mm
*v i *WiM$
" ; ' ¦'1B ¦:
¦ ' ¦ . V. . ' ...
1 ! I
f wk; f
¦ rt Htot rmumtWiB
Executive Summary for the East Fork Poplar Creek
— Sewer Line Beltway Remedial Investigation
Report
c. 01JAIM94
m m wm.:
iSMa
r
K' ViwmMGm
-\V
III pi
1 jl 1\1 ul
Oak Ridge Reservation
Environmental Health Archives
(ORREHA)
Document Number
XVI
$%m4i
¦yfi-p'-'
•• ' ./• 1 ' . .V| " -V;'
'¦•¦•V ''' '¦ ¦ •' '¦ • %¦ . ¦¦¦ -¦¦¦£:¦''¦¦•
. X.^: v %V ;
IflHHGw1
Hi I
4 feHlSB * r4 * , . /
9H9B
P m BBr
i*p
•';• -v.. •" *••'• -vA" j^.• ••; V' . -1 >
P«,m^ 'M J#.
fy: ''i.: ip-oii®1#
-------
k - Sewer line belcway t>2
n report. Executive
RECEIVED
AUG ^ 11994
DOE/OR/02-1119&D2
INFORMATION
ore^'^T
Executive Summary
for
the East Fork Poplar Creek — Sewer Line Beltway
Remedial Investigation Report
US EPA REGION 4 LIBRARY
AFC-TOWER 9™ FLOOR
— —«
61 FORSYTH STREET SW
ATLANTA, GA. 30303
CLEARED FOR
PUBLIC RELEASE
OAK RiDGE ROOM
' H'Of.E PU8UC LIBRARY
-------
DOE/OR/02-1119&D2
East Fork Poplar Creek - Sewer Line Beltway
Remedial Investigation Report
Executive Summary
January 1994
Prepared by
Science Applications International Corporation
800 Oak Ridge Turnpike
P.O. Box 2502
Oak Ridge, Tennessee 37831
Submitted to the
U.S. Department of Energy
-------
SCIENCE APPLICATIONS INTERNATIONAL CORPORATION
contributed to the preparation of this document
and should not be considered an eligible contractor
-------
CONTENTS
Page
FIGURES ES-v
TABLES ES-vii
ACRONYMS ES-ix
ES.l INTRODUCTION ES-1
ES.2 SITE CHARACTERIZATION ES-4
ES.2.1 Environmental Setting ES-4
ES.2.2 Applicable or Relevant and Appropriate Requirements ES-11
ES.2.3 Nature and Extent of Contamination ES-11
ES:2.3.1 RI implementation and results ES-13
ES.2.3.2 Data quality assessment ES-31
ES.3 RISK ASSESSMENT ES-34
ES.3.1 Human Health Risk Assessment ES-34
ES.3.1.1 Data collection and evaluation ES-36
ES.3.1.2 Exposure assessment ES-38
ES.3.1.3 Toxicity assessment ES-40
ES.3.1.4 Risk characterization ES-41
ES.3.2 Ecological Risk Assessment ES-73
ES.3.2.1 Problem formulation ES-75
ES.3.2.2 Exposure characterization ES-88
ES.3.2.3 Effects characterization ES-90
ES.3.2.4 Risk characterization ES-100
ES.3.2.5 Spatial distribution of ecological risk ES-111
ES.3.2.6 Future exposures and risks ES-114
ES.3.2.7 Comparison of ERA results with assessment endpoints ... ES-115
ES.4 REMEDIATION GOAL OPTIONS ES-118
ES.5 RECOMMENDATIONS FOR FS-EIS DEVELOPMENT ES-121
ES.5.1 Surface Water ES-121
ES.5.2 Groundwater ES-122
ES.5.3 Soils ES-122
ES.5.4 Sediments ES-123
ES.5.5 Air Pathway ES-123
ES.5.6 Direct Radiation ES-123
ES.6 REFERENCES ES-125
92-164PSQ/011494/RI
-------
92-I64PSQ/011494/RI
ES-iv
-------
FIGURES
Page
ES. 1 Location map for EFPC and SLB ES-5
ES.2a Present and future land use by segment for EFPC and its floodplain ES-8
ES.2b Total land use percentages, present and future . . ES-9
ES.3 Mean mercury concentrations (rag/kg) in the surface layer of
EFPC floodplain soils ES-21
ES.4 Mean total uranium concentrations (pCi/g) in the surface layer of
EFPC floodplain soils ES-22
ES.5 Mean uranium activity and mercury concentrations in EFPC floodplain soils
by horizon ES-24
ES.6 Mercury concentrations (mg/lcg) in EFPC sediments ES-27
ES.7 Total uranium concentrations (pCi/g) in EFPC sediments ES-28
ES.8 Hazard index for noncarcinogens for adults under current land use scenario . ES-45
ES.9 Hazard index for noncarcinogens for adults under future land use scenario . . ES-47
ES.10 Excess cancer risk for carcinogens for adults under
current land use scenario ES-49
ES.ll Excess cancer risk for carcinogens for adults under
future land use scenario ES-51
ES.12 Excess cancer risk for radionuclides for adults under
current land use scenario ES-53
ES.13 Excess cancer risk for radionuclides for adults under
future land use scenario ES-55
ES.14 Hazard index for noncarcinogens for children under
current land use scenario ES-57
ES.15 Hazard index for noncarcinogens for children under
future land use scenario ES-59
ES.16 Excess cancer risk for carcinogens for children under
current land use scenario ES-61
ES.17 Excess cancer risk for carcinogens for children under
future land use scenario ES-63
ES.18 Excess cancer risk for radionuclides for children under
current land use scenario ES-65
ES.19 Excess cancer risk for radionuclides for children under
future land use scenario ES-67
ES.20 Framework for three-phase, four-step ecological risk assessment ES-74
ES.21 Food web relationships of aquatic biota sampled (clear boxes) or
modeled (shaded boxes) for EFPC ecological risk assessment ES-79
ES.22 Food web relationships of terrestrial biota sampled (clear boxes) or
modeled (shaded boxes) for EFPC ecological risk assessment ES-80
ES.23 Map of EFPC ecological sampling sites ES-81
ES.24 Hinds Creek ecological reference sampling sites ES-84
ES.25 Mill.Branch ecological reference sampling sites ES-85
ES.26 Percentage of captured fish species from EFPC and Hinds Creek
classified as tolerant of degraded water quality, October 7-12, 1991 ES-97
92-164PSQ/0U 594/RI
-------
FIGURES (continued)
ES.27 Observed fish species diversity and maximum possible fish diversity at
sites in EFPC and Hinds Creek, October 7-12, 1991 ES-98
ES.28 Total family richness, EPT richness, and density of benthic macroinvertebrates
collected from EFPC and Hinds Creek, October 22-29, 1991 ES-99
ES.29 Ecological risk for aquatic and terrestrial resources under current conditions ES-112
ES.30 Risk-based RGOs for mercury in soils of EFPC: protection of human health ES-120
-------
92-164PSQ/011494/RI
-------
TABLES
Page
ES.l COPCs by environmental medium ES-14
ES.2 EFPC RI - media sampling by location ES-18
ES.3 Volumes of mercury-contaminated soils in EFPC and SLB ES-25
ES.4 Risk measures and equations for risk ES-43
ES.5 EFPC ERA endpoints ES-76
ES.6 Measurements made for the EFPC ERA ES-87
ES.7 Dominant mode of exposure of indicator organisms to contaminated
source media in EFPC ES-89
ES.8 Summary table of trends for whole-body concentrations of contaminants in
aquatic biota collected from EFPC and Hinds Creek, October 7-29, 1991 . . ES-91
ES.9 Summary table of trends for whole-body contaminant concentrations in
terrestrial biota collected from EFPC and reference site in late 1991 ES-93
ES.10 Classification of risk quotients for aquatic exposures by range and location . ES-102
ES. 11 Classification of risk quotients for terrestrial exposures by
range and location ES-104
ES.12 Patterns of measurement endpoints for three ecological risk assessment
segments ES-113
ES.13 Comparison of ERA results with assessment endpoints ES-117
92-164PSQ/011494/RI
-------
ACRONYMS
ARAR
applicable or relevant and appropriate requirement
AWQC
ambient water quality criteria
BMAP
Biological Monitoring and Abatement Program
BRA
baseline risk assessment
BTF
biotransfer factor
CDI
chronic daily intake
CERCLA
Comprehensive Environmental Response, Compensation, and Liability Act
CLP
Contract Laboratory Program
CMS
corrective measures study
COC
contaminant of concern
COPC
contaminant of potential concern
DOE
U.S. Department of Energy
DOE-OR
DOE Field Office - Oak Ridge
DQO
data quality objective
EDXA
energy dispersive X-ray analysis
EFPC
East Fork Poplar Creek
EIS
environmental impact statement
EPA
U.S. Environmental Protection Agency
EPT
Ephemeroptera, Plecoptera, and Trichoptera
ERA
ecological risk assessment
FDA
U.S. Food and Drug Administration
FFA
Federal Facilities Agreement
FS
feasibility study
HEAST
Health Effects Assessment Summary Tables
HI
hazard index
HQ
hazard quotient
IRIS
Integrated Risk Information System
MCL
maximum contaminant level
MLE
most likely exposure
NAA
neutron activation analysis
NCR
nonconformance report
NEPA
National Environmental Policy Act
NOAA
National Oceanic and Atmospheric Administration
NPL
National Priorities List
ORAU
Oak Ridge Associated Universities
ORNL
Oak Ridge National Laboratory
ORR
Oak Ridge Reservation
OU
operable unit
PAH
poiycyclic aromatic hydrocarbon
PARCC
precision, accuracy, representativeness, completeness, and comparability
PCB
polychlorinated biphenyl
QA
quality assurance
QC
quality control
92-164PSQ/011494/R1
-------
ACRONYMS (continued)
RCRA
Resource Conservation and Recovery Act
RfC
reference concentration
RfD
reference dose
RFI
RCRA facility investigation
RGO
remediation goal option
RI
remedial investigation
RME
reasonable maximum exposure
SAIC
Science Applications International Corporation
SAP
Sampling and Analysis Plan
SDWA
Safe Drinking Water Act
SEM
scanning electron microscopy
SLB
Sewer Line Beltway
TCLP
toxicity characteristic leaching procedure
TDEC
Tennessee Department of Environment and Conservation
TVA
Tennessee Valley Authority
UCL
upper 95% confidence limit
USGS
U.S. Geological Survey
92-164PSQ/011494/R1
-------
EXECUTIVE SUMMARY
ES.l INTRODUCTION
On December 21, 1989, the U.S. Environmental Protection Agency (EPA) placed the U.S.
Department of Energy's (DOE's) Oak Ridge Reservation (ORR) on the National Priorities List
(NPL). On January 1, 1992, a Federal Facilities Agreement (FFA) between the DOE Field
Office in Oak Ridge (DOE-OR), EPA Region IV, and the Tennessee Department of Environment
and Conservation (TDEC) went into effect. This FFA establishes the procedural framework and
schedule by which DOE-OR will develop, coordinate, implement, and monitor environmental
restoration activities on the ORR in accordance with applicable federal and state environmental
regulations. The DOE-OR Environmental Restoration Program for the ORR addresses the
remediation of areas both within and outside the ORR boundaries, including Oak Ridge National
Laboratory (ORNL), the former Oak Ridge Gaseous Diffusion Plant (K-25 Site), the Y-12 Plant,
Oak Ridge Associated Universities (ORAU), the Clinch River, and East Fork Poplar Creek
(EFPC).
This report focuses on the remedial investigation (RI) of the stretch of EFPC flowing from
Lake Reality at the Y-12 Plant, through the city of Oak Ridge, to Poplar Creek on the ORR and
its associated floodplain. Both EFPC and its floodplain have been contaminated by releases from
the Y-12 Plant since the mid-1950s. Because the EFPC site—designated as an ORR operable unit
(OU) under the Comprehensive Environmental Response, Compensation, and Liability Act of
1983 (CERCLA)—is included on the NPL, its remediation must follow the specific procedures
mandated by CERCLA, as amended by the Superfund Amendments and Reauthorization Act in
1986. Because EFPC involves off-site release of contaminants from the Y-12 Plant, its
remediation also must conform with the procedures of Sect. 3004(v) of the Resource
Conservation and Recovery Act of 1980 (RCRA), as amended by the Hazardous and Solid Waste
Amendments of 1984. Because the actions taken to remediate EFPC may affect the environment,
the potential environmental impact of those actions must be publicly addressed in accordance with
the National Environmental Policy Act (NEPA). The RI of EFPC, conducted from 1990 to 1993,
integrated the requirements of these three primary federal regulations, as outlined in the FFA for
the ORR, as well as those of other federal and Tennessee state regulations.
The primary steps included in a CERCLA RI are (1) to collect data to characterize, or
describe, site conditions; (2) to determine the nature and extent of contamination at the site; and
(3) to assess current and future risks to human health and the environment if no remediation
occurred. The first two steps are collectively referred to as the site characterization, and the
92-1MPSQ/011494/R]
-------
ES-2
third step is termed the baseline human health and ecological risk assessment. RCRA calls for
similar activities to be performed under a RCRA facility investigation (RFI). After the RI is
completed, work begins on the feasibility study (FS) - environmental impact statement (EIS).
The purpose of the FS-EIS is to provide decision makers with sufficient information to adequately
compare the alternatives for site cleanup, to select a remedy for the site, and to demonstrate
compliance with CERCLA remedy selection requirements. During the FS, risk assessment
activities initiated under the Rl are continued to determine whether the proposed alternatives will
significantly reduce the risk to human health and/or the environment. The objectives of the
RCRA-required corrective measures study (CMS) are very similar to those of the CERCLA-
required FS. NEPA, however, requires that adequate information on the environmental effects
of the alternatives be available to aid decision making. In addition, the NEPA "no action"
alternative is always considered as a basis to judge the effectiveness and feasibility of all other
alternatives. An integrated FS-CMS-EIS considers (1) effects not only on the site itself but also
on surrounding communities; (2) socioeconomic impacts; and (3) interactive or cumulative effects
resulting from activities at other sites.
This report presents the results of the RI performed for the EFPC OU, which includes the
lower 23 km (14.5 miles) of EFPC, its 2.71-km2 (670-acre) floodplain, and the Oak Ridge Sewer
Line Beltway (SLB). EFPC originates on the Y-12 Plant site. From its point of origin at the
Y-12 Plant to where it feeds into Lake Reality, the creek is referred to as Upper EFPC.
Beginning at the outfall of Lake Reality, Lower EFPC travels through the city of Oak Ridge
before re-entering the ORR and joining Poplar Creek. As a result of discharges from the Y-12
Plant, EFPC and its floodplain were contaminated with mercury, other heavy metals,
radionuclides, and organic compounds. In addition, the SLB, which was constructed between
1982 and 1983, received soil from the floodplain as fill and topsoil.
A work plan for the EFPC OU, which incorporates both CERCLA and RCRA requirements,
was submitted as a two-part Sampling and Analysis Plan (SAP) in December 1991. Part I
included a conceptual model and data quality objectives (DQOs), which had not been included
in an RFI Plan prepared before the ORR was placed on the NPL. Part II presented the traditional
SAP, prepared in accordance with CERCLA guidelines. Field sampling, laboratory analysis, and
data evaluation began in the autumn of 1990 and continued through the summer of 1992. A
public awareness and participation program, including briefings to the public and to the staff of
Region IV EPA and TDEC, was implemented and maintained throughout the RI.
The EFPC RI was conducted in two segments, Phase la and Phase lb. Phase la of the RI
was designed to determine the nature of contamination and to identify the contaminants of
-------
ES-3
potential concern (COPCs)—primarily mercury and other metals, radionuclides, and organic
compounds, including polychlorinated biphenyls (PCBs) and pesticides. Completion of this phase
of the RI required installation of 12 groundwater monitoring wells and the quarterly sampling of
22 wells to characterize the groundwater quality and hydrogeology of the floodplain. More than
500 surface water, creek sediment, and surface and subsurface soil samples were taken along the
watershed and at a noncontaminated reference site and analyzed for 182 inorganic, organic, and
radionuclide analytes as well as geotechnical parameters and soil chemistry related to treatability
studies.
Phase lb, which began in the summer of 1991, was designed to establish the extent and level
of contamination. More than 3000 field samples were collected and analyzed, and soil gas
surveys were conducted to monitor mercury volatilization. The work conducted under Phases
la and lb of the RI was successful in that it enabled determination of the nature and extent of
contamination, definition of the potential for migration, and definition of the exposure pathways.
The numbers of field and quality control (QC) samples collected during the two phases of the RI
are summarized below.
Sample
Type
Number of Samples
Phase
Total
la
lb
Field
569
3445
4014
QC
304
405
709
Total
873
3850
4723
As part of the RI, baseline human health and ecological risk assessments were also completed
and are documented in this report. The primary objectives of the baseline risk assessments
(BRAs) are to determine whether current and future exposure potential presents an "imminent and
substantial" endangerment to human health and the environment, and to evaluate the need for site
remediation. The BRAs examine the (1) presence of chemicals and radionuclides in EFPC,
(2) potential routes of exposure to human and ecological receptors, and (3) likelihood of adverse
health or ecological effects resulting from contact with contaminated environmental media.
-------
ES-4
Results of the site characterization and the baseline human health and ecological risk
assessment for the EFPC OU are presented in the following sections. The EFPC site
characterization, including a discussion of the nature and extent of contamination, is presented
first. The assessment of risk is discussed in two sections, the first describing the human health
BRA and the second describing the ecological BRA. This executive summary closes with a brief
discussion of the remediation goal options for the EFPC OU, which were developed from the
primary results of the RI and BRA, and with recommendations to be considered in developing
the follow-on FS-EIS.
ES.2 SITE CHARACTERIZATION
ES2.1 Environmental Setting
EFPC is a perennial stream located in Anderson and Roane Counties in Oak Ridge,
Tennessee, ~40 km (25 miles) west of Knoxville (see Fig. ES.l). Its headwaters are contained
in 137- to 183-cm (54- to 72-in.) underground collection pipes that extend from the west end to
the central area of the Y-12 Plant, where the above-ground portion of the creek begins. From
the Y-12 Plant site, EFPC flows northward through a gap in Pine Ridge and enters Gamble
Valley and the city of Oak Ridge. From there, the stream flows northwestward along Illinois
Avenue through commercial and light industrial areas in Oak Ridge, then trends generally
westward, parallel to Oak Ridge Turnpike in East Fork Valley, through primarily residential,
agricultural, and undeveloped forest areas, until it joins Poplar Creek. EFPC waters, after
entering Poplar Creek, flow into the Clinch River, which is impounded behind Watts Bar Dam.
From the point at which it exits Lake Reality to its confluence with Poplar Creek, EFPC
measures —23 km (14.5 miles). Stream depths range from <1 m to 3 m (3 to 9 ft). The
100-year floodplain bounding the creek varies in width from several meters in its upper reaches
to -500 m (1640 ft) and encompasses -2.71 km2 (670 acres). The course and streambed of
EFPC have been modified as a result of Oak Ridge development; the creek has been channelized
in some sections of town, and riprap has been added to protect the banks. Box culverts and
bridge piers are present in EFPC at roadway crossings and numerous drainage ditches, and lateral
culverts traverse the floodplain and discharge to the creek. Major tributaries to EFPC include
Tuskegee Branch, Mill Branch, Gum Hollow Branch, Pinhook Branch, and Bear Creek. The
average flow for EFPC over a 25-year period of record is 1.5 m3/sec (51.4 ftVsec), with artificial
flows from the Y-12 Plant and Oak Ridge Sewage Treatment Plant contributing some 40% of that
-------
rx .
APPROXIMATE
WATERSHED
BOUNDARY
;I^VrV
SEWAGE
TREAT
PLANT
J ^TREATMENT
ORGDP
SEWER LINE
BELTWAY
ROAD
eft HWytfl
BEAR CREEK
J UPPER EAST FORK POPLAR CREEK
/
ORNL
pdoeI£22 "
MELTON
HILL DAM
2 MILES
LEGEND
I | EFPC floodplaln
1 Sewer Line Beltway
02 041593107
-------
ES-6
flow. These contributions and the general urban and agricultural discharges to the watershed
impact the water quantity and quality of the EFPC OU.
The EFPC R1 includes the entire SLB, which encompasses over 16 km (>10 miles) of
sanitary interceptor sewers and force mains. Construction of the SLB involved use of sediments
and soils from EFPC and its floodplain as fill material for portions of the SLB. Contaminated
fill was used in pipe trenches in some locations down to 15 cm (6 in.), and floodplain soils were
used as topsoil dressing along the line in many areas. The eastern portion of the SLB consists
of two branches, one paralleling Emory Valley Road and the other paralleling Warehouse Road.
The two branches converge near the Daniel Arthur Rehabilitation Center into a single system that
parallels Oak Ridge Turnpike to the Oak Ridge Sewage Treatment Plant (Welch 1987). Because
the SLB from Illinois Avenue to the treatment plant falls within the EFPC floodplain, it is not
distinguished in this report but rather is incorporated as part of the floodplain discussion. Only
that portion of the SLB located west of Illinois Avenue is discussed independently. Several
locations along the SLB have already been subjected to remedial actions to allow for new
construction; in these cases, contaminated soils/sediments were removed and appropriately
packaged for disposal.
EFPC, for about 25% of its length, traverses DOE property. The remainder of the creek
and all of the SLB traverse properties owned by private citizens or the local government.'
Although the land that falls within the 100-year flood boundaries of EFPC is mostly undeveloped,
developed areas occur on adjacent land just outside the floodplain. As a result, land use in the
vicinity of the site is diverse, including residential, commercial, agricultural, open (undeveloped)
and unclassified (roadways and the creek itself) uses. The residential category includes single-
and multi-family housing units and schools. The commercial category includes offices, retail
stores, restaurants, meeting places, and other establishments. Agricultural land use within the
floodplain itself is limited to grazing of livestock and horses, with no current use for raising crops
or family gardens. Open land use areas are those areas involving only incidental use, either
through recreation or trespass. (Although the exposure pathways for agricultural and open land
use are so similar that they could have been considered together in the baseline human health risk
assessment, the BRA took into account the ingestion of crops or livestock raised on the
floodplain, which represents a far more conservative case.)
As a basis for investigation, the EFPC and its floodplain were divided into nine segments
based on geography, uniformity of land use, and similarity of contaminant levels (see Figs. ES.8
through ES.19 for the location of each segment). Present land use for the nine segments is
estimated to be 7% residential, 2% commercial, 52% agricultural, and 31% open land. The
-------
ES-7
remaining 8% of the EFPC site, including the creek itself and all roads and rights-of-way, is
unclassified. The total number of individual residences within 150 m (500 ft) of the EFPC and
SLB is 1062, with an estimated population of 2622. The Oak Ridge 1988 Comprehensive Plan
and current trends toward commercialization and residential development, however, suggest that
areas within and adjacent to the floodplain now designated as open space are likely to convert to
residential land use in the future. As a result, projected future land use of the EFPC floodplain
is estimated to be 29% residential, 12% commercial, 42% agricultural, 8% open, and 9%
unclassified. Figures ES.2a and ES.2b show the present and future land use by segment for
EFPC and its floodplain.
The EFPC watershed and SLB are located in the Valley and Ridge Physiographic Province
of the Appalachian Mountains. This geologic structure produced a topography dominated by
narrow elongated ridges and valleys, which trend to the northeast/southwest. The prominent
valleys and ridges traversed by EFPC and the SLB are, from north to south: East Fork Valley,
East Fork Ridge, Gamble Valley, Emory Valley, Pine Ridge, and Bear Creek Valley. Elevations
within the study area range from 400 m (1300 ft) at Pine Ridge to 225 m (740 ft) at the mouth
of EFPC. Four different geologic units directly underlie EFPC along its course. These units are,
in order of encounter downstream, the Conasauga Group, the Rome Formation, the Chickamauga
Group, and the Knox Group. Soil types mapped in the EFPC floodplain belong to the Newark,
Newark Variant, Hamblen, Sequatchie, Pope, and Roane Series.
