For other analytes (e.g., toluene, ethylbenzene, xylenes, naphthalene, MtBE), the EPA
PVI database includes data for fewer sites, although the dataset for aromatic and aliphatic
hydrocarbons is substantial due to recent data from 11 sites in Maine. Although indoor air data
are included for a few sites, this information was not used in the analysis because of the limited
number of indoor air data points and the known effect of background indoor (non-PVI) PHC
sources in overestimating soil vapor-to-indoor air attenuation factors for PHCs.
The database includes:
• 893 benzene soil vapor records;
• 655 oxygen soil vapor records; and
• 829 records with paired benzene soil vapor and groundwater data.
Additional information on each site in the EPA PVI database is provided in Tables B-2 and B-3
in Appendix B and in the database. Many of the original sources of data in the database (e.g.,
reports, journal articles, figures, data tables) are referenced and linked to a full set of electronic
document files organized by site.
Lead Scavengers. Ethylene dibromide (EDB) and 1,2-dichloroethane (1,2-DCA) are synthetic
organic chemicals that were historically used as gasoline additives to prevent lead deposits that
foul internal combustion engines. For this reason, they are commonly referred to as lead
scavengers. The EPA PVI database does not contain any soil gas data for these chemicals, but
Appendix F provides a comprehensive evaluation of their historical use, toxicity, transport, and
fate.
6. EPA PVI Database Analysis Approach and Methods
The data analysis began with an evaluation of whether the soil gas data at the site were obtained
in an area of LNAPL or dissolved-phase groundwater contamination. Given the importance of
the contamination source type on soil vapor concentrations, the analyses were conducted
separately for the LNAPL and dissolved sources. The data analysis consisted of three main parts:
• Exploratory data plots to identify data trends and relationships, discussed in Section
6.2.1;
• Estimation of vapor concentration attenuation distance using the vertical distance
method, developed by Lahvis et al. (In-prep.) and discussed in Section 6.2.2; and
• Estimation of non-contaminated vertical soil thickness needed for concentration
attenuation using the clean soil method, developed by Davis (2009) and discussed in
Section 6.2.3.
29
-------
6.1 Source Zone Identification (LNAPL versus Dissolved Indicators)
Several indicators were adopted for identifying whether the soil gas probe was located in an area
of LNAPL or dissolved-phase contamination (Table 4).
The primary indicator was direct evidence of LNAPL, such as a sheen or measurable
accumulations of product in a nearby monitoring well, borehole logs indicating a sheen or
significant hydrocarbon staining in soil, or a site investigation report indicating the soil gas probe
was installed in an LNAPL source zone. The direct indicators in Table 4 were the determining
factors for approximately 80 percent of the sites identified as having LNAPL contamination.
Table 4. Potential LNAPL Hydrocarbon Indicators
Type
Indicator
Measures and Screening Values
Adopted for this analysis
Direct
Indirect
Indirect
Current or historic presence of LNAPL in
groundwater or soil
Individual PHC compound and/or TPH
concentrations approaching (>0.2 times)
effective solubilities or effective soil
saturation concentrations (Csat concentration)
Proximity to source area likely to be impacted
with LNAPL
Laboratory and/or field observations, sheens,
results of paint filter, dye, and shake tests
Groundwater
- benzene > 5 mg/L
- TPH > 30 mg/L (gasoline)
Soil
- benzene > 10 mg/kg
- TPH > 250 mg/kg (gasoline)
Soil gas probes located near (within 20 ft [6.1 m])
or within former UST fields or fuel dispenser areas
Other potential indicators
Indirect
Indirect
Indirect
Indirect
Fluorescence response in LNAPL range
Organic vapor analyzer (e.g., photoionization
detector)
PHC vapor, O2 and CO2 profiles
Elevated aliphatic soil gas concentrations
UV, LIF, or UVIF fluorescence above background
levels (visual observation)
>500 ppmV
PHC vapor and CO2 concentrations in soil gas that
show no decrease (or O2 concentrations that show
no increase) or remain relatively constant with
distance from contamination source
For example, hexane soil gas concentrations more
than approximately 100,000 ug/m3 suggest LNAPL
because dissolved plumes are primarily composed
of soluble aromatic hydrocarbons (Lahvis et al., In
prep.)
Note: For two sites, #6-046 and #102 Chevron, there were long dissolved plumes (several hundred feet long) with
elevated benzene concentrations (up to 12 mg/L) in groundwater that exceeded the above criteria, but there was
no evidence for LNAPL at locations where the elevated benzene concentrations were measured. For these sites,
the above criteria were overridden (i.e., site was designated as a dissolved source).
Indirect or secondary indicators were:
• Groundwater Concentration Data. Benzene and/or TPH groundwater concentration
from which the presence of LNAPL near the soil gas probe was inferred. This was the
determining indicator for approximately 13 percent of the sites.
30
-------
• Soil Concentration Data. Benzene and/or TPH soil concentration from which the
presence of LNAPL near the soil gas probe was inferred. This was the determining
indicator for approximately 1 percent of the sites (one site).
• Proximity to Fuel Storage/Dispensing Facilities. Determined by soil gas probes
installed within 20 ft (6.1 m) of the tank field or dispenser. This was the determining
indicator for approximately 5 percent of the sites.
For sites with limited data, the secondary indicators were important to enable appropriate
classification with respect to LNAPL versus dissolved sources. However, a sensitivity analysis
showed that the exclusion distances were not sensitive to the benzene and TPH groundwater
concentration thresholds because direct indicators were the determining factor for 80 percent of
the site data (see Section 7.2.1).
6.1.1 Groundwater Concentration Data
Concentrations of chemicals that approach their effective solubility are indirect evidence for
LNAPL. For example, Bruce et al. (1991) suggest groundwater concentrations greater than the
effective solubility multiplied by 0.2 as possible evidence for LNAPL. For gasoline, when a
benzene mole fraction of 0.01 was assumed, the threshold was 3 mg/L, assuming a ratio of 0.2.
Given the uncertainty in these estimates, a slightly higher threshold for the benzene groundwater
concentration (5 mg/L) was adopted for identification of LNAPL sites. A TPH threshold
groundwater concentration of 30 mg/L was adopted based on the calculated approximate average
ratio of benzene to TPH groundwater concentrations in the database. An LNAPL source site was
identified based on either the benzene or TPH groundwater concentration exceeding the
threshold.
6.1.2 Soil Concentration Data
Concentrations of chemicals in soil that approach an estimated LNAPL saturation concentration
are indirect evidence for LNAPL. The soil saturation concentration is highly dependent on
chemical and soil properties. Concentrations representative of possible LNAPLs suggested in the
literature include a gasoline range organics (GRO) concentration greater than the range of 100 to
200 mg/kg and a diesel range organics (DRO) concentration greater than 10 to 50 mg/kg (e.g.,
ASTM, 2006; Alaska Department of Environmental Conservation, 2011).
The thresholds adopted for identifying LNAPL sites are a benzene soil concentration of
10 mg/kg and a TPH (gasoline) soil concentration of 250 mg/kg. The benzene concentration
(10.7 mg/kg rounded down to 10 mg/kg) was estimated from the equation for soil saturation
(Csat) and the default input parameters in Equation 9 of the EPA Soil Screening Guidance (U.S.
EPA, 1996, p. 28). A TPH soil concentration of 250 mg/kg was adopted to provide a slightly more
conservative screening basis (i.e., more sites are included as dissolved sites with higher
thresholds) than the ranges reported in the literature cited above. An LNAPL source site was
identified based on either the benzene or TPH soil concentration exceeding the threshold.
-------
6.1.3 Proximity to Fuel Storage/Dispensing Facilities
Soil gas probes located near or within former UST tank nests or fuel dispenser areas are
considered to have a high probability of being within LNAPL zones. Soil gas probes within 20 ft
(6.1 m) lateral distance of the tank nests or dispensers were categorized as being located within a
LNAPL source. Fifteen sites included data for soil gas probes that were within 20 ft of tank nests
or dispensers, but as indicated above, this criterion was the determining LNAPL indicator for
only 5 percent of the sites.
6.2 Data Analysis Methods
The data analysis consisted of an exploratory data analysis followed by the vertical distance and
clean soil methods for evaluating vapor attenuation. The analysis using the vertical distance
method focuses on benzene, given its importance for risk evaluations; however, the vertical
distance method was also performed for select other compounds, including those analyzed at the
Maine sites where full-spectrum hydrocarbon analyses (i.e., data on aliphatic and aromatic
fractions) were available. Given that the EPA PVI database was set up to allow for filtering, the
influence of site type (UST, fuel terminal, petroleum refinery, petrochemical), soil properties,
and surface cover (i.e., building, pavement, or ground cover ) was evaluated.
6.2.1 Exploratory Data Analysis
The exploratory data analysis evaluated the trends between groundwater and soil vapor
concentrations for different vapor constituents. Furthermore, the relationships between oxygen
and hydrocarbon concentrations and benzene and methane concentrations were assessed with
respect to what would be expected given the conceptual model for aerobic biodegradation (Davis
et al., 2009; Sweeney, 2012).
6.2.2 Vertical Distance Method
The vertical distance method involved plotting soil vapor concentration versus distance above a
source and estimating the probability for the soil vapor concentration to be less than a given
concentration threshold for different distances above the contamination source. The conditional
probabilities (P) were estimated as follows:
P (Cv < Qhreshoid/z > d, Contamination (z = 0) = LNAPL or dissolved)
where Cv is the soil vapor concentration, Qhreshoid is the soil vapor concentration threshold, z is
the vertical direction, d is the vertical distance from the top of the contamination to the soil gas
probe, and source contamination either is characterized as an LNAPL or a dissolved source.
First, the data were sorted in a cumulative distribution of specified vertical separation distances
from the source (e.g., > 0, > 2, ..., > n ft). The conditional probabilities were calculated for two
benzene vapor concentration thresholds (50 and 100 ng/m3) using two statistical techniques:
7 e.g., ground that is gravel-surfaced, grassy, or dirt-covered
32
-------
• Probability P = N [Cv < CWeshoidJ/N [total] where N [Cv < Qhreshoid] is the number of
benzene vapor concentrations less than the threshold and N [total] is the total number
of concentration measurements. For this analysis, concentrations below the reporting
limits were replaced with half the reporting limit, a common first approximation for
non-detect measurements, i.e., below specified reporting limits; and
• Probability was estimated from the concentration distribution calculated by the non-
parametric Kaplan-Meier method (Kaplan and Meier, 1958).
Approximately 31 percent and 57 percent of the paired data points for benzene vapor
concentration and distance associated with LNAPL sources and dissolved-phase sources,
respectively, were non-detects. The Kaplan-Meier method is a robust, non-parametric method for
considering data below reporting limits, particularly when there are multiple reporting limits
(Helsel, 2005; 2006). The Kaplan-Meier method, which has been shown in recent literature to be
the preferred method in many cases for estimating statistical parameters (e.g., mean, median,
standard deviation), does not rely on underlying assumptions about the data and can be used with
multiple reporting limits (Helsel, 2005).
Next, the depth to contamination was estimated. This quantity is important for an accurate
estimation of the thickness of clean soil needed for attenuation of soil vapor concentrations. For
LNAPL sites, this depth was estimated from boring logs and indications of LNAPL zones (e.g.,
observations of product, high headspace organic vapor concentrations above 500 to 1,000 parts
per million, and soil chemistry data); in the absence of data, the depth to contamination was
assumed to be the seasonal high water table where multiple monitoring events were available (on
average, there were about two events per site). For dissolved sites, the depth to contamination
was the depth to the water table closest to the time the soil gas data were obtained.
6.2.3 Clean Soil Method
The clean soil method (Davis, 2009; 2010) consists of an analysis of the thickness of un-
impacted clean soil (i.e., soil without NAPL) required for soil vapor benzene concentrations to
attenuate to below a defined threshold, which for this analysis is 100 ng/m3. This analysis
enabled comparison to published exclusion distances previously reported by Davis (2009; 2010)
that were based on this method. A clean soil thickness was calculated except when the vertical
distance between soil gas probes was greater than 10 ft (3.0 m), because in those cases there was
insufficient resolution (i.e., spacing between probes) for meaningful estimation. Two procedures
were used to estimate the clean soil thickness:
• Procedure 1: Distance to first soil gas probe with benzene Cvapor < 100 ng/m3 where:
— Lower depth = DI = Depth to top of contamination;
— Upper depth = Du = Depth to first probe with benzene Cvapor < 100 ng/rn3; and
— Distance = DI - Du.
• Procedure 2: Interpolated distance between a soil gas probe with benzene Cvapor
> 100 ng/rn3 and a soil gas probe with Cvapor < 100 ng/m3 where:
— Lower depth = DI = Depth to top of contamination;
33
-------
— Upper depth = Du = Interpolated as halfway between the depths to a probe with
benzene Cvapor < 40 ng/m3 and Cvapor > 100 ng/m3; however, if the lower
concentration is greater than 40 ng/m3, then Du = depth to first probe with
benzene Cvapor < 100 ng/m3; and
— Distance = DI -Du, subject to minimum thickness of 0.5 ft (0.15 m) because a
minimum distance is required for concentration attenuation to occur.
For both procedures 1 and 2, benzene concentrations below reporting limits were replaced in the
EPA PVI database with a concentration equal to half the method reporting limit for the sample.
(Replacement with half the detection limits is a common assumption.) An example of a clean soil
thickness calculation is presented in Figure 10. For the soil gas profile data shown, the
procedure 1 (left side of figure) distance is 15 ft (4.6 m) and the procedure 2 (right side of figure)
distance is 12.5 ft (3.8 m).
A lower concentration threshold was considered warranted for procedure 2 because of the
potential for the halfway distance interpolation to be non-conservative when the lower
concentration is much greater than 100 ng/m3 and the upper concentration is just less than
100 |-ig/m3. The 40 ng/m3 threshold is subjective, but when the upper benzene vapor
concentration is less than this threshold, the halfway interpolation procedure is more accurate.
Although more complicated and possibly more accurate interpolation rules could have been
developed (such as log-linear plots), procedure 2 was intended as a simple, approximate
technique. For any shallow concentration above 40 ug/m3, the procedure provides for a
conservative estimate of the attenuation distance because it uses the full distance to this vapor
sample location.
For locations where the measured soil vapor benzene concentration does not attenuate to less
than 100 |ag/m3, a clean soil thickness cannot be calculated, but a minimum clean soil thicknes:
is reported as the distance between the shallowest soil gas probe and the top of contamination.
34
-------
Clean Soil Thickness
Procedure 1: 25'- 10' = 15'
Feet Below
Grade
Vapor
Sample
Probe 1
First Depth with
Benzene Concentration
<100ng/m3
10
12.5'
20'
7'
25'
Clean Soil Thickness
Procedure2: 25'-{[15'+10']/2) = 12.5'
LEGEND
19 % - Oxygen cone. (%)
5 - Benzene vapor cone, (ng/mj
Linear Interpolation
Between These Two
Points
Figure 10. Example calculation of clean soil method distances.
35
-------
6.3 Soil Vapor Concentration Thresholds
This section describes the development of risk-based soil vapor concentration thresholds for
comparison with measured soil vapor concentrations to determine when PVI may be of potential
concern. The soil vapor concentration threshold was based on the expected attenuation of vapor
concentrations between shallow soil vapor and indoor air and the toxicity of the chemical under
consideration.
6.3.1 Sub-slab to Indoor Air Attenuation Factors
The processes that affect the vapor concentration in indoor air for a shallow soil vapor source are
primarily soil gas advection and ventilation that causes mixing of the chemical vapor within the
enclosed space (Hers et al., 2003; Johnson, 2005). Some additional biodegradation and sorption
could also occur between a shallow vapor source and an indoor environment. Several modeling
studies provide insight on the attenuation factor8 for a typical residential house. Yao et al. (2011)
reported attenuation factors between 2 x 10"4 and 7 x 10"3 for a numerical modeling study.
Johnson (2005) in a modeling study using the Johnson and Ettinger (1991) model calculated
attenuation factors between 4.4 x 10"3 and 7.3 x 10"3 for a shallow soil vapor source and
representative input parameters for a residential house. A similar study by Hers et al. (2003)
included a sensitivity analysis where the maximum attenuation factor for a range of conditions
was 9 x io~3, and a relatively good comparison (within an order of magnitude) was obtained
between Johnson and Ettinger model predictions and measured attenuation factors for
chlorinated solvent chemicals. Although this comparison is for chlorinated solvent chemicals, it
is relevant here because it identifies typical attenuation factors between shallow or sub-slab
vapor and indoor air, irrespective of possible biodegradation processes.
EPA has assembled a database of empirical slub-slab-to-indoor vapor attenuation factors (U.S.
EPA, 2012b). Although most of the data in this EPA database is for chlorinated solvents, the
sub-slab attenuation factors can be applied to PHCs because little bioattenuation is expected
between sub-slab and indoor air. Sub-slab attenuation factors in this U.S. EPA (2012b) database
vary over several orders of magnitude because of spatial and temporal variability in both indoor
air and sub-slab vapor concentrations and background sources of chemicals in indoor air. In the
U.S. EPA (2012b) database, the 50th and 95th percentiles of the sub-slab attenuation factor are
5.0 x 10"3 and 1.8 x 10"1, respectively, when the data are limited to indoor air concentrations
above a 90th percentile background concentration found in the literature. When the data are
limited to sub-slab vapor concentrations greater than 100 times the literature background, the
50th and 95th percentiles of the sub-slab attenuation factor are 2.5 x 10"3 and 2.0 x 10"2
respectively (U.S. EPA, 2012b). Therefore, a shallow soil vapor-to-indoor air attenuation factor
of 0.01 was considered a reasonably conservative attenuation factor.
6.3.2 Risk-based Concentration Thresholds
Risk-based indoor air concentrations for a residential scenario are provided in Table 5 for the
chemicals of potential concern considered in this analysis. The risk-based indoor air
concentrations assume a residential scenario, continuous lifetime exposure to vapors, and no
The attenuation factor is defined as the concentration of a chemical in indoor air divided by its concentration in soil
gas under a foundation slab (sub-slab attenuation factor) or deeper in the soil beneath the house.
36
-------
exposure amortization (e.g., calculation of average exposure rate from a less than continuous or
lifetime exposure).
For chemicals other than benzene, a risk-based soil vapor concentration (RBCV) was calculated
as 100 times the risk-based air concentration. The thresholds adopted for benzene, 50 and
100 |-ig/m3, were based primarily on practical considerations relating to the detection limit (i.e.,
the frequency of non-detects increases as the benzene concentration decreases) and for
consistency with previous database evaluations by Davis (2009) and Lahvis et al. (In prep.). For
comparison, assuming a generic sub-slab attenuation factor of 0.01, the calculated benzene
threshold is 29 |ag/m3 for an incremental lifetime cancer risk (ILCR) of 1 x 10"6, and 290 |-ig/m3
for an ILCR of 1 x 10"5. Given the uncertainty in the attenuation factor and benzene toxicity, the
adopted threshold of 50 ng/m3 is not considered substantively different from the 1 x 10"6 ILCR
threshold of 29 ng/m3shown in Table 5.
Table 5. Risk-based Indoor Air Concentration for Primary Chemicals of Potential Concern
Chemical
Benzene
Toluene
Ethylbenzene
Xylenes
Naphthalene
n-Hexane
MADEP Aliphatic C5-8
MADEP Aliphatic C9-1 8
MADEP Aromatic C9-18
Toxicity Endpoint
Carcinogenic
Non-carcinogenic
Non-carcinogenic
Non-carcinogenic
Non-carcinogenic
Non-carcinogenic
Non-carcinogenic
Non-carcinogenic
Non-carcinogenic
Risk-Based Indoor Air
Concentration (ng/m3)3' b
2.9(1 x 1fj5ILCR);
0.29(1 x irj6ILCR)
5,000 (RfC)
1,000(RfC)
100 (RfC)
3 (RfC)
700 (RfC)
200
200
50
Source
EPA IRIS
EPA IRIS
EPA IRIS
EPA IRIS
EPA IRIS
EPA IRIS
MADEP (2003)
MADEP (2003)
MADEP (2003)
3 The risk-based indoor air concentration assumes a residential receptor and continuous exposure over a lifetime.
b The benzene risk-based air concentration is based on the midpoint of the toxicity factor range provided in the EPA
IRIS (Integrated Risk Information System) database.
ILCR = Incremental lifetime cancer risk; RfC = reference concentration.
EPA IRIS database accessed February 2012.
MADEP = Massachusetts Department of Environmental Protection TPH method
7. EPA PVI Database Analysis Results
7.1 Exploratory Data Analysis
7.1.1 Comparison of Groundwater and Soil Vapor Concentrations
The cumulative distributions of benzene concentrations in groundwater are plotted in Figure 11
for hydrocarbon sources classified as dissolved phase and LNAPL. The groundwater benzene
concentrations for LNAPL sites are higher than for dissolved sites, but the difference in the
distribution between LNAPL and dissolved sites is smaller than expected. This may be due to
spatial variability in groundwater concentrations, highly weathered residual-phase LNAPL that is
37
-------
relatively depleted of benzene, variable well screen intervals relative to the water table, and
vadose zone LNAPL sources.
Benzene Groundwater Concentrations
100,000
• LNAPL (all)(N=198)
•Dissolved(N=98)
0%
20% 40% 60%
Cumulative Frequency
80%
100%
Figure 11.
Groundwater concentrations measured near soil vapor sampling locations for
dissolved and LNAPL source zones (all refers to UST, fuel terminal, refinery, and
petrochemical sites).
The relations between benzene concentrations in groundwater and deep soil vapor probes (within
3 ft [0.9 m] of the contamination source) for dissolved and LNAPL sources are shown in
Figure 12. There is no apparent correlation for dissolved source data based on visual
observation. The reason for the lack of correlation for dissolved source data and relatively
consistent, low soil vapor concentrations is consistent with attenuation by biodegradation.
For dissolved-source data, the measured deep benzene vapor concentrations are, in almost all
cases, at least an order of magnitude and, in many cases two orders of magnitude, less than the
predicted soil vapor concentration based on Henry's Law partitioning and the measured
groundwater concentration. This is reasonably attributed to biodegradation. A dimensionless
Henry's Law constant of 0.14 was used for benzene, which is based on a groundwater
temperature of 15 C, a representative value for the United States based on the groundwater
temperature map in U.S. EPA (2004).
38
-------
Benzene - GW vs SV - Dissolved (IM=44)
J
o 1 E+03
o, i E-t-02
^ 1 E+01 !
#
l.E+00
l.E-01
,„'*
„ ** ^-^
^ •" _ -* '
^ f •
** ^ •
-* ^ "
* *
t.'-' t «• % * *
* * * •* \ «« **
* s* t*
* *
1 10 100 1,000 10,000 100,000
Groundwater Cone. (nfi/L)
Benzene - GW vs SV - NAPL (all) (N=63)
1 Ei 07
u
O
5 l.E+01
*-.'
* ,*?«**
^ «*l *Z
Henry's Law --^ ^ -
prediction,^-5*-
•--•- . **
:•« * .
»
* * * *
* * * *
•
1 10 100 1,000 10,000 100,000
Groundwater Cone. (ng/L)
Figure 12. Groundwater concentrations measured near soil vapor sampling locations for
dissolved and LNAPL source zones (all refers to LIST, fuel terminal, refinery, and
petrochemical sites). Only detectable benzene vapor concentrations shown.
A very weak proportional relation based on visual observation (R2 = 0.08) exists between
groundwater and soil vapor concentrations for LNAPL source data. For the LNAPL source data,
the measured benzene vapor concentrations are significantly less than predicted for benzene
groundwater concentrations less than approximately 1 mg/L. For concentrations greater than
1 mg/L, the benzene vapor concentrations for some data points are close (within a factor of 2 or
3) to the vapor concentrations predicted by Henry's Law multiplied by the groundwater
concentration, which is consistent with an LNAPL source and limited attenuation between the
source and deep soil vapor sample.
The poor correlation between groundwater and soil vapor concentrations may be due to several
factors including variable well screen intervals relative to the water table, variable
biodegradation between the groundwater and lowermost soil gas sampling location, spatial
variability and differences in dissolved-phase concentrations at groundwater and soil gas
sampling locations, and sampling errors.
