HYDROGEOLOGIC INVESTIGATION REPORT
SPARTA AQUIFER VULNERABILITY ASSESSMENT AT THE
FORMER SHUMAKER NAVAL AMMUNITION DEPOT
EAST CAMDEN, ARKANSAS
United States
Environmental Protection
Agency
EPA 906/R-10/001
-------
HYDROGEOLOGIC INVESTIGATION REPORT
SPARTA AQUIFER VULNERABILITY ASSESSMENT AT THE
FORMER SHUMAKER NAVAL AMMUNITION DEPOT
EAST CAMDEN, ARKANSAS
United States
Environmental Protection
Agency
By:
Scott Ellinger, P.G.
U.S. EPA Region 6
May 3, 2010
-------
ACKNOWLEDGEMENTS
This investigation was accomplished with the support of numerous individuals and
organizations. I would like to thank all the citizens, corporations, and other organizations
residing in the Shumaker area, and to the Office of the Calhoun County Judge, for kindly
providing assistance during this project. I want to thank the U.S. Geological Survey and the
Arkansas Department of Environmental Quality for providing assistance and advice. EPA
staff Tim Townsend and Linh Nguyen provided much needed help to accomplish difficult
fieldwork tasks. Finally, I want to thank EPA managers Ben Banipal, P.E., Troy Hill, P.E.,
and Laurie King for their guidance and overall support.
Sparta Aquifer A mA United States
Vulnerability Assessment Page ii ^P&^^iTr^ Environmental Protection
-------
SPARTA AQUIFER VULNERABILITY ASSESSMENT AT THE
FORMER SHUMAKER NAVAL AMMUNITION DEPOT
EAST CAMDEN, ARKANSAS
TABLE OF CONTENTS
ACKNOWLEDGEMENTS ii
LIST OF FIGURES v
LIST OF TABLES v
1. INTRODUCTION 1
A. Overview 1
B. Water Resources 2
C. Shumaker History 4
2. PURPOSE OF INVESTIGATION 5
A. Background 5
B. Goals 7
C. Project Planning 8
3. GOAL 1: GEOLOGIC INVESTIGATION 10
A. Geologic Setting 10
B. Subsurface Investigation 14
C. Extent of the Cook Mountain Formation 14
D. Sedimentary Particle Analysis 17
E. Deeper Sparta Sands 18
F. Summary of Findings 19
4. GOAL 2: SPARTA AQUIFER MONITORING 21
A. Procedures Summary 21
B. Analytical Results Summary 22
C. Sparta Aquifer Flow Directions 23
Sparta Aquifer A mA United States
Vulnerability Assessment Page ill ^P&^^iTr^ Environmental Protection
-------
5. GOAL 3: SURFACE WATER INVESTIGATION 27
A. Streams and Lakes 27
B. Sampling Results 28
6. GOAL 4: PERCHLORATE AT LOCUST BAYOU 30
A. Background 30
B. Monitoring Results 32
C. Flow Directions 33
D. Perchlorate Fingerprinting 35
i. Introduction 35
ii. Field Sampling 36
in. Isotopic Results 38
E. Most Likely Sources of Perchlorate 39
i. Imported Chilean Nitrate Fertilizer 40
ii. Synthetic Perchlorate 40
in. Other Possible Sources and Uncertainties 43
7. KEY OBSERVATIONS AND UNCERTAINTIES 46
8. CONCLUSIONS 53
9. REFERENCES 54
Sparta Aquifer A mA United States
Vulnerability Assessment Page iv ^P&^^iTr^ Environmental Protection
-------
LIST OF FIGURES
Figure 1: Location of Shumaker NAD and Study Boundary 2
Figure 2: Location of Most (5 of 6) Shumaker Public Service Company Wells 3
Figure 3: Generalized Geologic Diagram 6
Figure 4: Exploratory Test Holes DW1 - DW7 15
Figure 5: Core Sample Photographs A-D 16
Figure 6: Photomicrographs of Sand and Clay within the Sparta Aquifer 18
Figure 7: Location of the 10 Sparta Aquifer Monitoring Wells 21
Figure 8: Groundwater Flow Direction In Sparta Aquifer 2008 24
Figure 9: Surface Water Sampling Locations 28
Figure 10: Locust Bayou and Vicinity 30
Figure 11: Location of Gravel Aquifer Monitoring Wells 31
Figure 12: Groundwater Flow Directions in Alluvial Aquifer Near Locust Bayou 35
Figure 13: Perchlorate Detections In Monitoring Wells Near Locust B ayou 37
Figure 14: Idealized Graph Showing Natural and Man-Made Perchlorate Data 39
LIST OF TABLES
Table 1: Generalized Geologic Column 11
Table 2: Screening Levels vs. Sample Concentrations 2007 22
Table 3: Screening Levels vs. Sample Concentrations 2008 23
Table 4: Water Level Measurements (mean sea level) 23
Table 5: Water Level Data For Well Clusters, 2008 25
Table 6: Screening Levels vs. Sample Concentrations, 2007 33
Table 7: Screening Levels vs. Sample Concentrations, 2008 33
Table 8: Water Level Elevations for Alluvial Wells 2007 and 2008 34
Table 9: Fingerprinting results for Locust Bayou and Aerojet 39
Sparta Aquifer
Vulnerability Assessment
Page v
vvEPA
United States
Environmental Protection
Agency
-------
1. INTRODUCTION
A. Overview
This report describes the U.S. Environmental Protection Agency (EPA)
groundwater contamination vulnerability assessment for the Sparta aquifer located at the
former Shumaker Naval Ammunition Depot (NAD) in south-central Arkansas. New
information is presented about the vulnerability of the Sparta aquifer to contamination
based on a geological and hydrogeological investigation conducted during 2007-2008.