The EFPC site is characterized by cleared areas, shrubs and herbaceous plants, and second-
growth trees, which form thick stands up to the creek banks in many locations. Because of
different land uses in the past, the woods are in various stages of ecological succession. The
floodplain contains areas that have been filled to allow commercial development, and the creek
has been channelized in places. Other portions of the floodplain contain agricultural tracts or
grass and old field habitats. Terrestrial biota such as deer, raccoons, birds, and insects occur in
these habitats. The creek contains several species of fish as well as benthic and other organisms
typical of aquatic habitats characterized by limestone rip-rap to smooth and muddy stream
bottoms.
Wetlands were inventoried by the U.S. Army Corps of Engineers along the entire length of
the EFPC floodplain. -Seventeen wetland areas, comprising a total of ~ 5 ha (12 acres), were
identified that exhibited all three regulatory criteria (COE 1992) used to define wetlands. Most
of those wetlands were < 1 acre in size. Habitats exist within the EFPC floodplain that could
support endangered and threatened species, if present.
-------
ES-8
Present Use by Segment
Segment
K Unclassified
~ Open
I Agricultural
59 Residential
OQ Commercial
Future Use by Segment
Segment
E2 Unclassified
~ Open
¦ Agricultural
S3 Residential
(H Commercial
-------
31%
ES-9
Total Present Use
8% 2% 1%
DD Commercial
S Residential
¦ Agricultural
~ Open
Z! Unclassified
Total Future Use
9% 12%
CD Commercial
S Residential
¦ Agricultural
~ Open
0 Unclassified
-------
ES-10
The primary climatic attribute affecting possible contaminant transport at the EFPC site is
precipitation (and the associated runoff). Precipitation in this area of the Tennessee Valley is
seasonally distributed, with peaks occurring in late winter, early spring, and mid- to late summer
and lows occurring in the fall. Runoff is greatest in the winter, when evapotranspiration is low
and precipitation is high. Precipitation not lost as evapotranspiration or as quick surface and
shallow-groundwater runoff percolates through the soil and eventually recharges the groundwater
system.
Sediment movement during storm events is the primary natural route of past, present, and
future contaminant transport at the EFPC she. Hie flow of surface water over and through areas
of contaminated soils and sediments provides a potential for further transportation and
redeposition of material within EFPC. It also provides for direct solubilization of contaminants
retained in the contaminated soils and sediments.
EFPC is the primary feature of a hydrologically interconnected network within the EFPC
watershed. Within this network, EFPC surface water can recharge the shallow groundwater in
a losing reach or, conversely, groundwater can recharge EFPC surface water in a gaining reach.
This interrelationship between surface water and groundwater is important because of the
potential for continuing movement of contaminants between the two systems.
The shallow aquifer in the EFPC floodplain encompasses both the soil horizon and upper
bedrock interval. The soil horizon consists primarily of alluvial silt and clay with lesser amounts
of sand and gravel. Thickness of the shallow aquifer ranges from essentially zero to as much as
6 m (20 ft). Water levels in the shallow aquifer fluctuate seasonally in response to variations in
recharge and evapotranspiration.
Bedrock is composed primarily of limestone that contains clay and shale partings, but also
includes lithologies composed of sandstone, calcareous siltstone, shale, and dolomite. These
lithologies typically have tightly bound matrices and insignificant intergranular permeability.
However, a moderately well-developed secondary permeability system provides for transmission
of water within the bedrock. This secondary system is composed of fractures that commonly
intersect one another, providing a significant degree of interconnection in the bedrock aquifer.
Although both lateral flow within the bedrock aquifer and flow from EFPC surface water
to groundwater are likely, historic sampling of groundwater monitoring wells within the EFPC
floodplain conducted by both the Y-12 Plant and by the U.S. Geological Survey (USGS) suggests
that groundwater flow is not a pathway of concern.
-------
ES-11
ES.2.2 Applicable or Relevant and Appropriate Requirements
To determine the regulatory status of existing conditions for the EFPC site, a comprehensive
listing of applicable or relevant and appropriate requirements (ARARs), was developed for the
EFPC OU. The ARARs listing is used as a guideline for ensuring that all local, state, and
federal laws and regulations affecting EFPC are adhered to. In addition to meeting CERCLA
requirements for adherence to the applicable federal or more stringent state environmental laws
to ensure protection of human health and environment, the ARARs listing developed for EFPC
meets all NEPA requirements.
CERCLA requires that both chemical- and location-specific ARARs must be developed for
the RI process. For the EFPC RI-RFI, chemical-specific ARARs were identified for (1)
hazardous contaminants in groundwater and surface water, (2) soil and sediments, and (3)
radioactive contaminants in all media. The chemical-specific ARARs for groundwater and
surface water at the EFPC site arise from the National Secondary Drinking Water Standards and
the Tennessee Water Control Act. These two regulations were used to develop chemical-specific
maximum contaminant levels, secondary drinking water standards, and water quality criteria for
the primary EFPC contaminants. Although no chemical-specific ARARs exist for soils and
sediments, TDEC proposed rules, EPA data for PCBs and lead, and Biological Effects Levels
from the National Oceanic and Atmospheric Administration (NOAA) for a variety of chemicals
were used to provide regulatory guidance as generic action levels. Similarly, guidance from EPA
and DOE was used to develop radioactive contamination at the EFPC site. Location-specific
ARARs were defined to cover floodplains; wetlands; aquatic resources and natural areas;
endangered, threatened, or rare species; and archaeological and historic resources. Details of the
specific guidance provided by Federal Executive Orders, the Clean Water Act, Tennessee Water
Quality Control Act, Fish and Wildlife Coordination Act, Endangered Species Act, and the
Archaeological Resource Recovery Act for these location-sensitive requirements are also provided
in this RI report.
ES.2.3 Nature and Extent of Contamination
Several past studies have yielded data concerning the type, extent, and levels of
contamination in the EFPC floodplain. Routine monitoring of the Y-12 Plant effluents and EFPC
sediments and groundwater has been conducted to varying degrees since 1977, and biological
monitoring has been provided at selected locations along the creek since 1984. An investigation
undertaken in 1983 as part of the Oak Ridge Task Force Study involved large-scale sediment,
soil, water, groundwater, and vegetation sampling in an early effort to understand the potential
-------
ES-12
impact of contamination within the EFPC watershed on human health. Although these and other
ancillary studies provide critical information, they lack the degree of rigor and documentation
now required by CERCLA. Thus, these study results have only been used for screening-level
assessments and for developing conceptual models to define sampling strategies, not as the basis
for the human health and ecological risk assessments.
The conceptual model for contaminant transport in the EFPC watershed is based on the
premise that soil contamination in the EFPC floodplain is closely linked with hydrologic events.
Contaminants from the Y-12 Plant were washed down EFPC during high-flow conditions
following rainstorms. At least some of the contaminants were adsorbed onto sediment particles
and were transported downstream in a suspended phase (TVA 1985). Other contaminants were
transported in a dissolved phase. During flood events, the creek overflows its banks and spreads
out across the floodplain, depositing contaminated sediments on vegetation and soils. Hence, the
EFPC RI focused on the evaluation of surface water, creek sediments, floodplain soils, and
groundwater as potentially affected media.
To determine the nature and extent of contamination at the EFPC site, sampling and
laboratory analysis were conducted in two phases. Phase la, which began in the autumn of 1990,
was designed to determine the nature of contaminants and to identify the COPCs. This phase
involved a screening-level assessment for 182 analytes identified from the primary groups of
metals, volatile and semivolatile organic compounds, pesticides/herbicides, PCBs, and
radionuclides. Phase lb, which began in the summer of 1991, was designed to establish the
extent and level of contaminants observed in Phase la. Both Phase la and lb were conducted
under Level IV DQOs (established for data used in the BRA), and all analytical analyses were
performed under the EPA Contract Laboratory Program (CLP) or approved equivalent.
The primary COPCs for the EFPC site were identified using Phase la sampling data and a
toxicity concentration scoring system, which ranks potential health effects of contaminants on the
basis of measured concentrations and established toxicity measures without making quantitative
assumptions about exposures and dose (EPA 1989a). Toxicity measures included both systemic
and cancer slope factors. A toxicity score was calculated for each chemical and each chemical
was ranked by its percentage of the total score for the group. Those highest-ranked compounds
that together represented 99% of the toxicity score for a particular group were considered
COPCs. In addition, chemicals were included if their maximum concentrations in groundwater
or surface water were greater than the maximum contaminant level established under the Safe
Drinking Water Act (SDWA) or federal or state ambient water quality criteria for the protection
of freshwater organisms established under the Clean Water Act. Also, several fission products,
-------
ES-13
such as 1J4Cs and 137Cs, and transuranics were suspected to be in process materials at the Y-12
Plant and were included as COPCs.
As shown in Table ES.l, the COPCs for all EFPC media are extensive. The contaminant
listing of 13 heavy metals, 9 polycyclic aromatic hydrocarbons (PAHs), 2 PCBs, and
11 radionuclides, while large in number of compounds, can be more easily interpreted if one
considers the relative toxicity percentage of those compounds. For the heavy metals, mercury
was by far the most significant contributor, with >85% of the total toxicity (Phase la).
Similarly, for radionuclides, total uranium ("U, MU, and WU) accounted for 98% of the total
activity. The organic compound groups of PAHs and PCBs were, in essence, indistinguishable
in terms of risk. Hence, the somewhat overwhelming listing of 35 COPCs can be more easily
managed by focusing on mercury, uranium, PAHs, and PCBs as representative of the primary
contaminant groups.
As noted in Table ES.l, the COPCs for soils and sediments were identical. The metal
COPCs for groundwater are similar, except that (1) silver was not detected and barium was
included; (2) PAH and PCB concentrations were not significant; and (3) total radium was found
at significant levels, but I37Cs was not. COPCs for surface water were developed independently
for base and storm flow conditions. Surface water under base flow conditions had contaminant
levels which yielded essentially no COPCs. All metals and organics detected are below existing
SDWA maximum contaminant levels (MCLs) and meet federal ambient water quality criteria for
the protection of freshwater organisms. Radionuclides were detected, as indicated in Table ES.l,
but the activity of uranium isotopes and 137Cs was low. Several other radionuclides were included
as COPCs because they were not measured in Phase la. COPCs for surface water during storm
flow included the radionuclides identified for base flow as well as many of the metals identified
for sediments.
ES.2.3.1 RI implementation and results
CERCLA requires that DQOs be developed during RI planning to serve as a guide in
selecting sampling methods and analytical procedures that will ensure that all data used in
developing RI conclusions are technically valid and legally defensible. Identification of the
intended data uses is an important part of the DQO development process because they are used
to determine the level of data quality required and to define the analytical test parameters. DQOs
for the EFPC RI were established during the initial scoping process and were revised prior to
Phase lb of the RI. Meetings were held in 1990 with the primary data users: Region IV EPA,
TDEC, DOE, Energy Systems, and the EFPC RI team. Historical data, obtained primarily from
-------
ES-14
Table ES.l. COPCs bj environmental medium
Contaminant
Medium J
-< Soil | Sediment,
! Surfac
Groundwater..! ..Baseflow
e water |
Storm flow |
Metals j
Arsenic
X
X
X
X
| Barium
X
Beryllium
X
X
X
Cadmium
X
X
X
Chromium
X
X
X
X
Copper
X
X
X
X
Lead
X
X ¦
X
X
Manganese
X
X
X
Mercury
X
X
X
X
Nickel
X
X
X
Silver
X
X
Vanadium
X
X
X
X
Zinc
X
X
X
X |
PAHs J
Benzo(a)anthracene
X
X
X |
Benzo(a)pyrene
X
X
X |
Benzo(b)fluonuitheae
X
X
X |
Benzo(k)fluonuithene
X
X
X I
Caxbazole
X
X
X I
Chrysene
X
X
X |
Dibenzo(«,h)antfaracene
X
X
X I
Indeno(l ,2,3-cd)pyrene
X
X
X I
Pyrene
X |
-------
ES-15
Table ES.l. (continued)
1
Medium |
' :
Surface water |
'fpnfaimiMmt'V
S63
Sediment
Groundwater
-Base flow *
Storm flow |
PCBs |
Aroclor-1254
X
X
Aroclor-1260
X
X
Radionuclides" |
Cesium-134
X
X
X
X
x D
Cesium-137
X
X
Neptunium-237
X
X
X
X
X I
Protafitinium-233
X
X
X
X
X I
Total Radium
X
X
X I
Thorium-228
X
X
X
X
X [
Thorium-230
X
X
X
X
X I
Thorium-232
X
X
X
X
X [
Uranium-234
X
X
X
X
X |
Uranium-235
X
X
X
X
X
Uranium-238
X
X
X
X
X •
*Radionuclidc* uc included as COPCi because they were not covered adequately during Phase la sampling and
analysis.
-------
ES-16
the Oak Ridge Task Force and the Tennessee Valley Authority (TVA) studies, were reviewed,
and a conceptual model was developed. The structure of a phased investigation was conceived,
and the overall project objectives were established.
Both qualitative and quantitative DQOs were established for the EFPC RI. Of the PARCC
parameters (precision, accuracy, representativeness, completeness, and comparability), analytical
precision and accuracy were controlled by adopting EPA CLP criteria and QC frequency and
types. Data verification and validation of the resulting analytical data packages ensured that the
laboratories produced an acceptable quality level for results. Sampling precision was evaluated
by the use of collocated samples (field replicates) and split samples (field duplicates).
Representativeness of data from the EFPC RI was accomplished by selecting sampling methods
and performing repetitive sampling events to accurately represent the characteristic population.
Intervals for soil sampling were chosen to obtain the strata with the highest concentrations of
contaminants in order to achieve the most conservative representation and to optimize the number
of samples required. DQOs for completeness for the EFPC RI were set at 90% for the
laboratory completeness for both Phase la and lb. Percent completeness for field sampling was
established at 90% for Phase la but only 70% for Phase lb. To achieve comparability, the EFPC
RI used one laboratory to perform its CLP analyses and applied the same sampling methods for
each medium.
Data uses were identified during the DQO development process as primary inputs into the
site characterization study, the human health and ecological risk assessments, the initial screening
of alternatives, the FS, and the EIS. A decision was made to obtain the highest quality data
(Level IV) possible for critical studies. The data collection program was compiled into a
sampling and analysis plan for each phase of the RI. During the design of the data collection
activities for Phase lb, assistance was obtained from the EPA-Environmental Monitoring Systems
Laboratory in Las Vegas. The adequacy of sampling density/data quantity to properly
characterize the floodplain was supported by EPA modeling studies using the historic data set.
Procedures were adopted or written to guide field activities and data management tasks, and
analytical methods were selected for each environmental sampling activity. Phase lb relied upon
neutron activation analysis (NAA), a non-CLP method, for the large-scale soil analysis of the
floodplain soil samples to determine the extent and distribution of contamination. DQOs for
NAA were established to meet the precision and accuracy of CLP methods but with an exception
for analytical sensitivity. The desired lower limit of detection for mercury in floodplain soils was
determined to be 11 mg/kg. This limit was established to meet any potential remedial action level
or risk-based need. -
-------
ES-17
The EFPC RI was conducted in two phases (la and lb) from 1990 to 1992. Phase la of the
RI involved collection of base flow surface water and sediment samples from EFPC and its
tributaries to define source contributions. Storm flow samples were collected from EFPC during
two flood events and from selected tributaries during one flood event. Analyses were performed
for filtered and unfiltered samples to distinguish contaminants attributable to suspended particles
from those in solution. Groundwater was sampled for four consecutive quarters from 22
monitoring wells. Soil samples for contaminant analyses and undisturbed geotechnical samples
for soil chemistry and engineering parameters were taken from three areas of known
contamination (NOAA, Bruner's Center, and Sturm sites). Corresponding sampling and analysis
were performed at a noncontaminated reference site, Hinds Creek, which was selected because
it possesses the same soil and bedrock types but is devoid of any contamination, including urban
contaminants. Sampling of the Hinds Creek site was necessary to provide for a true comparison
of conditions at the EFPC site with conditions at a noncontaminated reference site, without
complications of municipal impacts.
Phase lb of the EFPC RI was initiated to determine the extent and distribution of COPCs
within the floodplain and to support the human health and ecological risk assessments. Extensive
sampling of all primary affected media was conducted along the entire length of the EFPC
floodplain and the SLB. As shown in Table ES.2, the Phase la sampling plan specified soil,
sediment, surface water, and groundwater sampling at 5 permanent creek sampling sites, 20
tributaries to EFPC, 3 primary floodplain study sites, and 2 locations on the Hinds Creek
reference site. Phase lb expanded the soil and sediment sampling by conducting soil/sediment
sampling in transects across the floodplain at 100-m (330-ft) intervals. Soil samples were taken
along each transect at 20-m (66-ft) spacing from the creek bank to the elevation of a 100-year
flood. Surface and subsurface core samples were taken to a depth of 123 cm, with >3000
samples collected and analyzed from the 159 transects. Stream sediment samples were collected
at odd-numbered transects. Every three sequential sediment samples were composited for
analysis to represent 600 m (2000 ft) of the creek. Twenty-seven sediment analyses numbered
from 003 to 152 provide results for the full suite of analytes. Soil along the SLB was sampled
during Phase lb, primarily to identify areas of elevated contamination. The results of this
extensive sampling and analysis effort along the EFPC and SLB are summarized here. The
results of investigations to determine the predominant form of mercury, in floodplain soils are
presented first, followed by the results of investigations of the principal media.
Mercury Form/Species. A significant effort was directed towards determining the
predominant form or species of mercury within the floodplain soils. Because the solubility,
-------
Table ES.2. EFPC RI—media sampling by location
Sample type
Tributaries
Creek stations*
Study sites* .
Stream segments'. '
Rcf.*
BC
CC
TA-TR
LA
LB
LC
LE
LR
NO
BR
Wa
I
3
„:4.
5
6
7
f
9
SLB
II
C
2
A
II
C
2
B
Soils
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Sediments
X
X
%
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Surface water
Base flow
Storm flow
X
X
X
x'
x«
X
X
X
X
X
X
X
X
X
X
Groundwater
X
X
X
X
Geotechnical
soils/sediments
>
X
X
X
X
X
X
X
X
TribuUriei: Major tribuUriei—Bear Creek (BC) and Civic Center (CC)
Minor tribuuriei—TA to TR (18 toul)
'Permanent creek sampling itationa: Lower East Fork LA, LB, LC, LE, and LR
'Floodplain (tudy litei: NOAA (NO), Bruner'i Center (BR), and Sturm (WA) iitei
'Segment* (traniecu): Segment 9 (traniecu 1-59), 8 (60-64), 7 (65-92), 6 (93-98), 5 (99-105), 4 (106-115), 3 (116-121), 2 (122-143), I (144-159)
'Hindi Creek reference site: HC2A and HC2B
'At tribuUriei TF, TJ, and TO
'At tribuUriei TF and TO
-------
ES-19
mobility, toxicity, and biological uptake of mercury change dramatically, depending upon its
form, this effort is crucial to the human health and ecological risk assessments. For example,
methylmercury is highly toxic and readily absorbed by organisms, whereas mercuric sulfide is
relatively nontoxic and has minimal biological uptake. However, direct investigative techniques
for determining the form of micron-sized mercury particles are limited. The primary technique
for providing direct evidence of a crystalline substance is X-ray diffraction. However, the size
of mercury particles in EFPC soils does not lend itself to this technique. Studies are currently
underway in the Environmental Sciences Division at ORNL to concentrate the mercury fraction
of EFPC soils to a richness which is amenable to X-ray diffraction.
Evidence for the occurrence of mercury in the form of mercuric sulfide in EFPC soils comes
from studies using scanning electron microscopy (SEM) and energy dispersive X-ray analysis
(EDXA). The combination of SEM and EDXA allows the presence of an elements) to be
mapped on the surface of a sample. A universal association of mercury and sulfur in EFPC soils
is shown by SEM/EDXA. This is not surprising considering the geochemical affinity for mercury
and sulfur to form a stable compound.
The strongest evidence to date showing that mercuric sulfide predominates in the EFPC soils
is from wet chemical extraction studies. The proportion of mercury species is determined by
compound-specific sequential extractions from soil samples by various chemical solutions. This
method indicates that approximately 85% of the mercury in the EFPC soils is in the form of
mercuric sulfide (Revis 1989). The EPA laboratories in Las Vegas have developed their own
chemical tests for mercury species, and a request is currently being processed for their assistance.
Direct analyses of methylmercury by gas chromatograph and total mercury by atomic
absorption were performed on a set of soil samples taken from the EFPC study areas." The
proportion of total methylmercury to total mercury averaged about 0.003% (which is in
agreement with similar published results), or ir; the 2 to 80 ng/g (parts per billion) range. As
part of the ecological study, the methylmercury attributable to the suspended sediment fraction
in surface waters of Lake Reality was analyzed. The proportion of methylmercury in the
suspended sediments to total mercury in unfiltered water samples (0.008%) was equivalent to the
proportion in soils. Levels of methylmercury observed in these tests indicate that it is not a
COPC.
Although the emphasis of these investigations has been to characterize the type of mercury
compound present in the soils, the final goal is to quantify the biological availability of the EFPC
mercury and the proper absorption factor to be used in the risk assessment. Several solubility
-------
ES-20
studies have been conducted, including application of the toxicity characteristic leaching
procedure (TCLP) and performance of bench-scale treatability tests with several common acids.
TCLP results were below the detection limit, and treatability extractions achieved <1%
dissolution of the mercury from the soils. Solubility studies simulating conditions in the human
gastrointestinal tract are now planned by ORNL. If confirmed by the ORNL tests, the realistic
absorption of EFPC mercury into an organism is expected to be much less than the 100% now
used as a default value.
Soils. Of all the media, the floodplain soils have the largest volume and highest
concentration of contaminants. While other COPCs were present in the soil, most attention was
directed toward mercury because it accounts for 85% of the total toxicity (from the Phase la
data). Mercury was found to be a predictor of the occurrence of die other metals and
radionuclides and was thus used as a surrogate to determine the distribution and extent of
contamination (with the exception of the organic compounds).
As illustrated in Fig. ES.3, mercury (and most other inorganic contaminants) decrease
significantly in the lower half of the creek, although elevated concentrations occur at isolated
sampling locations throughout the floodplain. In general, however, mercury and the other
inorganic contaminants are situated in defined areas of the floodplain and not randomly scattered
throughout its length. The creek bank within a few meters of the water has elevated contaminant
concentrations along most of its entire length, but the concentrations along the bank decrease as
distance from the Y-12 Plant increases. Only four broad areas within the floodplain (NOAA site,
Bruner's Center site, Sturm site, and Grand Cove Subdivision) display significantly elevated
concentrations of mercury and other metals. Areas that are frequently inundated, such as
upstream reaches of the creek behind culverts and roadway underpasses, also show elevated-
levels. For the uranium isotopes, activity levels were relatively low throughout the length of the
creek (see Fig. ES.4), with little discernible pattern of distribution. PCB and PAH concentrations
also were found to be low [mean concentrations per segment of < 2 mg/kg (ppm) and < 5 mg/kg
(ppm), respectively]. Their presence is more pervasive and widespread than that of other
contaminants, possibly as a result of municipal contributions in addition to the Y-12 Plant source.
PCBs and PAHs occur most often in the upper segments of the floodplain near commercial
development and urban runoff.