7.1.2 TPH Vapor versus Oxygen Concentrations
The relation between co-located oxygen and TPH vapor concentrations was evaluated to provide
insight on biodegradation processes and a possible lower oxygen limit for occurrence of aerobic
biodegradation (Davis et al., 2009; Sweeney, 2012). The expectation is lower oxygen
concentrations when there are higher hydrocarbon vapor concentrations and higher oxygen (near
atmospheric) concentrations at lower hydrocarbon vapor concentrations (Figure 1). An
exception to this trend is where there is high natural oxygen demand (e.g., peat sites). At such
sites, low concentrations for both PHC vapors and oxygen would be expected. (There are few
such sites in the database.) Although a more robust paradigm for interpretation of oxygen and
hydrocarbon data would be based on mass fluxes, this requires evaluation of site-specific vertical
concentration profiles and soil property data which are more detailed than the data available for
many sites in the EPA PVI database.
A plot of co-located oxygen versus TPH vapor concentration data is a useful approximate
indicator of trends, and the results shown on Figure 13 are generally consistent with expected
39
-------
behavior. For dissolved sources, there are no data indicating oxygen depletion, which is
consistent with expected low oxygen demand associated with the vapor flux from dissolved
sources. The data in the EPA PVI database are consistent in this regard; there are no data
indicating depleted oxygen in association with dissolved sources. For LNAPL sources, relatively
low oxygen concentrations (defined here as less than 4 percent) occur when TPH vapor
concentrations exceed approximately 1 x 106 |ig/m3 (1 mg/L), but both oxygen and TPH vapor
concentrations are elevated for a few data points. To provide context for interpretation of this
data, it is important to recognize that:
• Relatively high TPH vapor concentrations are required before the oxygen depletion is
resolvable based on the stoichiometric demand associated with TPH oxidation.
Ambient oxygen (21 percent vol/vol or 280 g/m3) has a measured resolution of about
2 percent vol/vol. With a 0.3 g-hydrocarbon (HC)/g-O2 consumption ratio based on
stoichiometric considerations, this suggests a hydrocarbon level at which significant
oxygen consumption should be resolvable of about 280 x (2/21) x 0.3 = 8 g/m3 or
8 x 106 |J,g/m3 (defined as the sensitivity threshold on Figure 13).
• Active soil gas sampling results in volume averaging of soil gas because of the
dimension of the soil gas probe and sand pack, which often ranges between 0.15 and
0.30 m. Volume averaging can result in elevated TPH vapor and oxygen
concentrations.
• Soil gas samples that are obtained either wholly within or that straddle the
biodegradation zone may have moderately elevated concentrations of both TPH
vapors and oxygen (5 to 10 percent). There are several case studies with detailed soil
gas profiles that demonstrate this behavior (Hers et al., 2000; Fischer et al., 1996;
Ririe et al., 2002).
• Some leakage of oxygen can occur through the process of sampling or analysis;
however, this is not considered to have caused a significant bias for the reasons
described below.
The upper right quadrant of Figure 13, where TPH vapor concentrations exceed the sensitivity
threshold and oxygen concentrations exceed 4 percent, contains only 26 data points (6 percent of
the data). Upon closer examination, many of these data were from Hal's site in Utah (Figure 13).
Approximately half of the Hal's site data from this quadrant were determined to be high quality
based on internally consistent oxygen versus TPH vertical profile data. The other half were
generally deep soil gas probes where oxygen was elevated, suggesting possible leakage (perhaps
due to a rise in the water table). Overall, the number of data points with possible concerns
relating to leakage was very small. In the case of Hal's site, the data with possible leakage were
for small exclusion distances; thus, they have no bearing on the overall conclusions with respect
to distances in this report.
40
-------
l.E»07
Oxygen vs. TPH - Dissolved (N=199)
l.E-06 :
l.ttOb
1 £*M
l.C>03
| l.t»02 ;
l.E'Oi ,
LCtOO
-*^-
-*-*
i
1ft IS 20
Oxygen Cone. C*, vol/vol)
!_£*»
Oxygen vs. TPH - NAPL (all) (N=442)
1£»00
10 IS 20
Oxyn»n Cone. (% uo(/»ol)
2-OEtOS
_16E+08
71.2E+08
5
S.8.0E+07
I
I
' 4.0E+07 , ,
O.OE*00
Oxygen vs. TPH - NAPL (all) (N=442)
—T
4% threshold
Hal's
A^.^»
A
2.0CI08
LGE*08
~ L2E*38
i
a &OE-K)7
Oxygen vs. TPH- NAPL (all) (N=442)
4.0E*07
aoe*DO
\
* ,
4% threshold
»
» *"
* K »
Sensitivity
thr
^.
•* -
10 IS 20 25
Oxygpn Cone. (%vol/uol)
10 15 20 25
Oxygen Cone (% vol/vol)
M
Figure 13. TPH vapor versus oxygen concentrations for dissolved and LNAPL source zones (all
refers to LIST, fuel terminal, refinery, and petrochemical sites). Data points shown are
where both TPH vapor and oxygen concentrations were above detection limits. Shaded
areas and ellipse encompass data that generally support the aerobic mineralization
paradigm.
7.1.3 Methane Concentrations
The EPA PVI database includes methane data for 27 sites. Methane concentrations exceeded
5 percent (lower explosive limit in air) at five sites (three UST sites and two petroleum refinery
sites) but were less than 1 percent at the remaining 22 sites. Three of the five sites with elevated
methane concentrations were investigated prior to 2000, so the methane at these sites was
probably not associated with ethanol in the gasoline leaked into the subsurface.
A plot of methane versus distance between the LNAPL source and soil gas probe indicates
methane concentrations at sites with profile data decreased to below 5 percent within
approximately 10 ft of the source (Figure 14a). The relationship between methane and benzene
vapor concentrations was evaluated because of concern that elevated methane concentrations
may result in reduced benzene vapor attenuation because of the oxygen demand represented by
methane. As a general relationship, higher benzene concentrations are expected when methane
concentrations are elevated because of the oxygen demand represented by methane oxidation
(Jewell and Wilson, 2011). There was no apparent correlation between methane and benzene
41
-------
vapor concentrations, possibly because of the limited number of sites with elevated methane
concentrations (Figure 14b). The data points with co-located elevated benzene vapor (greater
than 1,000 ng/m3) and methane (greater than 1 percent) were found to have distances between
the benzene contamination and the soil gas probe of less than 7.5 ft. The relatively small
distances do not suggest enhanced benzene vapor transport because of methane.
a
l.E+08
~ l.E+07
c
0. l.E+02
c 1 E+01
| 1-E+0°
l.E-01
(
Methane versus Benzene Vapor - NAPL (all)
u*
•
'. «
* * *
>
•
) 20 40 60 80 100
Methane (%)
b Methane versus Distance - NAPL (all)
90
80
70
X 60
g 50
0 I
l.Ei
•
• *c
(ttfirTrftrVrnfr « <* + *
00 l.E+01 2.E+01 3.E+01 4.E+01 5.E+01 6.E+01
Distance between Soil Vapor Probe and Contamination (ft)
Figure 14. Methane concentrations versus distance and benzene vapor concentrations (all refers
to LIST, fuel terminal, refinery, and petrochemical sites).
7.1.4 Comparison between Benzene and TEX Vapor Concentrations
As an initial screening step, benzene concentrations were compared with toluene, ethylbenzene,
and xylenes (TEX) concentrations to evaluate whether benzene is a risk driver relative to the
TEX compounds. The relationships between benzene and ethylbenzene and between benzene
and total xylenes concentrations in soil gas are shown in Figure 15. Qualitatively, there was a
relatively good correlation between these two analyte pairs, although the comparison of benzene
to xylenes indicates that concentrations of xylenes were generally up to two orders of magnitude
higher than benzene concentrations. The comparison indicated that further analysis of xylenes
was warranted and that benzene was a reasonable surrogate for toluene and ethylbenzene.
Benzene versus Ethylbenzene (>DL) - All Data
Benzene versus Xylenes (>DL) -All Data
l.E-01
l.E-01 l.E+00 l.E+01 l.E+02 l.E+03 l.E+04 l.E+05 l.E+06
Benzene Vapor (ng/ms)
l.E-01 l.E+00 l.E+01 l.E+02 l.E+03 l.E+04 l.E+05 l.E+06
Benzene Vapor (ng/m!)
Figure 15. Relationship between benzene and ethylbenzene (left) and benzene and xylenes (right)
vapor concentrations. (All refers to LIST, fuel terminal, refinery, and petrochemical
sites.) Data points are shown where both compounds were above detection limits.
42
-------
7.2 Vertical Distance Method
For the vertical distance method, the soil vapor concentrations of the PHC compound assessed
(excluding benzene) were compared with the risk-based vapor concentration (RBCV), which is
the risk-based indoor air concentration multiplied by a dilution factor of 100 (i.e., attenuation
factor of 0.01). For benzene, the probability that the soil vapor concentration is less than a
defined threshold (50 and 100 ng/m3) for varying source-separation distances was estimated.9
For definition of inclusion distances, a probability greater than 95 percent was considered a
reasonable threshold based on regulatory precedence.
7.2.1 All Data
Data analysis was conducted for dissolved sources, LNAPL sources at UST sites, and LNAPL
sources at fuel terminal, petroleum refinery, and petrochemical (non-UST) sites (Figures 16
through 20). There are only PHC fraction and hexane data for UST sites; hence, there is not a
figure for the all-LNAPL sites category for these compounds. The following compounds were
evaluated: benzene, xylenes, hexane, 2,2,4-TMP, 1,2,4-TMB, naphthalene, MtBE, and
Massachusetts Department of Environmental Protection (MassDEP) hydrocarbon fractions (C9-
10 aromatics, C5-8 aliphatics, and C9-12 aliphatics).
For the benzene analysis, non-detects were addressed both by the common practice of
substituting half the detection limit and by using the more statistically robust Kaplan-Meier
method. The Kaplan-Meier method resulted in similar or slightly (0 to 7 percent) higher
probabilities than the substitution method. The probability of the soil vapor concentration being
less than a threshold was estimated for 5 and 30 mg/L groundwater concentration thresholds for
benzene and TPH, respectively, which are part of the LNAPL hydrocarbon indicators (presented
previously in Table 4). A sensitivity analysis was conducted where the threshold for the benzene
groundwater concentration was varied between 1 and 10 mg/L and the TPH groundwater
concentration was varied between 10 and 50 mg/L. The resulting variation in probability
(estimated using the substitution technique) for this range was less than 0.5 percent, indicating
the results are not sensitive to the concentration thresholds (possibly because groundwater is a
poor predictor of soil vapor concentrations).
9 Negative contamination source-probe distances indicate the soil gas probe was installed below the top of the
contamination zone.
43
-------
Benzene vs. Distance - Dissolved
Distance between Soil Vapor Probe and Contamination (ft)
100
Benzene Conditional Probability- Dissolved
86
84
82
80
-Probability<100(l/2DL)
-Probability<50(l/2DL)
Probability<100(KM)
-Probability < 50 (KM)
1 2 3 4 5
Distance between Soil Vapor Probe and Contamination (ft)
Oxygen vs. Distance- Dissolved
Xylenes vs. Distance - Dissolved
l.E-01
-10 0 10 20 30 40
Distance between Soil Vapor Probe and Contamination (ft)
-10 0 10 20 30 40
Distance between Soil Vapor Probe and Contamination (ft)
Figure 16. Vertical distance method: benzene (a), oxygen (c), and xylenes (d) data for dissolved-
source sites (KM = Kaplan-Meier). Panel b shows the benzene probability data.
44
-------
1 a C9-C12 Aliphatic vs. Distance- Dissolved
C9-C12 Aliphatic Vapor Cone, ((ig/m3)
l-» I-" h-1 M I-1 t-
+ + + + + -
S 2 S S g £
0 5 ft RBCV
^
A
• >DL (N=26)
D DL(N=32)
DDL(N-32)
» D
-------
Benzene vs. Distance - NAPL (UST only)
-10 O 10 2O 30 10 SO
Distance between Soil Vapor Probe and Contamination (ft)
Benzene Conditional Probability - NAPL (UST only)
P rob ability < 100 (1/2DL)
P rob ability < 50 (1/2 DL)
P rob ability < 100 (KM)
P rob ability < 50 (KM)
5 10 15 20 25 30
Distance between Soil Vapor Probe and Contamination (ft)
Oxygen vs. Distance- NAPL (UST Only)
Xylenes vs. Distance - NAPL (UST Only)
-10 0 10 20 30 4O 50
Distance between Soil Vapor Probe and Contamination (ft)
-10 0 10 20 30
Distance between Soil Vapor Probe and Contamination (ft)
Figure 18. Vertical distance method: benzene (a, b), oxygen (c), and xylenes (d) data for LNAPL
sources at UST sites (KM = Kaplan-Meier).
46
-------
a
i
u
O
0.
s
u
0.
H
3
C9-12 Aliphatic vs. Distance- NAPL (UST only) b C5-C8 Aliphatic vs. Distance - NAPL (UST only)
+ 15ft
*
«>DL (N=35)
D DL (N=28)
4-
00 10 20 30 40 50
Distance between Soil Vapor Probe and Contamination (ft)
u
•c
{2
!
u
l.E+00
-1
^ 15ft
?
»>DL (N=52)
n + ^^ ^
00 10 20 30 40 50
Distance between Soil Vapor Probe and Contamination (ft)
1 d 1 Hexanevs. Distance- NAPL (UST only)
-- 1 E+05 -
O
a.
(U
l.E-01
-1
* * 15ft
^*
.... —•- - -
J *
** «
*
. n m .
^.cad*3 "ff
n
RBCv
« > DL (N=40)
D
-------
a
•^2 1 E+07 |
ro
re
">•
3
-i
c
1 E+05
t
5 . p n.
1
i
1
r-J
rt
-1
2,2,4 TMP vs. Distance - NAPL (UST only)
\ 0 DL(N=13)
» \
*5*
n
00 10 20 30 40 50 6
Distance between Soil Vapor Probe and Contamination (ft)
1,2,4-TMB vs. Distance - NAPL (UST only)
+ ^ «>DL (N=13)
•
D^ Q < DL (N=25)
D &•
D
*
n n n \
n o D D
n an
LTD n
00 10 20 3
Distance between Soil Vapor Probe and Contamination (ft)
0
0
b
f
U
s
1 E+01
-1
d
1 E+05
d
c
1
5
1
Z
-1
MtBE vs. Distance - NAPL (UST only)
A
DDL(N=21)
*
n
n
n
» *
*nirf n n a
— H — — ^N —
n cj] rrg«B^] an
n mi 111 n 1 1 HI H^ n rm
00 10 20 3
Distance between Soil Vapor Probe and Contamination (ft)
Naphthalene vs. Distance - NAPL (UST only)
11 ,.
a D CD
n n *>DL (N=14)
u
. Pn
DD a n
- D^Di^D
D 8: n o> an n
ii mi in nti i i iiiiiii ^ n nti
00 10 20 3
Distance between Soil Vapor Probe and Contamination (ft)
0
0
Figure 20. Vertical distance method: 2,2,4-trimethylpentane (TMP), MtBE, 1,2,4-trimethylbenzene
(TMB), and naphthalene data for LNAPL sources at UST sites.
48
-------
Table 6. Summary of Results for Vertical Distance Method
Dissolved Source
LNAPL Source—UST Sites
LNAPL Source — Refinery,
Fuel Terminal, Petrochemical
Sites
Oxygen
Benzene
(100|jg/m3
threshold)
Benzene
(50 |jg/m3
threshold)
Xylenes
Hexane
C5-8 Aliphatic
C9-12
Aliphatic
C9-10
Aromatic
PKM = Probability
Most 62 cone. > 4%, and no
O2<1%
PKM > 97% for 0 ft separation
increasing to 99% at 5 ft (1 .5
m)
PKM > 94% to 95% for 0 ft to
5 ft (1.5m)
One vapor concentration >
RBCV for separation distance
of 3 ft (0.9m)
All vapor concentrations <
RBCV 0 ft
Two vapor concentrations >
RBCV for separation distance
up to 3 ft (0.9 m)
All vapor concentrations <
RBCv 0 ft
All vapor concentrations <
RBCv 0 ft
estimated using Kaplan-Meier
Many data points with O2 <
4%, andO2< 1% to 6 ft (1.8
m) separation
PKM > 61% for 0 ft separation
increasing to ~ 95% at 15 ft
(4.6 m)
PKM > 57% for 0 ft separation
increasing to ~ 93% at 15 ft
(4.6 m)
>10 vapor concentrations >
RBCv for separation distance
up to 11 ft (3. 4m)
Five vapor concentrations >
RBCv for separation distance
up to 4 ft (1.2m)
Five vapor concentrations >
RBCV for separation distance
up to 3 ft (0.9 m)
Eight vapor concentrations >
RBCV for separation distance
up to 2 ft (0.6 m)
Four vapor concentrations >
RBCV for separation distance
up to 2 ft (0.6 m)
method for exceeding threshold.
Many data points with O2 < 4%,
and O2 < 1 % to 1 1 -ft (3.4-m)
separation, greater O2
depletion than UST only
P-IGDL > 22% for 0-ft separation
increasing to ~ 90% at 18 ft
(5.5 m)
P-IGDL > 22% for 0 ft separation
increasing to ~ 90% at 18 ft
(5.5 m)
>10 vapor concentrations >
RBCv for separation distance
up to 12 ft (3. 7m)
N/A
N/A
N/A
N/A
Pi/2Di_ = Probability estimated using
• For LNAPL sources at UST sites, the findings were:
— Approximately 95 percent of the benzene soil vapor concentrations are less than
100 |-ig/m3 and 93 percent are less than 50 ng/m3 at a source-separation distance
of approximately 15 ft (4.6 m) (Table 6 and Figures 18 to 20).
- For all compounds, soil vapor concentrations decrease rapidly between about 5
and 15 ft (1.5 m and 4.6m).
- For compounds other than benzene, the available data indicate that measured soil
vapor concentrations are less than the risk-based concentrations beyond 11 ft
(3.4m).
- There are significant differences in source soil vapor concentrations at small
source-probe separation distances. The approximate maximum soil vapor
concentrations were 3xl07 ug/m3 for 2,2,4 TMP; 7xl07 ug/m3 for benzene,
hexane, and MtBE; 20,000 ug/m3 for 1,2,4 TMB; and 180 ug/m3 for naphthalene.
— The elevated MtBE vapor concentrations are mostly for one site (Stafford) with
residual LNAPL above the water table, and there was evidence for a complete
MtBE vapor intrusion pathway at a house with a 6.5-ft LNAPL-basement
separation.
49
-------
• For LNAPL sources at refinery, fuel terminal, or petrochemical (non-UST) sites, the
findings were:
— There are limited data for benzene beyond a 20-ft source-separation distance and
limited data at all distances for other compounds.
— Approximately 90 percent of the benzene soil vapor concentrations are less than
the thresholds (50 and 100 ug/m3) at a source-separation distance of
approximately 18 ft (5.5 m). The probability does not substantially increase above
90 percent beyond 18 ft because data are limited for larger separation distances.
The benzene vapor concentration versus distance plot indicates benzene vapor
concentrations generally decrease by orders of magnitude between 10 and 30 ft
(Table 6 and Figures 21 and 22)
— For compounds other than benzene, the available data indicate that measured soil
vapor concentrations are less than the risk-based concentrations beyond 12 ft
(3.6m).
- At the Mandan site, a very large diesel release (over 1 million gallons) led to
elevated deep, near-source methane concentrations10 (Breyer and Cowart, 2004).
The least attenuation between deep and shallow soil gas (8-ft distance) was
observed for methane and naphthalene, for which the ratios of deep to shallow
concentrations were 16 and 19, respectively. The greatest attenuation was
observed for benzene where the ratio was about 260. Also, deep naphthalene
vapor concentrations were up to 20,000 ug/m3, approximately two orders of
magnitude greater than the maximum naphthalene concentration measured at
other sites in the database.
— At a refinery with a very large gasoline release (over 1 million gallons) and deep
LNAPL source (about 55 ft [16.4 m] depth), the 2,2,4-TMP concentration
remained elevated for separation distances as great as 45 ft (15 m).11
The analysis results indicate benzene is the risk driver, with exceedances of the risk-based vapor
concentrations occurring at larger contamination source-building separation distances compared
to other compounds evaluated. There is less attenuation for 2,2,4-TMP compared to benzene, but
2,2,4-TMP is not a risk driver because there is no U.S.EPA toxicity factor for it.
10 There were no oxygen data in the information reviewed.
11 The reason for the 2,2,4-TMP trend at this site is not known, but the available data indicate significant temporal
fluctuations in PHC vapor and oxygen concentrations that may be related to water table fluctuations and operation of
a remediation system (ASTDR, 2011). In addition, Mickelski et al. (2010) in a review of data from this site suggests
there may also be near surface-contamination sources, although the significance of such possible shallow sources on
the 2,2,4-TMP concentrations is not well understood.
50
-------
a
Benzene Vapor Cone.
c
u
c
l.E+Oi
l.E+cr
l.E+Of
l.E+0!
l.E+CB
I.E+O:
I.E+O;
l.E+0.
l.E+OC
I.E-O:
25
10 -
0 -
-1
i 1
)
L •
-1
Benzene vs. Distance - NAPL (non-UST)
* ' * 15ft «>DL(N=196)
*\ t D
i "*
5
l.E+C
l.E+C
Benzene Conditional Probability - NAPL (non-UST)
/^-*-
— —
~^
/^~ Probability < 100 and
^^^ < 50 are identical
J*
_S
r-
— •— Probabili
_ .
ty<100{l/2DL)
ty < 50 (1/2DL)
II
) 5 10 15 20 25 30
Distance between Soil Vapor Probe and Contamination (ft)
Xylenes vs. Distance - NAPL (non-UST)
3 -
0 -
-]
15ft
* $ *
• >DL (N=ll)
* 1
RBCv
D UP
00 10 20 30 40
Distance between Soil Vapor Probe and Contamination (ft)
Figure 21. Vertical distance method-benzene, xylenes, and oxygen data for LNAPL sources at
fuel terminal, refinery, and petrochemical (non-UST) sites. Red plots over blue.
51
-------
2,2,4 IMP vs. Distance - NAPL (non-UST)
l.E+06 _
l.E+05 i
l.E+04 i
l.E+03 i
l.E+02 i
l.E+01 -
l.E+00
-10
10
20
30
40
50
Distance between Soil Vapor Probe and Contamination (ft)
b 1,2,4-TMB vs. Distance - NAPL (non-UST)
l.E+06
~
I i-
°1E+04
0)
G
~ l.E+03
X
I l.E+02
4 l.E+01
l.E+00
c Naphthalene vs. Distance - NAPL (non-UST)
l.E+05
l.E+04
-10 0 10 20 30
Distance between Soil Vapor Probe and Contamination (ft)
3 l.E+
03
l.E+02
' l.E+01
D< DL (N=0)
» > DL (N=4)
RBCV
-10 0 10 20 30
Distance between Soil Vapor Probe and Contamination (ft)
Figure 22. Vertical distance method-2,2,4-trimethylpentane (TMP), naphthalene, and 1,2,4-
trimethylbenzene (TMB) data for LNAPL sources at fuel terminal, refinery, and
petrochemical (non-UST) sites.
7.2.2 Influence of Surface Cover
The potential influence of a surface cover was evaluated through comparison of the probabilities
of benzene soil vapor concentrations that are less than 100 ug/m3 for varying soil-separation
distances and through analysis of oxygen concentrations for three different surface covers:
(1) building concrete foundation, (2) pavement, and (3) ground cover12 (Figures 23 and 24). The
datasets evaluated were limited to LNAPL sources because oxygen is not limiting for dissolved-
source sites.
For LNAPL sources at UST sites, the probabilities of benzene soil vapor concentrations less than
100 ng/m3 were highest for the building scenario, second highest for the ground-cover scenario,
and lowest for the pavement scenario (Figure 23). Oxygen data are another important indicator
of the possible effect of surface cover on aerobic biodegradation. The oxygen concentration
trends were qualitatively similar for the pavement and ground-cover scenarios and indicated low
(less than 2 percent) oxygen concentrations were limited to small separation distances (less than
10 ft distance). Oxygen concentrations for the below-building scenario were higher than the
' e.g., ground that is gravel-surfaced, grassy, or dirt-covered
52
-------
pavement and ground cover scenarios for similar distances, but the dataset was relatively small
(N=35).