The investigation involved exploratory drilling, core sampling, logging, and groundwater
and surface water sampling including special isotopic analysis. This report also discusses
the origin of perchlorate detected at Locust Bayou.
This investigation was undertaken because of concerns about potential impacts to
the Sparta aquifer from past military operations and current industrial activities, and
because of detections of perchlorate contamination in drinking water at the nearby
community of Locust Bayou. The area's main drinking water supplies come from
groundwater within the Sparta aquifer and the overlying shallow alluvial aquifer. Any
contamination impacting the Sparta aquifer or the alluvial aquifer by organic and
inorganic contaminants above health based levels could pose a public health hazard and
result in greater water treatment costs and limit the future utility of groundwater.
Shumaker is about 75 miles south of Little Rock, Arkansas, near the City of
Camden. The study area boundary (fig. 1) is the same as the Shumaker NAD boundary,
except for the southern side, which extends southward to include residences near state
highway 278 (Locust Bayou community). The investigation covered approximately
73,000 acres including the former Shumaker NAD, the City of East Camden, the Locust
Bayou area, an industrial complex known as Highland Industrial Park, and timber
production lands.
Site access includes state roads 203, 205, 274, 278, and Calhoun County road 95,
which are all paved roads. Other roads are logging roads and other unimproved roads on
timber company lands. A minor number of gravel roads are maintained by county
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 1 ^jlj^^^^Environmenta! Protection
-------
governments. A number of small communities surround the Shumaker area including
East Camden, Harmony Grove, Eagle Mills, Bearden, and Woodberry.
\ State of Arkansas
Little Rock
Camtlen
•
I Shumaker
Figure 1: Location of Shumaker NAD and Study Boundary
Boundary of investigation includes area within solid black line, and
dashed line south of Shumaker NAD (Locust Bayou community).
The area of investigation encompasses numerous industrial facilities operated by
defense contractors, including facilities regulated by the Resource Conservation and
Recovery Act (RCRA). These facilities utilize many of the buildings, ammunition
magazines, and other structures that were previously part of Shumaker but are now
privately owned. Shumaker NAD operated from 1944-1961 and was used for the
manufacture, testing, distribution, destruction, and storage of ordnance and naval rockets.
B. Water Resources
There are 4 public water systems near Shumaker that utilize groundwater as their
primary water source. These include: Bearden Waterworks and the Harmony Grove
Water Association in Ouachita County; and the Locust Bayou Water Association and
Shumaker Public Service Company in Calhoun County (EPA, 2008a). These utilities
obtain their water supplies from the Sparta aquifer.
The most significant of these water supply companies is the Shumaker Public
Service Company located at Highland Industrial Park (fig. 2). The company's 2007
annual report, prepared for the Arkansas Public Service Commission (2008), reports that
the annual average domestic water demand is 18,731,000 gallons per month, serving a
Sparta Aquifer
Vulnerability Assessment
Page 2
x>EPA
United States
Environmental Protection
Agency
-------
community of 5,000. The Shumaker Public Service Company obtains groundwater from
6 Sparta aquifer wells consisting of 2-18 inch wells and 4-12 inch wells averaging 180-
250 feet deep. Pumping rates are 500 gallons per minute (GPM) each. Water treatment
consists of aeration, chemical treatment, pre and post chlorination, sand filtration, and
sedimentation. Three samples of water are submitted to the Health Department each
month and there have been no health based violations, and no monitoring, reporting, or
other violations in the last 10 years (EPA, 2008a).
I #3
Well #2
WTP #1
Well #5
Shumak'er Public.'"'