Geotechnical investigations were undertaken to determine engineering parameters, soil
chemistry, and hydraulic conductivity for the various soil horizons and to define relationships
between particle size distribution and contaminants. The particle size distribution results revealed
that mercury is fairly evenly distributed across all size fractions rather than being concentrated
-------
HC 9876 54321
FLOODPLAIN SEGMENT
Fig. ES.3. Mean mercury concentrations (mg/kg) in the surface layer of EFPC
floodplaln soils. Values less than detection were set to zero before averaging. The value
above each bar is the number of samples taken. Floodplain segment numbers are plotted
going upstream from left to right (see Map 2), and HC refers to the Hinds Creek reference
-------
FLOODPLAIN SEGMENT
Fig. ES.4. Mean total uranium concentrations (pCi/g) in the surface layer of EFPC
floodplain soils. Values less than detection were set to zero before averaging. The value
above each bar is the number of samples taken. Floodplain segment numbers are plotted
going upstream from left to right (see Map 2), and HC refers to the Hinds Creek reference
-------
ES-23
in the fine fraction as previous investigations reported (TVA 1985). As shown in Fig. ES.5,
mercury levels are highest in Horizon 2 and decrease in Horizon 1, probably reflecting the
decreased output of mercury into the EFPC system by the Y-12 Plant in recent years. Similarly,
no relationship between uranium concentrations and particle size was apparent. The highest
uranium concentrations occur in Horizons 1 and 2, with uranium highest in the fine fraction in
Horizon 1 but highest in the coarse fraction in Horizon 2. Uranium levels for the deepest three
horizons are relatively constant with particle size classes.
The spatial variability of contaminant concentrations within the floodplain soil was so great
that little direct correlation could be made between the overall contaminant concentration of an
area and any individual sampling point. Thus, to characterize the level of mercury in floodplain
soils, the extent or limit of contamination was first bounded by drawing contours of mercury
concentrations for 50 and 200 mg/kg (ppm). These contaminant levels were selected to
correspond to remedial goals, which are discussed later. To assess the distribution of mercury
concentrations within the limits of contamination, a geostatistical method called kriging was then
used to interpolate the concentrations between sampling points. The kriging process calculated
mercury concentrations for 20 x 20-m blocks by aggregating the nearest sampling data according
to a distance-variance relationship.
By combining the results of the kriging with the hand-drawn mercury contours, volumes of
soils containing mercury concentrations > 50 ppm and > 200 ppm were estimated for the EFPC
floodplain. Volumes are based on the limits of contamination established for 50- and 200-ppm
mercury and for three 41-cm (16-in.) intervals (0-41, 41-82, and 82-123 cm). The total volume
of all intervals within the EFPC floodplain containing >50 ppm mercury is estimated at more
than 150,000 m' (5 million ft1). A summary of the volume estimates for mercury-contaminated
soils, which easily exceeds the range of volumes for other contaminants, is provided in
Table ES.3.
Investigations of mercury concentrations in soils along the SLB found elevated concentrations
(>50 ppm) to be isolated in three principal areas (Tulane, Fairbanks, and Emory Valley Road
areas). The volumes of soil at those sites [286.7 m3 (10,123 ft3)] are small compared to the
EFPC floodplain volumes. As expected, the measured contaminant types and levels were similar
to the distribution found along the creek. Table ES.3 provides estimates of the contaminated soil
volumes found at the three primary locations.
Sediments. As would be expected, creek sediments contain the same contaminants as
floodplain soils but at lower concentrations. Because of the transient nature of sediments, the
-------
ES-24
MERCURY (mg/kg)
SOIL HORIZON
I i 1 1 1 1 1 1
0 100 200 300 400 500 600 700 800
URANIUM (pCi/g)
SOIL HORIZON
0 10 20 30 40 50 60 70
fine medium 1 1 coarse
Fig. ES.5. Mean uranium activity and mercury concentrations in
-------
ES-25
Table ES3. Volumes of mercury-contaminated soils in EFPC and SLB
Soil Volume (1000 mVlOOO fP) |
EFPC intervals
fcSOppmHg
£200 ppm Hg |
1 (0-41 an/0-16 in.)
123.8/4371.2
23.6/833.3 1
2 (41-82 an/16-32 in.)
26.4/932.1
9.6/338.9 1
3 (82-123 cm/32-48 in.)
0 to
0/0 1
| Total
150.2/5303.3
33.2/1,172.2 |
SLB areas
Lengtb 01ft) ' _v;V".
Vohnu? 1
(mW) f
Contaminated
<:/% '".f
Tulane
15/49
91/298
17%
3.5 |
Fairbanks
396/299
1615/5298
25%
92.0
Emory Valley
823/2700
2438/7999
34%
191.2
Total
1234/4048
4144/13,596
30%
286.7/10,123 |
'Assumptions: depth 0.152 m (6 in.); width 1.52 m (5 ft)
-------
ES-26
distribution of metals in EFPC is not as predictable as that in soils. The same general patterns
do exist, however. The upper reaches of the creek show somewhat elevated levels of the various
metals (especially mercury) compared to other sections of the creek. Figure ES.6 shows the
measured mercury concentrations by transect sampling location, with most locations exhibiting
< 10 mg/kg (10 ppm) in the creek sediments. The few locations showing higher levels of
mercury still had contaminant levels below the respective floodplain soils levels (see Fig. ES.3).
For total uranium, sediment levels were consistently low (all < 12 pCi/g), with a less discernible
pattern of contaminant location along the stream length (see Fig. ES.7). The significantly higher
concentration of metals (including mercury) and uranium in the EFPC sediments compared to the
tributary sediments seems to confirm a Y-12 Plant source for this in-stream contamination. The
results of the PCB and PAH sampling indicate contaminant source(s) in the upper third of the
creek, immediately downstream of the Y-12 Plant and along the commercial areas of Oak Ridge.
The results are indicative of virtually every urban environment and do not necessarily point to
the Y-12 Plant as the sole source of these low-level organic contaminants. A sampling results
anomaly is observed immediately downstream of the confluence of Bear Creek and EFPC. This
area displays slightly elevated concentrations of several contaminants compared to adjacent
segments of the creek. This occurrence is thought to result from either the input of Bear Creek
or the "pooling" of the sediments from creek back flow. Back flow of EFPC occurs when the
Watts Bar Reservoir level is raised and temporary reverse flow takes place in the lower reaches
of EFPC.
To determine the location and distribution of sediments within EFPC, sediment
accumulations were mapped in the spring of 1992. Field crews walked the creek from Lake
Reality to the confluence of EFPC and Poplar Creek. They identified and measured depths at
locations of sediment accumulation. Many reaches of EFPC were found to be barren of sediment
and to have a bedrock base. The accumulations of sediments that do exist are primarily coarse-
grained in nature. Apparently, the normal stream velocity and frequent storm events carry fine-
grained sediments downstream to Poplar Creek. Because of the transient nature of sediments
within the creek, the data derived from this mapping represent only a gross estimation of
sediment location and thickness at a single point in time. Based on this study, an estimated
5.4 x 10* m3 (1.9 x 106 ft3) of sediments are present-in the creek, only a portion of which
would be considered contaminated based on the sampling information presented earlier.
Surface Water. Surface water was sampled during the RI under two scenarios, base and
storm flow, to determine present contaminant source contributions to the EFPC watershed. This
sampling included locations (1) at the discharge of Lake Reality to measure Y-12 Plant input to
-------
100
H
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
C
0
0
1
1
2
2
3
3
4
5
6
6
6
7
7
8
8
9
0
0
1
1
2
3
4
4
5
3
9
2
8
3
9
5
8
4
0
6
2
7
3
8
8
9
6
0
6
2
8
4
4
0
6
2
CREEK TRANSECT
Fig. ES.6. Mercury concentrations (mg/kg) in EFPC sediments. Each bar represents
a composite of three samples taken —200 meters apart. The bar is labeled with the transect
number of the central sample of the composite. Transect numbers increase moving
upstream. The average of two samples from the reference station is labeled "HC". Results
-------
H
0
0
0
0
0
0
0
0
C
0
0
1
1
2
2
3
8
8
9
2
8
8
9
5
8
0000000000
4556677889
4062788896
CREEK TRANSECT
1
1
1
1
1
1
1
1
1
0
0
1
1
2
3
4
4
5
0
6
2
8
4
4
0
6
2
Fig. ES.7. Total uranium concentrations (pCi/g) in EFPC sediments. Each bar
represents a composite of three samples taken -200 meters apart. The bar is labeled with
the transect number of the central sample of the composite. Transect numbers Increase
moving upstream. The average of two samples from the reference station is labeled "HC".
Results less than the detection limit were set to zero and marked with an M*".
-------
ES-29
the system, (2) at the confluence of EFPC with Poplar Creek to measure output from the system,
and (3) at one location adjacent to an area of known high floodplain soil contaminant
concentrations. Also, surface water samples were obtained for the six ecological study sites.
Except for uranium and one detection of mercury, no COPCs were observed during base flow
conditions at any of these creek stations. Three sites also were sampled to characterize potential
runoff of pesticides, herbicides, and residual hydrocarbon products from selected residential and
commercial areas of Oak Ridge. In addition, two major tributaries—an unnamed tributary (which
drains a residential and commercial area near the Civic Center) and Bear Creek (which drains the
western half of the Y-12 Plant site)—were sampled. Concentrations at these sites were very
similar to the low values obtained for the EFPC sites.
Collection of surface water samples from three EFPC sites during two storm events and from
three residential and commercial tributaries during one storm event allowed comparison of base
flow and storm flow concentrations of total and dissolved contaminants. Several metals (arsenic,
chromium, copper, lead, manganese, mercury, nickel, vanadium, and zinc) were detected at
EFPC locations in much higher concentrations during storm flow than during base flow.
Detection of no or minimal dissolved concentrations of most metals indicates primarily particle-
bound transport. Total activity of uranium isotopes was observed in detectable quantities only
in EFPC, but it was present during both base flow and storm flow. Noteworthy by their absence
in either storm or base flow, in total or dissolved concentrations, are beryllium, cadmium, and
PCBs. No pesticides, herbicides, or high concentrations of heavy metals were detected in
adjoining watersheds. Also noteworthy are the presence of chromium, copper, lead, vanadium,
and zinc in storm flow samples at the Civic Center tributary and the absence of these
contaminants at the other tributaries. These findings indicate impacts on EFPC water quality
from urban sources.
Groundwater. Groundwater was investigated as a potential pathway for contaminant
migration within the EFPC floodplain. The conceptual model for groundwater indicates that the
EFPC valley is an alluvial aquifer receiving surface water from Bear Creek Valley and recharge
from the watershed along Black Oak and East Fork ridges. Deep groundwater flow paths are not
postulated to cross surface water divides.
Twenty-two wells were sampled for the full suite of analytes for four consecutive quarters.
Two types of wells, soil horizon wells and bedrock wells, were installed for the EFPC Rl. The
soil horizon wells were installed to evaluate the contaminant transfer from soil to the groundwater
and to assess the concentration of contaminants within the groundwater in proximity to soil
contamination and the creek. Thus, these soil horizon wells, screened from roughly 0.9 to 2.4 m
-------
ES-30
(3 to 8 ft) at depth, were intended to study contamination in specific areas of concern, not to
characterize contaminant migration pathways for human exposure via groundwater. No potable
water supply wells now exist within the floodplain, and it is highly unlikely that a future resident
would attempt to establish a reliable water supply from a nonproductive (< 1 gpm recharge rate)
clay interval at a depth of 0 to 6.1 m (0 to 20 ft). The potential for a groundwater exposure
pathway seems even more remote given that (1) a well within the confines of a narrow floodplain
would be submerged several times a year due to flooding and (2) a municipal water supply exists
throughout the area.
Groundwater field parameters and geochemical analyses were examined to define whether
a soil horizon aquifer exists separately from an underlying bedrock aquifer. Hydrogeochemical
modeling demonstrated two findings: (1) because no difference could be ascertained between
"shallow" and "deep" groundwaters, no aquitard is inferred to exist between the two horizons;
and (2) the geochemical environment of the EFPC groundwaters is such that significant mercury
concentrations are not likely to persist in a dissolved phase and will precipitate out, most likely
as a mercury sulfide.
Filtered and unfiltered groundwater samples were analyzed to compare total and dissolved
concentrations of metals because contaminant transport in groundwater is mainly concerned with
those contaminants that are found in solution under ambient pH, temperature, and other
environmental conditions. As for surface water, most of the detectable concentrations of COPCs
found in groundwater were associated with total rather than dissolved analyses. Total
concentrations exceeding SDWA levels were found for beryllium, copper, lead, manganese,
mercury, nickel, radium, and vanadium, but only manganese, which is a common element in
local groundwaters, had dissolved concentrations exceeding the SDWA levels. This association
of contaminants with total analyses suggests that the particle-reactive contaminants are attached
to fine-grained particles in the total sample and are not necessarily mobile in groundwater.
Equilibrium geochemical modeling of the analytical results from the groundwater monitoring
wells using MINTEQA2 (EPA 1987) supported the contention that mercury concentrations
measured in groundwater samples are related to particulate-bound mercury.
Biomonitoring. A number of historical and ongoing studies, especially the Biological
Monitoring and Abatement Program (BMAP), have measured population densities and
contaminant body burdens within the aquatic and terrestrial biota of the EFPC watershed. In
general, these studies have found elevated levels of mercury and PCBs in fish, with the highest
concentrations in the upper reaches of EFPC nearest the Y-12 Plant. Relative abundance
estimates of fish collected in the creek indicated that greater densities were present in the upper
-------
ES-31
reaches of the stream, in the area most accessible to the public and also where the highest
contaminant concentrations were observed. Other related studies of garden plots and mouse
populations have added to the accumulating knowledge of heavy metal (particularly mercury)
interactions in the environment. This information base was the starting point for the
/
comprehensive EFPC ecological risk assessment performed for the RI. Further discussion of the
scope and results of that assessment is provided in Sect. ES.3.2 of this Executive Summary.
Air Pathway. Atmospheric contaminant concentrations occurring in the general environment
around the ORR and the surrounding region are monitored or sampled continuously by an air
monitoring network. Measurements of air concentrations (as reported in the annual ORR
Environmental Report) of 15 radioactive parameters, fluorides, sulfur dioxides, total suspended
particulates, and mercury indicate that the ORR operations are not measurably impacting the
regional air quality (including the EFPC floodplain). Ambient air monitoring of mercury also
has been conducted by ORNL researchers at EFPC-specific locations selected for known high
concentrations of mercury in soil. This monitoring included both high-sensitivity and high-
volume measurements. Concentrations were similar to levels at a noncontaminated reference site
(within the Y-12 reservation), indicating that mercury concentrations in the air above the
floodplain cannot be distinguished from background levels.
Direct Radiation. Historically, most of the large-area radiation survey information on the
ORR and in the surrounding area has been provided by the Aerial Measuring System, an aerial
radiological surveillance capability maintained by DOE. The most recent Aerial Measuring
System survey of the EFPC watershed was conducted during the period of March-April 1992.
Measurements were made from a helicopter at an altitude of 46 m (150 ft) and calibrated to
terrestrial exposure rates at 1 m above ground level. Gross gamma counts were recorded to
determine the ground level exposure rates, and spectral windows were used for photopeak count
rates to identify specific man-made radionuclides. Typical background exposure rates (which
include the cosmic ray contribution of 3.8 /zR/h) vary from 7 to 11 /xR/h. Total gamma exposure
rates (including background) for the majority of the floodplain range from < 8 to 20 /iR/h. Only
one site within the floodplain (Bruner's Center site) recorded a level in the 13 to 20 /iR/h range,
and it is subject to a ±30% precision error as determined by the aerial and ground-based control
correlation.
ES.2.3.2 Data quality assessment
A primary goal of the quality assurance (QA) program for the EFPC RI was to ensure that
the results of analysis of all environmental samples were fit for their intended use. To this end,
-------
ES-32
a QA Program Plan, a QA Project Plan, and field procedures were compiled to guide the
investigation. Audits and surveillances were conducted to determine the adequacy of field and
laboratory performance against the QA plan and procedures. For the project as a whole,
85 surveillances were performed covering field activities as well as procedural activities such as
training and document review. In total, 564 nonconformance reports (NCRs) were written
concerning quality deficiencies, and all but a few (which do not affect data use) were resolved.
Analytical data generated under this project have been subjected to a rigorous process of data
verification, validation, and review. After data reports were received from the analytical
laboratory, verification staff systematically examined the reports following standardized data
checklists to ensure the content, presentation, and administrative validity of the data.
Discrepancies identified during this process were recorded and documented using the QA
Program NCR system.
During the data validation phase, data were subjected to a systematic technical review by
examining all analytical QC results and laboratory documentation, following the appropriate
functional guidelines for laboratory data validation. The primary objective of this phase was to
assess and summarize the quality and reliability of the data for the intended use and to document
factors that may affect the usability of the data. As an end result of this phase of the review, data
were qualified on the basis of a technical assessment of the evaluation criteria. Qualifiers were
generated for each analytical result to indicate the usability of the data for intended uses. For the
overall purposes of this investigation, however, data were either accepted or rejected as meeting
Level IV quality requirements. With certain exceptions (such as defining the extent of
contamination), data from this RI have been used only if they meet the highest qualification
criteria established for the human health risk assessment.
To date, procedures for radiological analyses and procedures for radiological data validation
have not been firmly established in the same manner as with the inorganic/organic analyses under
CLP. EFPC radiochemical analytical data were validated using "Laboratory Data Validation
Guidelines for Evaluating Radionuclide Analyses" (SAIC, Rev 4, 1992) and "Laboratory Data
Validation Guidelines for Evaluating Neutron Activation Analyses" (SAIC, Rev 2, 1992). These
documents were developed with the intent of providing data validation guidelines for
radioanalytical data equivalent to those provided by EPA under CLP guidance. These documents
provide a systematic process for reviewing the data against the DQOs to provide assurance that
the data are adequate for their intended use.
-------
ES-33
On a project-wide basis, 96% of the data were accepted as usable based on the formal
validation process. Of the major analytical groupings, only TCLP and NAA have overall
rejection rates exceeding 10%. NAA provided acceptable data for 89% of the analyses, while
TCLP provided 82%. Specific analytes that were below the 90% level included:
Group
Compound
% acceptable
Anions
Sulfide
68%
NAA
Antimony
Arsenic
87%
69%
Semivolatiles
2,4-Dinitrophenol
89%
TCLP
Phenols
Other organics
Silver
25%
66-75%
85%
Meeting the 90% level of acceptance for these analyses was not considered critical to the
investigation. Sulfide was part of the water analyses and not a COPC. A large number of
acceptable antimony and arsenic results (including CLP analyses) are available for soils, even
with the low (87% and 69%) acceptance rate for NAA results. The single semivolatile organic
compound listed above is only one of several phenols and was not a COPC. The organics from
the TCLP samples were not compounds of particular interest. In summary, only a very limited
number of analyses did not meet the laboratory objective, and no assessments for the investigation
were restricted due to lack of data.
Overall, precision for soils analysis as well as water analysis is considered adequate for the
EFPC project. Accuracy on a project-wide basb was deemed acceptable for both CLP analysis
and NAA. While some of the data were qualified as usable but estimated, or unusable due to
poor matrix spike recoveries, the accepted data are sufficient for use by all phases of the project.
The radiological isotopes of most interest were the uranium series. From the combined use
of NAA and alpha spectroscopy, large amounts of uranium data are available for these isotopes.
Of all the radiological results, those for thorium had the highest rejection rates, with <50% of
the analyses validated as usable data. However, the available results do not indicate that a
problem exists, and additional thorium results were not required. Moreover, thorium should be
found in the same general areas as uranium; therefore, uranium can act as a surrogate for
remedial decisions concerning thorium.
-------
ES-34
The numbers of samples taken for analysis to characterize the primary media within the
EFPC OU are summarized as follows:
Medium
No. samples
groundwater
159
surface water
77
floodplain soils
4181
creek sediments
195
miscellaneous
111 1
TOTAL
4723 I
ES3 RISK ASSESSMENT
Risk assessment is an essential component of the RI/FS process. A BRA is conducted as
part of the RI to assess site conditions in the absence of remedial actions. As part of the FS
process, risk assessment is used to evaluate the acceptability of proposed remedial actions and
as a tool in developing remediation objectives (target cleanup levels). The primary objectives of
a BRA are to determine whether there is an "imminent and substantial" endangerment to human
health and ecological receptors based on current and future exposure potential and to evaluate the
need for site remediation. Separate human health and ecological risk assessments were conducted
for the EFPC OU and the SLB; the results are presented in this report and summarized here.
Both assessments examined the presence of chemicals in the EFPC OU and SLB attributable to
release from the Y-12 Plant, the potential routes of exposure, and the likelihood of adverse effects
following contact with contaminated environmental media.
ESJ.l Human Health Risk Assessment
A phased approach to human health risk assessment was adopted in conducting the BRA for
EFPC. This hierarchical approach to risk assessment facilitates derivation of the most scientifi-
cally valid estimates of the potential for adverse effects. The primary objective was to focus the
evaluation on the receptors and exposure pathways of principal concern and to quantitatively
characterize the uncertainty surrounding all assumptions and the resulting risk estimates. Risk
assessment has been conducted in three tiers:
-------
ES-35
• Tier I: screening-level deterministic assessment using Phase la monitoring data from
locations of (projected) highest concentrations,
• Tier II: deterministic assessment using the full (OU-wide) data set (Phase la and lb), and
• Tier III: probabilistic assessment and quantitative uncertainty analysis using the full data set
and focusing on pathways that drive the overall risk assessment.
The human health risk assessment process, as outlined by EPA, is divided into four
fundamental component analyses: data collection and evaluation, exposure assessment, toxicity
or hazard assessment, and risk characterization. Although an analysis of uncertainty is conducted
throughout the risk assessment, uncertainty analysis is presented as a fifth component of the risk
assessment process and is included in a separate section.
Data collection and evaluation are the first step in the risk assessment process. All data
on chemicals present in environmental media at EFPC were evaluated with the objective of
organizing the data into a form appropriate for the BRA. From the full listing of chemicals
identified in EFPC, a subset was identified that was of sufficient quality to be used in risk
assessment. According to EPA guidance, it is unnecessary to evaluate all chemicals that are
found in environmental media. Instead, representative "highest risk" chemicals (COPCs) are
selected for study on the basis of (1) environmental concentration and distribution, (2) toxicity
or degree of hazard, and (3) mobility and persistence of the chemical in the environment.
Exposure point concentrations are then derived for each COPC. Key summary statistics are the
arithmetic mean and upper 95% confidence limit of the mean (UCL).
Exposure assessment is the second step in the risk assessment process. The objectives of
the exposure assessment are to (1) delineate exposure pathways; (2) identify receptors at risk; and
(3) measure or estimate for each receptor the intensity, duration, and frequency of the exposure.
EPA has specified that actions at hazardous waste sites should be based on an estimate of the
reasonable maximum exposure (RME) expected to occur under both current and future land-use
conditions (EPA 1989a). EPA defines the RME as the highest exposure that is reasonably
expected to occur at a site. RMEs are estimated for individual pathways and are combined across
exposure routes if appropriate.
The third step, toxicity assessment, evaluates the inherent toxicity of the COPCs with regard
to carcinogenicity and systemic effects. The objectives are to identify and select toxicity
measures for use in evaluating the significance of exposure. In the development of these toxicity
measures, available dose-response data are reviewed on the adverse effects to human and
-------
ES-36
nonhuman receptors. Toxicity measures used in risk assessment are principally obtained from
the EPA Integrated Risk Information System (IRIS), a data base of EPA verified measures.