Benzene Probability Different Source Cover - NAPL (UST)
100
•Probability < 100 u.g/m3 - Ground Cover Scenario
•Probability < 100 u.g/m3 - Pavement Scenario
Probability < 100 u.g/m3 - Building Scenario
40
5 10 15 20 25
Distance between Soil Vapor Probe and Contamination (ft)
30
Oxygen for Different Surface Cover - NAPL (UST)
*O2- Building (N=35)
• 02-Ground (N=66)
02- Pavement (N=193)
1
-10
0
10
20
30
40
50
Distance between Soil Vapor Probe and Contamination (ft)
Figure 23. Comparison of probability for benzene soil vapor concentrations to be less than
threshold and oxygen concentrations for different surface covers for LNAPL sources
at UST sites. Below detection limit concentrations replaced with half the detection limit
for analysis.
53
-------
For LNAPL sources at fuel terminal, refinery, and petrochemical sites, the dataset is relatively
small and the benzene probability analysis is variable. Therefore, it is not possible to infer
potential differences between the scenarios (Figure 24). The oxygen results indicate the lowest
concentrations for the building scenario, with seven data points with oxygen concentrations that
were below 1 percent between 3 and 11 ft distance. Qualitatively, the oxygen concentrations for
the pavement scenario were similar but somewhat higher than for the building scenario, and the
concentrations for the ground-cover scenario were significantly higher than for the building
scenario.
54
-------
Benzene Probability Different Surface Cover - NAPL (all)
Probability < 100 |ig/rrr - Building Scenario
Probability < 100 ng/m3 - Ground Cover Scenario
Probability < 100 ug/m3 - Pavement Scenario
0 5 10 15 20 25 30
Distance between Soil Vapor Probe and Contamination (ft)
35
o
"o
u
I
o
M
z
o
Oxygen for Different Surface Cover - NAPL (all)
10
+ 02- Building (N=56)
«02-Ground(N=20)
02-Pavement (N=60)
*'
-10
0 10 20 30 40
Distance between Soil Vapor Probe and Contamination (ft)
50
Figure 24. Comparison of probability for benzene soil vapor concentrations to be less than
threshold and oxygen concentrations for different surface covers for LNAPL sources
at all sites (LIST, fuel terminal, refinery, and petrochemical). Below detection limit
concentrations replaced with half the detection limit for analysis.
55
-------
In summary, the type of surface cover did not appear to affect the probability analysis for
benzene concentration thresholds (e.g., it did not indicate a consistently lower probability for
data obtained below buildings). For UST sites, oxygen concentrations were lower at paved sites
but not below buildings. For fuel terminal, refinery, and petrochemical sites, there was greater
frequency of depleted oxygen for the below-building scenario, compared to the ground cover and
pavement scenarios, which may be a result of larger petroleum releases at such sites and
consequent greater oxygen demand, compared to typical releases at UST sites.
7.2.3 Influence of Soil Type
The potential influence of soil type was evaluated through comparison of the probabilities of
benzene soil vapor concentrations less than 100 |ag/m3 for varying soil-separation distances and
for two general soil types: fine grained and coarse grained (Figure 25). The probabilities of
benzene soil vapor concentrations less than 100 ng/m3 were similar for dissolved-source sites for
the two soil types. For LNAPL source sites, the probabilities were between 6 and 16 percent
greater for coarse-grained soils than those for fine-grained soils for small separation distances,
but at larger separation distances, there is a reversal in the trend. The evaluation of data trends is
limited by absence of data for fine-grained soils beyond a 14-ft separation distance. The analysis
did not identify whether soil type has an influence on benzene soil vapor concentrations and
probabilities of exceedances.
.< Threshold
D U3 U3 O
* Oi 00 O
Probability Benzene Vapor Cone
§00 CO CO 00 UD ID I.
N> £> cn oo o NJ
Benzene Probabilityfor Different Soil Types -
Dissolved
— ^^* ^>^
— ^-i
^ — -^
—*— Probability Cv < 100
^^Probability C <100
Fine Grained
117145
Distance between Soil Vapor Probe and Contamination (ft)
100
-a
o 90
| 80
v 70
o 60
Probability Benzene
t-^ ro uj .
0 0 0 0 C
Benzene Probabilityfor Different Soil Types - NAPL
(all)
S+^ '
7 f-^
-X //
^ f
^^^
*^^__
—•—Probability C^ < 100
Coarse Grained
-•-Probability^ 100
) 5 10 15 20 25 30
Distance between Soil Vapor Probe and Contamination (ft)
Figure 25. Comparison of probability for benzene soil vapor concentrations to be less than the
threshold for different soil types (coarse and fine grained). Below detection limit
concentrations replaced with half the detection limit for analysis.
Given the data available, it was not possible to conduct an empirical analysis of the potential
effect of layered soil systems where, for example, fine-grained soils (that may be wet) overlie
coarse-grained soils. The evaluation of case studies indicated one site where there may have been
reduced vertical concentration attenuation of PHC vapors because of a geologic profile where a
surficial clay layer overlay coarser-grained soil (Luo et al., 2010).
7.3 Clean Soil Method
The clean soil method (Davis, 2009; 2010) consists of an analysis of the thickness of un-
impacted clean soil (i.e., soil not impacted by LNAPL) required for soil vapor benzene
56
-------
concentrations to attenuate to below 100 ng/m . As described previously in Section 6, two
interpolation procedures were used as part of the estimation process. Procedure 2 is considered a
more representative method. The results of the analyses, shown in Table 7 and Figures 26
through 28, are summarized as follows:
• For dissolved-source sites, the 95th percentile clean soil thicknesses (calculated using
Excel) for procedures 1 and 2 were 10 ft (3.0 m) and 5.4 ft (1.6 m), respectively
(Figure 26). There was no trend between clean soil distance and dissolved benzene
groundwater concentrations. This result was expected because of the observed poor
correlation between groundwater and deep soil vapor concentrations described in
Section 7.1.1.
• For LNAPL sources at UST sites, the 95th percentile clean soil thicknesses
(incorporating all the data) for procedures 1 and 2 were 13.9 ft (4.2 m) and 13.5 ft
(4.1 m), respectively. There was an increase in the clean soil thicknesses for benzene
groundwater concentrations greater than approximately 5 mg/L. For a small
percentage of the data points (4 percent), an attenuation distance could not be
calculated (green symbols on Figure 27). However, the omission of that small
amount of data was not considered significant to the overall result.
• For LNAPL sources at fuel terminal, refinery, and petrochemical sites, there were
insufficient data to estimate percentiles (Figure 28).13 The estimated maximum clean
soil thickness was approximately 20 ft (6.1 m). Approximately 26 percent of the data
points represent vertical profile data where the shallowest benzene soil vapor
concentration was greater than 100 ng/m3 (green symbols on Figure 28).
Table 7. Summary of Results for Clean Soil Method
Source Scenario
and Facility Type
Dissolved
LNAPL- UST sites
LNAPL -Fuel
Terminal, Refinery,
and Petrochemical
Sites
Number Sites
47
53
60
Number Data
Points
170
172
216
95th Percentile Clean Soil Thickness
Procedure 1
10.0 ft (3.0m)
13.9 ft (4.2m)
20.0ft (6.1 mf
Procedure 2
5.4 ft (1.6m)
13.5 ft (4.1 m)
16.2 ft (4.9 mf
Note: The above statistics include site data when no benzene groundwater concentration was available.
a Values in italics are maximums because percentiles could not be calculated for non-UST LNAPL sites.
There were additional data where a clean soil thickness was calculated, but no benzene groundwater data were
available near the soil vapor probe, and therefore these data could not be plotted.
57
-------
Thickness Clean Soil Method - Dissolved
o
Q.
O
30
25
.-,
,S £.20
* Thickness Clean Soil Attenuate
Benzene < Threshold - Procedure 1
• Thickness Clean Soil Attenuate
Benzene < Threshold - Procedure 2
Thickness Soil Where Benzene Not
Attenuated < Threshold
10 100 1,000 10,000 100,000
Benzene Groundwater Cone. (fig/L)
Figure 26. Results of clean soil method for dissolved-source sites. 47 sites, N = 170.
o
Q.
O
+j
E
o
30
25
Thickness Clean Soil Method - NAPL (UST only)
E £20
Thickness Clean Soil Attenuate
Benzene < Threshold - Procedure 1
Thickness Clean Soil Attenuate
Benzene < Threshold - Procedure 2
Thickness Soil Where Benzene Not
Attenuated < Threshold
r *>»*
10 100 1,000 10,000
Benzene Groundwater Cone. (u,g/L)
100,000
Figure 27. Results of clean soil method for LNAPL sources at UST sites. 53 sites, N = 172.
58
-------
Thickness of Clean Soil (from top of
contamination) (ft)
h-> h-> M M UJ
O Ui O Ui O Ui O
Thickness Clean Soil Method - NAPL (non-UST)
* Thickness Clean Soil Attenuate
Benzene < Threshold - Procedure 1
• Thickness Clean Soil Attenuate
Benzene < Threshold - Procedure 2
AThickness Soil Where Benzene Not
Attenuated < Threshold
A 4, ^
A • •
A A
L 10 100 1,000 10,000 100,000
Benzene Groundwater Cone. (ug/L)
Figure 28. Results of clean soil method for LNAPL sources at fuel terminal, refinery, and
petrochemical (non-UST) sites. 60 sites, N = 216.
8. Discussion
8.1 Conceptual Site Model and Mathematical Models
The CSM for PHC vapor is the basis for identifying exclusion distances and inclusion behavior
criteria. At sites with dissolved PHC contamination in groundwater, aerobic biodegradation is
expected to result in the attenuation of PHC vapors, such that there is limited potential for a
complete PVI pathway, except for sites with very shallow contamination. For sites with LNAPL
contamination, there is greater potential for oxygen limitations below buildings and a complete
PVI pathway, depending on site conditions such as source concentrations, depth to source, and
building characteristics. Case studies reviewed (Section 2) suggest that the potential for a
complete PVI pathway may exist at fuel terminal, petroleum refinery and petrochemical sites
(referred to as terminal and refinery sites) with large volume LNAPL releases, particularly where
there are large buildings or a capping effect based on geologic conditions. The empirical data
analysis of surface type scenario indicated an apparent increase in frequency of depleted oxygen
below building foundations for fuel terminal, refinery, and petrochemical sites, but not for UST
sites.
The mathematical modeling studies reviewed support the empirical analysis for dissolved PHC
sources in that model simulations sources predict very low vapor attenuation factors, except for
small separation distances (i.e., less than about 5 ft [1.5 m]) between vapor sources and
buildings. For LNAPL vapor sources, three-dimensional model simulations for a residential
59
-------
house scenario and homogeneous soil conditions predict that a vertical separation distance on the
order of 23 ft (7.0 m) is required for aerobic reaction front development within the unsaturated
zone below the building foundation. There are both potentially conservative aspects associated
with the modeling studies reviewed (e.g., simulations did not include oxygen transport through
the foundation) and non-conservative aspects (e.g., limited evaluation of layered soil deposits or
larger buildings). The modeling results indicate further evaluation of factors potentially
influencing oxygen supply and demand, such as source vapor concentration, source size,
building size, surface cover, soil layer properties, and natural soil oxygen demand, is warranted
for the LNAPL source scenario.
8.2 Methods and Characteristics of the Database
The results of the analysis may be used to derive inclusion distances based on the probability of
vapor concentrations being less than defined thresholds for various separation distances between
a benzene source and an overlying building and qualitative comparisons of soil vapor
concentrations to risk-based soil vapor concentrations for other compounds. The clean soil
method is also an acceptable method for defining inclusion distances, but is less accurate when
the vertical concentration attenuation is poorly resolved (i.e., when soil gas probes are more than
about 5 to 10 ft apart).
The database is representative of abroad range of environmental site conditions, climatic
conditions (including relatively dry areas such as Utah), soil types, and land-surface covers that
may be found at UST sites. Although the data analysis suggests that the type of surface cover
(e.g., building foundation, pavement, open ground) can have an effect on the attenuation of
benzene vapor and oxygen concentrations, the database includes sub-slab or vapor data from
deeper distances below buildings for 39 sites with small to medium sized buildings. The dataset
is considered sufficiently large and robust such that exclusion distances derived from the analysis
will include the potential influence of surface cover.
The sources of uncertainty associated with the empirical analysis, and for which additional
validation studies should be considered, include the following:
• Influence of methanogenesis on oxygen demand and specifically the effect of ethanol
content in gasoline on methane generation rates and aerobic biodegradation of PHC
vapors;
• Effect of extensive high organic matter content soils (e.g., peat) with potentially high
natural oxygen demand;
• Effect of possible capping through either large buildings and/or certain geologic
conditions (wet surface clay underlain by coarse-grained soils) where there may be
increased potential for oxygen limitations;
• Limited knowledge of vapor attenuation behavior in fractured bedrocks (although the
Australian database appears to indicate less attenuation than for soil geologic settings,
as described in Appendix F).
• Limited soil vapor data for fuel terminal, refinery, and petrochemical sites and vapor
attenuation behavior of aliphatic compounds such as 2,2,4-TMP; and
60
-------
• Absence of soil vapor data for lead scavengers, EDB and 1,2-DCA.
8.3 Data Analysis Results
For dissolved sources, the vertical distance method indicates that the probability of benzene
vapor concentrations being less than the defined concentration thresholds (50 and 100 |ag/m3) is
94 to 97 percent (Kaplan-Meier statistics) for small source-separation distances (as little as 0 ft),
meaning that PVI is unlikely to occur unless a dissolved source is very close to a building
foundation. For development of exclusion distances for dissolved sites, allowance should be
made for uncertainty in water table elevations due to seasonal variability. For LNAPL sources at
UST sites, the vertical distance method indicates that there is a high probability that benzene
vapor concentrations will be less than the defined thresholds at distances greater than about 15 ft
(4.6 m) when just UST facilities are considered. For fuel terminal, refinery, and petrochemical
sites, the data indicate larger distances are required for benzene vapor concentrations to attenuate
to similarly low concentrations.
The data indicate a weak correlation between benzene concentrations in groundwater and deep
soil vapor for both dissolved and LNAPL sources. However, for the clean soil method, an
approximate trend was observed where the clean soil thicknesses needed for benzene vapor
attenuation increased when dissolved benzene concentrations were above approximately 5 mg/L
(i.e., indicative of LNAPL source zones). The implication of this result is that there is a basis for
utilizing groundwater concentrations to identify LNAPL source zones, but that groundwater
concentrations cannot be correlated to inclusion distances.
8.4 Exclusion Distance Assessment Framework
The assessment framework for vertical exclusion distances requires identification of the PHC
source type (dissolved phase or LNAPL) based on a sufficiently intensive and comprehensive
site investigation and multiple lines of evidence approach for LNAPL indicators similar to that
described earlier in this report (see Table 4), consisting of:
• Collection of continuous soil cores in inferred LNAPL zones;
• Field tests consisting of headspace vapor monitoring, shake and dye tests, visual
observations, etc.
• Groundwater monitoring for presence of LNAPL;
• Monitoring of depth to groundwater including seasonal monitoring (needed to
identify potential top of the contamination zone);
• Soil and groundwater chemistry testing;
• Geophysical methods such as ultraviolet induced fluorescence where available.
• Soil gas monitoring data where available or warranted.
It is recommended that LNAPL sites be identified primarily by direct indicators (i.e., LNAPL
presence). Groundwater and soil concentration thresholds for benzene and potentially other
chemical parameters may also be useful, but given the uncertainty in the relation between
groundwater and soil vapor concentrations, groundwater and soil chemistry should not be the
61
-------
primary factor for LNAPL indicator screening. If soil gas data are available, elevated aliphatic
concentrations (e.g., hexane) may also be indirect indictor for the presence of LNAPL (e.g., see
Appendix E and Lahvis et al., In prep.).
Buffer distances to account for uncertainty in the empirical data and site investigation should be
considered (i.e., relative to inclusion distances based on this analysis). There is uncertainty for all
source scenarios considered, but it is greatest for fuel terminal, refinery, and petrochemical sites.
Given that the thickness of clean soil (with no LNAPL present) is an important factor, the
assessment framework should address data requirements for LNAPL source zone assessment and
collection of data that may affect the inclusion distances, such as water table fluctuations, the
height of the capillary fringe, and building location. The inclusion of soil gas testing and
minimum oxygen thresholds in the framework would reduce the uncertainty in the assessment
process but may not be warranted provided there is sufficient rigor in the site characterization
approach to delineate PHC sources, define clean soil zones, and apply robust methods for
identifying LNAPL versus dissolved sites.
An inclusion distance approach for LNAPL sources should also include criteria designed to
capture sites that fall outside of the findings of the above analysis, including
• fuel terminal, refinery, and petrochemical sites (related to size of petroleum release),
• high organic-rich soils (e.g., peat),
• large building foundations (e.g., associated with apartment complexes or
commercial/industrial buildings),
• fractured bedrock, and
• subsurface utilities that act as significant preferential pathways (i.e., where the utility
connects a LNAPL-impacted soil zone with the building).
Future research may indicate certain ethanol contents in gasoline may also warrant inclusion and
PVI assessments because of their tendency to generate methane, and consume oxygen that is
needed for the biodegradation of PHCs.
Because the observed differences in vapor attenuation between UST and non-UST (e.g., fuel
terminal, refinery, or petrochemical) sites are inferred to be associated primarily with the volume
of the petroleum release, it might seem reasonable to try and define a threshold release volume of
concern that would apply to UST sites. However, there are limited data that would enable
definition of such a threshold, and a simpler approach based on site type is recommended given
that releases at UST sites are typically much smaller than those at fuel terminal, refinery, or
petrochemical sites.
8.5 Lateral Inclusion Distances
Greater attenuation of PHC vapors is expected when hydrocarbon sources are offset laterally
from buildings compared with sources that are directly below buildings. The modeling studies
reviewed in this report indicate greater hydrocarbon vapor attenuation in the lateral compared to
vertical direction. In concept, the modeling results suggest similar lateral inclusion distances
could be applied as the vertical distances estimated from the current analysis. However, from a
62
-------
practical standpoint, the uncertainty associated with delineating a PHC source near existing or
future buildings because of typical data collection density may warrant larger lateral distances
than those based on the vertical distance analysis. Recommendation of lateral distances is beyond
the scope of this report.
8.6 Comparison with Other Studies
Four different PHC data analysis efforts were conducted in roughly the same time frame as this
one:
• Davis [2009],
• Lahvis et al. [In prep.],
• Wright [2011], and
• Peargin and Kolhatkar [2011]).
All five analysis efforts (including this one) support essentially the same inclusion distances for
PHC UST sites, in spite of differences in the base data analyzed and each study's approach to the
analysis. This suggests an underlying consistency in mechanisms for PHC biodegradation in the
subsurface and supports the protectiveness of the use of these distances as inclusion criteria for
UST petroleum release sites.
9. Findings and Conclusions
Approaches for assessing PVI that do not account for aerobic biodegradation processes are
overly conservative because they do not take into account a proven mechanism for attenuating
PHC concentrations in the subsurface. The statistical analysis of soil gas data drawn from the
74 sites14 presented in this report and accompanying database, along with four similar but
distinct efforts, provides an opportunity for developing a screening approach for PHC
compounds. The inclusion distance approach is based on the observed attenuation of PHCs over
a characteristic separation distance beyond which there is limited potential for a complete PVI
pathway. The focus of this analysis was primarily on evaluating PVI at UST sites with petroleum
fuel releases, although data from other types of sites (fuel terminals, petroleum refineries,
petrochemical facilities) were also considered.
Findings from analysis of dissolved sources in the PVI database include:
• For the vertical distance method, approximately 97 percent of the benzene soil vapor
concentrations are less than 100 ng/m3 and 94 percent of the concentrations are less
than 50 ng/m3 for contamination source-to-building separation distances as small as 0
ft. For other compounds evaluated, measured soil vapor concentrations are less than
the risk-based concentrations for separation distances greater than 3 ft (0.9 m).
• For the clean soil method, the 95th percentile vertical clean soil thickness for benzene
vapor attenuation to below 100 ng/m3 is approximately 5.4 ft (1.6 m).
14 Some analyses and findings are based on fewer than 74 sites, see Appendix B for details on data.
63
-------
• The analysis indicates there is a low probability of exceeding risk-based
concentrations even for small separation distances.
Findings from analysis of LNAPL sources at UST sites in the PVI database include:
• Approximately 95 percent of the benzene soil vapor concentrations are less than
100 |-ig/m3, and 93 percent of the concentrations are less than 50 ng/m3 at a
contamination source-to-building separation of approximately 15 ft (4.6 m). For other
compounds evaluated, measured soil vapor concentrations are less than the risk-based
concentrations beyond 11 ft (3.4 m).
• For the clean soil method, the 95th percentile vertical clean soil thickness for benzene
vapor attenuation is approximately 13.5 ft (4.1 m).
Findings from analysis of LNAPL sources at fuel terminal, refinery, and petrochemical (non-
UST) sites in the PVI database include:
• For the vertical distance method, approximately 90 percent of the benzene soil vapor
concentrations are less than the thresholds for a contamination source-to-building
separation distance of approximately 18 ft (5.5 m); the probability does not increase
beyond 90 percent beyond this distance because data are limited for larger separation
distances. For other compounds evaluated, measured soil vapor concentrations are
less than the risk-based concentrations beyond 12 ft (3.6 m).
• For the clean soil method, there are insufficient data to estimate percentiles, but the
maximum vertical clean soil thickness for benzene vapor attenuation is approximately
20 ft (6.1m).
Other conclusions from this work include:
• The available data indicate benzene is the risk driver for the sites evaluated, with
exceedances of the risk-based vapor concentrations for benzene occurring at larger
contamination source-building separation distances than observed for the other PHCs
with EPA toxicity values.
• There was significantly less attenuation in vapor concentrations for the aliphatic
hydrocarbon 2,2,4-TMP compared with benzene, although data were relatively
limited. However 2,2,4-TMP does not have a toxicity benchmark and so cannot be
evaluated in the vertical distance or clean soil method.
• The data analysis indicates a poor correlation between benzene concentrations in
groundwater and deep soil vapor taken above a groundwater source. The implication
is that a screening approach for vapor intrusion based on groundwater concentrations
is not appropriate for PVI sites. However, groundwater concentrations can be used as
an approximation to identify LNAPL sources.
• The analysis of surface cover indicated:
- For LNAPL sources at UST sites, there were lower oxygen concentrations and
less benzene vapor attenuation below paved surfaces, but not below buildings,
compared to bare ground cover, and
64
-------
— For fuel terminal and refinery sites, there were lower oxygen concentrations
below buildings but not below paved surfaces. This finding may result from larger
petroleum releases and consequent increased oxygen demand at such sites,
compared with typical LNAPL releases at UST sites.
These results are variable and not conclusive as to the effect of surface cover, but they
suggest that there can be reduced oxygen availability below hard surfaces (pavement
or building foundations) for the sites evaluated.
• Because the vertical distance method evaluation includes soil vapor concentration
data from below buildings at 39 sites, the results are considered reasonably robust
with respect to the potential influence of surface cover (although further evaluation of
this factor is recommended).
Critical factors affecting PVI, and important metrics for site screening, include:
• The facility type (UST versus fuel terminal, refinery, or petrochemical site), which
influences the size of the release;
• The PHC source type (dissolved versus LNAPL); and
• The vertical separation distance between the source and receptor (building
foundation).
The mathematical modeling studies reviewed strongly support the empirical analysis and
inclusion distances for dissolved sources. For LNAPL sources, the modeling generally supports
the empirical analysis, but further evaluation of factors potentially influencing oxygen supply
and demand is warranted. These factors include source vapor concentration, source size, building
size, surface cover and soil layer properties, and natural soil oxygen demand.
The findings of this report have important implications for PHC screening approaches based on
the observed attenuation in PHC vapor concentrations and an inclusion distance approach.
Inclusionary criteria or conditions not analyzed in this database, but where more detailed PVI
assessment is considered warranted, include non-UST facilities, high organic-rich soils (e.g.,
peat), large building foundations (e.g., associated with apartment complexes or commercial or
industrial buildings), and significant subsurface preferential pathways.