Service Co.
Figure 2: Location of Most (5 of 6) Shumaker Public Service Company Wells
Privately owned domestic water supplies are obtained by individual homeowners
from water wells in the shallow gravel aquifer or the Sparta aquifer. Most private
domestic supplies are probably from the shallow Quaternary sediments because the
sediments contain a basal gravel layer (gravel aquifer) capable of producing useable
quantities of groundwater. The gravel aquifer is relatively shallow, thus reducing drilling
and production costs. Some residences at Locust Bayou are known to utilize water from
private domestic wells. There are approximately 40 residential properties at Locust
Bayou, and of these, approximately 14 use private wells as their main drinking water
source and do not have access to a municipal drinking water supply (ATSDR, 2007).
A number of additional residences along county roads in the vicinity of Locust Bayou
may also utilize private wells. During 2005, EPA and ADEQ collected water samples
from Locust Bayou residences utilizing water from private domestic wells, and detected
perchlorate concentrations up to 2.2 j^g/1. Assuming the state-wide average of 3 people
Sparta Aquifer
Vulnerability Assessment
Page 3
United States
Environmental Protection
Agency
-------
per household and 14 residences, approximately 42 people living at Locust Bayou may
have potentially been exposed to perchlorate contamination.
C. Shumaker History
Shumaker NAD existed from 1944 to 1961 for the manufacture, testing,
distribution, destruction, and storage of ordnance and naval rockets (U.S. Army Corps of
Engineers, 2003a). Defense related industrial activity still takes place today, but
Shumaker no longer exists as a military installation and there is no land within the study
area owned by the Navy. Local residents reported during fieldwork that the area was
primarily used for farming prior to the existence of Shumaker, beginning in 1944.
Shumaker NAD was operational during periods of WWII through the Korean War, with
employment levels varying between 20,000 in 1945, to 3,900 in 1951. In 1956, the Navy
announced plans to close Shumaker, and the property was sold to private owners from
1959 to 1961. These owners included International Paper, and Brown Engineering
(which became Highland Resources and was later renamed Highland Industrial Park).
Shumaker finally closed in 1961 (U.S. Army Corps of Engineers, 2003b).
Currently, the eastern two-thirds of the former Shumaker site is heavily forested
and used for timber production and hunting. The western one-third contains numerous
industrial facilities which are operated by, or have been operated by, defense contractors.
Some of these defense contractors include Lockheed-Martin, Loral Vaught Systems,
Aerojet, BEI Defense Systems, Tracer Aerospace, Hughes Missiles Systems, National
Testing Service, Olin Industries, Camden Ordnance, Hitech Incorporated, Armtec
Defense Products, and Austin Powder.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 4 ^ij^Environmenta! Protection
-------
2. PURPOSE OF INVESTIGATION
A. Background
When this study began, there was no significant information indicating whether
the Sparta aquifer had been impacted by contamination, or whether it was likely for the
Sparta aquifer to become impacted. A number of organizations including the EPA, the
Arkansas Department of Environmental Quality (ADEQ), the Army Corps of Engineers,
and industrial facilities have been collecting and evaluating shallow groundwater
contamination data and conducting other types of site evaluations at Shumaker for years,
but have not collected much information on the Sparta aquifer until now. Industrial
facilities routinely conduct site specific groundwater investigations and corrective actions
as required by RCRA permits, but these investigations are typically for shallow
groundwater and not the Sparta aquifer.
In 2005, EPA and ADEQ agreed that a groundwater study of the entire Shumaker
footprint was needed to evaluate Sparta aquifer vulnerability and to establish a
groundwater monitoring system. Groundwater investigations at RCRA regulated
facilities were fairly detailed on a site-specific basis, but covered only about 3% of the
entire Shumaker footprint, based on an estimated comparison of facility areas to total
Shumaker area. Conversely, regional groundwater studies conducted by government
agencies (e.g., the U.S. Geological Survey, Arkansas Geological Survey, etc.) covered
such large geographic areas that their large scales lack sufficient detail for making
decisions about Shumaker. Other previous studies at Shumaker were conducted by the
Army Corps of Engineers and involved the investigation of unexploded ordnance and
possible sources of contamination at certain areas including the former rocket test range,
fuse test range, rocket burn area, TNT burn area, a buried drum area, and some possible
ordnance disposal wells.