The final step, risk characterization, is the process of integrating the results of the exposure
and toxicity assessments; estimates of dose are compared with appropriate toxicity measures to
determine the likelihood of adverse effects in exposed populations. Risk is characterized
separately for carcinogenic and noncarcinogenic effects because organisms typically respond
differently to exposure to carcinogenic agents than they do to exposure to noncarcinogenic agents.
For noncarcinogens, toxicologists recognize the existence of a threshold of exposure below which
there is only a very small likelihood of adverse health impacts in an exposed individual. It is
current EPA policy, however, that any level of exposure to carcinogens is considered to carry
a risk of adverse effect.
ES3.1.1 Data collection and evaluation
Concentrations of COPCs for the EFPC OU were aggregated and statistically evaluated to
derive a meaningful estimate of the exposure point concentrations. In human health risk
assessment, the exposure point concentration is most commonly an estimate of lifetime average
daily intake or dose. The average daily intake or dose was calculated, as required by EPA, from
two summary statistics for concentrations of COPCs in environmental media: (1) the arithmetic
mean and (2) the UCL. The UCL is the basis for deriving RME estimates.
The EFPC OU is quite large and characterized by a number of different land use types and
habitats. To obtain meaningful results from the EFPC BRA, analysis focused on (1) key
receptors at potential risk of exposure and (2) the land uses and circumstances under which
exposure was most likely to occur. Thus, RME estimates were determined for specific "exposure
units" along the length of the EFPC floodplain. Exposure units are geographic areas within
which a receptor would realistically be expected to spatially and temporally receive average
exposure to contaminants.
To develop exposure point concentrations for each EFPC exposure unit, data obtained from
the RI were aggregated using different methodologies for different environmental media. Data
for surface water and sediments were aggregated into a single set, representing the entire length
of the creek. Data were aggregated in this way based on the assumption that residents of EFPC
are free to wander the creek and may be exposed to surface water and sediments at any location
along the system. However, in an effort to address isolated areas of contamination, two different
sets of exposure point concentrations were used to calculate risk: one includes the UCL on the
arithmetic mean as the RME exposure point concentration, and the other uses only the maximum
-------
ES-37
concentration as the exposure point concentration. The analysis was conducted using two
different sets of exposure point concentrations at the request of EPA Region IV.
For groundwater, data from the bedrock wells and the soil horizon wells were treated
separately, thereby dividing the data into two subgroups. Note that no one is currently using
groundwater from either horizon as a source of drinking water, nor is anyone likely to do so in
the future.
Three different methods were used to aggregate soil data for EFPC. Although thousands
of soil samples were collected, the data set cannot support the derivation of exposure point
concentrations for very small exposure units within each land use area (e.g., a quarter of an
acre). Therefore, three approaches were adopted to achieve a higher level of spatial resolution
for soils.
First, the EFPC floodplain was divided along the length of the creek into nine segments on
the basis of geography, uniformity of land use, and similarity of contaminant levels. Each
segment was effectively treated as a homogeneous unit with regard to data aggregation and
estimation of exposure point concentrations. In this manner, all data within a given segment were
combined to calculate a high-end (RME) exposure point concentration, which would
conservatively characterize the level of contamination encountered anywhere in the segment.
The second approach was to aggregate the soil data within each segment according to land
use. More than one land use often is defined within each segment. Data from each land use area
were aggregated so that exposure point concentrations could be calculated separately for each land
use area within each segment. The land use area exposure point concentrations were then
compared to the segment-wide exposure point concentrations.
The third approach to data aggregation for the EFPC floodplain soils uses geostatistical
7 "
interpolation methods for spatial evaluation of contaminant concentrations (i.e., kriging). Kriging
is a weighted moving average interpolation method that uses the best linear unbiased estimator
by weighing the adjacent sample values to calculate an average value for a given region or block.
Two different approaches were used to evaluate levels of chemicals in fish tissue. First, data
over the entire.length of ihe creek were aggregated, and a mean and UCL value were derived for
the full data set. In the second approach, data were aggregated into three groups representing
upper, middle, and lower portions of the creek. This division of the creek into three areas was
based on a statistical analysis of difference between sampling locations. The objective of the
-------
ES-38
second method was to ensure that the most conservative exposure point concentrations were
developed for use in the BRA.
The SLB was divided into three areas based on observed levels of mercury as an indication
of the nature and extent of contamination. These areas are Emory Valley Road, Fairbanks
Avenue, and Tulane Avenue. Concentrations of COPCs were aggregated for each of these three
areas.
ES3.1.2 Exposure assessment
The purpose of the exposure assessment is to identify die receptors at potential risk from
contact with contaminants, to determine the exposure pathways of importance, and to quantify
intake or dose for all contaminants and pathways of concern. The exposure scenarios for the
EFPC BRA, based on land use type, include (1) current and future exposure in the agricultural
setting, (2) current and future exposure to residential populations, (3) future exposure in the
commercial setting, and (4) current and future exposure resulting from occasional use of open
land. The receptor groups at greatest risk of exposure were assumed to be children and adults
who reside in the community along EFPC. For each exposure scenario and receptor group, the
intensity, duration, and frequency of exposure were characterized.
The exposure evaluations were based on RME exposure assumptions. EPA Region IV
requested that RME values be used as the principal basis for risk assessment of EFPC. MLE
(most likely exposure) estimates were also calculated as a point of comparison but were not
included in the risk characterization section of the report (see Appendix M). The RME estimate
is not a "worst case" measure, but a high-end, conservative estimate of exposure in the
population at potential risk. In addition to RME point estimates, Monte Carlo simulations were
used to generate probabilistic estimates of exposure and risk that were used in uncertainty analysis
and to supplement the single-point RME estimates.
Current and future exposure pathways for residential receptors are as follows:
• dermal exposure to surface water while swimming,
• dermal exposure to surface water while wading,
• incidental ingestionof. surface water while swimming,
• dermal exposure to sediments while wading,
• dermal exposure to soil,
• incidental ingestion of soil,
• ingestion of groundwater and inhalation of groundwater vapors (future exposure only),
-------
ES-39
• ingestion of homegrown produce, and
• ingestion of recreationally caught fish.
Current and future exposure pathways for the agricultural setting include those identified for
residential exposure plus the following:
• ingestion of home-raised beef,
• ingestion of home-produced milk (future exposure only), and
• inhalation of particulates while mowing.
Risks to both children and adults were assessed for both the current and future exposure
scenarios. Two age groups of children were assessed: 3 to 12 years for RME and 6 to 9 years
for MLE.
The open land use scenario is designed to account for occasional exposure to EFPC soils,
surface water, and sediment by individuals who are not residents of the EFPC community.
Exposure was assessed for both children and.adults, with the assessment of children focusing on
ages 9 to 18 for the RME and on ages 12 to 15 for the MLE. For children, exposure pathways
included only the first six pathways identified above for residential receptors and were of shorter
durations and lesser frequencies than for adults.
All commercial zones within the vicinity of the EFPC OU occur on paved areas located at
a distance from the creek and outside the floodplain. Therefore, the BRA assumed no exposure
of commercial receptors to contaminants under the current land use scenario. Future use of
commercial land assumes exposure only via the groundwater pathway. For the business
community, the only pathway evaluated was ingestion of groundwater by adult receptors. This
is a very conservative assumption because the availability of municipal water makes it unlikely
that groundwater will be a source of drinking water, either now or in the future.
Ideally, exposure factors should be derived from estimates of site-specific activities and
behavior patterns of receptor groups at potential risk of exposure. Because of the size of EFPC,
such information could be meaningfully developed only through a large-scale survey of Oak
Ridge residents and was not developed for the EFPC BRA. In the absence of such data, EPA
guidance was used whenever possible in selecting or deriving values for exposure variables. The
principal sources of information used are Superfimd Exposure Assessment Manual (EPA 1988),
Exposure Factors Handbook (EPA 1989b), Risk Assessment Guidance for Superfimd: Human
Health Evaluation Manual (EPA 1989a), Human Health Evaluation Manual, Supplemental
Guidance: Standard Default Exposure Factors (EPA 1991), and Dermal Exposure Assessment:
-------
ES-40
Principles and Applications (EPA 1992a). Exposure factors for adults were assumed to remain
relatively constant over the duration of exposure. However, because exposure factors vary as
children grow to adulthood, weighted-average values were used to characterize RME estimates
for children over a specified period of time.
To estimate uncertainty in the exposure assessment, a probabilistic method using Monte
Carlo simulation was applied. Exposure variables were characterized by probability density
functions that reflect the inherent variability in the exposure factors. In looking at the probability
density functions for the exposure variables, one sees that the RME point estimates are generally
quite conservative and fall within the upper tail (>90%) of the probability distribution. This
pattern is expected, given the responsibility of the regulatory agencies (in light of the inherent
variability and uncertainty in risk assessment) to ensure protection of human health.
ES.3.13 Toxicity assessment
The purpose of the toxicity assessment is to evaluate the inherent toxicity of the COPCs and
to identify and select toxicity measures for use in estimating risk to receptors. The BRA for
EFPC required the selection of toxicity measures for both chemical compounds and radionuclides.
The toxicity measures used for chemical compounds were:
• reference doses (RfDs) for oral exposure — acceptable intake values for subchronic and
chronic exposure (noncarcinogenic effects),
• reference concentrations (RfCs) for inhalation exposure - acceptable intake values for
subchronic and chronic exposure (noncarcinogenic effects),
• cancer slope factors (i.e., cancer potency) for oral exposure, and
• cancer slope factors for the inhalation route.
EPA derives RfDs for noncarcinogens from estimates of the no-observable-adverse-effect
level or lowest-observable-adverse-effect level in humans or test animals. The inhalation RfC is
also derived from the no-observable-ad verse-effect level, but requires conversion of the levels
observed in animals to human equivalent concentrations before data sets and effects levels can
be evaluated and compared.
The assessment of the potential for noncarcinogenic effects (i.e., the use of RfDs and RfCs
to assess risk) is based on the assumption of a threshold below which health effects are not
expected to occur. Carcinogenesis, however, is thought to be a phenomenon for which the
presumption of threshold effects is inappropriate (EPA 1989a). Therefore, rather than estimate
an effects threshold for these chemicals, EPA first assigns a weight-of-evidence classification and
-------
ES-41
then calculates its cancer potency, or slope factor. Weight-of-evidence classifications include
(1) A, human carcinogen; (2) Bl, probable human carcinogen based on availability of limited
human data; (3) B2, probable human carcinogen based on availability of sufficient data in animals
but inadequate or no data for humans; (4) C, possible human carcinogen; (5) D, not classifiable
as to human carcinogenicity; and (6) E, evidence of noncarcinogenicity for humans. EPA
develops a cancer slope factor, which is a plausible upper-bound estimate of the slope of the
dose-response curve in the low-dose range, for all chemical compounds that have been classified
as definite, probable, or possible human carcinogens.
EPA has not developed RfDs or slope factors for dermal exposure. In the absence of these
factors, the toxicity measures for oral exposure were used in the EFPC BRA to calculate dermal
exposure, as recommended by EPA (EPA 1992a). Information for identifying and selecting
toxicity measures was obtained from the two sources recommended by EPA, IRIS and Health
Effects Assessment Summary Tables (HEAST), with priority given to IRIS. Toxicity measures
used for all COPCs and the toxicological information used to select those measures are provided
in the BRA.
No EPA-verified oral and inhalation RfDs are currently available for mercury in the IRIS
data base. EPA is evaluating the potential noncarcinogenic effects of mercury and has, in the
interim, published a tentative oral RfD of 0.0003 mg/kg (ppm) per day for inorganic mercury
based on the effects of mercuric chloride in rats. Because mercury is the primary contaminant
of concern (COC) at EFPC, information on the bioavailability and physicochemical properties
of mercury is critical to accurately characterizing risk to public health.
During the RI, chemical analysis of the soil samples collected from the EFPC floodplain
indicated that mercury was predominantly present in an insoluble form, probably as mercuric
sulfide (Revis et al. 1989). The toxicity and bioavailability of mercury closely parallel its
solubility in the aqueous media. The existing RfD for mercury is based on exposure of
laboratory animals to mercuric chloride, which is a highly soluble form of mercury. Mercuric
sulfide is a less bioavailable, less toxic form of mercury. However, the EFPC BRA
conservatively assumed that all mercury in the EFPC OU was present in the most bioavailable,
most highly toxic form. This assumption resulted in very conservative (highly protective)
estimates of risk to human health.
ES3.1.4 Risk characterization
Risk characterization integrates the results of the exposure and toxicity assessments by
combining estimates of dose with appropriate toxicity measures to determine the likelihood of
-------
ES-42
adverse effects in exposed populations. Risks are calculated for each chemical or radionuclide,
each pathway, each receptor group (adults and children), each current land use scenario, and each
future land use scenario. These individual risks are then summed over chemicals or radionuclides
and pathways to obtain a total risk for each receptor group and for current and future land use
scenarios. These total risks are calculated differently for each of three contaminant types: (1)
chemical carcinogens; (2) chemical noncarcinogens; and (3) radionuclide carcinogens. For each
of these contaminant types, there is a different equation for estimating risk, but all integrate the
exposure and toxicity assessments. COPCs for which calculated risks exceed EPA targets are
classified as COCs.
Table ES.4 identifies the exposure and toxicity measures, the equations for calculating risk,
the methods for combining results in deriving an estimate of total risk, and EPA target risk
levels. For the EFPC BRA, risks were presented for individual pathways and exposure routes
and were also summed across multiple pathways and exposure routes. Note that the combined
risk estimates are very conservative because they are simple summations of the results across all
pathways to a single receptor. It is unlikely that a human receptor would aggregate exposure in
this manner.
A tiered approach to risk characterization was implemented for the EFPC BRA, beginning
early in the RI process. The first tier was a screening-level deterministic risk assessment using
data from Phase la from the known areas of high mercury contamination. This assessment was
performed to estimate the magnitude of potential risks and to help identify important pathways
and COCs. The second tier was a deterministic risk assessment using all validated data. The
final tier was a probabilistic risk assessment and quantitative uncertainty analysis using Monte
Carlo simulations. Because the first tier was only useful during the early stages of the RI and
is essentially a subset of the Tier II analysis, it is not discussed here.
Tier n. None of the Tier n risk estimates indicate an imminent or substantial endangerment
to human health requiring immediate, short-term measures. The most substantial risk levels are
associated with exposures that are hypothetical and highly uncertain (e.g., the food chain
pathways). The Tier II results indicate that cancer risks from radionuclides on the EFPC site
consistently fall either below or within the EPA target cancer risk range of 10* to 10"\ In the
case of nonradionuclide chemicals, however, both noncancer and cancer risk estimates exceed
EPA targets. Results of the Tier n assessment are illustrated in Figs. ES.8 through ES.19.
These maps show summed RME risks that account for multiple contaminants and pathways for
each receptor group (adult and child), for both current and future land use scenarios, within the
nine EFPC segments.
-------
Table ES.4. Risk measures and equations Tor risk
Contaminant type
Rbk
measure .
. Exposure
measure
ToxSdty
measure ,
Individual contaminant
risk equation :
Total risk '.: ¦
EPA guidance on
ateepttble risk
Chemical
carcinogens
Upper bound
estimate of excess
lifetime cancer risk
to an individual
CDI - chronic daily
dose averaged over a
70-year period
(CDI = chronic daily
intake)
CSF 95% upper
bound estimate of
slope of dole-response
relationship
(CSF = cancer slope
factor)
Risk, » CDI X CSF
i = chemical
n p
C E Riik^
i=l j = l
where i = chemical
j «s pathway
Risk £ lO4 or
10* £ Risk £ lO"1
Chemical
noncarcinogens
HQ *= hazard
quotient: potential
for adverse
noncarcinogenic
e fleets
CDI = chronic daily
dose averaged over
the exposure period
RID = acceptable
daily intake for
subchronic and
chronic oral exposure
or RfC =¦ acceptable
daily intake for
subchronic and
chronic inhalation
exposure
HQ = CDI/RfD
HQ = CDI/RfC
n
HI o E HQ
i = l
i = chemical
(HI ~ hazard index)
HI < 1
Radionuclides
(all carcinogens)
Excess lifetime
cancer risk to an
individual
IR = daily intake rate
and
EF = exposure days
per year
SF "= age- and sex-
specific coefficients
for individual organs
receiving radiation
doses combined with
organ-specific dose
conversion factor
(SF = slope factor)
Risk, =
C x IR x EF x SF
where
C» concentration
i =» radionuclide
n p
E E Riikj
i«=l j">l
where i = radionuclide
j «« pathway
Risk £ 104 or
-------
ES-44
-------
'hi mwmmsssr
RME = Reasonable maximum exposure estimate (95% upper confidence limit on mean)
Segment numbers are shown as labeled boxes on EFPC
-------
RME = Reasonable maximum exposure estimate (95% upper confidence limit on mean)
Segment numbers are shown as labeled boxes on EFPC
-------
Incomplete
Pathway
Scalei 1*=5500'
RME = Reasonable maximum exposure estimate (95% upper confidence limit on mean)
Segment numbers are shown as labeled boxes on EFPC
-------
RME < 1 x 10
RME 1 1 x 10"'
~ Incomplete
Pathway
Scale" l'=5500'
RME = Reasonable maximum exposure estimate (95% upper confidence limit on mean)
Segment numbers are shown as labeled boxes on EFPC
-------
KEYi
¦ RME < 1 x 10"
¦ RME 1 1 x'lO~*
Incomplete
Pathway
Scale" l'=5500'
RME = Reasonable maximum exposure estimate (95% upper confidence limit on mean)
Segment numbers are shown as labeled boxes on EFPC
Fig. ES.12. Excess cancer risk Tor radionuclides Tor adults under current land use scenario.
E
-------
¦ RME 1 1 x <10
Scalet l'=5500'
RME = Reasonable maximum exposure estimate (95% upper confidence limit on mean)
Segment numbers are shown as labeled boxes on EFPC
-------
Scale' r=5500'
RME = Reasonable maximum exposure estimate (95% upper confidence limit on mean)
Segment numbers are shown as labeled boxes on EFPC
Fig. ES.14. Hazard index for noncarcinogens for children under current land use scenario.
Q
i
-------
n
Scale. l'=5500'
RME = Reasonable maximum exposure estimate (95% upper confidence limit on mean)
Segment numbers are shown as labeled boxes on EFPC
-------
KEY
Scalei l'=5500'
RME = Reasonable maximum exposure estimate (95% upper confidence limit on mean)
Segment numbers are shown as labeled boxes on EFPC
Fig. ES.16. Excess cancer risk for carcinc
-------
Incomplete
Pathway
RME = Reasonable maximum exposure estimate (95% upper confidence limit on mean)
Segment numbers are shown as labeled boxes on EFPC
-------
Incomplete
Pathway
RME = Reasonable maximum exposure estimate (95% upper confidence limit on mean)
Segment numbers are shown as labeled boxes on EFPC
-------
Inconplete
Pathway
RME = Reasonable maximum exposure estimate (95% upper confidence limit on mean)
Segment numbers are shown as labeled boxes on EFPC
-------
ES-69
For the current land use scenario, the risk estimates are consistently below or within EPA
targets for most of the 271-ha (670-acre) floodplain. Figures ES.8, ES.10, ES.14, and ES.16
indicate that four areas, three of which are residential (in Segments 3, 6, and 8) and one of which
is agricultural (in Segment 7), have both noncancer and cancer risks exceeding EPA targets.
Risks for open land use are consistently either below or within EPA noncancer and cancer risk
targets. In all cases, the pathway contributing the greatest risk is the foodchain (in particular
produce), and the contaminants contributing the greatest risk are mercury (noncancer) and arsenic
(cancer). All contaminants identified as COCs on the basis of risk estimates [e.g., individual
substances with a hazard quotient (HQ) > 1 or a cancer risk > ICT*J are associated with the food
chain pathway only and occur only in those segments that include the food chain pathway
(Segments 3, 6, 7, and 8).
Risks projected for future land use are above EPA targets in all segments, for both children
and adults, as indicated in Figures ES.9, ES.ll, ES.15, and ES.17. The exceedances are due
primarily to food chain exposures and to the hypothetical use of groundwater as a future source
of drinking water. In many cases, future risks are greater than current risks because of projected
increases in exposure resulting from change over time to a more conservative land use (i.e., from
open to residential). Exceedance of the EPA target for noncancer risks for future commercial
land use, however, was entirely due to groundwater exposures. Noncancer and cancer risks for
both children and adults exceeded EPA targets in all instances in which the food chain pathway
was evaluated (i.e., all nine segments). COCs (e.g., individual substances with an HQ > 1 or
cancer risk > 1&4) identified on the basis of the produce pathway include arsenic, cadmium,
manganese, mercury, silver, and vanadium. When the beef and dairy pathways are included,
COCs also include Aroclor-1260, barium, chromium, copper, and selenium. For soil exposures
in the future, only the noncancer target was exceeded, and only for Segments 1 and 4. The
primary COC based on the risk estimates for soil exposures is mercury.
The primary health concern for both current and future soil exposures is noncancer effects
related to exposures to mercury. When considering both current and future total risk estimates
(e.g., summed across multiple substances), the soil pathway risks exceeded EPA targets at
Segments 1, 3, 4, 5, and 8.
As noted above,-EPA targets for-both noncancer and-cancer effects were exceeded anywhere
the food chain pathway was considered (all nine segments). To put the food chain pathway in
perspective, risks were estimated for the noncontaminated reference site at Hinds Creek using the
same agricultural exposure assumptions that were used for the EFPC OU. Despite the absence
of substantial mercury contamination in the Hinds Creek soils, the produce, beef, and dairy
-------
ES-70
ingestion pathways all had risks exceeding EPA targets for both noncancer and cancer effects.
For the produce and beef ingestion pathways, arsenic and manganese account for more than 90%
of the risk. For the dairy ingestion pathway, arsenic, barium, copper, manganese, and zinc also
account for over 90% of the risk. The magnitude of the food chain risk estimates for both EFPC
and the Hinds Creek site is primarily attributable to the very conservative biotransfer factors that
were used to model contaminant uptake from soil into the food chain.
No one is currently using groundwater in the vicinity of EFPC as a source of drinking water
or is likely to do so in the future. However, this pathway was included in the BRA to comply
with EPA requirements to comprehensively evaluate the groundwater pathway. Risks of
hypothetical exposure to groundwater were evaluated separately for soil horizon wells and for
water from the underlying bedrock. Both unfiltered and filtered samples were evaluated.
Groundwater risks do not vary by segment because (1) the groundwater sample data were
aggregated over the entire creek rather than by segment and (2) the same exposure assumptions
were used for both the agricultural and residential scenarios. Commercial use of groundwater
in the future results in risks in excess of EPA targets. The COCs for the groundwater ingestion
pathway were arsenic, beryllium, manganese, and mercury.
Analysis of unfiltered groundwater data indicated levels of mercury and manganese
exceeding EPA targets. A comparative risk assessment conducted using filtered groundwater
samples resulted in lower hazard index (HI) scores overall as well as an HQ score for mercury
that is lower by a factor of nearly 20. The HQ for manganese, however, remained > 1 in the
filtered sample. Other groundwater analyses showed that contaminant levels in groundwater from
the soil horizon wells differed from those in groundwater from the underlying bedrock wells.
Both noncancer and cancer risks are generally greater in the soil horizon wells, and this remains
true when comparing unfiltered and filtered samples. Whether unfiltered or unfiltered, or from
soil horizon or bedrock, the groundwater noncancer and cancer risks are all above EPA targets.
Manganese is the primary chemical driving the elevated risk estimates for groundwater.