Releases of certain ethanol blends of gasoline may also warrant additional consideration for
inclusion and PVI assessments, although further research is required to determine the
significance of ethanol content with respect to inclusion distances. Biodegradation of ethanol
may generate methane at a greater rate than gasoline without ethanol, consuming oxygen that
would otherwise be available for biodegradation of PHCs and thus increasing the potential for
PVI.
Inclusion criteria may not apply at sites where there is significant methane generation because of
the potential for safety hazards, advective soil gas transport and reduced biodegradation of other
PHCs (due to oxygen demand represented by methane). High methane generation potential has
been documented at large diesel and gasoline spills at non-UST sites.
65
-------
10. References
Abreu, L.D., R. Ettinger, and T. McAlary. 2009. Simulated Soil Vapor Intrusion Attenuation
Factors including Biodegradation for Petroleum Hydrocarbons. Ground Water
Monitoring and Remediation. 29:105-177.
Abreu, L.D., and P.C. Johnson. 2005. Effect of Vapor Source, Building Separation and Building
Construction on Soil Vapor Intrusion as Studied with a Three-Dimensional Numerical
Model. Environmental Science & Technology. 39:4550-4561.
Abreu, L.D., and P.C. Johnson. 2006. Simulating the Effect of Aerobic Biodegradation on Soil
Vapor Intrusion into Buildings: Influence of Degradation Rate, Source Concentrations.
Environmental Science & Technology. ¥0:2304-2315.
Alaska Department of Environmental Conservation. 2011. Hydrocarbon Risk Calculator User
Manual. January 4. Prepared by Lawrence Acomb, Geosphere, Inc.
American Petroleum Institute (API). 2012. BioVapor Indoor Vapor Intrusion Model. Available
at http://www.api.org/Environment-Health-and-Safety/Clean-Water/Ground-
Water/Vapor-Intrusion/. Accessed February 27, 2012.
ASTM (ASTM International). 2006. Standard Guide for Development of Conceptual Site Models
and Remediation Strategies for Light Nonaqueous-Phase Liquids Released to the
Subsurface. E2531-06el. ASTM International, West Conshohocken, PA. Available at
www.astm.org.
Atlas, R.M. 1981. Microbial degradation of petroleum hydrocarbons: an environmental
perspective. Microbiological Reviews. ¥5(1): 180-209.
Branco, F.A., and J. de Brito. 2004. Handbook of Concrete Bridge Management. American
Society of Civil Engineers (ASCE), ASCE Press, Reston, VA.
Breyer, L., and J.B. Cowart. 2004. Attenuation factors for hydrocarbons associated with a diesel
spill. Presented at U.S. EPA Vapor Intrusion Attenuation Workshop: A Study of Observed
Vapor Intrusion Attenuation. 14th Annual AEHS West Coast Conference on Soils,
Sediments and Water. San Diego, CA. March 15-18.
Bruce, L., T. Miller, and B. Hockman. 1991. Solubility versus Equilibrium Saturation of
Gasoline Compounds: A Method to Estimate Fuel/Water Partition Coefficient Using
Solubility or Koc. In: Proceedings of the NWWA/API Conference on Petroleum
Hydrocarbons in Ground Water, p. 571-582, by National Water Well Association,
Dublin, OH. Available at http://info.ngwa.org/gwol/pdf/910155295.PDF.
California EPA. 2011. Final Guidance for the Evaluation and Mitigation of Subsurface Vapor
Intrusion to Indoor Air (Vapor Intrusion Guidance). Department of Toxic Substances
Control. California Environmental Protection Agency. Sacramento, CA. October.
Davis, R.V. 2009. Bioattenuation of Petroleum Hydrocarbon Vapors in the Subsurface: Update
on Recent Studies and Proposed Screening Criteria for the Vapor-Intrusion Pathway.
LUSTLine. 67:11-14.
66
-------
Davis, R.V. 2010. Evaluating the Vapor Intrusion Pathway: Subsurface Petroleum Hydrocarbons
and Recommended Screening Criteria. 22nd Annual U.S. EPA National Tanks
Conference, Boston, MA. September 20-22.
Davis, R.V. 201 la. Evaluating the Petroleum Vapor Intrusion Pathway: Studies of Natural
Attenuation of Subsurface Petroleum Hydrocarbons & Recommended Screening Criteria.
AEHS 21st Annual West Coast International Conference on Soil, Sediment, Water &
Energy, Mission Valley, San Diego, CA. March 15.
Davis, R.V. 201 Ib. Attenuation of Subsurface Petroleum Hydrocarbon Vapors: Spatial and
Temporal Observations. AEHS 21st Annual West Coast International Conference on
Soil, Sediment, Water & Energy, Mission Valley, San Diego, CA. March 16.
Davis, G.B., B.M. Patterson, and M.G. Trefiry. 2009. Evidence for Instantaneous Oxygen-
Limited Biodegradation of Petroleum Hydrocarbon Vapors in the Subsurface.
Groundwater Monitoring and Remediation. 29:126-137.
DeVaull, G.E. 2007a. Indoor Vapor Intrusion with Oxygen-Limited Biodegradation for a
Subsurface Gasoline Source. Environmental Science & Technology. ¥7:3241-3248.
DeVaull, G.E. 2007b. Indoor Air Vapor Intrusion: Predictive Estimates for Biodegrading
Petroleum Chemicals. Presented at Air and Waste Management Association (A&WMA)
Specialty Conference: Vapor Intrusion: Learning from the Challenges, Providence, RI.
September 26-28.
DeVaull, G.E. 2010. Vapor Intrusion from Subsurface to Indoor Air: Biodegradable Petroleum
Vapors versus Recalcitrant Chemicals. Presented at Vapor Intrusion 2010: Air & Waste
Management Association (A&WMA) Conference, Chicago, IL. September 29-30.
DeVaull, G.E. 2011. Biodegradation Rates for Petroleum Hydrocarbons in Aerobic Soils: A
Summary of Measured Data. Presented at International Symposium on Bioremediation
and Sustainable Environmental Technologies, Reno, NV. June 27-30.
DeVaull, G.E., R.A. Ettinger, J.P. Salinitro, and J.B. Gustafson. 1997. Benzene, Toluene,
Ethylbenzene, and Xylenes (BTEX) Degradation in Vadose Zone Soils during Vapor
Transport: First Order Rate Constants. Presented at NWWA/API Conference on
Petroleum Hydrocarbons and Organic Chemicals in Ground Water: Prevention,
Detection, and Remediation, National Water Well Association, Houston, TX. November
12-14.
Fischer, M.L., AJ. Bentley, K.A. Dunkin, A.T. Hodgson, W.W. Nazaroff, R.G. Sextro, and J.M.
Daisey. 1996. Factors Affecting Indoor Air Concentrations of Volatile Organic
Compounds at a Site of Subsurface Gasoline Contamination. Environmental Science &
Technology. 30:2948-2957.
Fitzpatrick, N.A., and JJ. Fitzgerald. 2002. An Evaluation of Vapor Intrusion into Buildings
through a Study of Field Data. Soil and Sediment Contamination. 77:603-623.
Friebel, E., and P. Nadebaum. 2011. Health Screening Levels for Petroleum Hydrocarbons in
Soil and Groundwater: Part 1: Technical Development Document. CRCfor
Contamination Assessment and Remediation of the Environment. Technical Report No.
10.
67
-------
Golder Associates. 2006. Investigation of Indoor Air Quality in Structures Located above VOC-
Contaminated Groundwater. Part 1: Evaluation of Soil Vapor Intrusion at Mount Holly
Site, New Jersey. Submitted to New Jersey Department of Environmental Protection.
July.
Grathwohl, P. and U. Maier. 2002. Natural Attenuation of Volatile Hydrocarbons in Unsaturated
Soil Zone. Agricultural Sciences. 7:9-15.
Helsel, D.R. 2005. Nondetects and Data Analysis, Statistics for Censored Environmental Data.
John Wiley and Sons, New York.
Helsel, D.R. 2006. Fabricating Data: How Substituting Values for Non-detects Can Ruin Results,
and What Can be Done about It. Chemosphere. 65:2434-2439.
Hers, I, J. Atwater, L. Li, and R. Zapf-Gilje. 2000. Evaluation of Vadose Zone Biodegradation
of BTX Vapours. Journal of Contaminant Hydrology. 46:233-264.
Hers, I, D. Evans, R. Zapf-Gilje, and L. Li. 2002. Comparison, Validation and Use of Models
for Predicting Indoor Air Quality from Soil and Groundwater Contamination. Journal of
Soil and Sediment Contamination. 77(4):491-527.
Hers, I, P. Jourabchi, M. Lahvis, P. Dahlen., E.H. Luo, P. Johnson, and U. Mayer. In
preparation. Cold Climate Study of Soil Vapor Intrusion at a Residential House above a
Petroleum Hydrocarbon Plume.
Hers, I, R. Zapf-Gilje, P. Johnson, and L. Li. 2003. Evaluation of the Johnson and Ettinger
Model of Indoor Air Quality. Ground Water Monitoring and Remediation. 23:119-133.
Holden, P. A., L.J. Halverson, and M.K. Firestone. 1997. Water Stress Effects on Toluene
Biodegradation by Pseudomonasputida. Biodegradation. 8:143-151.
Illinois Department of Public Health. 2010. Letter Health Consultation, Vapor Intrusion Event,
Hartford Residential Community, Hartford, IL. February 24.
Jewell, K.P. and J.T. Wilson. 2011. A New Screening Method for Methane in Soil Gas Using
Existing Groundwater Monitoring Wells. Groundwater Monitoring and Remediation.
37(3): 82-94.
Johnson, P.C. 2005. Identification of Application-Specific Critical Inputs for the 1991 Johnson
and Ettinger Vapor Intrusion Algorithm. Groundwater Monitoring & Remediation.
25(l):63-78.
Johnson, P.C., and R.A. Ettinger. 1991. Heuristic Model for Predicting the Intrusion Rate of
Contaminant Vapors into Buildings. Environmental Science & Technology. 25:1445-
1452.
Jourabchi, P., N. Sihota, U. Mayer, I. Hers, and G. DeVaull. 2012. Effects of Ethanol Blended
Gasoline Release on Methane Generation and Soil Vapor Transport: Preliminary Results
of Biogas Generation Experiments and Soil Vapor Modeling. Presentation at 2012
National Tanks Conference & Expo, March 1-21, St. Louis, MO.
Jury, W.A., W.F. Spencer, and W.J. Farmer. 1983. Behavior Assessment Model for Trace
Organics in Soil: I. Model Description. Journal of Environmental Quality. 72:558-564.
68
-------
Kaplan, E.L., and P. Meier. 1958. Nonparametric Estimation from Incomplete Observations.
Journal of the American Statistical Association. 53:457-481.
Kobayashi, K., and K. Shuttoh. 1991. Oxygen Diffusivity of Various Cementitious Materials,
Cement and Concrete Research. 27:273-284.
Lahvis, M.A., and A.L. Baehr. 1996. Estimating Rates of Aerobic Hydrocarbon Biodegradation
by Simulation of Gas Transport in the Unsaturated Zone. Water Resources Research.
32:2231-2249.
Lahvis, M.A., and A.L. Baehr. 1999. Quantification of Aerobic-Biodegradation and
Volatilization Rates of Gasoline Hydrocarbons near the Water Table during Natural-
Attenuation Conditions. Water Resources Research. 35:753-765.
Lahvis, M.A., I. Hers, R. Davis, J. Wright, and G.E. DeVaull. In preparation. Vapor Intrusion
Screening Criteria for Application at Petroleum UST Release Sites.
Leahy, J.G., and R.R. Colwell. 1990. Microbial Degradation of Hydrocarbons in the
Environment. Microbiological Reviews. 5¥(3):305-315.
Leeson, A., and R.E. Hinchee. 1996. Principles and Practices of Bioventing. Volume 1:
Bioventing Principles and Volume 2: Bioventing Design. Battelle Memorial Institute.
September.
Luo, H., P. Dahlen, and P.C. Johnson. 2010. Hydrocarbon and Oxygen Transport in the Vicinity
of a Building Overlying a NAPL Source Zone. Proc. of Air Waste Management Vapor
Intrusion Conference, Chicago, IL. September 28-30.
Luo, H., P. Dahlen, P.C. Johnson, T. Peargin, and T. Creamer. 2009. Spatial Variability of Soil-
Gas Concentrations Near and beneath a Building Overlying Shallow Petroleum
Hydrocarbon-Impacted Soils. Ground Water Monitoring & Remediation. 29:81-91.
Luo, H., C. Holton, and P.C. Johnson. 2012. Field Study and Numerical Simulation of Indoor Air
Sources Effects on Soil Gas Chemical Concentrations at a Vapor Intrusion Study Site.
The 22nd Annual International Conference on Soil, Water, Energy, and Air. March 19-
22, San Diego, CA.
McAlary, T., P. Nicholson, D. Bertrand, H. Groenevelt, and R. Ettinger. 2010. A Case Study on
the Influence of Aerobic Biodegradation on Vapor Intrusion at a Former Refinery
Property. Chapter 13 in Vapor Emission to Outdoor Air and Enclosed Spaces, S.
Saponaro, E. Sezenna, and L. Bonomo, eds. Nova Science Publishers, Inc., Hauppauge
NY.
Madigan, M.T., J.M. Martinko, D.P. Clark, and D.A. Stahl. 2010. Brock's Biology of
Microorganisms. Prentice Hall, NJ.
Massachusetts Department of Environmental Protection (MADEP). 2003. Updated Petroleum
Hydrocarbon Fraction Toxicity Values for the VPH/EPH/APH Methodology. November.
McHugh, T.E., R. Davis, G. DeVaull, H. Hopkins, J. Menatti, and T. Peargin. 2010. Evaluation
of Vapor Attenuation at Petroleum Hydrocarbon Sites: Considerations for Site Screening
and Investigation. Soil and Sediment Contamination. 79:1-21.
69
-------
McHugh, T., P.C. De Blanc, and RJ. Pokluda. 2006. Indoor Air as a Source of VOC
Contamination in Shallow Soils Below Buildings. Soil & Sediment Contamination.
75:103-122.
Michalski, P., S. Thompson, C. DeWolf, P. Nicholson, and T. McAlary. 2012. Effects of
Alternate Petroleum Hydrocarbon Sources in the Vadose Zone on the Vapor Intrusion
Pathway beneath a Residential Community. Proceedings, AWMA Vapor Intrusion
Specialty Conference, Denver, CO. October.
Nelson, D.K., T.M. Lapara, and PJ. Novak. 2010. Effects of Ethanol-Based Fuel Contamination:
Microbial Community Changes, Production of Regulated Compounds, and Methane
Generation. Environmental Science and Technology. 44:4525-4530.
NJ DEP (New Jersey Department of Environmental Protection). 2012. Vapor Intrusion
Technical Guidance. Version 2.0. January. Available at
http://www.nj.gov/dep/srp/guidance/vaporintrusion/vig.htm.
Patterson, B., and G. Davis. 2009. Quantification of Vapor Intrusion Pathways into a Slab-on-
Ground Building under Varying Environmental Conditions. Environmental Science and
Technology. ¥3:650-656.
Peargin, T. and R. Kolhatkar. 2011. Empirical Data Supporting Groundwater Benzene
Concentration Exclusion Criteria for Petroleum Vapor Intrusion Investigations. Proc. of
Battelle Presentation at International Symposium on Bioremediation and Sustainable
Environmental Technologies, Reno, NV, June 27-30.
Ririe, T. and R. Sweeney. 1995. Fate and Transport of Volatile Hydrocarbons in the Vadose
Zone. Proceedings of 1995 Petroleum Hydrocarbon and Organic Chemicals in
Groundwater, Houston, TX, pp. 529-542.
Ririe, G.T., R.E. Sweeney, and S.J., Daugherty. 2002. A Comparison of Hydrocarbon Vapor
Attenuation in the Field with Predictions from Vapor Diffusion Models. Soil and
Sediment Contamination. 77:529-544.
Robinson, N.I. and L. Turczynowicz. 2005. One- and Three-Dimensional Soil Transportation
Models for Volatiles Migrating from Soils to House Interiors. Transport Porous Media.
59:301-323.
Roggemans, S., C.L. Bruce, and P.C. Johnson. 2001. Vadose Zone Natural Attenuation of
Hydrocarbon Vapors: An Empirical Assessment of Soil Gas Vertical Profile Data. Soil
and Groundwater Research Bulletin 75. American Petroleum Institute, Washington, DC,
December.
Sanders, P.F., and I. Hers. 2006. Vapor Intrusion into Homes over Gasoline-Contaminated
Ground Water in Stafford, New Jersey. Groundwater Monitoring and Remediation.
2d(l):63-72.
Spalding, R.F., M.A. Toso, M.E. Exner, G. Hattan, T.M. Higgins, A.C. Sekely, and S.D. Jensen.
2011. Long Term Groundwater Monitoring Results at Large, Sudden Denatured Ethanol
Releases. Ground Water Monitoring and Remediation. 37(3):69-81.
70
-------
Sweeney, B. 2012. Temporal Effects on Aerobic Biodegradation of Hydrocarbons in Soil Vapor.
AEHS West Coast Conference. San Diego, CA. March. Available at
https ://iavi .rti .org/Workshops AndConferences. cfm.
Tittarelli, F. 2009. Oxygen Diffusion through Hydrophobic Cement-Based Materials. Cement
and Concrete Research. 39:924-928.
U.S. EPA (Environmental Protection Agency). 1996. Soil Screening Guidance-User's Guide.
EPA Document Number: EPA540/R-96/018. U.S. Environmental Protection Agency,
Washington, DC. March.
U.S. EPA (Environmental Protection Agency). 2002. Draft Guidance for Evaluating the Vapor
Intrusion to Indoor Air Pathway from Groundwater and Soils (Subsurface Vapor
Intrusion Guidance). Office of Solid Waste and Emergency Response, Washington, DC.
Available at http://www.epa.gov/osw/hazard/correctiveaction/eis/vapor.htm.
U.S. EPA (Environmental Protection Agency). 2004. Superfund User's Guide for Evaluating
Subsurface Vapor Intrusion into Buildings. Prepared by Environmental Quality
Management, Inc. for EPA. Contract Number: 68-W-02-33.
U.S. EPA (Environmental Protection Agency). 2009. Composition and Behavior of Fuel
Ethanol. EPA/600/R-09/037. U.S. Environmental Protection Agency, Washington, DC.
April.
U.S. EPA (Environmental Protection Agency). 2012a. Petroleum Hydrocarbons And
Chlorinated Hydrocarbons Differ In Their Potential For Vapor Intrusion. U. S.
Environmental Protection Agency, Washington, DC. September.
U.S. EPA (Environmental Protection Agency). 2012b. EPA 's Vapor Intrusion Database:
Evaluation and Characterization of Attenuation Factors for Chlorinated Volatile
Organic Compounds and Residential Buildings. U.S. Environmental Protection Agency,
Washington, DC. Report 530-R-10-002. February.
U.S. EPA (Environmental Protection Agency). 2012c. Conceptual Model Scenarios for the
Vapor Intrusion Pathway. U.S. Environmental Protection Agency, Washington, DC.
Report 530-R-10-003. February.
Wiedemeier, T.H., M. A. Swanson, J.T. Wilson, D.H. Kampbell, R. Miller, and I.E. Hansen.
1996. Approximation of Biodegradation Rate Constants for Monoaromatic Hydrocarbons
(BTEX) in Ground Water. Ground Water Monitoring and Remediation. Summer: 186-
194.
Wilson, J.T., J.W. Weaver, and H. White. In press. An Approach for Developing Site-Specific
Lateral and Vertical Inclusion Zones within which Structures Should be Evaluated for
Petroleum Vapor Intrusion due to Releases of Motor Fuel from Underground Storage
Tanks. EPA Issue Paper, Office of Research and Development, Ada, OK.
Wisconsin Department of Natural Resources. 2010. Addressing Vapor Intrusion at Remediation
& Redevelopment Sites in Wisconsin. Madison, WI: Department of Natural Resources.
PUB-RR-800. December.
Wright, J. 2011. Establishing Exclusion Criteria from Empirical Data for Assessing Petroleum
Hydrocarbon Vapour Intrusion. Program and Proceedings of the 4th International
71
-------
Contaminated Site Remediation Conference—2077 CleanUP. Adelaide, South Australia,
September 11-15. p. 142-143.
Wright, J. 2012. Evaluation of the Australian Petroleum Hydrocarbon VI Database: Exclusion
Criteria. Presented at: Recent Advances to VI Application & Implementation—A State-
of-the-Science Update. AEHS West Coast Conference. San Diego, CA. March. Available
at
https://iavi.rti.orgAVorkshopsAndConferences.cfm?PageID=documentDetails&AttachID
=549.
Yao, Y., R. Shen, K.G. Pennell, and E.M. Suuberg. 2011. Comparison of the Johnson-Ettinger
Vapor Intrusion Screening Model Predictions with Full Three-Dimensional Model
Results. Environmental Science & Technology. ¥5(6):2227-3 5.
Zobell, C.E. 1946. Action of Microorganisms on Hydrocarbons. Bacteriological Reviews. 10(1-
2): 1-49.
Zwick, T.C., A. Leeson, R.E. Hinchee, L. Hoeppel, and L. Bowling. 1995. Soil Moisture Effects
During Bioventing in Fuel-Contaminated Arid Soils. Third International In-Situ and On-
Site Bioreclamation Symposium. In-Situ Aeration, v. 3, Batelle Press, San Diego, CA.
72
-------
Appendix A. Review of Exclusion/Inclusion Distances in Existing
Vapor Intrusion Guidance
Many state regulatory approaches exclude sites from the need for PVI assessments if they are
more than a specified distance from the source of vapor contamination. Distances applied in this
context are referred to as inclusion distances in this report.15 A default distance adopted by many
states is 100 ft (30 m) in the lateral, and in some cases, vertical direction. However, guidance for
New Hampshire, New Jersey, Connecticut, and Massachusetts specifies a 15- to 30-ft (4.6- to
9.1-m) exclusion or inclusion distance for aerobically biodegradable chemicals.16 New
Hampshire and New Jersey specify this distance applies laterally and vertically, whereas
Connecticut more genetically indicates inclusion distances as extending from contaminated
groundwater. The criterion for Massachusetts is dependent on whether volatile LNAPL is
present. The technical justification for exclusion distance criteria is relatively limited or not
provided in the guidance by these jurisdictions. This is understandable given that detailed
evaluations of empirical data and/or modeling studies to support inclusion or exclusion distances
are a recent development.
Several states are in the process of developing or have recently developed guidance for PHCs
based on a pathway exclusion or inclusion distance approach. A summary of guidance from New
Jersey, California, and Wisconsin follows:
• The New Jersey Department of Environmental Protection (NJ DEP, 2012)
recommends a PVI investigation based on a 30-ft (9.1-m) critical distance criterion
for PHC-related groundwater contamination and a 100-ft (30-m) criterion for PHC-
related free product contamination. As an alternative approach, NJ DEP (2012)
allows for an exclusion distance approach based on benzene concentrations for
gasoline contamination. A PVI investigation is not necessary if the vertical separation
distance between the water table (all references to water table are for seasonal high
conditions) and building slab is:
- At least 10 ft (3.0 m) for a benzene groundwater concentration < 1,000 ug/L;
- At least 5 ft (1.5 m) for a benzene groundwater concentration < 100 ug/L; or
— At least 5 ft (1.5 m) for oxygen > 2 percent (v/v) in the unsaturated zone and a
benzene groundwater concentration < 1,000 ng/L.
The gasoline exclusion criteria apply only when all of the following four
conditions are met: 1) The building is relatively small, 2) The area around the
building is not extensively paved, 3) Clean soil exists between the water table and
15 States may use the term "inclusion distance" or "exclusion distance" in their guidance. Either term defines a safe
distance between the petroleum source and likely receptors in buildings. In the main body of this document, we use
inclusion distance, but in this appendix we may use either exclusion or inclusion, depending on which term was used
in the state guidance we are discussing.
16 The U.S. EPA's 2002 Draft Guidance for Evaluating the Vapor Intrusion to Indoor Air Pathway from
Groundwater and Soils specified vertical and lateral exclusion distances of 100 ft (30 m), based on empirical
observations of the approximate distance from the interpolated edge of chlorinated solvent plumes where indoor
vapor detections were observed. This guidance did not address vapor intrusion from petroleum releases.
A-l
-------
the building, and 4) NAPL is not present within 30 ft (9.1 m) of the building
(vertically and horizontally) (see NJ DEP [2012] for additional details).
• The California EPA (2011) presents an exclusion distance approach to managing
retail petroleum sites, characterized as a low-threat closure scenario. The exclusion
distances were based on a review of empirical data (primarily Lahvis [2011] and
Davis [2009]) and modeling studies (primarily Abreu et al. [2009]). Four scenarios
are defined with the following benzene exclusion distance criteria:
- Scenario 1—Unweathered LNAPL on groundwater:
• A 30-ft (9.1-m) vertical bioattenuation zone between an unweathered LNAPL
(residual or free-phase) source and a building foundation.
— Scenario 2—Unweathered LNAPL in soil:
• A 30-ft (9.1 m) lateral and vertical separation distance between an
unweathered LNAPL (residual or free-phase) source in soil and a building
foundation.
— Scenario 3—Dissolved phase benzene concentrations in groundwater:
• With no oxygen measurements:
— A 5-ft (1.5 m) vertical separation distance between a dissolved-phase
benzene source < 100 |ag/L and a building foundation.
- A 10-ft (3.0 m) vertical exclusion distance for a dissolved-phase benzene
source < 1,000 |ag/L and a building foundation.
• With oxygen > 4 percent:
— A 5-ft (1.5 m) vertical separation distance between a dissolved-phase
benzene source < 1,000 ng/L and a building foundation.
- Scenario 4—Direct measurement of soil gas concentrations:
• Application of a bioattenuation (additional attenuation) factor of 1,000 times
to risk-based soil gas criteria (i.e., vapor sources) located within 5 ft (1.5 m) of
a building foundation.
• Wisconsin's Department of Natural Resources (2010) in their guidance states that
where no petroleum odors are detected, PVI can be ruled out at most petroleum
release sites with low source concentrations where there is 5 ft (1.5 m) in the
horizontal and vertical directions of clean, unsaturated soil with an oxygen content >
5 percent between the residual petroleum and the building. Larger exclusion distances
are specified when free product is present (30 ft [9.1 m]) or benzene concentrations in
groundwater exceed 1 mg/L (20 ft [6.1 m]). When these distance thresholds and other
criteria (e.g., no preferential pathways, no fractured bedrock) are met, a PVI
assessment is not required.
A-2
-------
Appendix B. Data Quality and Database Content
Table B-l provides data quality information by site in the PVI database. Each site was scored
considering conceptual site model (CSM) robustness using a three-point ranking (3 highest, 1
lowest) and data quality (DQ) using a five-point ranking (5 highest, 1 lowest). CSM robustness
rankings included:
• CSM-3: Well-developed CSM, appropriately located soil gas probes, vertical soil gas
profiles, well-characterized contamination source (NAPL and dissolved), and
available ancillary data (e.g., soil properties).
• CSM-2: Less well-developed CSM with well-located probes but with more limited
soil gas locations (e.g., single location), and a reasonably well characterized
contamination source.
• CSM-1: Limited data to develop CSM and to evaluate appropriateness of soil gas
probe locations and results, generally a single soil gas location, limited or no CSM-
related information, and/or inadequate data to perform clean soil thickness analyses.
DQ rankings were defined as follows:
• DQ-5: Very high quality data, with fully documented QA/QC, permanent probes,
leak tracer and/or pneumatic testing, and fixed gas data consistent with hydrocarbon
vapor concentrations. In some cases, a site's data have been given a 5 ranking when
not all of these aspects were met, but when there is well-developed CSM and the
research is peer reviewed.
• DQ-4: High quality data, with QA/QC mostly documented, generally permanent
probes and leak tracer testing, and fixed gas data that are consistent with hydrocarbon
vapor concentrations.
• DQ-3: Moderate quality data, with some QA/QC documentation and fixed gas data
that may be limited in quantity or inconsistent with hydrocarbon vapor concentration
data.
• DQ-2: Low to moderate data quality, limited QA/QC documentation (but typically
still collected under state program oversight), limited data documentation, no fixed
gas results. Minimum data quality for inclusion in database analysis.
• DQ-1: Low quality data, unacceptable data quality indicators or methods. Data
excluded from all analyses.
A detailed data quality table can also be found in the PVI database that accompanies this report.
The PVI database contents are summarized in Tables B-2 and B-3. Table B-2 provides basic site
information, including facility type, the type of release, soil type, building information, and what
media were sampled (e.g., soil, groundwater, soil gas, indoor air). Table B-3 provides counts of
by site and analyte of soil vapor analyses in PVI database.
B-l
-------
Table B-1. Data Quality and Conceptual Site Model Robustness Information in PVI Database
Site Name
Alameda Naval
Air Station
Coachella
Huntington Beach
Mission Valley
Terminal
Newport Beach
Port Hueneme
Former Chevron
Station
Dave's Amoco
NYM
Jacobsen
Residence
Larsons 66
D&E Sales
Moen Oil
Johnsons Auto
Midi own Service
John's Garage
Buchannon
Nursing
State
CA
CA
CA
CA
CA
CA
CA
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
CSM
Rank3
3
3
3
2
3
3
3
1
1
1
1
1
1
1
1
1
1
Data
Quality
Rank"
5
4
4
4
4
4
5
2
2
2
2
2
2
2
2
2
2
Probe Type
Permanent
Driven
Driven
Driven/
Permanent
Driven
Permanent
Permanent
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Leak
Tracer0
N
N
N
Y
N
N
Y
N
N
N
N
N
N
N
N
N
N
Purging
Procedure?
Y
Y
Y
Y
Y
N
Y
N
N
N
N
N
N
N
N
N
N
Pneu-
matic
Testing?
Y
Y
Y
N
Y
N
N
N
N
N
N
N
N
N
N
N
N
Fixed
Gas
Data?
Y
Y
Y
Y
Y
Y
Y
N
N
N
N
N
N
N
N
N
N
Fixed Gas
andHC
Vapor
Data Con-
sistent?
Y
Y
Y
Y
Y
Y
Y
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Accept-
able VOC
Analysis
Method?
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
VOC Analysis
Method
GC/PID/FID/TCD
Field GC/PID
Field GC/PID
EPA8260B/TO-15
Field GC/PID
TO-15, ASUGC/FID
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
Labor-
atory
QA/QC?
Y
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Peer
Review or
Research
Program
Journal
Journal
Journal
N
Journal
Journal
N
N
N
N
N
N
N
N
N
N
N
Federal or
State
Program
N
N
N
CA DTSC
N
N
CA DTSC
MNPCA
MNPCA
MNPCA
MNPCA
MNPCA
MNPCA
MNPCA
MNPCA
MNPCA
MNPCA
Comment
Geoprobe method used for
driven probes
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
(continued)
-------
Table B-1. Data Quality and Conceptual Site Model Robustness Information in PVI Database (continued)
Site Name
Red & White
Service
Side Lake Store
Ossippe Store
AC Oil
Schmunks
Kennys Oil
Settes Garage
Tilson Auto
Rogers Mobile
Rub-a-Dub
Long Shot
Trucking
Eggens Oil
Chillum
Reuben's Market
Cumberland
Farm 1803
Cumberland
Farm 181 7
State
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MD
ME
ME
ME
CSM
Rank3
1
1
1
1
1
1
1
1
1
1
1
1
2
3
3
3
Data
Quality
Rank"
2
2
2
2
2
2
2
2
2
2
2
2
2
3
4
4
Probe Type
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Permanent
Permanent
Permanent
Permanent
Leak
Tracer0
N
N
N
N
N
N
N
N
N
N
N
N
N
Nd
Nd
Nd
Purging
Procedure?
N
N
N
N
N
N
N
N
N
N
N
N
N
Y
Y
Y
Pneu-
matic
Testing?
N
N
N
N
N
N
N
N
N
N
N
N
N
Y
Y
Y
Fixed
Gas
Data?
N
N
N
N
N
N
N
N
N
N
N
N
N
Y
Y
Y
Fixed Gas
andHC
Vapor
Data Con-
sistent?
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
N
Y
Y
Accept-
able VOC
Analysis
Method?
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
VOC Analysis
Method
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
Labor-
atory
QA/QC?
N
N
N
N
N
N
N
N
N
N
N
N
N
Y
Y
Y
Peer
Review or
Research
Program
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Federal or
State
Program
MNPCA
MNPCA
MNPCA
MNPCA
MNPCA
MNPCA
MNPCA
MNPCA
MNPCA
MNPCA
MNPCA
MNPCA
EPAR3
MEDEP
MEDEP
MEDEP
Comment
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Data from MN PCA; limited
documentation
Elevated TPH vapor cone.
and 02 cone.
(continued)
-------
Table B-1. Data Quality and Conceptual Site Model Robustness Information in PVI Database (continued)
Site Name
Twin Bridge
Market
Cumberland
Farm 1806
Cumberland
Farm 1805
Cumberland
Farm 1839
Cumberland
Farm 1822
7-Eleven
Cumberland
Farm 1836
Cumberland
Farm 1829
Cumberland
Farm 1842
Cumberland
Farm 1834
Mandan
BP Paulsboro
Hulme Street
Stafford
BP Akron
BP Columbiana
BP Conneaut
BP Kent
State
ME
ME
ME
ME
ME
ME
ME
ME
ME
ME
ND
NJ
NJ
NJ
OH
OH
OH
OH
CSM
Rank3
3
3
3
3
3
3
2
2
1
1
2
3
3
3
2
2
2
2
Data
Quality
Rank"
4
4
4
4
4
3
4
4
4
4
4
4
4
5
3
4
4
4
Probe Type
Permanent
Permanent
Permanent
Permanent
Permanent
Permanent
Permanent
Permanent
Permanent
Permanent
Permanent
Permanent
Permanent
Permanent
Permanent
Permanent
Permanent
Permanent
Leak
Tracer0
Nd
Nd
Nd
Nd
Nd
Nd
Y
Nd
Nd
Nd
Y
N
Y
Y
N
N
N
N
Purging
Procedure?
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
Y
Y
N
N
N
N
Pneu-
matic
Testing?
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
N
N
N
N
N
N
N
Fixed
Gas
Data?
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Fixed Gas
andHC
Vapor
Data Con-
sistent?
Y
Y
Y
Y
Y
N
Y
Y
Y
Y
Y
Y
Y
Y
N
Y
Y
Y
Accept-
able VOC
Analysis
Method?
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
VOC Analysis
Method
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-1 5,10-1 3
TO-15, ASUGC/FID
TO-15
TO-15
TO-15, ASUGC/FID
TO-15, ASUGC/FID
TO-15, ASUGC/FID
TO-15, ASUGC/FID
Labor-
atory
QA/QC?
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
Y
Y
N
N
N
N
Peer
Review or
Research
Program
N
N
N
N
N
N
N
N
N
N
Conference
Thesis
NJDEP
NJDEP
Thesis
Thesis
Thesis
Thesis
Federal or
State
Program
MEDEP
MEDEP
MEDEP
MEDEP
MEDEP
MEDEP
MEDEP
MEDEP
MEDEP
MEDEP
N
N
NJDEP
NJDEP
N
N
N
N
Comment
Elevated TPH vapor cone.
and 02 cone.
Site not analyzed; soil gas
data are limited
Site not included; soil gas
data are limited
AEHS presentation; good
notes
(continued)
-------
Table B-1. Data Quality and Conceptual Site Model Robustness Information in PVI Database (continued)
Site Name
Former Refinery
Beaufort
Bountiful Bicycle
Gas & Go #7
Gold Cross
Ambulance
Hal's Chevron
Handi Mart
#102 Chevron
Logan Food Mart
Price Rental
Property
Salina Cash
Saver
Jenkins
Wheel-In Market
Teasdale Country
Store
Tesoro #40
7-Eleven #23387
Refinery Site
Chatterton
Research Site
Ottawa
North Battleford
State
OK
SC
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
na
BC
ON
SK
CSM
Rank3
3
3
3
3
3
3
1
3
3
2
3
3
3
3
3
3
3
3
1
3
Data
Quality
Rank"
4
4
4
4
3
5
2
4
3
3
5
5
3
2
5
4
4
5
1
5
Probe Type
Permanent
Permanent
Permanent
Permanent
Permanent
Permanent
Driven
Permanent
Permanent
Permanent
Permanent
Permanent
Permanent
Driven
Permanent
Permanent
Permanent
Permanent
Permanent
Permanent
Leak
Tracer0
Y
N
Y
ye
N
Y
N
Y
N
N
Y
N
N
N
N
Y
N
N
N
Y
Purging
Procedure?
Y
Y
Y
N
Y
Y
N
Y
Y
Y
N
Y
N
N
Y
Y
N
Y
N
Y
Pneu-
matic
Testing?
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Y
N
Y
Fixed
Gas
Data?
Y
Y
Y
Y
N
Y
Y
Y
N
N
Y
Y
N
Y'
Y
Y
Y
Y
N
Y
Fixed Gas
andHC
Vapor
Data Con-
sistent?
Y
Y
Y
Y
N/A
Y
Y
Y
N/A
N/A
Y
Y
N/A
N/A
Y
Y
Y
Y
N/A
Y
Accept-
able VOC
Analysis
Method?
Y
Y
Y
Y
Y
Y
Unknown
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
Y
VOC Analysis
Method
TO-15,TO-3,Landtec
TO-15
TO-15
TO-15
TO-15
Unknown
EPA8260B
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15
TO-15, ASUGC/FID
Field GC/PID
Field detector
TO-15
Labor-
atory
QA/QC?
Y
N
Y
Y
Y
Y
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
Y
N
Y
Peer
Review or
Research
Program
Book
Journal
N
N
N
USEPA
ORD
Thesis
N
N
N
N
N
N
N
N
N
Thesis
Thesis
N
Journal
Federal or
State
Program
N
N
UTDEQ
UTDEQ
UTDEQ
UTDEQ
UTDEQ
UTDEQ
UTDEQ
UTDEQ
UTDEQ
UTDEQ
UTDEQ
UTDEQ
UTDEQ
UTDEQ
N
N/A
N/A
N/A
Comment
Chapter in peer-reviewed
book
Utah State Univ, Ryan
Dupont
(continued)
-------
Table B-1. Data Quality and Conceptual Site Model Robustness Information in PVI Database (continued)
Site Name
Peace River
Fort Ware
Perth
State
AB
BC
AUWA
CSM
Rank3
3
3
3
Data
Quality
Rank"
5
4
4
Probe Type
Permanent
Permanent
Permanent
Leak
Tracer0
Y
Y
N
Purging
Procedure?
Y
Y
N
Pneu-
matic
Testing?
Y
N
N
Fixed
Gas
Data?
Y
Y
Y
Fixed Gas
andHC
Vapor
Data Con-
sistent?
Y
Y
Y
Accept-
able VOC
Analysis
Method?
Y
Y
Y
VOC Analysis
Method
TO-15
TO-15
GC/MS/TCD
Labor-
atory
QA/QC?
Y
Y
N
Peer
Review or
Research
Program
N
N
Journal
Federal or
State
Program
N/A
N/A
N/A
Comment
td
Notes
a CSM-3: Well-developed CSM, appropriately located soil gas probes, vertical soil gas profiles, well characterized contamination source (NAPL vs. dissolved) and ancillary data (e.g., soil properties)
CSM-2: Less well-developed CSM with well located probes, but with more limited soil gas locations (e.g., single location), reasonably well characterized contamination source
CSM-1: Limited data to develop CSM on and to evaluate appropriateness of soil gas probe locations, generally single soil gas location
b DQ-5: Very high quality data, QA/QC is fully documented, includes permanent probes, leak tracer and pneumatic testing, and fixed gas data that is consistent with hydrocarbon vapor concentrations. In some cases, site
data has been given a 5 ranking when not all aspects met, but when there is well-developed CSM and peer-reviewed research program.
DQ-4: High quality data, QA/QC is mostly documented, generally includes permanent probes and leak tracer testing, includes fixed gas data that is consistent with hydrocarbon vapor concentrations.
DQ-3: Moderate quality data, some QA/QC documentation, generally fixed gas data, but may be limited in quantity or inconsistent with hydrocarbon vapor concentration data.
DQ-2: Low to moderate data quality, limited QA/QC documentation (but typically still collected under state program), limited data, no fixed gas results. Minimum data quality for inclusion in database analysis.
DQ-1: Low quality data, unacceptable data quality indicators or methods. Data excluded from all analysis.
c Leak test results acceptable except as noted.
d While Maine sites did not include leak tracer testing, relatively extensive fixed gas analyses were performed. In addition, the fixed gas concentrations before and after collection of the Summa canisters were obtained and
reviewed for consistent concentrations
e One in seven samples failed the leak tracer test (iso-propanol>10,000 jjg/m3)
f Limited data (one sample).
CA DISC = California Department of Toxic Substances Control; ME DEP = Maine Department of Environmental Protection; MN PCA = Minnesota Pollution Control Authority;
EPA R3 = U.S. EPA Region 3; NJ DEP = New Jersey Department of Environmental Protection; UT DEQ = Utah Department of Environmental Quality;
Y = yes; N = no; N/A = not applicable; na = not available
-------
Table B-2. Summary of Site Information in PVI Database
Site Name
Alameda Naval
Air Station
Coachella
Huntington
Beach
Mission Valley
Terminal
Newport Beach
Port Hueneme
Former Chevron
Station #9-5669
Dave's Amoco
NYM
Jacobsen
Residence
Larsons 66
D&E Sales
Moen Oil
Johnsons Auto
Midtown Service
John's Garage
City
Alameda
Coachella
Huntington
Beach
San Diego
Newport Beach
Port Hueneme
South San
Francisco
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
Country
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
State or
Province
CA
CA
CA
CA
CA
CA
CA
MN
MN
MN
MN
MN
MN
MN
MN
MN
Contami-
nation
Source
G
G
G
G, D, J, E
G
G
G
G*
G*
G*
G*
G*
G*
G*
G*
G*
Vapor
Source
Type
LNAPL
LNAPL
LNAPL
LNAPL
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
Soil Type
Coarse
Coarse
Coarse
Coarse/Fine
Coarse
Coarse
Coarse
Coarse
Coarse
Fine
Coarse
Coarse
Coarse
Coarse
Coarse
N/A
Site Type
UST
Terminal
UST
Terminal
UST
UST
UST
UST
UST
UST
UST
UST
UST
UST
UST
UST
Building Use
Number Buildings
1
2
Residential
Smaller Commercial
0
0
Larger Commercial
0
£
11
Q_
"o
£
0>
_c
m
538
1,500-
7,200
Foundation
Type
Basement
Slab on Grade
0
0
Crawlspace or Dirt Floor
Media Sampled
Groundwater
0
0
0
0
0
0
0
0
0
0
0
0
o
Q_
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Sub-slab Vapor
0
0
'&
0
0
0
0
0
0
0
0
0
0
0
(continued)
-------
Table B-2. Summary of Site Information in PVI Database (continued)
Site Name
Buchannon
Nursing Home
Red & White
Service
Side Lake Store
Ossippe Store
AC Oil
Schmunks
Kennys Oil
Settes Garage
Tilson Auto
Rogers Mobile
Rub-a-Dub
Long Shot
Trucking
Eggens Oil
Chillum site
Reuben's Market
Cumberland
Farm 1803
City
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
Maryland
Milo
Sandford
Country
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
State or
Province
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MD
ME
ME
Contami-
nation
Source
G*
G*
G*
G*
G*
G*
G*
G*
G*
G*
G*
G*
G*
G
G
G
Vapor
Source
Type
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL
Soil Type
Coarse
Coarse
Coarse
Coarse
Coarse
Coarse
Fine
Fine
Fine
Fine
Coarse
Fine
Fine
Coarse
Coarse/Fine
Coarse
Site Type
UST
UST
UST
UST
UST
UST
UST
UST
UST
UST
UST
UST
UST
UST
UST
UST
Building Use
U)
O>
erBuildit
_Q
3
Z
1
1
ro
o
T3
•55
2
0
ro
o
E
o
o
1_
"<5
E
C/3
0
as
P
r Comme
c
^
m
N/A
N/A
Foundation
Type
"c
0
1_
0
o
u_
•c
Q
0
8
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
^_
o
Q_
£
_Q
ra
U)
0
3
C/3
oe
0
3,
0
0
0
0
0
0
0
0
0
0
0
(continued)
-------
Table B-2. Summary of Site Information in PVI Database (continued)
Site Name
Cumberland
Farm 181 7
Twin Bridge
Market
Cumberland
Farm 1806
Cumberland
Farm 1805
Cumberland
Farm 1839
Cumberland
Farm 1822
7-Eleven
Cumberland
Farm 1836
Cumberland
Farm 1829
BP Paulsboro
Hulme Street
Stafford
Mandan
City
Berwick
Leeds
South Portland
Portland
Portland
Saco
Lewiston
North Windham
Augusta
Paulsboro
Mount Holly
Stafford
Mandan
Country
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
State or
Province
ME
ME
ME
ME
ME
ME
ME
ME
ME
NJ
NJ
NJ
ND
Contami-
nation
Source
G
G
G
G
G
G
G
G
G
G
G
G
D
Vapor
Source
Type
LNAPL/Dis
LNAPL/Dis
LNAPL
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
LNAPL/Dis
Dis
LNAPL
Dis
LNAPL/Dis
LNAPL
Soil Type
V.Coarse
Coarse
Coarse
V.Coarse/Coarse
Coarse/Fine
Coarse
Coarse/Fine
Coarse
V.Coarse
Coarse
Coarse
Coarse
Fine
Site Type
UST
UST
UST
UST
UST
UST
UST
UST
UST
Terminal
UST
UST
Refinery
Building Use
Number Buildings
1
1
2
1
1
3
5
13
Residential
0
0
0
0
0
Smaller Commercial
0
0
Larger Commercial
0
0
0
Building Footprint (ft.2)
3,900
1,500-
2,000
5,000
N/A
400
600-
800
varies
Foundation
Type
Basement
0
0
0
0
0
Slab on Grade
0
0
0
Crawlspace or Dirt Floor
0
0
Media Sampled
Groundwater
0
0
0
0
0
0
0
0
0
0
0
0
0
o
Q_
£
£
0
0
0
0
0
0
0
0
0
0
0
0
0
Sub-slab Vapor
oe
0
oe
0
0
0
0
'&
0
0
0
0
0
0
0
(continued)
-------
Table B-2. Summary of Site Information in PVI Database (continued)
Site Name
BP Akron
BP Columbiana
BP Conneaut
BP Kent
Former Refinery
Beaufort
Bountiful Bicycle
Gas & Go #7
Gold Cross
Ambulance
Hal's Chevron
Handi Mart
#102 Chevron
Logan Food Mart
Price Rental
Property
Salina Cash
Saver
Jenkins Oil
Wheel-In Market
City
Akron
Columbiana
Conneaut
Kent
not specified
Beaufort
Bountiful
North Salt Lake
Salt Lake City
Green River
Midvale
Jacksons
Logan
Price
Salina
Santa Clara
Salt Lake City
Country
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
USA
State or
Province
OH
OH
OH
OH
OK
SC
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
UT
Contami-
nation
Source
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
Vapor
Source
Type
LNAPL/Dis
LNAPL
LNAPL
LNAPL
LNAPL
LNAPL
LNAPL
LNAPL/Dis
Dis
LNAPL/Dis
LNAPL
Dis
Dis
LNAPL
LNAPL
LNAPL/Dis
LNAPL
Soil Type
Coarse
Fine
Coarse
V.Coarse
Fine
Fine
Fine
Fine
Fine
Fine
Coarse
Coarse
Fine
Coarse
Coarse
Fine
Fine
Site Type
UST
Terminal
UST
UST
Refinery
UST
UST
UST
UST
UST
UST
UST
UST
UST
UST
UST
UST
Building Use
Number Buildings
1
2
2
4
1
1
1
1
1
Residential
0
0
Smaller Commercial
0
0
0
0
0
0
0
Larger Commercial
0
0
£
11
Q_
"o
£
0>
_c
m
775
1,500-
10,000
625-
2,500
N/A
N/A
N/A
2,700
1,350
N/A
Foundation
Type
Basement
0
0
Slab on Grade
0
0
0
0
0
0
0
Crawlspace or Dirt Floor
Media Sampled
Groundwater
0
0
0
0
0
0
0
0
o
Q_
0
0
0
0
0
0
0
0
0
0
0
0
0
Sub-slab Vapor
0
0
0
0
0
0
0
oe
'&
0
0
0
0
0
0
0
0
0
0
(continued)
-------
Table B-2. Summary of Site Information in PVI Database (continued)
Site Name
Teasdale
Country Store
Tesoro #40
7-Eleven #23387
Refinery Site
Peace River
Chatterton
Research Site
Fort Ware
Ottawa
North Battleford
Perth
City
Teasdale
Salt Lake
Murray
Hooven
Peace River
Delta
Fort Ware
Ottawa
North Battleford
Perth
Country
USA
USA
USA
USA
Canada
Canada
Canada
Canada
Canada
Australia
State or
Province
UT
UT
UT
-
AB
BC
BC
ON
SK
WAus
Contami-
nation
Source
G
G
G
G,D
G
BTX
D
G,D
G
K
Vapor
Source
Type
LNAPL
Dis
LNAPL
LNAPL
LNAPL
LNAPL
LNAPL
Dis
LNAPL/Dis
LNAPL
Soil Type
Coarse
Fine
Fine
Coarse
Fine/Coarse
Coarse
V. Coarse
Fine
Coarse
Coarse
Site Type
UST
UST
UST
Refinery
UST
Refinery
UST
UST
UST
Refinery
Building Use
U)
O>
erBuildit
_Q
3
Z
1
2
0
1
0
1
1
ro
o
T3
•55
2
0
0
0
ro
O
E
o
O
i_
"<5
E
C/3
0
as
P
r Comme
c
^
m
N/A
2,200
610
2,700
Foundation
Type
"c
8
0
0
0
0
0
0
0
^_
o
Q_
£
_Q
ra
U)
0
3
C/3
0
0
0
0
3,
0
0
td
For contamination type: G = gasoline, D = diesel, J = jet fuel, E = ethanol, K = kerosene; G* for MN sites inferred to be gasoline-impacted sites, but no confirmatory data provided.