Risks were assessed for three areas along the SLB but were limited to soil exposures for
open land use because commercial, agricultural, or residential land use is considered unlikely
within the sampled portions of the SLB. Little or no change in land use was assumed for the
SLB over time, and exposures at the SLB do not overlap those considered at EFPC. All of the
noncarcinogenic or carcinogenic risk estimates for the SLB are within or below EPA targets. All
His are < 1, and the greatest combined cancer risks are on the order of 10*.
-------
ES-71
Risk characterization for exposures to lead from the EFPC OU required an alternative
approach in which lead uptake into children's blood is estimated. EPA has developed a computer
program to evaluate blood lead uptake in children less than 6 years old, the most sensitive
receptors of lead exposures. Under current land use assumptions, only Segments 3, 6, 7, and
8 include exposures to children less than 6 years old.
None of the blood lead levels associated with current exposure at any of these segments
approach the EPA target for blood lead levels. When considering future exposures, blood lead
levels are generally within EPA targets. For children residing at Segments 3, 4, 5, and 8, the
target was exceeded only when soil and groundwater exposures were combined. This is a
marginal exceedance and includes soil, unfiltered groundwater, and built-in exposures from air
and dietary sources of lead that are not related to the EFPC OU. Soil exposures did not approach
the target when considered independently of groundwater exposures. When risk assessment was
conducted using data from the underlying bedrock (instead of the soil horizon wells), blood lead
levels did not exceed EPA targets. Only unfiltered groundwater exposures are associated with
blood lead levels exceeding the EPA target, and the remaining exposures are well below the EPA
target. As an additional point of comparison, the soil lead concentrations at each segment fall
substantially below EPA soil cleanup guidelines for lead.
Tier m. The third tier, or probabilistic risk assessment, focused on exposure pathways that
contributed the most to the overall risk, including inadvertent ingestion of soil, the produce
ingestion pathway, and the hypothetical ingestion of groundwater. The outcomes of the Monte
Carlo simulation include both receptor groups (adult and child) for each pathway and for both
current and future land use. For the soil-related pathways (soil and produce ingestion), results
are presented only for Segment 4. An illustrative uncertainty analysis for the soil pathways is
presented for the segment with the highest observed soil concentration in Sect. 5 of this RI report.
The uncertainty analysis is presented only for the single segment because the exposure
assumptions for the residential and agricultural scenarios are the same across land use segments.
The only difference between segments is the exposure point concentrations. Therefore, an
analysis of a single segment can be used to illustrate the magnitude of uncertainty in the risk
estimates.
The RME point estimates are generally quite conservative and fall within the upper tail of
the probability distributions. In some cases, the RME estimates exceed the 95th percentile
values. Of particular concern are the conservative point estimates used for the food chain
biotransfer factors (BTFs). These factors predict the biotransfer of each contaminant from soil
to plant tissue. These data and the results of an experimental garden study suggest that the BTF
-------
ES-72
for mercury obtained from the literature and used in the BRA is probably one to two orders of
magnitude too high.
The Tier II and m results, the results of the garden uptake study, and the risk assessment
of Hinds Creek all indicate that there is considerable uncertainty in the risk estimates associated
with the food chain pathways. Given this and the fact that very limited agricultural or gardening
activity is occurring along EFPC (or is likely to occur in the future), the results for the food
chain pathways should be considered only as hypothetical, "high-end" or "bounding" estimates
of risk to human health.
Although risk estimates for the groundwater ingestion pathway exceed the target range
established by EPA, this pathway does not present a substantial risk to human health. No one
is currently using groundwater in the vicinity of EFPC as a source of drinking water or is likely
to do so in the future because all residents are provided with city water. Even though
contaminant concentrations were considerably reduced in filtered samples, the overall HI score
for the pathway is still > 1 by a small margin. This is principally due to dissolved concentrations
of manganese. The presence of arsenic is responsible for the excess lifetime cancer risk estimate
exceeding 10"4. Manganese is present in background soils at levels comparable to that in EFPC
soils and, therefore, occurs naturally in this geographic area. Although levels of arsenic in
floodplain soils exceed those at the reference location, the presence of arsenic in groundwater
cannot clearly be attributed to releases from the Y-12 Plant. The results of the Tier III analysis
of groundwater indicate a range of uncertainty of less than an order of magnitude based upon
variability in exposure assumptions.
The results of the EFPC BRA indicate that RME estimates of risks of adverse
noncarcinogenic effects associated with inadvertent ingestion of soils exceed the EPA target range
at locations throughout the floodplain. The risk is due primarily to mercury concentrations in
soil, with cadmium and arsenic contributing to a lesser degree. Concentrations of mercury,
cadmium, and arsenic were significantly higher than those found at the noncontaminated reference
site. The Tier in assessment demonstrates that there is less than an order of magnitude of
uncertainty in the RME point estimates for soil ingestion.
The risk from the soil ingestion pathway was estimated using an RfD for mercury of
0.0003 mg/kg per day. The RfD used was based on toxicity testing using soluble mercury
species (mercuric chloride) in laboratory animals, rather than the less soluble forms (mercuric
sulfide) believed to predominate in EFPC floodplain soils. The BRA conservatively assumed that
all mercury in EFPC is present in its most bioavailable form.
-------
ES-73
ESJ.2 Ecological Risk Assessment
The purpose of the EFPC ecological BRA was to determine baseline levels of risk to various
animal and plant populations and habitats in EFPC. This information was then used to determine
whether contaminants pose an imminent and substantial danger to the health of various ecological
resources and whether and where site remediation may be needed. The ecological risk
assessment (ERA) process differs from the human health risk assessment process in that it focuses
on populations and communities rather than on individuals. Individuals are addressed only if they
are protected under the Endangered Species Act. Rather than attempt to examine every species
living in and around EFPC, a weight-of-evidence assessment was conducted to determine the
effects of contaminants on indicator species (representative species from each major aquatic and
terrestrial habitat) and to extrapolate from them to project the overall effects of other and
unmeasured contaminants on the ecosystem.
The ERA, as described in Framework for Ecological Risk Assessment (EPA 1992c), involves
four basic steps occurring in three primary phases (Fig. ES.20): (1) problem formulation, (2)
exposure characterization, (3) effects characterization, and (4) risk characterization. These
interrelated activities are defined as follows.
Problem formulation establishes the goals, breadth, and focus of the assessment. It
provides a preliminary characterization of (1) chemical and physical stressors present in the
ecosystem, (2) the components and especially indicator organisms of the ecosystem likely to be
at risk, and (3) the potential ecological effects. This preliminary characterization, along with a
selection of assessment and measurement endpoints, is used as the basis for developing a
conceptual model of stressors, components, and effects. Assessment endpoints are values that,
if exceeded, suggest the need for remediation.
Exposure characterization evaluates the interactions of the stressors with the ecosystem
attributes and describes the biotic and abiotic ecosystem attributes, along with the route,
magnitude, frequency, duration, trend, and spatial pattern of exposure of each indicator
population or habitat component in relation to a chemical or physical stressor.
Effects characterization evaluates the ecological response to chemical and physical stressors
in terms of the selected assessment and measurement endpoints. Depending on the parameters
of exposure, it results in a profile of response to stressors at concentrations or doses or other
units of stress to which populations and habitats are exposed. Data from both field observations
and controlled laboratory studies are used to evaluate ecological effects.
-------
ES-74
Discussion
Between the
Risk Assessor
and
Risk Manager
(Planning)
Ecological Risk Assessment
PROBLEM FORMULATION
VO .X"' ¦ .><•N\'V ":'sC^-V-.Vs\s\':< >s-\\ %'VvJ"% w\" :\v
n^:/*'<'* •% \^>\£V'*'v> y% n>%\'\'\n -ST '¦* '" " ' " N v
¦V %:T~| ->}*&$»?'1>f"' £ > &\U< v>
• "v\ \ |# \\%i/> .'ww' %*>'»'»%•• \ . % N<\\^Yi'v':>',V'N ">.'¦¦> .
S« A\S\SJ\S\vN'>\^f*>> «, /VyNV.f\\s\>\
>><»>
~>>>«^ >JN»»
••'lit-- •>-'.'>X>v.^y »• -Y
5 " • :V:;-^V; *y.
»»>£ •>»»»>:»»5>»
I
Characterization! Characterization
• .,.v.\. S . ¦ %¦-.••' \ .\\* ¦A,\
j%yy^j'Ss^wNvN>',V',/''\
of
Exposure
of
Ecological
Effects
wmtm
.-... ..s" ••••: • ••-:"• v> • .s •
'': x vi' ~< 'v.- :'>;. <&$%>$£#&:?&'& •
, \f a^\\W . ^\»W\ . v\v .
v;y,y-sv. ;>\v* ...•>" • •••¦«-¦... • V. -
>/*/>\. >>»\-^>> >>:•>}•
s\ ¦ v 'a -* *. " % > )W/(,X\\W
<•¦ ¦¦ v ws * \% •":•'•
> ••;¦> •rw^vNs
fe i3if
;»»W/ y>
••> s'< • v^"
. N . ->•.%>•»} W, ;v .-
>2v..
•. > % Msw^tv w»»)S<> >--- a
s;:; •. •; wv <
»>>»5>*»:>x>x>»
>^>>y»»»»»>:
O
»
ST
>
n
ss
c
5!
mm »
o
3
2
Ei
r»
tt
MM*
©
3
GS
3
a
o
3
mm*
o
2,
3*
90
Discussion Between the
Risk Assessor and Risk Manager
(Results)
»>m
>i-2&Fyy>»
>>&<<$?<
Fig. ES JO. Framework for three-phase, four-step ecological risk assessment
-------
ES-75
Risk characterization integrates the effects of exposure and stressor response on indicator
populations or habitat components, using risk quotients (a ratio of exposure value to effect value);
interprets these data according to the weight-of-evidence approach; and then further interprets the
ecological significance, including the potential for ecosystem recovery.
ES.3.2.1 Problem formulation
Critical to the performance of an ecological assessment is the early establishment of goals
and objectives for the study. For the EFPC ERA, seven primary goals were developed, along
with 12 associated assessment/measurement endpoints (Table ES.S). These goals reflect
government environmental policy, such as NEPA compliance or ARARs attainment, as well as
important characteristics of the specific populations and habitats within the EFPC OU that are
potentially at risk.
Also important to the initial formulation of the problem and to the development of conceptual
models and sampling plans. is a thorough understanding of the historical information on
contaminant types, locations, and the respective ecological receptors of these contaminants.
Contaminants within the EFPC environment have been described and characterized in several
previous studies (Van Winkle et al. 1984; TVA 1985; Gist 1987; Carmichael 1989; Loar 1992;
Hinzman 1992; Turner et al. 1992 and Appendix Q). The ORAU (Gist 1987) study, conducted
between 1983 and 1987, included surveys for mercury in the surface soil along 100-m transects
across the floodplain from the Y-12 Plant to the Oak Ridge city limits in Roane County, grid
sampling for mercury within a highly contaminated floodplain property, sampling for mercury
from vertical profiles at various sites, and sampling of selected lawns and gardens in Oak Ridge.
Vegetation, sewage sludge, groundwater, and two road-killed deer were sampled and analyzed
for radionuclides, mercury, and other metals. The TVA study, performed during 1984, sampled
soils, sediments, surface water, fish, amphibians, crustaceans, and reptiles, which were then
analyzed for mercury and other metals, radionuclides, organics, pesticides, and PCBs. The Y-12
Plant BMAP, begun in 1985, performs semiannual sampling of natural populations of fish, which
are then analyzed for mercury, PCBs, and cesium (Loar 1992; Hinzman 1992). Bioaccumulation
studies of PAHs and PCBs have been conducted annually using caged clams. The USGS
conducted a survey of groundwater contamination by metals, radionuclides, and organics in 1987
(Carmichael 1989). This study sampled 14 shallow-aquifer wells at sites of known mercury
contamination within the floodplain. Other one-time studies include the assessment of mercury
and strontium in small mammals by Energy Systems and a 1982 rapid assessment survey of
mercury contamination in multiple media (i.e., sediments, mosses and liverworts, grasses, and
fish) (Van Winkle et al. 1984).
-------
Table ES.5. EFPC ERA endpoints
Asscssinmt tiidpolid
% Meaiorttnent endpolnt ' |
1. Conservation of threatened, endangered, and rare
species and their critical habitats
la. No harm to any threatened and endangered ipecies
and their critical habitat of the EFPC floodplain
la. Surveys of presence/absence of species and 1
their habitats on the EFPC floodplain
lb. Maintenance of plant community composition and/or
structure required for rare plant/animal and support
ipecies
lb. Abundance and distribution of plant/animal
species that would support threatened,
endangered, and rare species
lc. Exposure of threatened and endangered specie* to
biomagnifiable contaminants through the food chain
le, The body burdens of contaminants in 8
selected species representative of lower 1
trophic levels of aquatic and terrestrial food I
chains ||
2. Protection of migratory birds
2a. No killing or harming of migratory birds as a result
of exposure to site-specific stressors
2a. Contaminant concentration! in selected 1
migratory birds and weight-of-evidence I
data |
3. Preservation of wetlands
3a. The presence and structure/function of wetlands in
relation to contaminants
3a. Wetlands survey and their contaminant 1
levels, if any, on the EFPC floodplain |
4. Existence of a fish community indicative of
undegraded conditions *
4a. Fish community in which the proportion of ipecies
tolerant of degraded water quality is <30%
4a. Proportion of tolerant species at the site
5. No advene effects from contaminants to aquatic
indicator organisms and/or predators that feed on
them
5a. Ratio of contaminant concentrations in surface water
to water quality criteria for protection of aquatic life
21
Sa. Contaminants in surface water
Sb. Aquatic indicator organisms contaminant body burden
ratio to toxicological efleets levels S1
5b. COPC concentrations in whole-body
sample* of aquatic indicator specie*
Sc. Fish contaminant body burden ratio to levels
protective of piscivorous biota Si
5c. COPC concentrations in whole-body
samples of aquatic indicator prey species
6. Existence of terrestrial animal community
indicative of undegntded conditions
6a. Terrestrial animals with diversity, abundance, and
distributions indicative of undegraded conditions
resulting in 220% decrease compared to the
reference
6a. Population abundance and distribution by
-------
Table ES.5. (continued)
, ', Goab
Assessment endpolni*:.
Measurement end point ¦.
7. No advene effects from contaminants to terrestrial
indicator organisms and/or predators that feed on
them
7a. Ratio of contaminant body burdens in terrestrial
indicator species to to ideological effects levels
£1
7a. COPC concentrations in whole-body
samples of terrestrial indicator species
7b. Ratio of contaminant body burdens in terrestrial
indicator species to levels protective of terrestrial
predators £1
7b. COPC concentrations in whole-body
samples of terrestrial indicator prey species
-------
ES-78
The conceptual model for the EFPC ERA is these early studies, which -indicated the
existence of two major exposure sources within EFPC and its floodplain. Water effluents from
the Y-12 Plant continually expose aquatic organisms to contaminants and can contribute to the
exposure of terrestrial organisms during floods through direct contact with biota or through
deposition of water-borne contaminants on floodplain vegetation and soil. The second source,
previously contaminated floodplain soil, also contributes to the exposure of terrestrial organisms
through direct contact, inhalation, and ingestion and to exposure aquatic organisms through
erosion or scouring and redeposition of soil into EFPC. Direct exposure routes exist through the
air, shallow groundwater, surface water, instream sediments, and soil. Indirect exposure occurs
through ingestion of contaminated forage or prey. Conceptual models of the food web
relationships for selected EFPC biota are depicted in Fig. ES-21 for aquatic biota and Fig. ES-22
for terrestrial biota. These show how contaminants in biotic media can be transferred to
predators.
ERA Sampling Locations and Sampling Design. The area in which organisms are exposed
to contaminants is defined by both their home range in habitats and the distribution of
contaminants. To facilitate evaluation of exposures in different EFPC habitats, the OU was
divided into sections. The sections were arbitrarily based on the EFPC sampling grid; 17
sections were created, each comprising 1 km from west to east on the sampling map. The first
section begins just below the confluence of EFPC with Poplar Creek, and section 17 is located
at the beginning of the study area below Lake Reality. Within these 17 sections, a wide variety
of surface water/sediment and soil/vegetation habitats was identified and mapped. For surface
water, four distinct aquatic habitats were defined (runs, riffles, pools, and other); for the
terrestrial habitats, 12 distinctions were made, including 4 forest categories, 5 field or relatively
cleared habitats, and 3 developed/commercial settings. Most of the terrestrial habitats were found
to be bottomland hardwoods, while the most prevalent aquatic habitat was the stream runs.
Seventeen wetlands were found to exist in or near the floodplain; some have mercury
contamination, but there is no visible evidence of negative impact on the wetlands.
Locations of the ecological sampling sites for the EFPC ERA were chosen to gain maximum
information on EFPC biota and their physical environment and to allow comparison and contrast
with historical data. The six aquatic sampling sites, as shown in Fig. ES.23, were selected to
represent a gradient of biological and physical conditions. Site 1 is located in Pine Ridge Gap
near the outfall from Lake Reality, and Sites 2 and 3 (NOAA and Bruner's Center sites) are
located near areas previously identified as having high mercury concentrations in soils.
Coordination with BMAP sampling personnel ensured that no adverse effects occurred to these
-------
ES-79
Herori and Mink L.
Stoneroller
I
i
k i
i
. . vv V.V • v. ._" V-Vv\ •_
largemoMi "
^ Bassr^
: ¦%^^:^Nn'\\'>v-.-xx->s:: .: :*: x*
t
Redbreast
Sunfish
Periphyton
Water Column
Sediment
Crayfish 1
i
k
Ben
Inverte
thic I
ibrate 1
Fig. ES.21. Food web relationships of aquatic biota sampled (clear boxes) or modeled
(shaded boxes) for EFPC ecological risk assessment.
-------
~ N.
Wren
Mouse
Vole
Insect
Shrew
Earthworm
Plant
Soil
Fig. ES.22. Food web relationships of terrestrial biota sampled (clear boxes) or modeled
(shaded boxes) for EFPC ecological risk assessment.
-------
ES-81
'onfl icnci
¦*—t> -finr pri fa-
Cra
e Laie
2Z
Gir s a
"5 fr
R£FE ?ENCI
MIL
NCH
REGIONAL MAP
Fish/Benlhos
jnsects
LECEND
EAST FORK
POPLAR CREEK
100 YEAR
FLOODPLAIN
Scffoce Aiphcoltons
International Corporohon
Birds
-------
ES-83
sampling locations during the RI sampling and facilitated comparisons with the BMAP results.
Site 3, or the Burner's site, also serves as an upstream reference for impacts of the Oak Ridge
Sewage Treatment Plant. Site 4, located below the plant outfall, was selected to enable the
assessment of impacts of the treated municipal wastewater on the biota. The last two sites are
located in the lower reaches of EFPC; Site 5 is upstream from Bear Creek, while Site 6 is
downstream. Where possible, terrestrial biota were sampled at the same sites. Sites were also
selected to allow easy as opposed to difficult access for sampling equipment and personnel.
Because EFPC originates on the Y-12 Plant site, an aquatic reference site located upstream
from the EFPC OU could not be obtained. Thus, a site on Hinds Creek near Norris, Tennessee,
was selected as the aquatic reference site (Figs. ES.24 and ES.25). The Hinds Creek site was
selected as the aquatic reference site because the available information on surface and subsurface
geology, hydrology, anthropogenic contaminant sources, and residue analyses in fish all indicated
that this site most closely approximated the EFPC watershed. The similarity between the Hinds
Creek and EFPC watersheds, coupled with the relative absence of anthropogenic contaminants
at Hinds Creek, qualified the site as adequate for making comparisons with the EFPC OU. In
addition, BMAP has been using a location in Hinds creek for its reference site for several years
with a allows data comparability with this ERA investigation.
Site 6 on Mill Branch (Fig. ES.25), a tributary within the EFPC watershed, was selected as
the terrestrial reference site. Several criteria were used in the selection process. All watersheds
with known or suspected contamination were automatically eliminated from consideration, and
an attempt was made to locate a site with similar vegetation. The Hinds Creek soil reference site
was judged to be unsuitable for use as the terrestrial vegetation reference site because the
vegetation surrounding the Hinds Creek site is all pasture, while most of the study sites along
EFPC are forested. The Mill Branch site, like the Hinds Creek site, was believed to have
received minimal anthropogenic chemical contamination. On this basis, the Mill Branch site was
judged to be an adequate reference site for terrestrial fauna for comparisons with the EFPC OU.
The location of the site within the EFPC watershed was also regarded as a positive factor.
Despite the advantages of the Mill Branch site, however, its use as a reference site for terrestrial
biota is somewhat limited by the fact that it supports a small, immature upland forest community
rather than the many acres of mature bottomland hardwood communities typically found along
EFPC. No noncontaminated reference site for terrestrial vegetation was found that was similar
to the EFPC floodplain with respect to species composition and therefore suitable for direct
comparisons.
-------
ES-84
J
ft
(0,c?/
/. -:
0 100 200 300 400 500
1 1 1 1 1 1
Scale in meters
TTT7
' ¦ V-'
HINDS CREEK ECOLOGICAL REFERENCE
SAMPLING LOCATIONS
LEGEND:
# Insect Sampling
¦¦ Aquatic Sampling
tb—nr
Science Applications
International Corporation
-------
ES-85
1000
VMill Branch
Scale in meters
MILL BRANCH ECOLOGICAL REFERENCE
SAMPLING LOCATIONS
LEGEND:
Bird Survey
Plant Survey
i
J Mammal Sampling
rm—nr
Science Applications
International Corporation
-------
ES-86
In addition to the aquatic and terrestrial species sampling conducted at the EFPC and
reference sites, grasses and browse were taken at three sites and browse was taken at one site
along the EFPC floodplain. Grasses and browse were also sampled at the Hinds Creek Reference
site. At each EFPC site, a sample of grass or browse wais taken at the creek edge and at a
distance of 100 m from the creek. Beets, kale, and tomatoes were grown at a special vegetable
plot located on the floodplain and a reference location outside the floodplain, and the plant roots,
leaves, and fruits were sampled. These vegetation measurements supplemented the historical
studies on woody species in EFPC. Also important to the ERA were the environmental sampling
results for soils, sediments, surface water, groundwater, and air discussed earlier in this
Executive Summary.
Indicator species were selected by a process, outlined in Habitat Evaluation Procedures ESM
102 (FWS 1980), that relies on groups of organisms that occupy a common environmental
resource (e.g., sediment, water column, vegetation). The indicator organisms were used as
surrogates for the groups of organisms mentioned in the assessment and measurement endpoints.
It was judged to be impractical to directly measure and/or model exposure, effects, and risk to
all the species represented by the assessment/measurement endpoints. Factors included in the
selection of indicator species included their presence in EFPC, ability to obtain measurable
responses, adequate sample size, phenology, and small home range within the EFPC OU to
ensure exposure. In addition, some species that did not necessarily meet these criteria, such as
threatened or endangered species and other species of concern (e.g., migratory birds), were
selected because they were important as assessment endpoints. Species of concern included those
that are vital to ecosystem energy and biogeochemical cycling, are particularly susceptible or
sensitive to contaminants, and are specific bioaccumulators. Also, plants were selected to
represent the lower trophic levels in the food chain to allow modeling up the food chain.
Aquatic indicator species included fish [common stonerollers (Campostoma anomalum),
redbreast sunfish (Lepomis auritus),largemouth bass (Micropterussalmoides)], sediment-dwelling
invertebrates (benthic macroinvertebrates and crayfish), and periphyton (algae and heterotrophic
microbes) attached to submerged surfaces in the substrate. Terrestrial indicator species included
small mammals [short-tailed shrew (Blarina brevicauda) and white-footed mouse (Peromyscus
leucopus)], birds [Carolina wren (Ihryothorus ludovicianus) and great blue heron (Ardea
herodias]), invertebrates (earthworms and adult terrestrial insects) living in or on the soil, and
plants (grasses and vegetables). These aquatic and terrestrial indicator species were targeted for
population surveys and body burden analyses to represent the key exposure points in the EFPC
ecosystem. A summary of the types of measurements and analyses conducted as part of the ERA
is provided in Table ES.6.