Dis = dissolved. An arbitrary threshold for smaller versus larger building was set as 2,500 ft2.
-------
Table B-3. Soil Vapor Analyses in PVI Database
Site Name
Alameda Naval Air Station
Coachella
Huntington Beach
Mission Valley Terminal
Newport Beach
Port Hueneme
Former Chevron Station
#9-5669
Dave's Amoco
NYM
Jacobsen Residence
Larsons 66
D&E Sales
Moen Oil
Johnsons Auto
Midtown Service
John's Garage
Buchannon Nursing Home
Red & White Service
Side Lake Store
Ossippe Store
City
Alameda
Coachella
Huntington
Beach
San Diego
Newport Beach
Port Hueneme
South San
Francisco
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
Benzene
2
27
26
84
20
0
27
4
4
4
4
4
4
4
10
4
4
4
5
4
Toluene
0
0
0
0
0
0
18
4
4
4
4
4
0
0
0
4
0
4
2
0
Ethylbenzene
0
0
0
0
0
0
18
4
4
4
4
4
0
0
0
4
0
4
2
0
|
-------
Table B-3. Soil Vapor Analyses in PVI Database (continued)
Site Name
Ossippe Store
AC Oil
Schmunks
Kennys Oil
Settes Garage
Tilson Auto
Rogers Mobile
Rub-a-Dub
Long Shot Trucking
Eggens Oil
Chillum site
Reuben's Market
Cumberland Farm 1803
Cumberland Farm 1817
Twin Bridge Market
Cumberland Farm 1806
Cumberland Farm 1805
Cumberland Farm 1839
Cumberland Farm 1822
7-Eleven
Cumberland Farm 1836
Cumberland Farm 1829
City
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
Maryland
Milo
Sandford
Berwick
Leeds
South Portland
Portland
Portland
Saco
Lewiston
North Windham
Augusta
Benzene
4
4
4
4
4
3
3
4
4
3
18
7
3
6
4
1
28
9
2
17
5
5
Toluene
0
0
0
0
0
0
3
0
2
1
0
7
3
6
4
1
28
9
2
17
5
4
Ethyl benzene
0
0
0
0
0
0
3
0
2
3
0
7
3
6
4
1
28
9
2
17
5
5
|
S1
o
0
0
0
0
0
0
0
0
0
0
0
7
5
6
4
1
28
9
2
17
5
5
c
o
_Q
(3
O
0
0
0
0
0
0
0
0
0
0
0
7
5
6
4
1
28
9
2
17
5
5
Methane
0
0
0
0
0
0
0
0
0
0
0
7
5
4
4
1
28
9
2
17
4
5
(continued)
-------
Table B-3. Soil Vapor Analyses in PVI Database (continued)
Site Name
Mandan
BP Paulsboro
Hulme Street
Stafford
BP Akron
BP Columbiana
BP Conneaut
BP Kent
Former refinery
Beaufort
Bountiful Bicycle
Gas & Go #7
Gold Cross Ambulance
Hal's Chevron
Handi Mart
#102 Chevron
Logan Food Mart
Price Rental Property
Salina Cash Saver
Jenkins Oil
City
Mandan
Paulsboro
Mount Holly
Stafford
Akron
Columbiana
Conneaut
Kent
OK (not
specified)
Beaufort
Bountiful
North Salt Lake
Salt Lake City
Green River
Midvale
Jacksons
Logan
Price
Salina
Santa Clara
Benzene
4
11
8
14
9
4
3
4
4
9
14
15
5
127
0
55
1
1
14
63
Toluene
0
11
8
14
0
0
0
0
0
6
14
15
0
14
0
44
0
1
14
63
Ethyl benzene
4
11
8
14
0
0
0
0
0
3
14
15
0
14
0
44
0
1
14
63
|
$
o
0
18
2
4
9
4
3
4
4
9
6
15
0
68
3
56
0
0
14
63
c
o
_Q
(3
O
0
18
1
0
9
4
3
4
3
9
7
15
0
66
3
56
0
0
14
62
Methane
2
7
0
0
0
0
0
0
2
0
0
15
0
31
0
56
0
0
0
63
(continued)
-------
Table B-3. Soil Vapor Analyses in PVI Database (continued)
Site Name
Wheel-In Market
Teasdale Country Store
Tesoro #40
7-Eleven #23387
Refinery Site
Peace River
Chatterton Research
Site
Fort Ware
Ottawa
North Battleford
Perth
Number Sites with Data
Total Number Analyses
City
Salt Lake City
Teasdale
Salt Lake
Murray
Unknown
Peace River
Delta
Fort Ware
Ottawa
North
Battleford
Perth
-------
[This page intentionally left blank.]
B-16
-------
Appendix C. Analysis of Australian PVI Database
C.I Introduction
Wright (2011, 2012) compiled a database of paired petroleum hydrocarbon soil vapor and
groundwater source concentrations over 3 years from sites in Australia, as summarized in Table
C-l. The data were obtained from 124 sites across Australia as geographically located in Figure
C-l. The contents of the Australian database are summarized in Table C-l.
Table C-1. Summary of petroleum hydrocarbon data within Australian database
Information
sources
Site types
Data reported
Geographic
Locations
Contamination
Sources
Soil types
1083 paired soil vapor and groundwater source concentrations from 124 sites
Vapor data from single and nested wells, 12% of which are sub-slab
120 USTs, 1 refinery, 2 pipeline leak sites, 1 fuel terminal
Benzene, toluene, ethylbenzene and xylenes (BTEX); volatile total petroleum hydrocarbon
fractions (TPH); and hexane (soil vapor only); measurements of oxygen, carbon dioxide, and
methane in the soil profile; source type and site characteristics include type of surface cover
(building, pavement, ground)
43% of sites from Victoria (VIC), 29% from New South Wales (NSW), 10% from South Australia
(SA), 7% from Australian Capital Territory (ACT), 4% from Queensland (QLD), 3% from Western
Australia (WA), and remaining 4% from Tasmania (TAS) and the Northern Territory (NT)
Dissolved phase (28% data) and NAPL (72% data) from petrol and/or diesel sources
The dominant soil types comprise sand (13%), clay (42%), and sandy clay (45%).
41% of all data sites include groundwater within a fractured rock system and where at least one
soil vapor probe was completed within fractured rock.
-' .
Northern
Territory
Western
Australia
Queensland
5
South
Australia
13
New South Wales
36
ACT g
Figure C-1. Geographical Distribution of Sites in Australian Database.
C-l
-------
C.2 Climatic Information
Victoria (VIC): The majority of the sites are located in suburbs of Melbourne, the capital city of
Victoria. Temperatures in Melbourne range from an average of 43 to 55°F (6 to 13°C) in winter
to 54 to 79°F (12 to 26°C) in summer. Melbourne can experience extreme heat, because of the
movement of hot dry air from central Australia, resulting in temperatures above 104°F (40°C).
Mean rainfall in Melbourne is 26 inches (650 mm) per year.
New South Wales (NSW): The majority of the sites are located in suburbs of Sydney, the capital
city of NSW. Sydney's weather is temperate, with the city's temperatures moderated by its
location close to the ocean. Temperatures in Sydney range from an average of 8 to 16°C in
winter to 43 to 77°F (13 to 25°C) in summer. Sydney's weather can be influenced by the
movement of warm/hot moist air from the north (Queensland) or from the movement of hot dry
air from central Australia. Mean rainfall in Sydney is 48 inches (1,213 mm) per year. Regional
areas of NSW experience hotter, drier conditions than Sydney.
Australian Capital Territory (ACT): The ACT, which largely comprises Canberra and urban
suburbs, is located within NSW, inland and south of Sydney. The climate has warm to hot
summers and cool to cold winters. Temperatures in Canberra range from an average of 32 to
55°F (0 to 13°C) in winter to 46 to 82°F (8 to 28°C) in summer. Mean precipitation in Canberra
is 24 inches (616 mm) per year.
Queensland (QLD): The majority of the sites are located in suburbs of Brisbane, the capital city
of QLD, located in the southeast corner of the state. Brisbane has a humid subtropical climate
with warm to hot humid summers and dry to moderately warm winters. Temperatures in
Brisbane range from an average of 48 to 70°F (9 to 2PC) in winter to 64 to 86°F (18 to 30°C) in
summer. Mean rainfall in Melbourne is 39 inches (986 mm) per year.
South Australia (SA): The majority of the sites are located in suburbs of Adelaide, the capital
city of SA. Adelaide has a hot Mediterranean climate resulting in cool, wet winters and hot, dry
summers. Temperatures in Adelaide range from an average of 45 to 59°F (7 to 15°C) in winter to
57 to 84°F (14 to 29°C) in summer. Adelaide can experience extreme heat and extended heat
wave conditions, because of the movement of hot dry air from central Australia, resulting in
temperatures above 104°F (40°C). Adelaide is the driest capital city in Australia, with highly
variable and unreliable rainfall. Mean rainfall in Adelaide is 21 inches (546 mm) per year. North
of Adelaide, in regional inland areas, the climate is hotter and drier.
Western Australia (WA): The majority of the sites are located in suburbs of Perth, the capital
city of WA. Perth has a temperate to Mediterranean climate resulting in cool, wet winters and
hot, dry summers. Temperatures in Perth range from an average of 45 to 57°F (7 to 14°C) in
winter to 61 to 88°F (16 to 31°C) in summer. Mean rainfall in Perth is 33 inches (850 mm) per
year. East, north, and northeast of Perth are drier desert areas that experience a more significant
and distinctive wet (winter) season and dry (summer) season.
The database includes some areas of Australia that are dry:
• Loxton, SA—mean annual rainfall is 10 inches (264 mm);
C-2
-------
• Barham NSW—mean annual rainfall is 15 inches (373 mm); and
• Cleve SA—mean annual rainfall is 16 inches (400 mm).
Note that from 2000 to 2010 Australia was in drought with most locations experiencing rainfall
much lower than average. All data in the Australian PVI database were collected between 2004
and 2010.
C.3 Data Evaluation and Filters
The criteria and filters applied by Wright (2011, 2012) for the data evaluation were
• presence of overlying clean soil,
• determination of the presence of NAPL,
• conceptual model qualifier, and
• data quality (DQ) qualifier.
The criteria used for determining the presence of overlying clean soil were the following:
• Available data do not suggest the presence of petroleum contamination in overlying
soil.
• Insufficient data are available to determine whether the overlying soil is
contaminated, so these locations may be affected by petroleum hydrocarbons in
overlying soil.
• Overlying soil is known to be affected by the presence of petroleum hydrocarbons or
the location is directly adjacent to operational fuel infrastructure where impacts are
present.
The following criteria were used to determine the presence of NAPL:
• Field observations noted as phase-separated hydrocarbon, free product or a sheen;
• Near effective solubility: consisting of benzene groundwater concentrations greater
than 3 to 10 mg/L, sum of BTEX greater than 20 mg/L (for petrol (gasoline) sites);
TPH (C6-14) greater than 30 mg/L (for petrol sites; value selected to account for
BTEX up to 20 mg/L and aliphatics concentrations of approximately 10 mg/L); and
TPH(C10-14) greater than 5 mg/L (for diesel sites);
• Methane in soil gas greater than 10 percent;
• Aliphatics (hexane) in soil gas >1,000 |J,g/m3; and
• Aliphatics (TPH) in soil gas >50,000 |J,g/m3 (where no hexane is reported).
Data characterized as LNAPL based solely based on elevated aliphatics in soil gas were not
included in the assessment where source concentrations are considered because these sites have
poor source (groundwater) data.
C-2
-------
The CSM Qualifier consisted of a five-point ranking (5-highest, 1-lowest) based on the following
criteria:
• 3-5: Data collected from appropriate locations on the site using nested soil gas (SG)
wells (preferred) or single point wells that include locations close to the source and
through the profile, with the further ranking between 3 and 5 based on quantity and
other factors (e.g., soil property data, hydrogeologic data).
• 2: Data collected from appropriate locations on the site using single or nested SG
wells that do not extend close enough to the source to enable an understanding of
subsurface attenuation.
• 1: Data collected either from inappropriate locations on the site or from depths that
are so shallow that no understanding of subsurface attenuation can be determined.
The DQ Qualifier consisted of a five-point ranking (5 = highest, 1 = lowest) based on the
following criteria:
• 3-5: Data quality considered to be moderate to high based on appropriate methods,
passing of tracer tests, appropriate documentation, appropriate QA/QC, and the
dataset included the measurement of fixed gases (some allocated level 3 lacked fixed
gases data).
• 2: Data quality was considered low because of limited QA/QC, limited analysis of
VOCs and/or no fixed gases.
• 1: Data quality was considered poor because of failure of the tracer test, breakthrough
of sample tubes, incorrect/inappropriate methodology used, no QA/QC, and/or
no/poor documentation.
C.4 Data Importation and Compilation Process
The Australian data received were incorporated into the PVI database but were analyzed
separately from the U.S. sites in this study. After importation into the Access version of the PVI
database, the Australian data were examined with respect to internal consistency and
inconsistencies and errors were corrected through communications with the database originator.
The data were then were exported in spreadsheet format for the analyses discussed in this
appendix.
C.5 Data Analysis
The Australian data analysis was conducted for dissolved and LNAPL sites, and different filters
were applied as described below. All analyses were conducted with data removed where there
was known petroleum hydrocarbon contamination in overlying soil. The data analyses scenarios
were as follows:
1. Scenario 1: All data (excluding where known petroleum contamination in soil).
2. Scenario 2: Fractured bedrock and non-UST sites filtered out.
C-4
-------
3. Scenario 3: Fractured bedrock, non-UST sites, and data with lower quality (DQ-1)
and confidence (CSM-1 and 2) filtered out.
Scenario 3 is considered most appropriate for use in estimating exclusion criteria.
The vertical distance method was employed to evaluate vertical separation distances in relation
to toxicity-based thresholds as described in the main body of this report for the above scenarios.
For Scenario 3, a further analysis was conducted to evaluate the effect on surface cover on the
results.
C.6 Data Results
The data analysis for key parameters (benzene, hexane, xylenes) is presented in Figures C-2
through C-4 for dissolved sites and Scenarios 1 through 3 and Figures C-5 through C-7 for
LNAPL sites and Scenarios 1 through 3. In addition, for benzene, the probability that the soil
vapor concentration is less than a defined threshold (50 and 100 ng/rn3) for varying source-
separation distances was estimated. For the benzene analysis, non-detects were addressed by the
common practice of substituting half the detection limit for non-detects.
The results for Scenario 3 (scenario appropriate for exclusion distance analysis) are summarized
in Table C-2. The results indicate significant differences between dissolved and LNAPL sites
and for different scenarios. The probability of benzene vapor concentration less than 100 ng/m3
for dissolved sites is 93 percent at 0 ft separation, increasing to 95 percent at 5 ft (1.5 m); while
the probability of benzene vapor concentration less than 100 ng/m3 for NAPL (UST) sites is
66 percent at 0 ft, increasing to 94 percent at 15 ft (4.6 m) separation.
A comparison of the U.S. EPA to Australian database indicates slightly lower probabilities for
both dissolved and NAPL sites (1 to 5 percent) for the Australian database for equivalent source-
separation distances. For NAPL sites, the oxygen concentrations were also depleted (less than
1 percent) for greater source-separation distances.
The analysis of the effect of surface cover on conditional probabilities for benzene
concentrations to exceed the 100 |J,g/m3 threshold and oxygen concentrations versus separation
distance did not suggest surface cover (building, pavement) resulted in an oxygen shadow. The
analysis also showed reduced attenuation for buildings and pavement compared with the ground-
cover scenario, although it is recognized that the number of sub-slab data points below buildings
were limited (Figure C-8).
C-5
-------
Table C-2. Summary of Results for Vertical Distance Method for Scenario 3
Oxygen
Benzene
(1 00 |jg/m3 threshold)
Benzene
(50 |jg/m3 threshold)
Xylenes
Hexane
Dissolved Source— UST Sites
Most O2 cone. > 4%, and no O2 < 0.5%
POOL > 93% for 0 ft separation increasing
to 95% at 5 ft (1.5m)
Pi/2Di_ > 90% for 0 ft separation increasing
to 95% at 5 ft (1.5 m)
All vapor concentrations < RBCV at 0 ft
All vapor concentrations < RBCV at 0 ft
LNAPL Source— UST Sites
Many data points with O2 < 4%, and
O2 < 1% to 25 ft (7.6 m) separation
Pi/2Di_ > 66% for 0 ft separation increasing
to ~ 94% for 15-ft (4.6-m) separation
Pi/2Di_ > 60% for 0 ft separation increasing
to ~ 88% for 15-ft (4.6-m) separation
Vapor concentrations > RBCV for separation
distance up to 12 ft (3.7 m)
Vapor concentrations > RBCV for separation
distance up to 12 ft (4 m)
PIGDL = Probability estimated using half detection limit method for exceeding
threshold.
C-6
-------
1 E 1 03 i
¥ V
c -^ p_+02 «H
- P-
§ I.
t 1 F+01 ••
1
i
Benzene vs. Distance - Dissolved
*>DL (N=75)
100 Hg/m3 ~' DL (N=35)
rmr» n D
§
(j
n
0
TPH vs. Oxygen
A v
• ^ "%4 + ^
n n 4. 4.
«>DL (N=112)DDL (N=98)
_______ — ^- _ _ _ _ .
D
-------
1 £+03 | —
Benzene vs. Distance - Dissolved
«>DL (N=64)
? ff 100 Hg/m3 n ^^
P n t *
F n n
0 20 40 60 80
Distance between Soil Vapor Probe and Contamination (ft)
1 E i O"1 *
J
E i p+m H
u
I S
B
5 i F+nn
X l.t+UU j+-
D
1 ^ 01 - ^-
Hexanevs. Distance- Dissolved
*>DL (N=31)
Hill | ^ DDL (N=93) DDL(N =77)
r\Dt— y
D
-------
1 Ei 03
*P 1 F+n? 1
zene Vapor Cone (jigy
•* h*
n m
+
3 2
Benzene vs. Distance - Dissolved (UST)
*>DL (N=56)
* 100ng/m3 C DL (N=28)
-WfWII% INHI 1 1 1 H * _D
n n ^^
n n a n n n
«>DL (N=84)D88
I 86
I 84
* 82
80
C
l.E+04
l"
c l.E+02
3
1
§ l.E+01
i
0)
>• l.E+00
Benzene Conditional Probability -
Dissolved (UST)
7 •
~f_
*^=i
^^ f /
^^^^ S/ » Probability <
k^^^ / —•— Probability <
100 (1/2 DL)
50 (1/2 DL)
1
12345
Distance between Soil Vapor Probe and Contamination (ft)
Xylenes vs. Distance - Dissolved (UST)
| RBCV
V7
R:QQ
w Dn
':
»>DL (N=69|
O
-------
Benzene vs. Distance - NAPL
Benzene Conditional Probability- NAPL
Probability < 100 (1/2 DL)
Probability < 50 (1/2 DL)
0 20 40 60 80
Distance between Soil Vapor Probe and Contamination (ft)
0 5 10 15 20 25 30
Distance between Soil Vapor Probe and Contamination (ft)
Hexanevs. Distance- NAPL
Xylenes vs. Distance - NAPL
0 20 40 60 80
Distance between Soil Vapor Probe and Contamination (ft)
0 20 40 60 80
Distance between Soil Vapor Probe and Contamination (ft)
TPH vs. Oxygen-NAPL
Oxygen vs. Distance- NAPL
10 15
Oxygen (% vol/vol)
0 20 40 60 80
Distance between Soil Vapor Probe and Contamination (ft)
Figure C-5. Vertical distance methods for NAPL Sites—all data (Scenario 1)
C-10
-------
Benzene vs. Distance - NAPL
O
o
Benzene Conditional Probability- NAPL
OO
o
-vI
o
ty V
5 6°
0 20 40 60 80
Distance between Soil Vapor Probe and Contamination (ft)
50
Probability < 100 (1/2 DL)
Probability < SO (1/2 DL)
0 5 10 15 20 25 30
Distance between Soil Vapor Probe and Contamination (ft)
Hexanevs. Distance- NAPL
Xylenes vs. Distance - NAPL
0 20 40 60 80
Distance between Soil Vapor Probe and Contamination (ft)
TPH vs. Oxygen - NAPL
0 20 40 60 80
Distance between Soil Vapor Probe and Contamination (ft)
Oxygen vs. Distance- NAPL
10 15
Oxygen (% vol/vol)
20
0 20 40 60 80
Distance between Soil Vapor Probe and Contamination (ft)
Figure C-6. Vertical distance method for NAPL LIST sites—fractured bedrock and non-UST sites
filtered out (Scenario 2).
C-ll
-------
Benzene vs. Distance - NAPL (UST)
Benzene Conditional Probability - NAPL (UST)
l.E-01
0 20 40 60 80
Distance between Soil Vapor Probe and Contamination (ft)
5 10 15 20 25 30
Distance between Soil Vapor Probe and Contamination (ft
1.E+'
Hexanevs. Distance- NAPL(UST)
Xylenes vs. Distance - NAPL (UST)
l.E-01
l.E-01
0 20 40 60 80
Distance between Soil Vapor Probe and Contamination (ft)
0 20 40 60 80
Distance between Soil Vapor Probe and Contamination (ft)
TPH vs. Oxygen - NAPL (UST)
Oxygen vs. Distance- NAPL (UST)
10 15
Oxygen (% vol/vol)
0 10 20 30 40
Distance between Soil Vapor Probe and Contamination (ft)
Figure C-7. Vertical distance method for NAPL UST sites—fractured bedrock, non-UST sites and
lower data quality and confidence (DQ-1 and CSM-1&2) filtered out (Scenario 3)
C-12
-------
Benzene Probability Different Surface Cover - NAPL (UST)
100
Probability < 100 u.g/m3 - Building Scenario
Probability < 100 u,g/m3 - Ground Cover Scenario
Probability < 100 ug/m' - Pavement Scenario
0 5 10 15 20 25
Distance between Vapor Probe and Contamination (ft)
30
Oxygen for Different Surface Cover - NAPL (UST)
- Building Scenario (N = 5)
• O2 -Ground Scenario (N = 98)
O2 - Pavement Scenario (N = 36)
10 20 30 40
Distance between Vapor Probe and Contamination (ft)
Figure C-8. Comparison of probability for benzene soil vapor concentrations to be less than
100 ug/m3 threshold (top panel) and oxygen concentrations for different surface covers
for NAPL UST sites (bottom panel). Below detection limit concentrations replaced with
half the detection limit for analysis.
C-13
-------
References
Wright, J. 2011. Establishing Exclusion Criteria from Empirical Data for Assessing Petroleum
Hydrocarbon Vapour Intrusion. Program and Proceedings of the 4th International
Contaminated Site Remediation Conference—2011 CleanUP. Adelaide, South Australia,
September 11-15. p. 142-143.
Wright, J. 2012. Evaluation of the Australian Petroleum Hydrocarbon VI Database: Exclusion
Criteria. Presented at: Recent Advances to VI Application & Implementation—A State-
of-the-Science Update. AEHS West Coast Conference. San Diego, CA. March. Available
at
https://iavi.rti.orgAVorkshopsAndConferences.cfm?PageID=documentDetails&AttachID
=549.
C-14
-------
Appendix D. PVI Database Data Dictionary
Attachment D-1. Data Structure as of Nov 30, 2012
Field Name
Type
Size
Description
Table: Building_Distances
buildingjd
locationjd
horz_dist_to_bldg
horz_dist_to_bldg_unit
horz_dist_comment
time_stamp
Long Integer
Long Integer
Double
Text
Text
Date
4
4
8
20
255
8
Linkto Buildings table
Link to Locations table
Closest horizontal distance from sampling location to
building (not applicable for indoor samples)
Measurement unit for horizontal distance from sample
location to building
Comment about building-to-location link
Date/time record was created
Table: Buildings
buildingjd
orig_bldg_id
sitejd
bldg_name
bldg_type
bldg_use
footprint_area
footprint_area_unit
foundationjype
fnd_depth_to_base
fnd_depth_to_base_unit
bldg_comment
time_stamp
Long Integer
Text
Long Integer
Text
Text
Text
Double
Text
Text
Double
Text
Memo
Date
4
50
4
60
50
50
8
20
50
8
20
0
8
Building identifier (aka, subsite)
Original ID number for building
Linkto Sites
Name of building
Physical description of building (links to
lt_Building_Types)
Use of building (residential, commercial, industrial,
school, etc.)
Footprint area of the building
Unit of measurement for footprint_area
Type of building foundation (lookup values in
lt_Foundation_Types)
Depth to base of foundation (below ground surface)
Unit of measurement for depth to base of foundation
Comment field
Date/time record was created
Table: Data_Provider
data_provider_id
data_provider
data_contact_name
data_contact_address1
data_contact_address2
data_contact_city
d ata_co ntact_state
data_contact_zipcode
data_contact_email
data_contact_phone
Long Integer
Text
Text
Text
Text
Text
Text
Text
Text
Text
4
60
30
40
40
20
2
10
60
60
Unique ID for data provider; links to Site table
Company, agency, or individual responsible for
submittal of PVI data
Name of contact associated with data_provider
Contact street address and/or box number
Site address, part two. Box number or other info.
City
Postal abbreviation for State
Zip code
Contact e-mail address
Contact phone number
D-1
-------
Field Name
time_stamp
Type
Date
Size
8
Description
Date/time record was created
Table: Documents
doc_id
sitejd
ref_id
doc_name
file_name
doc_year
doc_desc
doc_date
doc_source
author_org
author_citation
author_name
Journal
Volume
Pages
author_phone
docjinks
doc_comments
public_yn
doc_original_format
date_QC_completed
time_stamp
Long Integer
Long Integer
Long Integer
Text
Text
Text
Text
Date
Text
Text
Text
Text
Text
Text
Text
Text
Memo
Memo
Text
Text
Date
Date
4
4
4
255
255
4
255
8
255
100
255
50
255
20
20
20
0
0
1
100
8
8
Unique ID for document
Link to Sites table
Link to Reference table
Document descriptive name
physical file name
Document year (for bibliography)
Description for the document
Document creation date
Document source
Document author's organization
Author name in citation formats
Document author's name
Journal name in citation format
Journal volume in citation format
Journal pages in citation format
Document author's phone number
Website address (i.e., URL) for documents available
on the Internet
Other comments about the document (e.g., use,
applicability)
Can this information be made available to the public?
(Y=yes, N=no)
Original format of document
Date that QC was completed
Date/time record was created
Table: DQ Table
—
sitejd
CSM_Rank
DQ_Rank
Probe_Type
Leak_T racer
Purging_Procedure
Pneumatic_Test
Fixed_Gas_Data
FG-HC_Consistent
Acceptable_VOC_Method
Long Integer
Integer
Integer
Text
Text
Text
Text
Text
Text
Text
4
4
4
25
3
3
3
3
3
10
Unique ID for site
Conceptual site model rank (1-3), 3 is best (see
lt_CSM_Rank table)
Data quality rank (1-5), 5 is best (see It DQ rank
table)
Type of soil gas probe (permanent, driven, unknown)
Was leak tracer test conducted? (Y/N) NOTE: ME
used extensive fixed gas data instead.
Was the probed purged before sampling? (Y/N)
Was a pneumatic test conducted? (Y/N)
Was fixed gas data collected? (Y/N)
Are the fixed gas and VOC data consistent? (Y/N)
Is the VOC method acceptable? (Y/N)
D-2
-------
Field Name
VOC_Method
Lab_QAQC
Peer_Review
Reg_Program
DQ_Comment
Type
Text
Text
Text
Text
Text
Size
25
3
25
25
255
Description
VOC analytical method (TO-15, EPA 8260, GC, PID,
FID, TCD, unknown, etc.)
Is laboratory QA/QC documented and acceptable?
(Y/N)
Was study peer reviewed? (journal, thesis, state/fed.
research, N)
Regulatory program oversight of study (reg. program,
NA)
Comment on data quality
Table: Links
LinkJD
Iocation_xy_id1
Iocation_xy_id2
distance_xy
distance_xy_units
Long Integer
Long Integer
Long Integer
Double
Text
4
4
4
8
10
Unique ID for the links of the specificed two
location_xy_id.
First location of the linkage: probe location_xy_id
Second location of the linkage: non-probe
location_xy_id
Lateral distance between the two linked xy locations
Unit of the lateral distance
Table: Locations
samplejocationjd
site_id
buildingjd
location_xy_id
importjocjd
loc_name
samp_loc_name
sample_depth
sample_depth_unit
loc_type
locjnt/ext
loc_desc
vz_soil_text_code
vz_alt_soil_desc
vz_alt_soil_desc_src
fractured_rock
vz_alt_soil_grade
vz_porosity
Long Integer
Long Integer
Long Integer
Long Integer
Text
Text
Text
Double
Text
Text
Text
Text
Text
Text
Text
Text
Text
Double
4
4
4
4
25
100
255
8
20
50
20
255
10
255
255
20
255
8
Location ID where sample was taken at 3 D level
(including depth z)
Link to Sites table
Linkto Buildings table
Unique ID for each location at 2D level, used for linking
locations laterally in the Links table
3D location XY ID assigned by RTI for import
(Example: M129-M130, M207, M208)
Location xy name at 2D level
More specific name of sampling location (e.g., port A
on Probe SV-2 or SV-2a); may vary by depth
Sample depth, below land surface
Unit of measurement for sample depth
Location type (Indoor air, outdoor air, probe, bulk soil,
or well)
Interior or exterior location
Additional description of location (e.g., floor,
designated use of room)
Vadose zone soil texture (Links to lt_Soil_Textures) in
code
Alternate soil description (may be more specific than
vz_soil_txt_code)
Description of soil between the sampling point and the
source
Indicates if there is fractured rock
Site soil gradation (V.Coarse, Coarse, Fine)
Vadose zone porosity
D-3
-------
Field Name
vz_porosity_unit
loc_comment
time_stamp
Table: lt_Building_Types
bldg_type
Type
Text
Memo
Date
Text
Size
20
0
8
50
Description
Unit of measurement for vadose zone porosity
Comment about latitude, longitude and vertical
elevation. Store information about the collection
method, post processing of the data (if GPS were
involved), or description of feature of the facility
represented by the coordinates.
Date/time record was created
Physical description of building
Table: ItjCountries
Country
country_name
Text
Text
5
25
Country short name
County name
Table: lt_CSM_Rank
CSM_Rank
CSM_Description
Integer
Text
4
255
Conceptual site model rank (1-3), 3 is best
Conceptual site model rank description
Table: lt_DQ_Rank
DQ_Rank
DQ_Description
Integer
Memo
4
0
Data quality rank (1-5), 5 is best
Data quality rank description
Table: lt_Foundation_Types
foundation_type
Text
50
Building foundation types (lookup values for Buildings
table)
Table: lt_Hydrogeologic_Settings
hydro_setting_desc
Text
255
General Hydrogeologic setting description
Table: lt_Parameters
parameterjd
parameter_abbrev
parameter_name
cas_number
parameter_class
organic_yn
HLC25
DeltaH
Tc
Tb
Comment
sort_name
Long Integer
Text
Text
Text
Text
Text
Double
Double
Double
Double
Text
Text
4
10
50
15
50
1
8
8
8
8
255
50
Unique ID for each measurement parameter
Short abbreviation for measurement parameter (e.g.,
MEK, BP)
Measurement parameter name (e.g., 2-butanone,
barometric pressure)
Chemical Abstract System number (where applicable)
Parameter class or grouping
Must be "Y" for organic constituents or"N" for
inorganic constituents
Henry's Law Constant at 25 degrees C (unitless)
Enthalpy of vaporization at the normal boiling point
(cal/mol)
Critical temperature (degrees Kelvin)
Normal boiling point (degrees Kelvin)
parameter name used for sorting
Table: lt_Sample_Media
Media
Text
50
Media sample type
D-4
-------
Field Name
Type
Size
Description
Table: lt_Soil_Textures
soil_txt_code
soil_txt_name
soil_txt_desc
Text
Text
Text
10
50
255
soil texture code (links to Locations table)
soil texture name
Description of soil texture from IAVI guidance
document Table 4, p. 35
Table: lt_Stat_Types
stat_type
Text
20
Statistic type
Table: lt_States
state_fips
state_name
state_abbrev
Text
Text
Text
2
50
4
State fips code
State name
State abbreviation
Table: ItJJnits
unit_type
unit_code
unit_desc
unit_pref
Text
Text
Text
Boolean
20
20
100
1
Type or category for which the units are applicable
(used to limit list in forms)
Reported unit (abbreviation)
Description of unit (unabbreviated)
Indicates which is the preferred unit for the unit_type
(used for setting default value)
Table: References
ref_id
References_text
time_stamp
Long Integer
Text
Date
4
255
8
Unique id for references
Description for the document
Date/time record was created
Table: Results
test_result_id
import_result_id
samplejd
parameter_id
parameter_name
result_value
result_unit
result_comment
lab_anl_method_code
report_detection
detect_flag_yn
value_type
Long Integer
Text
Long Integer
Long Integer
Text
Double
Text
Text
Text
Text
Text
Text
4
25
4
4
50
8
15
255
35
20
1
12
Unique ID for test result
Result ID assigned by RTI for data imports
Sample ID that this test result is for - linked to Samples
Link to lt_Parameters. Measurement parameter that
result measures.
Measurement parameter name (e.g., 2-butanone,
barometric pressure)
Analytical result, field measurement, or statistical
calculation
Units of measurement for the result (and
result_error_delta)
Result-specific comments
Laboratory analytical method code
report detection limit
Must be either "Y" for detected analytes or "N" for
non_detects
Value type for result_value ("actual", "estimated",
"interpolated", or "calculated").
D-5
-------
Field Name
stat_type
st at_o b s_d ate_f i rst
stat_obs_date_last
test_result_comment
fixed_gas_method
time_stamp
Type
Text
Date
Date
Memo
Text
Date
Size
20
8
8
0
255
8
Description
Statistic type reflected in the result value (links to
lt_Stat_Types)
Earliest date of sample used to determine result_value
Latest date of sample used to determine result_value
Comment field
Method for fixed gases
Date/time record was created
Table: Samplejinks
Sample_link_ID
sample_id_sg_in/outdoor_air
sample_id_gw
sample_id_soil
Long Integer
Long Integer
Long Integer
Long Integer
4
4
4
4
Unique ID for each sample link; used to pair samples
Sample ID for soil gas, indoor air, or outdoor air
sample
Sample ID for groundwater sample
Sample ID for soil sample
Table: Samples
samplejd
samplejocationjd
original_sample_id
sample_medium
sample_start_date
sample_comment
time_stamp
Headspace_yn
gw_temp
gw_temp_units
ground_cover
leak_test_yn
vz_moisture_content
vz_moisture_content_unit
Soil_TPH_paired_result_value
Soil_TPH_paired_result_unit
sample_DQ
sample_confidence
Long Integer
Long Integer
Text
Text
Date
Memo
Date
Text
Double
Text
Text
Text
Double
Text
Double
Text
Text
Text
4
4
40
20
8
0
8
1
8
10
50
1
8
20
8
15
20
20
Unique ID for each sample
Location ID where sample was taken at 3 D level
(including depth z)
Sample ID in original source
Medium within which measurement was taken (links to
lt_Sample_Media)
Date sample collection began in (MM/DD/YYYY)
format
Comments related to the sample
Date/time record was created
Soil sample only: is this a headspace
measurement?(Y, N)
Groundwater samples only: groundwater temperature
at time of sampling
Groundwater samples only: units for groundwater
temperature
Soil Gas samples only: Surface cover (paved, grassy,
etc.)
Soil Gas samples only: Has the vapor probe been leak
tested? (Y, N)
Vadose zone moisture content (measured value)
Unit of measurement for vadose zone moisture content
Paired soil TPH analytical results (links by 3D location
and sample date)
Units of measurement for the paired soil TPH result
(and result_error_delta)
Sample data quality (1 to 5): 1 = poor quality and 5 =
high quality (Aus. data)
Conceptual site model qualifier confidence (1 to 5): 1 =
inappropriate locations and 5 = appropriate locations
(Aus. data)
D-6
-------
Field Name
Type
Size
Description
Table: Sites
sitejd
data_provider_id
original_site_id
site_city
s ite_state_a b b rev
site_country
site_hydrology
site_vapor_src_type
site_vapor_src_origin
public_yn
time_stamp
Long Integer
Long Integer
Text
Text
Text
Text
Text
Text
Text
Text
Date
4
4
45
20
4
50
255
50
255
1
8
Unique ID for site
Links site to the Data_Provider table.
Site identifier in original source
City of site
State abbreviation for State of site (links to lt_States)
Country name (links to lt_Countries)
Hydrogeologic Setting (links to
lt_Hydrogeologic_Settings)
Type of contamination (e.g., gasoline)
Origin of the vapor source (UST, spill, landfill, etc.)
Can this information be made available to the public?
(Y=yes, N=no)
Date/time record was created
Table: Sources
sourcejd
samplejd
site_id
NAPL_direct_indication
NAPL_reported
NAPL_inferred_prox
NAPL_inferred_other
NAPL inferred other comme
nt
depth_to_water
depth_to_water_unit
depth_to_src
depth_to_src_unit
Source_type_calculated
Thickness_Clean_Soil_Benze
ne_100_ug/m3
Thickness_Clean_Soil_Benze
ne _100_ug/m3 (Less than)
Thickness_Clean_Soil_Benze
ne_100_ug/m3 (Both)
Thickness_Clean_Soil_Benze
ne _100_ug/m3 (Refined
estimate)
Thickness_Clean_Soil_Benze
ne _100_ug/m3 (Greater than)
Thickness_unit
Long Integer
Long Integer
Long Integer
Boolean
Boolean
Boolean
Boolean
Text
Double
Text
Double
Text
Text
Double
Double
Double
Double
Double
Text
4
4
4
1
1
1
1
255
8
20
8
20
255
8
8
8
8
8
20
Unique ID for sources
Linkto Sample table
Link to Sites table
NAPL based on direct indication
NAPL based on report
NAPL inferred from proximity
NAPL inferred from other references
Comments on NAPL inferred from other references
Depth to water table (below land surface) at time of
sampling
Unit of measurement for depth to water
Depth to vapor source at time of sampling (= GW
depth for GW samples)
Unit of measurement for depth to src
Source type (NAPL or Dissolved)
Method 1: Thickness Clean Soil Benzene 100 |ig/m3
Criteria
Method 1: Thickness Clean Soil Benzene 100 |ig/m3
Criteria (less than)
Method 1: Thickness Clean Soil Benzene 100 |ig/m3
Criteria (both)
Method 2: Thickness Clean Soil Benzene 100 |ig/m3
Criteria (refined estimate)
Thickness Clean Soil Benzene 100 |ig/m3 Criteria
(greater than)
Unit of measurement for thickness
D-7
-------
Field Name
Benzene GW>5000(ug/L)
TPH GW>30000(ug/L)
Benzene Soil> 10(ug/g)
TPH Soil>250(ug/g)
overlying_clean_soil
source_comments
Type
Boolean
Boolean
Boolean
Boolean
Text
Text
Size
1
1
1
1
20
255
Description
Calculated Benzene GW indicator by 3D and date
linkages
Calculated TPH GW indicator by 3D and date linkages
Calculated Benzene Soil indicator by 3D and date
linkages
Calculated TPH Soil indicator by 3D and date linkages
Indicates if there is overlying clean soil between the
source and the sample
Comments related to the source
D-8
-------
Appendix E. PVI Database Entity Relationship Diagram
E-l
-------
PVI Database Diagram
November 21, 2012
data provider fd(E2) lt_States
data_provid@r(G)
data contact addr
data contact city (
data contact state
data contact zipco
data___contact_phor
P\ )
It BuildingJTypes
bldg_type
(0)
ssl O)
ss2 O)
0}
de(0)(IEl) ^^"
(0)
e(0)
o- - --
It J ound at) on_Ty pes
state abbrev
state fips O)(IE1)
state name(O) '
It Countres
country
country name (O)
.^ Buidngs
bui d ng_id
O9
»
09
!t_Hydrogeo!ogc_Sett ngs
Sjtes hydro_sett ng_desc
^- — — <^J
site id
data provider d (O)(FKJE1, ES)
ongna_site_d(0)[E4)
site_city (O
site hydrology(O)(FKJE2, E5)
site_vapor_src_type (O)
site vapor src orign (O)
timbe «amp° 0) ' O- -• D°— ts
Y Y ste_idfO)(FK,IE5,IE4)
orig_bldg_ld(0)(IES)
site id (FKJE1JE6)
foundation typeilEl) • £>- -fl bldgjiarne(O)
_ ^ , "SX^ ^ bide type fQ1i'FK,IE21
bldg_use (0)
footprnt area O)
BuildingJDi stances
tootpr nt__area__
§ «
bui!ding_id (0)(FKJ E2JE1
ocationjd (Q)[IE3
horz_dtst_to_bldg(O)
horz_dist_tojhldg_u-tit{0
horz_dist_commerrt O)
tma stamp(O)
soiljxt_code
soiMsS: jiame(Q)
soil_tKt_dese (O)
Links
^ .f
Link_ID
ocation_>;yjdl(FK, El) ^^~~~~ ™
ocation_xyJd2(IE2
distance_*y (0)
distance *y units (O)
Samplejinks
^% fnd_depth_tojb
time_stamp(0
«
,.,,., A
nit(O)
O)(FK,IE3, E
ase (0)
ase.umt (O
0)
doc narn fOl
doc_year (O
doc_date(0)
doc__source O)
author_c tat on (O)
author name (O)
•) v":[lo
Pages (0
author phone (O)
doc comments (O) |
doc_orignai_format (O
date_QC_cornpleted (O)
t me_stamp (O)
samplejocationjd (IE6)
site td(O)(FK,lEl,lEB
bui ding_d (O) (FK,JE3,IE2)
import loc id (O)(IE4
loc_name(0)
samp Joc_name(O)
sampie_depth (O)
sample__depth___unit (C
loc _type (0)
loc_nt/ext (O)
ioc_dssc (O)
vj_alt_so _desc [0)
vz at so grade (O)
V2_porosity (O)
vz_porosity_unit (O)
loc_comment (0)
time_stamp(O)
vz alt so desc src
fractured rock (O)
)
0)
<>
Samples ^ <
^> „
Samp!e_link_lD
samp!ejd_sg_in/c
samp!e_id_gw(O'
sample id soil (0
«
utdoor sr (O1 (FK)
_ sample id
sample Jo
ongnaJ__s
sample__m
sample st
sample c
time_starr
Headspac
gwjemp
*.-**>* S*-temP-
J-"Vx ground c
leak_test_
v?_rnotstL
Soil_TPH_
Soil_TPH_
sampleJD
Sources
It DQ R
DQ Ran
DQ Table
site id(FK,IE3) ]
CSM Ran
DQ Rank
Probe Typ
Leak Trac
Purging P
Pneumatic
Fixed Gas
Acceptable
VQC Metr
<(0)(FK,IE1
0) (FKJE2)
6(0)
er(0)
ocedure (0)
_Test (0)
_Data (O)
nsistent (0)
od(O)
Lab QAQC(O)
Peer Review (O)
Peg_Frogram (0)
DQ Comment (0)
ank /\
<
DQ_Description (0)
lt_
C
C
¥
CSM_Ran/S
iM_Rank
3M_Description (0)
References
\ -£*> ^ 3QOQ
Benzene Soil>
|"""" TPH Soil >250 i
| overlying_csar
\
J R6Sul,s
test resut id
cationjd (0
ample id (O)
edium(0)(F
art _d ate [O)
mment (0
p(0)
2 yn(0)
0)
units (0)
ver (0)
yn(0)
re_content (
paired_resu t
paired resu t
Q(0)
nfidence (O
(FK, E4.E1 import result id(O)(E4 ™
samp e_d (hKJtAltb ^
P „ ( )( • i }
result vaue (O) pa
£ result comment (O) pa
report detection (O) ^XS*. Ca
detect fag_yn(0) P~"**^? Pa
vaue type(O) or
Dj stat type(0)(FK) HL
nit (0) stat_obs_date_first (O) De
_value 0) stst_obs_datejast [O Tc
_unt O test_result_comment (0 Tb
fixed__gas__method (O) co
time stamp (O) so
(0)
QOO(ug/L)(O)
(ug/L)(0)
Oug/B(0)
g/g)(0)
_soil (0)
Parameters
100_ug/rn3(O)
lQO_ug/m3(Lessthan)(Q
100_ug,'m3(Both)(O
100 ug/m3(Reftnedestimate)(O)
10D ug/m3(Greaterthan)(0)
amsterjd (IE 2
ameter_abbrev [O)(IE1)
ameter_name(O
number (O
janic yn (O)
C25(0)
taH (0)
(0)
(0)
•nment(O)
t_name(0)(E3
^^** lt_Stat_Types
statjype
It Unts
unit_type
unit_code
unit desc (O)
unit_pref (O)
E-2
-------
Appendix F. Analysis of Lead Scavengers: Ethylene Dibromide
and 1,2-Dichloroethane
F.I Historical & Current Uses
F.I.I Gasoline Additives
Ethylene dibromide (EDB or 1,2-dibromoethane) and 1,2-dichloroethane (1,2-DCA) are
synthetic organic chemicals that were historically used as gasoline additives to prevent lead
deposits that foul internal combustion engines. For this reason, they are commonly referred to as
lead scavengers. Addition of EDB and 1,2-DCA to gasoline began in significant amounts from
the mid-1920s and continued until leaded gasoline was phased out beginning in late 1980s.
The U.S Environmental Protection Agency (EPA) began a phase-down program in 1973 to
reduce the lead content in gasoline. Since the early 1940s and until that phase-down, leaded
gasoline contained EDB and 1,2-DCA with molar portions of Pb:Cl:Br of 1:2:1. On-road uses of
leaded gasoline were banned in 1996 (Falta, 2004; U.S. EPA, 2006).). Prior to 1974, the average
EDB and 1,2-DCA concentrations in U.S. automotive gasoline were as high as about 0.320 g/L,
which decreased to about 0.180 g/L in late 1970s and further down to about 0.060 g/L by early
1980s. Use of leaded gasoline in on-road vehicles has been banned since 1996 (U.S. EPA, 2006).