-------
ES-87
Table ES.6. Measuranents made for the EFPC ERA
Population measures
¦ J
Matrix
sampled: .
Relative -
abundance
Bl/rmnot
Diversity
Contaminant
. concentration
Colonization
rate
Growth
rate
Aquatic
Fish
X
X
X
X
Benthic macro-
invertebrates
X
X
X
Periphyton
X
X |
Water
X
Sediments
X
1
Terrestrial
Small numnuls
X
X
X
X
Birds
X
X
X
Insects
X
X
X
X
Earthworms
X
X
X
Vegetation
X
X
X
Soil
X
|
-------
ES-88
The field investigations were conducted from the late summer of 1991 through the late
summer of 1992. The aquatic ecosystem is considered to be representative under conditions of
stable base flow. The late summer and early fall were selected for intensive field surveys to
satisfy the need for low rainfall conditions typically experienced during these seasons. Because
of the present knowledge of the area, abundance of existing biological data on EFPC, and the
schedule-driven constraints for the EFPC RI, as specified in the FFA, no initial field screening
was performed. The schedule and funding level allowed the one-season approach. In addition,
fall sampling allowed data comparability with BMAP's fall sampling of fish and benthic
macroinvertebrate communities. Communication and coordination with BMAP sampling
personnel ensured that no adverse effects occurred to the BMAP sampling locations or BMAP's
long-term studies. Without this communication with BMAP personnel, sampling crews could
have inadvertently physically damaged or sampled existing riffles used by BMAP for their benthic
macroinvertebrate studies, thereby potentially leading to benthic community data that were not
representative of the site(s). Data from the BMAP—a multi-season and multi-year study—were
used to round out the temporal aspects of the characterization.
ES.3.2.2 Exposure characterization
Exposure environments in EFPC and its floodplain consist of a heterogeneous mixture of
flow regimes, streambed conditions, and terrestrial habitats varying from commercial through
agricultural to woodland. Surface water is a primary medium by which aquatic organisms are
exposed to contaminants currently being released to EFPC via surface water from the Y-12 Plant
and to contaminants in floodplain soils that are transported via erosion, surface runoff, and
shallow groundwater. Soil is the primary medium of exposure for terrestrial biota. Stream
sediments are a second major exposure route for benthic invertebrates and for terrestrial
consumers by way of flood deposits on soil and vegetation. In both aquatic and terrestrial
communities, indirect exposure of organisms via the food chain is important.
Methylation/demethylation of mercury in sediments and stream water and biodegradation of
organic contaminants are the most important chemical transformations. Methylmercury is
assumed to be the predominant form of mercury in aquatic biota. In the absence of remediation,
the exposures of EFPC communities to contaminants are not expected to change dramatically
from those levels identified through the Phase la and lb sampling of the soils, sediments, surface,
and groundwater along EFPC and the SLB. Table ES.7 summarizes the expected dominant mode
of exposure for the EFPC ecosystem, based on the identified media of exposure and the
respective indicator species.
-------
ES-89
Table ES.7. Dominant mode of exposure of indicator organisms
to contaminated source media in EFPC
| Indicator biota *
Medium.. J
Surface
-"-water'
Tntfn»nm
sediment
' 'Groundwater"
sba
Biota |
Fish
D
D/I
I
*
D
| Benthic invertebrates
D
D
I
•
D
Aquatic insects
D
D
I
*
D
Crayfish
D
D
I
»
D
Periphyton
D
D
I
*
* 1
.Srn«" mammals
D
*
I
D
D |
Birds
D
I
I
I
D |
Terrestrial insects
*
*
I
D
D 1
Earthworms
*
*
I
D
* 1
Plants
D
D
I
D
* E
D = Direct cxpoiurc
* = Incomplete pathway or considered negligible
I = Indirect exposure
-------
ES-90
Exposure conditions of the indicator species were characterized by implementing the ERA
sampling plan and incorporating historical data where appropriate to fill in daca gaps. The ERA
sampling results are summarized in Table ES.8 for aquatic biota and in Table ES.9 for terrestrial
biota.
COPCs in biota were screened to eliminate combinations of analytes and biota in which body
burdens in EFPC and floodplain samples were undetectable, were below the levels in reference
site biota, or were less than five times the concentration in blank samples. Metals, PCBs,
chlordane as a representative of pesticides, and PAHs were retained as COPCs for various biota.
COCs were chosen after risk characterization for these analytes as described in die following
sections.
Both historical and current studies of bioaccumulation show (1) higher body burdens of
contaminants (mercury, PCBs, and various pesticides, PAHs, or metals) in common stonerollers,
redbreast sunfish, crayfish, earthworms, and terrestrial insects at EFPC sites than at
noncontaminated reference sites and (2) generally decreasing body burdens with increasing
distance downstream from the Y-12 Plant. Current body burdens are generally lower than those
from the 1980s, possibly as a result of prior remediation activities at the Y-12 Plant. Evidence
of ecological recovery in the upper part of EFPC above and below Lake Reality has been
documented by BMAP. Whole-body mercury concentrations found in redbreast sunfish during
the BRA were 30 to 50% less than concentrations found in bluegill (Lepomis macrochirus) and
redbreast sunfish taken from similarly located sites in studies conducted by ORNL in 1982 (Van
Winkle et al. 1984) and by TVA in 1984 (TVA 1985). Other notable findings of the EFPC
ecological BRA, which indicated a change in the general patterns documented by historical
studies, were the high present body burdens of mercury in redbreast sunfish from a site located
6.4 km (3 miles) downstream from Lake Reality and increasing body burdens of PCBs in sunfish
at this site and other sites downstream.
ES J.23 Effects characterization
Data on the toxicity of inorganic and organic mercury compounds to fish demonstrate a
positive relationship between dose and effect. Toxicity tests on a variety of aquatic and terrestrial
organisms using EFPC surface water, soils, or sediments have shown deleterious effects, and in
three studies the effects were greatest at the sites nearest the Y-12 Plant. The results of surveys
of indicator organisms at EFPC also indicate effects that most likely occurred from exposure to
contaminants released from the Y-12 Plant. The number of indicator fish species, benthic
macroinvertebrate families, and insect adult individuals at EFPC generally increased with
-------
Table ES.8. Summary table of trends for whole-body concentrations of contaminants
in aquatic biota collected from EFPC and Hinds Creek, October 7-29, 1991
PCBj
=^sssaaaBs=sssexs=5ss
Taxon
Mercury
Other Inorganics
Aroclor 1260
Other
Pesticides
PAH$
Redbreast
• In samples from all six
• Uranium was
• In samples from all
• All other
• Most were less than
• All individual PAHs
sunfish
sites in EFPC and
below sample
six sites in EFPC
mixtures were
sample detection
were below sample
Hinds Creek.
detection at all
and Hinds Creek.
below sample
limit.
detection limits
• Maximum at Site 3.
sites,
• Maximum level at
detection limits.
* Chlordane, dieldrin,
except for
• Decreasing body
Site 3.
and heptachlor were
acenaphthene.
burdens downstream
• Decreasing body
in samples from all
• Acenaphthene body
from Sites 3 to 6.
burdens.
sites, including
burden was greatest
• All samples in EFPC
Downstream from
Hinds Creek.
at Site 1, then
exceeded concentrations
Sites 3 to 6.
• Decreasing body
nearly constant at
in sample from Hinds
• All samples in
burdens from Sites 1
remaining EFPC
Creek.
EFPC exceeded
to 2, then increasing
sites.
concentrations in
to maximum at
* All samples in
sample from Hinds
Sites 4 or 5.
EFPC exceeded
Creek.
• Nearly all sites in
concentrations in
EFPC exceeded
samples from Hinds
Hinds Creek.
Creek.
Common
• In all six samples from
• Uranium in
• Sufficient sample at
• All the mixtures
• Same sites as for
• Same sites as for
stoneroller
EFPC and Hinds
samples from
Sites 1 to 5 and
were below
PCBs.
PCBs.
Creek.
all sites.
Hinds Creek.
sample detection
• Maximum at Site 1.
* Maximum at Site 1.
• Maximum at Site 2.
• Maximum
* Maximum at
limit.
• General trend of
• Same trend as for
• Decreasing body
body burdens
Site 1.
decreasing body
pesticides.
burdens downstream
at Site 1.
• Decreasing body
burdens from Sites 1
• Nearly all sites in
from Sites 2 to 6.
• Decreasing
burdens
to 4.
EFPC exceeded
• All samples in EFPC
body burdens
downstream from
• All samples in EFPC
sample from Huids
except Site 6 exceeded
downstream
Sites 1 to 4.
exceeded
Creek.
Hinds Creek sample.
from Sites 1
• All samples in
concentrations in
to 6.
EFPC exceeded
samples from Hinds
• Sites 1 to 4
concentrations in
Creeks.
exceeded
samples from
Hinds Creek
Hinds Creek.
sample.
-------
Table ES.8 (continued)
PCBj
Tixon
Mercury
Other inorganics
Aroclor 12(0
Other
Pesticides
PAHs
Bcnthic
• Sufficient sample only
• Same sites as
• Sufficient sample
• All other
* Same sites as for
* Same sites as for
macro-
at 4 sites in EFPC and
for uranium
only at 2 sites in
mixtures below
PCBs.
PCBs.
inverte-
Hinds Creek.
and mercury.
EFPC (Sites 1
sample detection
• Nearly all were
• Most individual
brates
• Maximum at Site 2;
• Same trends as
and 2).
limit.
below sample
PAHs were similar
other sites similar to
for mercury.
• Maximum at Site 1
detection limits.
to concentrations in
each other.
(6-fold greater than
samples from the
• All samples from
at Site 2).
two sites.
EFPC exceeded sample
• Fluoranthene was
from Hinds Creek.
maximum at Site 1.
• Benzo(a)pyrene was
maximum at Site 2.
Crayfish
• Sufficient sample at 4
• Same sites as
• Same sites as for
• All other
• Same sites as for
• Sufficient sample
sites in EFPC and
for mercury.
mercury.
mixtures below
mercury.
only at Sites 1 to 3,
Hinds Creek.
• "ranium was
• Same trends as for
sample detection
• Nearly all were
and Hinds Creek.
• Maximum at Site 1.
below sample
mercury.
limit.
below sample
• Maximum at Site 3.
• Decreasing body
detection limit
detection limits.
• Most individual
burdens downstream.
at most sites.
PAHs increased
• All samples from
• Selenium body
downstream from
EFPC exceeded
burdens were
Sites 1 to 3.
concentrations in
greatest in
• Most PAHs at the
samples from Hinds
samples for
3 sites in EFPC
Creek.
Site 3, and all
EFPC samples
exceeded
concentrations
in samples
from Hinds
Creek.
exceeded
concentrations in
sample from Hinds
-------
Table ES.9 Summary table of trends for whole-body contaminant concentrations in terrestrial biota
collected from EFPC and reference site in late 1991
PCBs
¦ '
Taxon
j _ ..
Mercury
Other inorganics
Aroclor 1260
Other
Pesticides
PAHs
White-footed
mouse
• Levels below
detection limit in
9 of 11
individuals.
• Maximum value
of 1.105 mg/kg
at Site 2.
• Antimony,
chromium, and
selenium
measured above
detection limit in
>50% of
individuals.
• Above detection
limits in all
samples except
reference.
• Range from
0.051 mg/kg to
0.480 mg/kg,
with highest at
Site 1.
• All others
below
detection
limit.
• Nearly all were
below detection
limits.
• One value for
DDT and one
for heptachlor
oxide above
detection limits.
• All levels below
detection limits
or twice
reference levels.
Short-tailed
shrew
• Maximum of
7.9 mg/kg at
Site 5.
• Uranium not
detected.
• Antimony,
chromium, and
selenium similar
to mice.
• Maximum of
1.4 mg/kg at
Site 5.
• Not detected.
• Nearly all were
below detection
limits.
• Not analyzed.
Woodland
vole
• Less than
detection limit.
• Chromium levels
similar to mice.
• Uranium not
detected.
• Reported at
0.046 mg/kg.
• Aroclor 1016
at 0.073
mg/kg.
• All below
detection limit.
• No unqualified
levels above
unqualified
-------
Table ES.9 (continued)
Taxon
Mercury
Other inorganics
PCBs
Pesticides
PAHs
Aroclor 1260
Other
Great blue
heron
• Breast feathers at
5.3 mg/kg
• Liver sample at
9.0 mg/kg.
• Chromium,
selenium, and
zinc above
detection limits in
feathers.
• Selenium and
zinc above
detection limits in
liver.
• Liver sample at
1.4 mg/kg.
• All others
below
detection
limit.
• Several reported
but most below
the detection
limit.
• No unqualified
levels above
detection limit.
Carolina
wren
a Levels at 3.5
mg/kg in both
samples.
• Antimony,
selenium, and
cadmium above
levels in
terrestrial insects.
• Average of
2.1 mg/kg is
10 times
average in
terrestrial
insects.
• No others
detected.
• Only DDE and
heptachlor
epoxide detected
above detection
limits.
• None above
detection limit.
Flying
insects
• Most samples
below detection
limit.
• Maximum levels
of 3.2 mg/kg at
Site 3.
• No correlation
between
measured
concentrations in
insects and soils.
• Uranium and
cadmium not
detected.
• Antimony,
chromium, zinc,
arsenic and
selenium noted
above detection
limits at 2 sites.
• Range of
0.02-0.35
mg/kg.
• Maximum at
Site 2 and
decreased
steadily with
distance from
Y-12 Plant.
• All levels
below
detection or
reference site
levels.
• DDE reported
at Site 3,
chlordanes at
Site 4.
• Others not
identified above
twice reference
site levels.
• Few detected;
highest value at
-------
Table ES.9 (continued)
PCBs
Taxdn:
. Mercury |
Other inorganics
Aroclor 1260
Other
Pesticides
PAHs
Earthworms
• Range from 5 to
33 mg/kg.
• Maximum
observed at Site
2, with a steady
decrease
downstream.
• No good
correlation
between
earthworm
composites and
soil levels.
• Arsenic,
cadmium,
selenium,
uranium, zinc,
and chromium
levels above
reference site.
• Not sufficient
sample for
analysis.
• Not sufficient
sample for
analysis.
• Not sufficient
sample for
analysis.
• Not sufficient
sample for
analysis.
Grass/
vegetables
• Geometric mean
grass/browse
concentration of
0.34 mg/kg, with
maximum of
17 mg/kg.
• Grasses near
creek showed
higher levels.
• Levels in
vegetables ranged
from <0.03 to
3.2 mg/kg.
• Data not
available.
• Not measured.
• Not
measured.
• Not measured.
-------
ES-96
increasing distance downstream from the Y-12 Plant. Compared with the Hinds Creek reference
site, EFPC had fewer species of fish, fewer families of benthic macroinvertebrates, and lower
mean numbers of individuals per sample of adult insects with terrestrial larval forms. The
number of fish species classified as tolerant of degraded conditions decreased downstream in
EFPC (Fig. ES.26). Taxonomic diversity (H*) of fish, benthic macroinvertebrates, and adult
insects increased downstream, except at Site 3, and was lowest at the three sites where surface
water mercury concentrations were greatest (Fig. ES.27). Taxonomic diversity for fish, benthic
macroinvertebrates, and adult insects with aquatic larval forms at EFPC sites was generally less
than at the reference site. The richness of families from sensitive orders of
arthropoda—Ephemeroptera, Plecoptera, and Trichoptera (EPT richness)—was lowest at the
EFPC site nearest the Y-12 Plant and was lower at all EFPC sites than at the reference site
(Fig. ES.28). In general, the results of the benthic macroinvertebrate sampling conducted to
support the ERA parallel those of the fish sampling, suggesting that both communities (1) are
dominated by large numbers of individuals from only a few of the total number of families
present, (2) are dominated by tolerant organisms, (3) have less total species/taxa or EPT richness
than the reference stream, and (4) have lower diversity indices than the reference stream. Taken
together, these factors indicate that the benthic and fish communities in EFPC, especially in the
upper part of the stream below Lake Reality, are experiencing environmental stress.
Although taxonomic richness (and presumably diversity) might be expected to increase in
a typical stream as its drainage area or stream order increases (Fausch et al. 1984), the argument
may be less applicable to EFPC because it does not possess the characteristics of a true headwater
stream due to the augmented flow from the Y-12 Plant. Therefore, the increased taxonomic
richness and diversity downstream from the Y-12 Plant is probably due more to reduction in
toxicant concentrations downstream that to the increase in drainage area.
Studies of periphyton colonization and growth, populations of small mammals, earthworms,
birds, and terrestrial vegetation showed less consistent patterns. A bird survey of EFPC in the
fall of 1991, as well as other bird surveys, identified more than 30 species of migratory birds,
which were likely ingesting contaminated worms and arthropods. White-footed mice were
captured at all EFPC sites except Site 6. Short-tailed shrews, a woodland vole (Microtus
pinetorum), and an eastern chipmunk (Tamias striatus) were captured at one or more of the sites
furthest from the Y-12 Plant. A limited vegetation survey revealed no patterns associated with
exposure to contaminants released from the Y-12 Plant. Earthworm contaminant body-burdens
generally decrease as a function of distance from the Y-12 Plant. The threatened and endangered
-------
00
90
80
70
60
50
40
30
20
10
0
Percent (%)
% Tolerants
Rel. Abund.
HC
4
Site
Fig. ES.26. Percentage of captured fish species from EFPC and Hinds Creek classified
-------
Diversity Index (H')
1.4
O H' max
0.8
0.4
0.2
0.0
6
HC
5
3
4
2
1
Site
Fig. ES.27. Observed fish species diversity and maximum possible fish diversity at sites
-------
Total and EPT richness (no. of families)
Individuals/m
HC
Total Richness
18 EPT Richness
S- Density (ind./ra*)
4
Site
Fig. ES.28. Total family richness, EPT richness, and density (individuals per m1) or
-------
ES-100
species survey, both literature and field, revealed no species requiring protection of their
populations or their support habitats.
ES3.2.4 Risk characterization
Risk characterization aggregates the effects of exposure and stressor response on indicator
organisms, summarizes risk according to the weight-of-evidence approach, and interprets the
ecological significance of these findings. This ERA applied two interrelated approaches to
ecological risk characterization for EFPC and its floodplain. The first was the application of the
quotient method to various criteria for protection of ecological resources. In this method,
measured concentrations are compared to threshold concentrations for protection of ecological
resources (e.g., assessment endpoints). The ratio of the observed contaminant concentration to
the protection thresholds indicates the severity of the impact or risk, with larger quotients
indicating greater potential impacts or risks. The second approach was to apply weight-of-
evidence arguments to evaluate the strength of the relationship between stressors and observed
effects on indicator species and the implications for assessment endpoints.
Weight-of-evidence helps to identify causes of observed ecological responses, using
arguments derived from human epidemiology. In this approach, a causal relationship between
a stressor and a response is proposed. Then, a series of questions or criteria is applied to the
proposition. Not all criteria must be satisfied to demonstrate that the proposition is true, but
weight is added to a conclusion by each criterion that is satisfied in the proposition(s).
Ultimately, professional judgment is used to establish the strength of the causal relationship. The
weight-of-evidence approach is especially useful when (1) there are multiple lines of evidence to
evaluate, (2) there are insufficient data for robust statistical analyses, (3) toxicity or other criteria
are uncertain, or (4) exposure models are not sufficiently precise for statistical testing. As
described by Suter and Loar (1992), the weight-of-evidence approach to ecological risk
assessment is analogous to a civil court case in which physical evidence, witness accounts, and
expert testimony are considered together to reach a verdict based on the preponderance of
evidencie. Rather than relying on a single line of evidence, the weight-of-evidence approach
ideally incorporates three categories of biological investigation. First, toxicity testing of water,
soil, and sediment to which biota are exposed provides direct measurement of effects on animal
and plant indicator species. Second, biological surveys assess the current state of biota on the
site (i.e., field-observed effects), and the results can be compared to a reference site or to the
observations expected in the absence of contamination. Finally, body burdens of site-related
contaminants can be measured as evidence of exposure and possible effects.
-------
ES-101
The ecological risks in this study are based on actual body burden data and surveys of
organisms occupying the floodplain, while human health risks are not based on actual
measurements of exposure in human bodies and are calculated by using standard exposure
scenarios and exposure to a form of mercury that is a small portion of the mercury in the
floodplain.
Quotient Methods Results. Risks to biota can be characterized by comparing exposure
concentrations to specific threshold concentrations intended to be protective of the organism or
its function in the ecosystem. Thus, criteria may be established to limit the concentrations of
chemicals in surface water in order to protect aquatic biota (TDEC, Chap. 1200-4-3), or limits
may be set on contaminant concentrations in tissues in order to protect organisms or their
predators. Thresholds may also be established by modeling the transfer of contaminants from
exposure media through the food chain to the various biotic receptors, thereby taking into account
bioaccumulation of contaminants. In the quotient method, if the ratio of exposure concentration
to threshold concentration is < 1, there is no problem or risk is considered insignificant, whereas
if the quotient is sufficiently > 1, more study or computation is needed to define the exact degree
of the risk that is inferred. Because of the uncertainty surrounding both sampling data and the
threshold concentrations themselves, this approach requires careful interpretation of quotients near
1, but it provides a clear indication of high 10) or low (<0.1) risks. The quotient method
requires consideration of the uncertainty of estimating both the exposure concentration and the
threshold values.
To address the assessment endpoints identified in Table ES.5 for the EFPC ERA, specific
criteria and threshold concentrations were established for (1) ambient water quality; (2) aquatic
biota and predator dietary uptake; (3) and body burdens in terrestrial mammals, birds,
earthworms, and insects and predator dietary uptake. Threshold values for specific COPCs (in
particular, mercury, other heavy metals, PCBs, rhlordane, and PAHs) were established for each
of these measures, and measured values in EFPC media and/or biota were compared to determine
a quotient range. Tables ES. 10 and ES. 11 present the results of the quotient method for aquatic
and terrestrial exposures, respectively. The tables identify those organisms at each EFPC site
that fall within five quotient ranges. As noted in the quotient ranges, risks 10) are being
predicted for both aquatic and terrestrial biota, primarily through mercury toxicity to predators
through the food chain. The higher quotients tended toward the upstream sample locations along
EFPC, with little ecological risk noted for the reference sites. Several additional findings are
highlighted below.
-------
ES-102
Table ES.10. Classification of risk quotients for aquatic exposures
by range and location1
Sample site
Quotient range*
Q>100
10<;Q<100
1^Q<10
0.1<;Q<1
Q<0.1
1
WQC, Hg
AI, pred.
SR, PCB tox.
SR, PCB pred.
CR, Hg tox.
CR, PCB pred.
RS, PCB tox.
CR, PCB tox.
RS, PCB pred.
WQC, Zn
SR, Hg tox.
RS, Hg tox.
2
WQC, Hg
SR, PCB tox.
SR, PCB pred.
SR, Hg tox.
CR, Hg tox.
AI, pred.
RS, PCB tox.
RS, PCB pred.
WQC, Zn
RS, Hg tox.
CR, Hg tox.
CR, PCB tox.
CR, PCB pred.
3
WQC, Hg
AI, pred.
SR, PCB tox.
RS, PCB tox.
SR, PCB pred.
RS, PCB pred.
WQC, Zn
SR, Hg tox.
RS, Hg tox.
CR, Hg tox.