Lead scavengers are still used as additives to aviation gasoline (avgas) and automobile racing
fuel. Avgas, however, does not use 1,2-DCA and has twice the amount of bromine with a molar
ratio Pb:Br of 1:2. Use of EDB as an additive to leaded gasoline accounted for more than 80
percent of its consumption in 1981. However, there are also other industrial and agricultural uses
of both EDB and 1,2-DCA.
F.I.2 Other Industrial & Agricultural Uses
EDB was used in agricultural applications as a soil fumigant and a pesticide from 1948 to 1983.
Its use as a pesticide was suspended in 1984 (Falta, 2004). According to U.S. EPA (2006), EDB
is currently used as a nonflammable solvent for resins, gums, and waxes and as a chemical
intermediate in synthesis operations. It is used most commonly to make vinyl bromide, which is
a flame retardant in modacrylic fibers. Other current applications of EDB are as an intermediate
in the preparation of dyes and pharmaceuticals.
U.S. EPA (2006) lists the historical uses of 1,2-DCA in varnish and finish removers, soaps and
scouring compounds, organic synthesis for extraction and cleaning, metal degreasers, ore
floatation, and paints, coatings, and adhesives. Its commercial production was first reported in
1922. Currently, 1,2-DCA is primarily used in the manufacturing of vinyl chloride.
F.2 Toxicity
The current EPA maximum contaminant level (MCL) in drinking water for EDB and 1,2-DCA
are 0.05 and 5 |ig/L, respectively. The maximum contaminant level goal (MCLG) for both
chemicals is zero, based on increased risk of cancer for 1,2-DCA and increased risk of cancer
F-l
-------
and "problems with liver, stomach, reproductive system, or kidneys; increased risk of cancer"17
for EDB. The MCLs for EDB and 1,2-DCA are "set as close to the health goals as possible,
considering cost, benefits and the ability of public water systems to detect and remove
contaminants using suitable treatment technologies."18 As a result, the incremental lifetime
cancer risk (TLCR) associated with these MCLs is higher than the generally acceptable 10~5 value
and some states (e.g. California, Florida, and Massachusetts) have lower MCLs for drinking
water (Falta, 2004). EPA's cancer risk-based regional screening levels (RSLs) set the tapwater
screening levels for 1,2-DCA and ethylene dibromide at 0.15 and 0.0065 |ig/L, respectively.
The contaminant source listed in the drinking water regulations is discharge from petroleum
refineries for EDB and discharge from industrial chemical factories for 1,2-DCA. According to
U.S. EPA (2006), the exposure pathways for EDB and 1,2-DCA include dermal absorption,
inhalation, and ingestion, and both chemicals are classified as probable human carcinogens. The
International Agency for Research on Cancer (IARC) has a classification of 2A (probably
carcinogenic to humans) for EDB and 2B (possibly carcinogenic to humans) for 1,2-DCA.
The Agency for Toxic Substances and Disease Registry (ATSDR) has a toxicity profile for
1,2-DCA, which specifies minimal risk levels (MRLs) for inhalation and oral exposure. The
inhalation MRL for 1,2-DCA is 0.6 ppm for chronic exposure (>365 days, but also protective for
intermediate exposure of 15-364 days), and an oral MRL of 0.2 mg/kg/day for intermediate-
duration (15-364 days) (ATSDR, 2001). Toxicity reference values for both EDB and 1,2-DCA
are available from EPA's Integrated Risk Information System (IRIS), with cancer driving the
risk for both the inhalation and ingestion exposure routes. Table F-l summarizes IRIS noncancer
and cancer toxicity reference values along with the corresponding EPA RSLs for residential
indoor air and tapwater. Values are provided for EDB and 1,2-DCA, along with benzene for
comparison, because benzene usually drives the risk at petroleum contamination sites. As can be
seen in Table F-l, EDB and 1,2-DCA have higher toxicity values and lower RSLs than benzene
and thus may be expected to contribute to risks at petroleum sites where they occur in high
enough concentrations.
Table F-1. Summary of US EPA Toxicity Reference Values and Regional Screening Levels for EDB,
1,2-DCA, and Benzene
Chemical
EDB
1,2-DCA
Benzene
Toxicity Reference Values3
SFO
(mg/kg-day)"1
2
0.091
0.055
IUR (Mg/m3)"1
6.4E-04
2.6E-05
7.8E-06
Regional Screening Levels'3
Residential Tapwater
(M9/L)
0.0065
0.15
0.39
Residential Air
(Mg/m3)
0.0041
0.094
0.31
SFO = oral cancer slope factor; IUR = Inhalation Unit Risk; EDB = ethylene dibromide or 1,2-dibromoethane;
1,2-DCA = 1,2-dichoroethane.
a Source: IRIS (2012). http://www.epa.qov/IRIS/Accessed December 2012.
b Source: Regional Screening Level summary table, http://vwvw.epa.qov/reqion9/superfund/prq/. RSLs reflect default
residential exposure assumptions and a 1><10"6 excess cancer risk.
http ://water.epa. gov/drink/contaminants/index. cfm#Organic
! http://water.epa. gov/drink/contaminants^asicinformation/ethylene-dibromide.cfm
F-2
-------
F.3 Physical and Chemical Properties
Select physical and chemical properties of the lead scavengers EDB and 1,2-DCA are
summarized in Table F-2. Due their moderately high aqueous solubility, both contaminants
readily partition into pore water and move downward to the water table with infiltration. This
results in greater potential for EDB and 1,2-DCA to move towards the water table as opposed to
being present as residual NAPL in the vadose zone, compared to PHCs, which can make them
harder to detect through soil gas surveys (Falta, 2004). Sorption to soil particles and organic
matter is not a significant process as evidenced by their relatively low octanol-water and soil
organic carbon-water partition coefficients (Kow and Koc). Retardation factors less than 2 are
generally used for typical aquifer conditions. Both compounds are considered to be more mobile
in groundwater than benzene.
Table F-2. Physical and Chemical Properties of EDB and 1,2-DCA
Property
Solubility (Ks)
Molar weight (Mw)
Gasoline-water partition coefficient (Kp)
Vapor Pressure (Vp)
Henry's constant (Kn)
Octonal water partition (Kow)
Organic carbon partition coefficient (Koc)
Specific gravity
Units
mg/L
g/mol
—
kPa
—
—
L/kg
—
EDB
4,300
187.88
152*
1.47
0.029
58
44
2.17
1,2-DCA
8,700
98.96
84
8.10
0.050
30
14
1.24
Reference
Falta (2004)
Falta (2004)
Falta (2004)
Falta (2004)
Falta (2004)
Falta (2004)
Falta (2004)
US EPA (2006)
Based on measurements, which differ from the value derived using Raoult's law by about a factor of 0.5.
F.4 Fate and Transport
F.4.1 Degradation Reactions
Table F-3 lists the abiotic and biotic degradation and transformation reactions for EDB and 1,2-
DCA.
Table F-3. Degradation Reactions of EDB and 1,2-DCA
Chemical
EDB
1,2-DCA
Abiotic
Hydrolysis, reactions with sulfur nucleophiles,
reactions with FeS, photochemical reactions
with hydroxyl radicals (in air)
Reactions with sulfides, reactions with FeS
Biotic
Aerobic cometabolism, anaerobic
dehalogenation
Aerobic cometabolism, anaerobic reductive
dechlorination
Estimated half-lives for abiotic hydrolysis reaction rates reported for EDB range from 1.5 to
15 years (Falta, 2004).
Both chemicals biodegrade aerobically in surface soils more readily than in deeper soils. EDB
degrades faster anaerobically in groundwater (15-50 days half-life) as compared to aerobically
(35-360 days half-life). On the other hand, 1,2-DCA degrades much more rapidly aerobically
(Falta, 2004):
F-3
-------
• 1,2-DCA aerobic degradation in soil: 52 days half-life;
• 1,2-DCA aerobic degradation in groundwater: 100 days half-life; and
• 1,2-DCA anaerobic degradation in groundwater: 400 days half-life.
U.S. EPA (2008) summarizes laboratory and field anaerobic biodegradation reaction rates for
EDB and 1,2-DCA and compares them to benzene (Table F-4).
Table F-4. Comparison of First-Order Rate Constants for Biodegradation of EDB and 1,2-DCA in
Anaerobic Aquifer Sediment to Rate Constants for Overall Removal with Ground Water Flow in
Anaerobic Aquifers
Material
First-Order Rate Constant for
Attenuation (per year)
EDB
DCA
Benzene
Reference
Microcosm studies in laboratory, all conducted with methanogenic material
Sediment from source zone of a spill of
leaded gasoline, South Carolina
Sediment from mid gradient zone of a spill of
leaded gasoline, South Carolina
Sediment impacted by leachate from a solid
municipal waste landfill, Norman, Oklahoma
Sediment impacted by leachate from a solid
municipal waste landfill, Norman, Oklahoma
Sediment from manufacturing site
contaminated with DCA in Louisiana
Sediment from manufacturing site
contaminated with DCA in Texas
1.5±1.0
5.4±0.3
17
1.3±0.3
0.3±0.1
1.7
4.4
1.2
1.4±0.2
3.5±0.8
2.6
Henderson et al., 2008,
SI
Henderson et al., 2008,
SI
Wilson etal., 1986
Kleckaetal., 1998
Kleckaetal., 1998
Kleckaetal., 1998
Field studies, flow path in aquifer
Spill of leaded gasoline, South Carolina
Spill of leaded gasoline, North Carolina
(1995 data)
Spill of leaded gasoline, North Carolina
(2004 data)
Leachate from municipal solid waste landfill,
Michigan
Fs-12 spill of aviation gasoline on Cape Cod,
Massachusetts
1.3
0.63
0.22
0.03
0.9
0.71
0.22±0.19
1.0
0.9
0.26
0.42±0.32
0.14
Henderson etal., 2008,
Supporting Information
Mayer, 2006
Mayer, 2006
Ravi etal., 1998
Falta, 2004
Source: Table 2.3 from U.S. EPA (2008).
The presence of hydrogen sulfide species (H^S and HS") enhances the hydrolysis breakdown of
both EDB and 1,2-DCA. Abiotic reaction rates are also sensitive to temperature, and the rates for
EDB are about an order of magnitude greater than for 1,2-DCA (U.S. EPA, 2008). Both EDB
and 1,2-DCA can react abiotically with iron(II) sulfide, analogous to the reaction involving
trichloroethylene (TCE). U.S. EPA (2008) conducted experiments to determine reaction rates of
EDB and 1,2-DCA with FeS, following the procedures described by Shen and Wilson (2007) on
TCE removal, but without the organic carbon source (plant mulch). The rates for EDB ranged
from 62.6 yr"1 to 94.8 yr"1, which are similar to the TCE removal rates from Shen and Wilson
F-4
-------
(2007). The rates for 1,2-DCA were lower by about an order of magnitude ranging from 6.7 yr"1
to lO.lyr"1.
Field results show much lower degradation rates and persistence in groundwater than observed in
the experiments described above. Falta (2004) presents two explanations for the lower field
degradation rates:
• Nonequilibrium sorption processes such as intraparticle diffusion and trapping in
micropores; and
• Biodegradation stops at lower threshold concentration.
Henderson et al. (2008) conducted microcosm study for the anaerobic biodegradation of EDB
and 1,2-DCA from an underground storage tank (UST) site to evaluate the effect of other fuel
hydrocarbons on dehalogenation reactions. They found that biostimulation by lactate can
enhance the degradation rates for EDB and that degradation of 1,2-DCA is much lower than that
of EDB (in agreement with field studies) and does not respond to the lactate biostimulation. The
highest EDB removal was measured in microcosms that produced the highest amounts of
methane.
Yu (2011) also conducted an experimental study of anaerobic biodegradation of EDB and
1,2-DCA and found that EDB is preferentially degraded when both compounds are present and
that the main process is the dihaloelimination to ethene.
F.4.2 Effect of Methane
Methanogenic bacteria can metabolize EDB and 1,2-DCA to ethylene (US EPA, 2008;
McKeever et al., 2012). However, the halogenated compounds can be harmful to the
methanogenic bacteria at high concentrations (e.g., 1,300 |ig/L for EDB and 11,000 |ig/L for
1,2-DCA). Bacteria strains of the dehalococcoides group can also metabolize EDB and 1,2-DCA
to ethyl ene.
McKeever et al. (2012) conducted microcosm studies using soil from an EDB contaminated
aquifer under aerobic and anaerobic conditions. They found that biostimulation by methane for
the aerobic microcosms increased the degradation rate by a factor of eight. In general, however,
they found that anaerobic degradation could lead to natural attenuation, while EDB is persistent
under aerobic conditions. They conclude that methane could be considered as an amendment to
EDB bioremediation in aerobic groundwater conditions.
F.4.3 Groundwater Data and Behavior
Studies of public drinking water systems in the U.S. have found that EDB concentrations are
greater than its MCL (0.05 |ig/L) for about 12 percent of systems serving the U.S. population
(third ranking amongst regulated contaminants). Likewise, the 1,2-DCA concentrations were
found to exceed its MCL (5 |ig/L) for 8.4 percent of the population (Falta, 2004, citing U.S.
EPA, 2003). Falta (2004) summarizes two case studies of fuel release containing lead
scavengers:
F-5
-------
• EDB plume associated with leaded gasoline at the Massachusetts Military
Reservation (MMR) contaminant release likely between 1940 and 1970; detached
EDB plume in 1999: 2,400m long, 360m wide and average 30m thickness,
beginning 30m below ground surface; and
• Service station in operation from 1953-1987; sampling in 1999 revealed
concentrations as high as 189 |ig/L and 111 |ig/L for EDB and 1,2-DCA,
respectively; extent of plume beyond the site not known.
While there is correlation in detection of EDB and 1,2-DCA in drinking water samples, co-
occurrence of lead scavengers and BTEX is not reported, which suggests that BTEX plumes can
separate from the plume of these gasoline additives. Only a few states have requirements to test
for lead scavengers at petroleum hydrocarbon contaminated sites, and therefore, U.S. EPA Office
of Underground Storage Tanks (U.S. EPA, 2010) recommended states, tribes, and EPA regions
to investigate lead scavengers at leaking UST sites.
U.S. EPA (2006) reports ranges of EDB and 1,2-DCA concentrations for PHC sites in three
states (Table F-5).
Table F-5. Range of Groundwater Concentrations of EDB and 1,2-DCA at Select Sites as Compiled
by U.S. EPA (2006)
State
Kansas
South Carolina
California
Number of sites
7
31
8
EDB concentration range
(M9/L)
0.05-8,200
0.013-1,140
0.084-65
1,2-DCA concentration range
(M9/L)
11 -1,310
—
0.4-101
F.4.4 Vadose Zone Studies
Fate and transport of EDB and 1,2-DCA in the vadose (unsaturated soil) zone were not the focus
of any studies obtained for this review. The focus is rather on groundwater contamination as
these compounds have high aqueous solubility and low soil organic carbon-water partitioning
coefficients (Koc) and are therefore mobile in soil pore waters and groundwater. However, EDB
and 1,2-DCA can volatilize from solution, and their volatilization from moist soil surfaces is
considered to be an important subsurface loss process (U.S. EPA, 2006).
F.5 Management Strategy
In a review of treatment technologies, U.S. EPA (2006) lists the most widely used groundwater
treatment technologies for EDB as air sparging, soil vapor extraction (SVE), and pump and treat
with granular activated carbon. Pump and treat is reported as the most widely used technology
for 1,2-DCA. U.S. EPA (2006) also reports that monitored natural attenuation has been used at
31 leaking UST sites in South Carolina, while seven leaking UST sites in Kansas use air
sparging and SVE or free product recovery for EDB remediation.
McGuire and Wilson (2010) present the results of SVE and air sparging treatment of a plume
containing BTEX, EDB, and 1,2-DCA in northwest Kansas. The results show that:
F-6
-------
• EDB and 1,2-DCA are degradable when oxygen is available, and they can be
remediated along with BTEX;
• Knowledge of hydraulic flow is needed to maximize the effectiveness of the
remediation;
• 1,2-DCA is more persistent than EDB and BTEX and slowest to clean up;
• In initial results, EDB and 1,2-DCA levels decreased by greater than 90 percent;
and
• Reevaluation of the remedial strategy followed by replacement of deeper sparge
wells with shallower ones and installation of new SVE wells resulted in
concentrations of benzene, 1,2-DCA, and EDB falling below 0.15, 0.15, and
0.0013 ng/L, respectively.
Davis et al. (2009) conducted a field scale bioremediation experiment for 1,2-DCA on a layered
silty and fine-sand anaerobic aquifer. Aerobic conditions were induced by air sparging and
estimated aerobic biodegradation rates (0.06 - 0.20 %/day) were greater than laboratory-based
studies. Air was injected for 50 days over a 12-month period, and 99 percent of the initial mass
was removed.
Henderson et al. (2009) use an analytical model (REMChlor) to evaluate the effectiveness of
partial source removal and plume remediation on EDB, 1,2-DCA and hydrocarbon plumes at
UST sites. They consider anaerobic biostimulation and two scenarios with long and short plume
lengths. First-order degradation rates are assigned for each compound in nine spatio-temporal
plume zones (three spatial zones with respect to distance from source and three temporal zones
with respect to NAPL release). The model results are assessed based on relative importance,
defined as the ratio of predicted concentration to applicable standard (i.e., the MCL). They found
that MtBE has the highest relative importance among the four compounds considered (benzene,
MtBE, EDB, and 1,2-DCA). The relative importance of 1,2-DCA is low near the source, but it
increases downgradient for both EDB and 1,2-DCA because these compounds have much slower
aerobic degradation rates than benzene. Henderson and colleagues concluded that if equilibrium
concentrations are more than a couple of orders of magnitude greater than the MCLs, a single
remediation technique may not be sufficient to treat plumes of lead scavengers, hydrocarbons,
and oxygenates.
F.6 Potential for Vapor Intrusion
Due to the relatively high aqueous solubility of EDB and 1,2-DCA, soil gas surveys are
complicated to perform (Falta, 2004; and personal communication, August 2012). However, both
chemicals are volatile with dimensionless Henry's constants that are about an order of magnitude
lower than benzene. Furthermore, EPA considers inhalation to be an exposure pathway for EDB
and 1,2-DCA and has developed inhalation toxicity reference values and indoor air RSLs for
both EDB and 1,2-DCA (see Table F-l).
Table F-6shows the predicted soil gas concentrations for groundwater concentrations at one-half
of the lowest detection limits available. The calculations assume that soil gas is in equilibrium
with groundwater.
F-7
-------
Table F-6. Predicted soil gas concentrations of EDB and 1,2-DCA
a,b
Chemical
EDB
1,2-DCA
Groundwater
Concentration3
(M9/L)
0.005
0.01
Predicted Soil Gas
Concentration
(Mg/m3)
0.1
0.5
Predicted Indoor
Air Concentration
(Mg/m3)
0.001
0.005
Regional Indoor Air
Screening Levelb
(Mg/m3)
0.004
0.09
a Using a value of one-half of the detection limit from US EPA method 8011 for EDB and US EPA method 8260B for
1,2-DCA.
b U.S. EPA Regional Screening Level (RSL), based on a 1*10"6 excess cancer risk. See Section F.2.
Assuming a shallow soil vapor-to-indoor air attenuation factor of 0.01 (based on the review and
conservative approach described in Section 6.3 of this report), the predicted indoor air
concentrations are 0.001 and 0.005 |ig/m3 for EDB and 1,2-DCA, which are below the RSLs for
residential indoor air discussed in Section F.2. The predicted indoor air concentrations for EDB
and 1,2-DCA (based on one-half of the detection limits in groundwater) are below the RSL for
residential indoor air by a factor of 4 and 20, respectively, with the RSLs based on an excess
cancer risk of 1 x 10"6. Therefore a screening approach is feasible where groundwater
concentrations are measured to determine the potential for vapor intrusion risks from EDB and
1,2-DCA.
F.7 References
Davis, G. B., B.M. Patterson, and C.D. Johnston. 2009. Aerobic bioremediation of 1,2
dichloroethane and vinyl chloride at field scale. Journal of Contaminant Hydrology 107:91-
100.
Falta, R. W. 2004. The potential for ground water contamination by the gasoline lead scavengers
ethylene dibromide and 1,2-dichloroethane. Ground Water Monitoring & Remediation 24:
76-87.
Henderson, J.K., D. Freedman, R.W. Falta, T. Kuder, and J.T. Wilson. 2008. Anaerobic
biodegradation of ethylene dibromide and 1,2-dichloroethane in the presence of fuel
hydrocarbons. Environmental Science & Technology, 42: 864-870.
Henderson, J. K., R.W. Falta, and D.L. Freedman. 2009. Simulation of the effect of remediation
on EDB and 1,2-DCA plumes at sites contaminated by leaded gasoline. Journal of
Contaminant Hydrology 108:29-45.
Klecka, G.M., C.L. Carpenter, and SJ. Gonsior. 1998. Biological transformations of 1,2-
dichloroethane in subsurface soils and groundwater. Journal of Contaminant Hydrology
34:139-154.
Mayer, R. 2006. Analysis of Ground-Water Contamination by Ethylene Dibromide and 1,2-
Dichloroethane at Leaded Gasoline Release Sites. M.S. Thesis. Clemson University. 188
pages.
McGuire, E., and J.T. Wilson. 2010. Successful application of air sparging to remediate ethylene
dibromide (EDB) in groundwater in Kansas. Presented at the 22" National Tanks
Conference, Boston, MA, September 20-22, 2010.
McKeever, R., D. Sheppard, K. Niisslein, K-H. Baek, K. Reiber, S.J. Ergas, R. Forbes, M.
Hilyard, and C. Park. 2012. Biodegradation of ethylene dibromide (1,2-dibromoethane
[EDB]) in microcosms simulating in situ and biostimulated conditions. Journal of Hazardous
Materials 209-210: 92-98.
F-8
-------
Ravi, V., J.S. Chen, J. T. Wilson, J. A. Johnson, W. Gierke, and L. Murdie. 1998. Evaluation of
natural attenuation of benzene and dichloroethanes at the KL landfill. Bioremediation
Journal 2:239-258.
Shen, H. and Wilson, J. T. (2007) Trichloroethylene removal from ground water in flow-through
columns simulating a permeable reactive barrier constructed with plant mulch.
Environmental Science & Technology 41:4077-4083.
U.S. EPA (Environmental Protection Agency). 2003. Occurrence Estimation Methodology and
Occurrence Findings Report of the Six-Year Review of Existing National Primary Drinking
Water Regulations. EPA-815/R-03-006, Washington, DC.
U.S. EPA (Environmental Protection Agency). 2006. Lead scavengers compendium: overview of
properties, occurrence, and remedial technologies. Available at
http://www.epa.gov/swerustl/cat/PBCOMPND.HTM.
U.S. EPA (Environmental Protection Agency). 2008. Natural Attenuation of the Lead
Scavengers 1,2-Dibromoethane (EDB) and 1,2-Dichloroethane (1,2-DCA) at Motor Fuel
Release Sites and Implications for Risk Management. Office of Research and Development,
National Risk Management Research Laboratory, Ada, OK, EPA 600/R-08/107, September.
U.S. EPA (Environmental Protection Agency). 2010. Recommendation for states, tribes and EPA
regions to investigate and clean up lead scavengers when present at leaking underground
storage tank (LUST) sites. Memorandum. Office of Underground Storage Tanks, Office of
Solid Waste and Emergency Response, Washington, DC, May.
Wilson, B. H., G. B. Smith and J. F. Rees. 1986. Biotransformation of selected alkylbenzenes
and halogenated aliphatic hydrocarbons in methanogenic aquifer material: a microcosm
study. Environmental Science & Technology 20(10):997-1002.
Yu, R. (2011) Biodegradation Kinetics for 1,2-Dichloroethane and Ethylene Dibromide in
Anaerobic Enrichment Cultures Grown on Each Compound, Master of Science Thesis,
Environmental Engineering and Science, Clemson University, SC, January, 2011.
F-9
-------
[This page intentionally left blank.]
F-10
-------
|