CR, PCB tox.
CR, PCB pred.
4
AI, pred.
SR, PCB tox.
RS, PCB tox.
SR, PCB pred.
RS, PCB pred.
WQC, Zn
SR, Hg tox.
RS, Hg tox.
5
HN, pred.
AI, pred.
HN, Hg tox.
SR, PCB tox.
RS, PCB tox.
SR, PCB pred.
RS, PCB pred.
WQC. Zn
SR, Hg tox.
RS, Hg tox.
HN, PCB tox.
6
SR, PCB tox.
SR, PCB pred.
WQC, Zn
CR, PCB tox.
CR, PCB pred.
SR, Hg tox.
RS, Hg tox.
CR, Hg tox.
CR, pred.
-------
ES-103
Table ES.10 (continued)
Sample site
Quotient range*
Q>100
lO^QclOO
lsiQclO
0.1^Q<1
Q<0.1
Reference
WQC, Zn
SR, Hg pred.
RS, Hg pred.
CR, Hg pred.
AI, Hg pred.
SR, Hg tox.
RS, Hg tox.
CR, Hg tox.
SR, PCB tox.
RS, PCB tox.
CR, PCB tox.
SR, PCB pred.
RS, PCB pred.
CR, PCB pred.
¦Key:
WQC
= Water quality criteria
SR
= Common stoneroller
RS
= Redbreast sunfish
CR
= Crayfish
Al
= Aquatic insect
tox.
= toxicity to indicator organism
pred.
= toxicity to predators
HN
— Great blue heron
k Quotients based on concentrations below detection limits were not included.
-------
ES-104
Table ES.ll. Classification of risk quotients
for terrestrial exposures by range and location"
Sample
ate
Quotient range*
Q> 100
lOsjQclOO
lsQclO
0.1s;Q
-------
ES-105
• Toxicological effects. Comparison of mercury body burden concentrations with
concentrations that have been proposed as protective against toxicological effects in the
indicator organisms [0.5 mg/kg (Eisler 1987)] showed that 2 of 7 crayfish samples and 1 of
10 common stoneroller composite samples had mercury concentrations exceeding
toxicological limits (5 mg/kg body weight). Body burdens of mercury in short-tailed shrews
were considerably above the toxicological limit (1.1 mg/kg body weight), and the feathers
and liver of a great blue heron also exceeded toxicological limits (1.1 mg/kg).
• Protection of predators. Comparison of body burdens of contaminants in aquatic biota with
concentrations expected to protect their predators from toxicological effects showed that
mercury concentrations in fish exceeded the threshold concentration for mercury by ten-fold
or more at sites 1,2, and 3. Concentrations in crayfish exceeded the threshold concentration
for mercury by at least ten-fold at the two sites closest to the Y-12 Plant, and mercury
concentrations in aquatic insects exceeded the threshold concentration by at least ten-fold at
two sites. Application of the same comparison to low-level terrestrial predators (small
mammals and song birds) preying on contaminated organisms showed that undepurated
earthworms at every site exceeded threshold concentrations for dietary mercury. Assuming
all the mercury in the undepurated earthworms is bioavailable, exposure was calculated to
be high, even if earthworms were only a small fraction of a predator's diet.
Bioaccumulation factors for mercury in earthworms may be on the order of 0.1, so that the
contribution of earthworm tissue to the predator's dietary exposure is probably much lower
than indicated by the undepurated body burden. Insectivores feeding exclusively or
predominantly in contaminated areas of the floodplain would be at risk from ingestion of
mercury. Excessive levels of mercury in short-tailed shrews and Carolina wrens likely result
from their consumption of earthworms and insects, respectively, and, in turn, contribute to
the risk of those animals that prey on them. Other metals, pesticides, and PAHs did not
appear to pose a risk to aquatic and low-level terrestrial predators. PCBs in common
stonerollers exceed threshold concentrations at sites 1 and 2 but did not appear to pose a risk
in other biota or in stonerollers at other sites.
Top predators, represented by hawks, owls, and foxes, were assumed to feed on a mix of
the biota sampled in three ecological risk zones, comprising risk assessment segments 1 through
4,5 through 8, and 9 (Fig. ES.8), respectively, as well as the reference areas at Hinds Creek and
Mill Branch. The high-end exposure was calculated with the assumption that the top predators
eat no uncontaminated prey. In this case, the composite prey body burdens of mercury exceeded
dietary threshold concentrations for hawks and owls by at least 25-fold in all three zones. The
-------
ES-106
mercury threshold for foxes was not exceeded in any zone, nor were thresholds foF other metals,
PCBs, pesticides, or PAHs. When the estimated exposures were adjusted to allow for dietary
intake of uncontaminated prey from outside the contaminated areas, dietary thresholds were not
exceeded for hawks or foxes and were exceeded for mercury by less than two-fold in the modeled
diet of owls.
Weight-of-evidence results. The weight-of-evidence analysis focused on risk to both aquatic
and terrestrial ecological receptors. Three propositions were advanced for aquatic biota; these
propositions were conclusions based on the results of the ERA field sampling as well as on
historical studies of EFPC. The three propositions each began with the statement, "Continual
releases of water-borne contaminants (including dissolved and particle-bound) such as mercury,
uranium, other inorganics, pesticides, PAHs, and Aroclor 1260 from the Y-12 Plant," and are
as follows:
Proposition 1 — Continual releases of water-borne contaminants . . . from the Y-12 Plant
are the largest source of the elevated whole-body burdens of these contaminants in fish,
benthic macroinvertebrates, and crayfish in EFPC, which result in risk from toxicological
effects to indicator organisms and risk to piscivorous predators feeding on the organisms.
Proposition 2 — Continual releases of water-borne contaminants . . . from the Y-12 Plant
impact fish community structure in EFPC, resulting in communities dominated by species
tolerant of degraded water quality conditions..
Proposition 3 — Continual releases of water-borne contaminants . . . from the Y-12 Plant
result in reduced taxonomic richness and diversity of the fish and benthic macroinvertebrate
communities in EFPC compared to the same measurements in communities at the reference
site.
The aquatic propositions were evaluated against five criteria (temporal association, spatial
association, stressor-response, strength of association, and plausibility) and qualitatively ranked
(as strong, moderate, or weak) based on the level of available evidence. Major observations from
the ERA field sampling and historical studies indicate that numerous COPC body burdens in the
aquatic indicator organisms (common stoneroilers, redbreast sunfish, crayfish, and benthic
macroinvertebrates) as well as several measures of community structure (fish and benthic
macroinvertebrates, taxonomic richness and diversity, and percentage of fish species classified
as tolerants) are greatest at locations closest to the Y-12 Plant, then decrease downstream in
EFPC. In addition, COPC body burdens and community impacts in biota from all sites in EFPC
usually were greater than equivalent impacts at the Hinds Creek reference site.
-------
ES-107
The data for mercury and PCBs indicate a clear temporal association between the release of
these two contaminants from the Y-12 Plant and their subsequent bioaccumulation in aquatic biota
downstream in EFPC. Several variables (such as the percentage of captured fish species that are
classified as tolerant of degraded water quality conditions, fish and benthic macroinvertebrate taxa
richness, and species diversity) have consistently indicated that aquatic biota in EFPC are
adversely impacted in comparison to biota in the Hinds Creek reference stream. However,
BMAP researchers have concluded that recent increases in taxa richness of fish and benthic
macro invertebrates, as well as increased survival and growth of clams in EFPC below Lake
Reality, indicate ecological recovery that is temporally related to improved control of toxicant
releases from the Y-12 Plant (Appendix R, pages R-36 through R-65).
Both current and historical evidence suggest that spatial associations are strong for all three
propositions. Body burdens of many COPCs were greatest in biota collected closest to the Y-12
Plant and decreased downstream. Also, the COPC body burdens in biota from most sites in
EFPC were greater than the body burdens in biota from the reference site. Similar trends were
observed for the percentage of tolerant fish species (Proposition 2). Fish and benthic
macroinvertebrate taxonomic richness and diversity (Proposition 3) generally increased
downstream from the Y-12 Plant, indicating that the most severe impacts occurred closest to the
plant.
Stressor-response associations were evaluated by determining linear correlation coefficients
between estimated COPC exposure concentrations (the stressor) and the magnitude of some
response variable (e.g., percentage of fish classified as tolerants). In the absence of COPC
concentration data in surface water, COPC body burden concentrations were used as an indicator
of the stressor exposure level.
Mercury in surface water was significantly correlated with mercury body burdens in
crayfish. Of the COPCs that were below detection limits in surface water, ten were present in
at least one of the aquatic indicator organisms at body burden concentrations that were maximum
in samples collected closest to the Y-12 Plant, then decreased downstream. Mercury
concentrations in surface water were correlated with the percentage of fish classified as tolerants,
but zinc concentrations in surface water were not correlated (Proposition 2). Five contaminant
body burdens in-common stonerollers were correlated with the percentage of fish classified as
tolerants. Evidence of stressor-response associations between COPC concentrations and
taxonomic richness or diversity (Proposition 3) were practically absent. Thus, the data suggest
a strong stressor-response association for Proposition 1, a moderate association for Proposition 2,
and a weak association for Proposition 3.
-------
ES-108
Strength of association for all three propositions is strong. For example, body burdens of
most contaminants decreased downstream from the Y-12 Plant, whereas urban runoff and point
source contaminants increased downstream. Habitat-altering activities, such as dredging or
impounding (which could have potential adverse impacts to fish and benthic macroinvertebrate
communities), have not been conducted extensively in EFPC. Concentrations of residual
chlorine, which is a potential chemical toxicant, did not exceed the EPA ambient water quality
criteria (AWQC) for protection of aquatic life (11 /xg/L) during BMAP 7-d media toxicity tests.
Loar (1992) stated that although episodic increases of chlorine concentration probably affect
benthos at EFK 24.4 (above forma1 New Hope Pond), the chlorine concentrations cannot account
for the low Sundance and diversity at EFK 23.4.
Likev. .:,e, plausibility for all three propositions is strong. The COPCs from this study are
known to bioaccumulate in most aquatic organisms. Although most COPC concentrations in
surface water were below detection limits, the large number of COPC body burdens displaying
longitudinal decreases in concentration downstream from the Y-12 Plant suggests that the Y-12
Plant is the source of the COPCs in die biota (Proposition 1). Propositions 2 and 3 are plausible
because the surface water mercury concentrations (0.54 to 0.32 ngfL) at the three sites closest
to the Y-12 Plant exceed the concentration known to cause toxic effects in fish (<0.1 ngfL).
This alone could account for the observed impacts to the fish and benthic macroinvertebrate
community structure and diversity.
The conclusions from the weight-of-evidence analysis have implications for the assessment
endpoints associated with Goals 4 (a fish community indicative of undegraded conditions) and 5
(no adverse effects from contaminants to aquatic organisms and/or predators that feed on them)
(see Table ES.5). For example, aqueous mercury concentrations in EFPC exceeding levels
known to produce toxicological impacts to sensitive (e.g., <0.1 ng/L in rainbow trout) as well
as tolerant species (e.g., 0.23 jig/L in fathead minnows) have been observed. This alone could
affect community structure (Goal 4 and its assessment endpoint), as demonstrated by the increased
proportion of tolerant fish species (Proposition 2) and the impacts to taxonomic richness and
diversity (Proposition 3). Elevated aqueous mercury concentrations exceeding AWQC also place
fish and benthic macroinvertebrate communities at risk, which links to Goal 5 and assessment
endpoint Sa. Elevated body burdens of mercury and other contaminants continue to place aquatic
organisms and their predators at risk (Goal 5, assessment endpoints 5b and 5c). The aquatic
weight-of-evidence analysis indicates that Goals 4 and 5 are not being met, most likely due to
ongoing releases of contaminants from the Y-12 Plant.
-------
ES-109
The weight-of-evidence analysis for the terrestrial biota makes two propositions concerning
EFPC soil contaminants. Common to both propositions is the premise that EFPC soil
contaminants were released from the Y-12 Plant and were transported by and deposited from
EFPC surface water. The propositions are as follows:
(1) EFPC soil contaminants have resulted in elevated body burdens of these contaminants
in terrestrial organisms residing on the floodplain, as indicated by field-observed
measurements of earthworms, small mammals, and birds.
(2) EFPC soil contaminants have resulted in reduced abundance of terrestrial organisms
residing on the floodplain, as indicated by field-observed measurements of terrestrial
populations.
The terrestrial weight-of-evidence analysis evaluated four criteria (spatial association,
stressor-response, strength of association, and plausibility). The propositions explicitly address
a subset of the EFPC terrestrial community—the animal indicator species associated with the
assessment and measurement endpoints. These species are generally ubiquitous and, for one or
more of their life stages, relatively sedentary inhabitants of the floodplain. The 10 families of
insects considered in detail in the following analysis are inhabitants of soil, soil surface, or
herbaceous vegetation, or are predators of other such inhabitants (Borror and DeLong 1964).
The spatial association for Proposition 1 is supported by the fact that concentrations of
mercury and five other metals in undepurated earthworms; cadmium, zinc, and Aroclor 1260 in
white-footed mice; and 6 metals, Aroclor 1260, and some pesticides in insects from most sites
in EFPC exceeded those in biota from the reference site. Also, body burdens of mercury in
undepurated earthworms generally decreased with increasing distance from the Y-12 Plant. This
also supports the stressor-response relationship for Proposition 1 because maximum mercury
concentrations in surface soils nearest the creek b?nk of EFPC decreased with increasing distance
from the Y-12 Plant, as did soil mercury concentrations at the biotic sampling sites (with the
exception of Site 1, which has only a narrow floodplain). The strength of association and
plausibility of Proposition 1 are high, given that the only known source of mercury and other
contaminants to terrestrial biota is the EFPC floodplain soils.
The spatial association and stressor-response relationship for Proposition 2 are supported by
the inverse relationship between soil mercury concentrations and abundances of ten families of
insects with terrestrial juveniles that represent a majority of all insects captured. There are no
such similar trends in abundance for small mammals, birds, and earthworms. Published dose-
response data for terrestrial annelids suggest that mercury concentrations in EFPC soils may be
-------
ES-110
high enough to adversely affect soil-dwelling insects. The strength of association and plausibility
of Proposition 2 are moderate. Insect groups showing a spatial association with mercury
concentrations in EFPC soils comprise between 70 and 84% of all terrestrial insects captured at
the six biotic sampling sites. The increase in abundance with increasing distance from the Y-12
Plant exhibited by the three dominant groups of insects is strong evidence that one or more
physical or chemical stressors associated with EFPC diminish with increasing distance
downstream. The most plausible cause of this pattern is the decrease in soil contaminant
concentration, although other stressors or subtle habitat differences may also influence the
observed pattern of insect abundance.
It is concluded that increased body burdens of mercury and other contaminants in
earthworms, mice, shrews, and wrens result from EFPC soil contaminants that originated at the
Y-12 Plant and were transported by and deposited from EFPC surface water onto the floodplain.
It is also concluded that the decreased abundances of 10 families of insects with terrestrial
juveniles at EFPC biotic sample sites nearest the Y-12 Plant (especially Sites 1, 2, and 3) result
from EFPC soil contaminants. While this suggests that insect abundances may have been reduced
by contaminants in EFPC floodplain soils, there is insufficient evidence to conclude that the
terrestrial insect community at any EFPC site is indicative of degraded conditions, as defined in
assessment endpoint 6a, because of habitat differences between the EFPC and Hinds Creek
sampling sites. Population densities of other terrestrial indicator organisms were typically higher
than at the reference site.
These conclusions have implications for the assessment endpoints. The propositions link
directly to Goal 7—the protection of terrestrial animals and their predators from the effects of
contaminants. Elevated body burdens of EFPC contaminants in earthworms and their predators
put still other unsampled and unanalyzed terrestrial predators on the EFPC floodplain at risk from
exposure to these contaminants. The propositions link indirectly to Goal 2 because terrestrial
organisms represent an essential resource for migratory birds passing through or residing
seasonally on the floodplain. Reductions in insect abundances may decrease the carrying capacity
of EFPC for numerous species of insectivorous birds that raise their young on the floodplain in
the summer. Combined with elevated body burdens of contaminants in insects, this represents
an even more potentially serious risk to bird populations. Thus, Goals 2 and 7, the protection
of migratory birds and the protection of terrestrial animals and their predators, are probably not
being met in some portions of the EFPC floodplain as a result of EFPC soil contaminants that
originated at the Y-12 Plant and were transported by and deposited from EFPC surface water
onto the floodplain.
-------
ES-111
ES.32.5 Spatial distribution of ecological risk
The floodplain was organized into nine risk assessment segments (Fig. ES.29), based on
geography, uniformity of land use, and similarity of contaminant levels. A more descriptive
explanation of the criteria for segment selection is found in Sect. 5 of the RI. Segments 1, 2,
3, and 4 are located in the upper reaches or part of the creek, below Lake Reality, whereas
Segments 5, 6, 7, and 8 are located further downstream along the east-west axis of the stream.
The longest segment (Segment 9) covers approximately the last 25% of the EFPC floodplain, and
its western edge is located at the confluence with Poplar Creek. The nine segments used for this
ERA are the same as those used for the human health risk assessment.
Ecological risk differs from segment to segment along the —23-km (14.5-mile) course of
EFPC. Risk varies according to noticeable trends in exposure to contaminants, as measured by
comparing contaminant body burdens relative to protective benchmarks (quotients) and weight-
of-evidence discussions on spatial associations. The decrease in most contaminant body burdens,
ecological community effects, and risk quotients with distance downstream from the Y-12 Plant
indicates that ecological risk is highest in Segments 1, 2, 3, and 4, which are the segments closest
to the source of contamination. Several contaminated wetlands with resident invertebrates,
amphibians, and other animals also occur in these segments. Ecological risk is lowest in
Segment 9, which is the segment farthest downstream from Y-12; contaminant body burdens and
other assessment endpoint measures are also lowest in Segment 9. Ecological risk is intermediate
in Segments 5, 6, 7, and 8, which are located in the midstream portion of EFPC. Table ES.12
summarizes these and related trends.
Along this overall gradient of ecological risk and within each segment, two types of
ecological resources are at risk: aquatic and terrestrial. Aquatic resources—fish, benthic
macro invertebrates, and primary producers—are at highest risk because they live in the water or
in sediments and are exposed continuously via body contact. Risks to aquatic biota are greatest
immediately downstream from the Y-12 Plant (EFK 20-23), where the highest aqueous mercury
levels were observed. Every ERA contaminant sampling location in EFPC provided
contaminated prey to fish-eating predators. Exposures to aquatic and piscivorous biota generally
decrease downstream. Aquatic ecological resources ar,e subjected to a greater risk, segment by
segment, than are the terrestrial resources in the corresponding segment.
Although biomagnification has in some cases resulted in body burdens well above the
benchmarks used for risk estimation, measurements of body burdens show that terrestrial
resources—birds, small mammals, earthworms, arthropods, and vegetation—generally exhibit less
-------
Ecological Risk
Highest
~ Medlun
H Lowest
Fig. ES.29. Ecological risk for aquatic and terrestrial resources under current
-------
Table ES.12. Patterns of measurement endpoints Tor the three ecological risk assessment segments
Media and measurement ..
endpoint*: .
Location in 23-km (144f-ml) long (loodplain
tipper caches
(Segments 1, 2,3, and 4)
Middle reaches
(Segments 5, tf, 7, and 8)
Lower reaches
(Segment 9)
Abiotic media
Water contaminants
Highest
Intermediate
Lowest
Soil contaminants
Highest
Intermediate
Lowest
Aquatic organisms
Fish
Tolerant species
Highest
Intennediate
Similar to reference
Species evenness
Lowest
Similar to highest
Highest
Contaminant body burden's
Highest
Intermediate
Lowest |
Benthic macroinverteb rates
Tolerant species
Highest
Intennediate
Similar to middle reaches
Species evenness
Lowest
Similar to highest
Highest
EPT richness
Lowest
Intermediate
Similar to middle reaches
Terrestrial organisms
Arthropod mean abundance
Lowest
Intennediate
Highest
Contaminant body burdens
for earthworms and mice
Highest
Intermediate
-------
ES-114
risk and less definitive trends than the aquatic resources for any particular segment within the
floodplain. Exposures to terrestrial animals are primarily via contaminated animal food, in which
contaminant levels are related to soil contaminant concentrations. Initial work shows that
terrestrial areas of highest mercury concentrations are sometimes found in the bottomland
hardwood forests, where flooding is most likely to have occurred. As explained in the risk
quotient and terrestrial weight-of-evidence sections, some available terrestrial measurements show
trends, particularly decreasing contaminant body burdens as distance from the Y-12 Plant
increases. Thus, both the aquatic and terrestrial organisms in the EFPC environment are at
differing risks as a function of distance from the Y-12 Plant. In turn, these ecological risks are
in three large patterns: (1) highest in Segments 1, 2, 3, and 4 (upper reach); (2) intermediate in
Segments 5, 6, 7, and 8 (middle reach); and (3) lowest in Segment 9 (lower reach).
Selection of COCs. On the basis of results described in the preceding risk characterization
subsections, mercury and PCBs are retained as COCs for both aquatic and terrestrial biota. Other
COPCs did not appear to provide exposures consistently above threshold concentrations used in
the risk characterization.
ES.3.2.6 Future exposures and risks
Many future exposure scenarios are possible. If there is no remediation at either the
Y-12 Plant or the EFPC floodplain, exposures of the biota are not likely to change significantly.
Bioaccumulation and biomagnification may cause an increase in contaminant load in long-lived
terrestrial biota. However, because most of the soil contaminants have been in place for 30 years
or more, most of the food taxa have probably reached steady-state body burdens. Cessation of
releases from routine Y-12 Plant operations would likely reduce chronic exposures in EFPC, but
releases into EFPC by stormwater runoff from the Y-12 Plant into EFPC would likely continue.
Changes in Y-12 Plant operations to reduce the volume of water, but not the mass of
contaminants, flowing into EFPC could increase exposure concentrations to aquatic biota.
Closure of the Y-12 Plant would alter the base flow of EFPC, causing disruption of many biotic
communities, especially in the upper reaches of the creek. Because ongoing releases of
contaminants from the Y-12 Plant into EFPC appear to dominate the effects on aquatic biota in
EFPC, changes in the baseline contaminant body burdens'that would result if the EFPC
floodplain and sediments were the only source of contaminants to surface water are not clear.
Studies are in progress to assess sediment toxicity and relationships between sediments and
transfer of contaminants to aquatic biota.
-------
ES-115
The future (i.e., assuming no change from current conditions) would result in the continued
exposure of (1) aquatic biota that directly contact and ingest EFPC surface water, sediment, and
food; (2) aquatic predators that prey on other contaminated aquatic organisms or terrestrial
insects; (3) terrestrial animals that dwell in or consume contaminated soil; and (4) terrestrial
predators that prey on contaminated aquatic or terrestrial animals in the EFPC floodplain. The
continued exposures of aquatic and terrestrial biota and their habitats to contaminants would result
in continued elevated body burdens of contaminants in these biota and risk to their predators,
along with impacts to community structure and taxa diversity.
Wetlands in contaminated areas of the floodplain would continue to receive contaminants
transported by storm flow. Contaminated soils could act as a source of contamination to wetlands
in the watershed. Contaminated soils could be eroded and transported as suspended load in the
high flow of flood events. As flow velocity decreases and flood waters recede, contaminants
transported as suspended solids drop out of suspension and are redeposited in wetlands in another
part of the floodplain. Because wetlands in the EFPC floodplain have formed in low-lying areas
of the floodplain subject to deposition, the wetlands could receive future contaminants. In this
manner, contaminants could migrate, and relatively uncontaminated soils in the floodplain
downstream from the source area could become contaminated. In the future, contaminated soils
in wetlands would remain a potential source of contamination to numerous other receptors,
including soil, sediment, surface water, groundwater, plants, animals, and humans.
The floodplain contains trees, some of which have cavities and other hiding/brood places for
bats. These trees are likely to remain, continuing to provide potential habitat to threatened and
endangered bat species. However, the endangered Indiana bat (Myotis sodalis) has major
population centers in Missouri and Kentucky and is not likely to reach eastern Tennessee. The
endangered gray bat (Myotis grisescens) is also concentrated in cave regions; the species is mostly
found in Arkansas, Missouri, Kentucky, Tennessee, and Alabama and forages primarily along
rivers or lake shores. This bat is not likely to expand its range to the small stream area of EFPC.
The Tennessee dace (Phonnus tennesseensis), a state-listed fish found in several tributaries to
EFPC, may move into EFPC.
ES.3.2.7 Comparison of ERA results with assessment endpoints
As documented in the Problem Formulation phase of this ERA (Table ES.5), 12 assessment
endpoints were identified that would help to determine whether existing site conditions provided
attainment of the goals set for the EFPC watershed. Following the conclusion of the Exposure
Characterization, Effect Characterization, and Risk Characterization phases, and within the limits
-------
ES-116
of the uncertainties associated with this type of study, the risk assessment can be compared to
these baseline assessment endpoints to get a clearer picture of the overall ERA results. Table
ES.13 provides a listing of the defined endpoints along with the respective ERA results from
direct surveys or risk characterization analyses (quotient or weight-of-evidence). A few key
results of this comparison and conclusions from the ecological risk follow:
• a few of the assessment endpoints (assessment endpoint la and possibly lb, lc, and 2a in
Table ES.13 were achieved with existing EFPC contaminant conditions;
• one assessment endpoint (6a) is achieved for some receptors at some places and not achieved
for one receptor at all places; and
• measurements or risk characterization method results indicate that the rest of the assessment
endpoints cannot be met in parts of the EFPC floodplain under existing conditions, i.e.,
many body burdens exceed die protective concentrations.
These and the following conclusions summarize the baseline ecological risk assessment for
EFPC:
• There is ongoing risk to ecological resources, particularly aquatic organisms in the upper
part of the creek, from exposure to contaminants in environmental and biological media of
EFPC. The results of the ERA do not indicate a need for immediate short-term action to
mitigate exposures or risks to ecological health.
• Direct contact with and ingestion of surface water, sediment, and sediment pore water are
primary exposure pathways for aquatic biota. The food chain is also a primary exposure
pathway for aquatic fauna. Releases from the Y-12 Plant are the primary source of water-
borne contaminants. Additional contributions, albeit much smaller, to ecological risk come
from the municipal sewage treatment plant and other point and non-point sources along the
creek. Experiments at ORNL on fish uptake in tanks containing mercury-laden waters from
Lake Reality, both with and without a contaminated sediment substrate, establish the minor
role of sediments as a source of exposure to aquatic organisms. Studies are in progress to
assess the toxicity of sediments from EFPC on aquatic organisms.
• The food chain is the most important exposure pathway for terrestrial fauna. Exposures to
terrestrial biota also come from contaminant deposits in the EFPC floodplain soils, via direct
contact and inhalation. These exposures result in ecological risk quotients in excess of 10
and 100 to predators of earthworms and insects. Predators of small mammals and birds are
-------
ES-1I7
Table ES.13. Comparison of ERA results with assessment endpoints
Awimwl wwtpoml
ERA rsait
1».
No harm to any threatened and endangered specie!
and their critical habitat of the EFPC floodplain
la.
Literature and field surveys did not reveal threatened and
endangered species living in EFPC floodplain. Habitat that could
support an expansion of the range of a threatened and endangered
species (e.g., Indiana and Cray bats) was found. However, a
range expansiooof such species is not likely.
lb.
Maintenance of plant community composition
and/or structure required for rare plant and animal
and support species
lb.
Literature and field surveys did not find that potential
habitat—plant community—for rare animal or support species was
banned by EFPC cornsmjnants. Wetlands may be an exception
depending on the remediation goal optioo.
lc.
No exposure to threatened and endangered specie*
through biomsgnifisble contaminants through the
food chain
lc.
Literature and field surveys and consultation with responsible
agencies did not reveal threatened and endangered species living in
the EFPC floodplain.
2a.
No lulling or harming of migratory birds as a result
of exposure to site specific streuon
2a.
Field survey evidence provides of more than 30 migrmat
bird species which are likely inflating cosaamsnaled worms aad 1
arthropods and, therefore, likely being affected by EFPC |
coouminftnu.
3*.
The presence and structure/function of wetlands in
relation to contaminants
3a.
Seventeen wetlands exist in the floodplain. some with mercury
contamination, but each appear* to exhibit uodegraded structure
and functions. (There is an ongoing contaminant sampling to
establish this).
4a.
Fish communities in which the proportion of
species tolerant of degraded water quality ia <30%
4a.
Continual releases of water-borne contaminants impact fish
community structure, resulting in communities dominated by
species tolerant of degraded water quality conditions at the 4 aites
closest to the Y-12 Plant.
5a.
Ratio of contaminant concentration in surface water
to water quality criteria for protection of aquatic
life £1. These criteria are 0.012 fig Hg/L and
0.001 fig PCB/L.
5a.
The quotient method indicated significant exceedances by factors
up to 45 from mercury concentrations above water quality criteria
of EFPC below Lake Reality. j
5b.
Aquatic indicator organisms contaminated body
burden ratio to lexicological effects levels S 1.
These levels are 5 mg Hg/kg and 0.4 mg PCB/kg.
5b.
Contaminant body burden measurements and quotients of 1 or less I
for mercury and up to 20 for PCBs were determined for common
stoneroUen, redbreast sunfish, and crayfish.
5c.
Fish contaminant body burden ratio to protect
piscivorous biota levels £ 1. These levels are
0.1 mg Hg/kg and 0.5 mg PCB/kg.
5c.
The quotient method identified numerous exceedances by factors
of 10 to > 100 for mercury and up to 16 for PCBs for fish
contaminant body burdena in a predator pathway.
6a.
Terrestrial animals with diversity, abundance, and
diitributiona indicative of uodegraded conditions
resulting in 220 or more percent decrease
compared to the reference.
6a.
Terrestrial insect sbundance was more than 20% lower at all
EFPC sites than at the reference site. Other findings suggest that
no 20% decreases occur for other indicator organisms.
7a.
Ratio of contaminant body burdens in terrestrial
indicator species to toxicological effects levels S1.
These levels are 1.1 mg Hg/kg and 16 mg PCB/kg.
7a.
Body burdens of Great blue heroo, Carolina wrens, and short- |
tailed shrews were found to exhibit quotient ranges between about I
2 and S for mercury in the ERA sampling.
7b.
Ratio of contaminant body burdens in terrestrial
indicator species to levels protective of terrestrial
predators £ 1. These levels are 0.05 mg Hg/kg
and 3.0 mg PCB/kg.
7b.
Contaminant body burden quotients in excess of 100 were
determined for Great blue herons, earthworms, and short-tailed
shrews for predator pathways. Quotients for mercury were as
high as 19 for white-footed mice and 70 for Carolina wren
predators.
-------
ES-118
at risk according to modeled pathways of contaminant uptake if those predators obtain the
majority of their food from the EFPC floodplain.
• Methylmercury concentrations were very low in EFPC soils and sediments; however,
methylmercury was assumed to be the predominant mercury species in aquatic biota,
including flying insects, based on data from the BMAP and other evidence. Consumption
of contaminated aquatic biota provides risk to terrestrial and aquatic piscivores.
• Evidence suggests that some ecological recovery of the aquatic ecological community has
been occurring in the upper reaches of the creek above and below Lake Reality as
documented by the BMAP. Dechlorination of the process water and other remedial activities
at the Y-12 Plant have lowered the concentrations of some water-borne chemical stressors;
thus, exposure has decreased, followed by partial ecological recovery. Nonetheless, elevated
contaminant body burdens and an excess of pollution tolerant species are still present.
• Water-borne contamination (especially mercury and PCBs) must be controlled for protection
of aquatic biota. Sediment or floodplain soil removal alone will probably not sufficiendy
reduce those contaminant concentrations in aquatic biota. Results from in-progress EFPC
studies on sediment toxicity and in-stream contaminant transfer will help clarify these
relationships. The implications of this and other findings will be rigorously and
systematically examined in the FS-EIS.
ES.4 REMEDIATION GOAL OPTIONS
Results of the characterization of the nature and extent of EFPC contamination and of the
BRA were used to develop remediation goal options (RGOs) for the FS. RGOs are initial
chemical-specific, numeric cleanup limits developed for each affected environmental medium.
These RGOs, derived from site-specific exposure assumptions, are protective of human health
and the environment and comply with ARARs. RGOs were developed for children and adults
using the assumptions and results of the BRA.
The process of deriving an RGO consists of establishing an acceptable target risk value for
exposure to a contaminant and back-calculating the corresponding concentration in the
environmental medium under evaluation. EPA requires the use of a 10"6 risk level as the point
of departure for determining remediation goals for carcinogens and an HI score of 1 for
noncarcinogens.
-------
ES-119
The list of RGOs presented in this report includes all human health COCs that were
evaluated in the Tier II risk assessment. However, this is a starting point, and remediation will
not be required for all of these substances. Although the list will be refined during the FS, the
RGOs for mercury in the agricultural setting (58 mg/kg for children and 198 mg/kg for adults)
will be used for the FS alternatives screening.
Based on the results of the human health BRA, the primary pathways driving the risk
assessment are the food chain pathway, the soil ingestion pathway, and the groundwater ingestion
pathway. These results indicated that mercury is the principal COC for EFPC. For the purposes
of comparison, Fig. ES.30 shows the relative magnitude of the human health RGOs for mercury
for all soil-related pathways. The RGOs are much lower for the produce pathway compared to
direct soil contact. This directly reflects the conservative exposure assumptions adopted for the
food chain pathways.
Groundwater is not currently a source of drinking water in Oak Ridge, nor is it likely to be
used as such in the future. Risk assessment of hypothetical exposure to groundwater was
conducted to comply with EPA requirements and to develop an additional measure of the
significance of observed levels of chemicals. Risk assessment examined exposure to both
unfiltered and filtered samples. Even if groundwater were used as a source of drinking water,
it is very unlikely that the unfiltered water would be representative of the water ingested by
homeowners. A more representative sample is the filtered groundwater. Although concentrations
of COCs in the filtered groundwater are below federal MCLs, the risk assessment indicates that
both the risk measures exceed the target levels, primarily because of the presence of arsenic and
manganese. High levels of manganese are naturally found within the area. Observed levels in
soils of EFPC are indistinguishable from background. Arsenic cannot be identified as originating
from DOE's Y-12 Plant, and the chemical is suspected to come from an area source. For the
sake of completeness, however, RGOs were calculated for chemicals in groundwater.
The ecological RGOs are based on (1) analytical data obtained during sampling of biota, (2)
physical media sampling carried out during Phase la and Phase lb, (3) published information
about the toxicity of mercury, cadmium, PCBs, chlordane, and PAHs, and (4) relationships
among these findings. Ecological RGOs were derived for physical media (water, sediments, and
soil) as sources of contaminants to biota by ingestion, by direct contact, or through the food
chain. For example, the soil RGO is the lowest value calculated to be protective of plants, soil
invertebrates, small mammals, birds, and top predators.
-------
1.00E + 03
1.00E + 02
1.00E + 01
1.00E + 00
1.00E-01
4.61E + 02
3.31E + 02
1.98E+02
~ Children
Adults
5.88E + 01
0.13E-01
3.04E-01
Ingestion
and Dermal:
Agricultural
Scenario
Ingestion
and Dermal:
Open Use
Scenario
Produce
Ingestion:
Agricultural
Scenario
Exposure Pathways
-------
ES-121
For mercury in sediments, an equilibrium partitioning equation was used to calculate the
RGO, which was set at 210 mg/kg.
For mercury in soils, four pathways were examined. Using equations for incidental
ingestion, a lower limit of 274 mg/kg was calculated. For inhalation of vapors from soil, the
lower limit was 1000 mg/kg. A lower limit of 230 mg/kg has been determined from food chain
equations used to calculate levels for protection from toxicity and in discussions of the food
chain, especially in relation to protection from predators that would feed on contaminated prey.
The RGO for mercury in soil to protect ecological resources associated with soil was set at
200 mg/kg.
To some extent, depending on flooding and erosion, sediments may be deposited and become
soil, or soil may be washed into the creek and become sediment. The similarity between soil and
sediment RGOs for mercury means that the distinction is not very important.
During the FS-EIS, these goals will be refined further to produce final remediation goals.
ES.5 RECOMMENDATIONS FOR FS-EIS DEVELOPMENT
Based on the results of the RI, BRA, and ERA, the following basic conclusions and
recommendations were put forward for consideration in the development of the follow-on FS-EIS
for EFPC. Although many other important findings and results are presented in earlier sections
of this report, this summary is meant to reduce the large volume of information associated with
this complex study into the principal guidance for remediation planning. This guidance is
presented by EFPC/SLB environmental media of concern for ease of incorporation into the
follow-on studies.
ES.5.1 Surface Water
There is-currently no excessive risk to human health associated with contaminants in surface
water in EFPC, and in general the mercury concentrations are below drinking water standards
(except on occasion in the upper reaches of the stream). Nonetheless, the preponderance of risk
to aquatic biota comes from water-borne contaminants currently being released from the Y-12
Plant (which does not fall within the EFPC OU). These releases appear to be (1) contributing
the major initial source of the elevated mercury and other contaminant levels in aquatic biota in
EFPC; (2) affecting the fish and benthic macroinvertebrate community structure, resulting in
communities dominated by species tolerant of degraded water quality conditions; and (3) resulting
-------
ES-122
in reduced taxonomic richness and diversity in fish and benthic macroinvertebrate communities
in EFPC compared to uncontaminated reference streams. This ongoing contaminant source must
be addressed to ensure protection and long-term recovery of the EFPC ecological community.
ES.5.2 Groundwater
Groundwater is not currently being used in the vicinity of EFPC as a source of drinking
water, nor is it likely to be used as such in the future. This pathway was included in the BRA
to comply with EPA requirements to comprehensively evaluate the groundwater pathway for
current and future scenarios. Risk of hypothetical exposure to groundwater was evaluated
separately for soil horizon wells and for water from the underlying bedrock (for both filtered and
unfiltered samples). The BRA demonstrates that unfiltered levels of inorganic contaminants are
primarily responsible for die risk levels that exceed EPA target ranges (manganese for noncancer
risk and arsenic and beryllium for cancer risk). It should be noted that levels of manganese and
beryllium in the soils are demonstrated to be indistinguishable from background (i.e., they are
clearly naturally occurring). Exposures to waters from the soil horizon and bedrock are currently
incomplete pathways. Furthermore, since the soil horizon cannot yield sufficient quantities of
water for domestic or commercial uses, it is not considered a complete future exposure pathway.
Remedial planning during the FS must consider whether cleanup is necessary for these inorganic
chemicals for groundwater in the bedrock.
ES.5.3 Soils
Floodplain soils contain the highest levels of contaminants of all the EFPC environmental
media. COPCs in the soils consist of inorganic trace metals, PCBs, PAHs, and radionuclides.
COPCs were identified by dividing the average concentrations of contaminants by their chronic
toxicity values for human health (or multiplying concentrations by cancer slope factors) and
ranking the results; contaminants that contributed to the top 99% of the scores were retained as
COPCs. Mercury was identified as the principal nonradiological contaminant through this
screening process, and uranium was identified as the primary radiological contaminant. Both of
these contaminants are directly attributable to the past operations of the Y-12 Plant.
Five million cubic feet of contaminated soil have been calculated to exist within the
floodplain and SLB. The results of the BRA indicate that soil ingestion is an exposure pathway
of principal concern to human health. The ERA also identified the soil pathway as a concern to
EFPC biota, primarily through terrestrial predator food pathways. RGOs have been derived for
the observed chemical contaminants in soil for use in the initial screening of remediation
alternatives in the FS-EIS. The ecological RGO for mercury in soil is 200 mg/kg. Final
-------
ES-123
remediation goals will be determined as part of the CERCLA process leading to a Record of
Decision. Incorporation of quantified ecological goals and approved human health goals, in
conjunction with the results of the soil contamination data, will identify locations where soil
remediation may be required.
ES.5.4 Sediments
Creek sediments contain the same contaminants as floodplain soils, but at lower
concentrations. Because of the transient nature of the sediments within the EFPC, however, the
distribution of the contaminants is less predictable than in the soils. The same general patterns
do exist, with the upper reaches of the creek showing somewhat elevated levels of the sediment
contaminants (primarily mercury). Due to the nature of the sediments and their generally lower
levels of contamination, the BRA identified little risk to human health via this environmental
medium. The RGO established for mercury in sediments for children was 61,000 mg/kg, a value
which essentially eliminates this medium from remediation concern for human health
considerations.
The ecological impact of the contaminated sediment pathway, while only partially understood
at this point, might be significant. Two studies are currently underway by DOE to assess EFPC
sediment toxicity and contaminant transfer from sediments to aquatic biota. Results from these
studies will help clarify the relationship between EFPC sediments and effects on aquatic
organisms. However, due to the ongoing contaminant releases into upper EFPC from the Y-12
Plant, assessment of remedial alternatives for sediments should be deferred from the FS-EIS until
such time as the releases from the Y-12 Plant are eliminated. Contaminant body burdens of
mercury in aquatic biota, particularly in benthic invertebrates and aquatic predators, indicate a
potential for significant bioaccumulation of contaminants in the food chain. The ecological RGO
has been determined to be 210 mg/kg for sediments.
ES.5.5 Air Pathway
No COPCs have been measured in air along the EFPC or within the vicinity of the ORR that
are of human health or ecological concern. This pathway, therefore, should be eliminated from
remediation consideration in the FS-EIS.
ES.5.6 Direct Radiation
Direct radiation measurements made along the EFPC floodplain documented the background
levels of direct exposure resulting from natural and man-made radionuclides. Therefore, risk
-------
ES-124
assessment was not conducted for irradiation exposure. Risk assessment was conducted for
radionuclides to evaluate exposure via ingestion and inhalation. No substantial human health or
ecological impacts from these background levels were determined from the BRA or ERA studies
(i.e., risks fell within the target ranges established by EPA).
-------
ES-125
ES.6 REFERENCES
Borror, D.J., and D.M. DeLong. 1964. An Introduction to the Study of Insects (revised ed.).
Holt, Rinhart, and Winston, New York.
Carmichael, J.K. 1989. An Investigation of Shallow Groundwater Quality Near East Fork Poplar
Creek. Oak Ridge, Tennessee Water-Resources Investigation Report 88-4219, U.S. Geological
Survey.
COE (U.S. Army Corps of Engineers). 1992. An Inventory of Wetlands in the East Fork Poplar
Creek Floodplain in Anderson and Roane Counties, Tennessee. Preliminary Draft Report to
die U.S. Department of Energy. July 1992. 21 pp. plus appendices.
EPA (U.S. Environmental Protection Agency). 1987. MINTEQA1, An Equilibrium Metal
Speciation Model: User's Manual. EPA/600/3-87/012, October.
EPA. 1988. Superfund Exposure Assessment Manual, EPA/540/1-88/001. U.S. Environmental
Protection Agency, Office of Remedial Response. OSWER Directive 9285.5-1.
EPA. 1989a. Risk Assessment Guidancefor Supetfimd: Human Health Evaluation Manual Part A.
EPA/540/1-89/002. U.S. Environmental Protection Agency, Office of Solid Waste and
Emergency and Remedial Response. OSWER Directive 9285.701A.
EPA. 1989b. Exposure Factors Handbook. EPA/600/8-89/043. U.S. Environmental Protection
Agency, Office of Health and Environmental Assessment.
EPA. 1991. Human Health Evaluation Manual, Supplemental Guidance: Standard Default
Exposure Factors. U.S. Environmental Protection Agency, Office of Emergency and Remedial
Response, OSWER Directive 9285.6-03.
EPA. 1992a. Dermal Exposure Assessment: Principles and Applications. EPA/600/8-91/001B.
U.S. Environmental Protection Agency, Office of Health and Environmental Assessment,
Interim Report.
EPA. 1992b. Integrated Risk Information System (IRIS). U.S. Environmental Protection Agency
On-line Database of Toxicity Measures. Office of Research and Development, Environmental
Criteria and Assessment Office, Cincinnati, Ohio. Electronic Mail Account information via
TOXNET, National Library of Medicine, Bethesda, Md.
EPA. 1992c. Framework far Ecological Risk Assessment. EPA/630/R-92/001.
U.S. Environmental Protection Agency, Risk Assessment Forum, Washington, D.C.
-------
ES-126
Fausch, K.D., J.R. Karr, and P.R. Yant. 1984. "Regional Application of an Index of Biotic
Integrity Based on Stream Fish Communities," Transactions of the American Fisheries Society
113, 39-55.
FWS (U.S. Fish and Wildlife Service). 1980. Habitat Evaluation Procedures (HEP) ESM102.
Gist, C.S. 1987. Soil Contaminant Uptake by Plants in the Terrestrial Food Chain in the
Floodplain of East Fork Poplar Creek. Oak Ridge, Tenn., April 28, 1987.
Hinzman, R.L. (ed). 1992. Second Report on the Oak Ridge Y-12 Plant Biological Monitoring
and Abatement Program for East Fork Poplar Creek. Draft Y/TS-888. Oak Ridge National
Laboratory, Oak Ridge, Tenn.
Loar, J.M. (ed.). 1992. First Report on the Oak Ridge Y-12 Plant Biological Monitoring and
Abatement Program far East Fork Poplar Creek. Y/TS-886. Oak Ridge National Laboratory,
Oak Ridge, Tenn.
Revis, N., G. Holds worth, G. Bingham, A. King, and J. Elmore. 1989. An Assessment of Health
Risk Associated with Mercury in Soil and Sediment from East Fork Poplar Creek, Oak Ridge,
Tennessee. Oak Ridge Research Institute. Final Report, pp. 1-58.
Suter, G.W., n, and J.M. Loar. 1992. "Weighing the Ecological Risk of hazardous Waste
Sites," Environ. Sd. Technol. 26, 432-438.
Turner, R.R., G.R. South worth, M.A. Bogle, M J. Peterson, and N.S. Bloom. 1992. Availability
of Sediment-Bound Mercury for Methylation and Bioaccumulation in Upper East Fork Poplar
Creek: Preliminary Results. Draft report to SA1C.
TV A (Tennessee Valley Authority). 1985. Instream Contaminant Study, Taskl: Water Sampling
and Analysis, Task 2: Sediment Characterization Task 3: Sediment Transport, Task 4: Fish
Sampling and Analysis, Task 5: Summary Report, Prepared for U.S. Department of Energy,
Oak Ridge, Tennessee, under Interagency Agreement No. DE-AI05-84OR21444.
Van Winkle, W., R. W., Counts, J. G. Doreey, J. W., Elwood, V. W., Lowe, Jr.,
R. McEIhaney, S. D., Schlotzhauer, F. G. Taylor, Jr., and R. R. Turner, 1984. Mercury
Contamination in East Fork Poplar Creek and Bear Creek, ORNL/TM-8894, Oak Ridge
National Laboratory, Martin Marietta Energy Systems, Inc.
Welch, S. H., 1987. RCRA Facility Investigation Plan, General Document, Y-12 Plant.
Y/TS-352/VI. Oak Ridge Y-12 Plant, Oak Ridge, Tenn.
-------
OAK RIDGE ROOM
OAK RIDGE PUBLIC LIBRARY
------- |