Understanding the vulnerability of the Sparta aquifer to contamination depends, in
part, on having sufficient information about the subsurface framework. Shumaker is so
large that it is not practical or economically feasible to place monitoring wells in all
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 5 ^ij^Environmenta! Protection
-------
unassessed areas. The geological framework then becomes critical to developing an
assessment of which areas are most and least vulnerable to contamination. Determining
the presence of thick clay, specifically clay of the Cook Mountain Formation, can reduce
concerns about where contamination may have impacted the Sparta aquifer because of
clay's ability to limit downward groundwater flow. The likelihood of impacts to the
Sparta aquifer would conceptually be less in areas where the Sparta aquifer was covered
by thick clays of the Cook Mountain Formation (fig. 3). Prior to this study, the extent of
the Cook Mountain Formation at Shumaker was a significant site uncertainty, and its
extent was unable to be assessed by using either site-specific facility data or by much
larger regional studies.
w
Sparta Recharge ?-
Alluvial Sand
Gravel
Unconformity
Cook Mountain
- or Sparta ?
,
Where does
Sparta begin
Vulnerability?
Impacts?
Sparta Aquifer
Unknown Characteristics
Figure 3: Generalized Geologic Diagram
Diagram depicts geologic uncertainties prior to study. They include (i) the western extent of
Cook Mountain Formation, (II) geologic characteristics of Sparta aquifer, and (Hi) elevation
of Sparta aquifer surface. Unconformity refers to missing strata (not deposited) between
Claiborne Formations and younger Quaternary sediments.
Also in 2005, EPA and ADEQ performed routine sampling of drinking water at
Locust Bayou to evaluate the human health environmental indicator referred to as CA-
725. This evaluation, required by the Government Performance and Results Act (GPRA)
of 1993, is a procedure for determining whether potential human health exposures to
Sparta Aquifer
Vulnerability Assessment
Page 6
x>EPA
United States
Environmental Protection
Agency
-------
contamination at RCRA facilities could be reasonably expected to be under control. For
the Locust Bayou area sampling, a total of 6 water samples were collected, and 5 of those
were found to contain low concentrations (0.5- 2.2 jig/1) of a contaminant known as
perchlorate. Perchlorate (C1CV) is used in explosives and rocket propellants including
mortars, grenades, flares, and solid rocket fuel. Perchlorate is also known to occur
naturally in limited numbers of geologic deposits used as feedstock for certain fertilizer
products. Perchlorate is readily soluble in water and may cause human health impacts
involving the thyroid gland. One of the goals of the current investigation is to determine
the source of perchlorate in water at Locust Bayou.
Another unknown involved the possibility of contamination of water in streams
and lakes. If streams contain contamination, contamination could rapidly spread across
and outward from Shumaker and downward through permeable sediments into the Sparta
aquifer. A surface water sampling program was needed to collect samples from all major
streams, lakes, and ponds at Shumaker. A particularly important element of stream water
sampling is that it makes a good reconnaissance tool for covering large areas because
individual sampling points can represent large-scale drainage basin runoff.
B. Goals
Investigation goals were centered on completing four interrelated project tasks
which were supported by two phases of fieldwork and sampling conducted during 2007-
2008. During this period, EPA conducted hydrogeologic field investigations and made
observations about Sparta aquifer vulnerability. Investigations for the vulnerability
assessment are mainly discussed under Goals 1-3. These involved collecting geological
information, installing wells, determining flow directions, and collecting and analyzing
samples of groundwater and surface water. During this period, EPA also performed a
focused investigation at Locust Bayou and vicinity to evaluate possible sources of
perchlorate contamination. The most likely sources of perchlorate at Locust Bayou are
described under results for Goal 4.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 7 ^ij^Environmenta! Protection
-------
Project Goals:
• Goal 1: Evaluate the vulnerability of the Sparta aquifer with respect to: (i) the
presence of confining clay overlying the Sparta aquifer, and (ii) by making
observations about the occurrence of clay and sand within the Sparta aquifer.
• Goal 2: Examine the vulnerability of the Sparta aquifer by: (i) installing permanent
groundwater monitoring wells, (ii) by monitoring for impacts to the aquifer, and (iii)
by examining groundwater flow directions.
• Goal 3: Examine aquifer vulnerability by conducting surface water sampling
including streams, lakes, and ponds over the extent of Shumaker, and the Locust
Bayou area.
• Goal 4: Determine the source of perchlorate detected in groundwater monitoring
wells and drinking water at Locust Bayou.
Completing these 4 goals benefits regulatory agencies and industries by providing
subsurface information leading to more consistent data interpretations from individual
facility investigations, and from other investigations at individual areas of concern
located at isolated sites across Shumaker. This study benefits the public by providing
current information on groundwater contamination and provides a groundwater
monitoring system to provide a level of protection for drinking water supplies.
C. Project Planning
Project planning took place during 2005-2006 and included a project proposal and
project scoping meetings. Field sampling and laboratory analysis were addressed in two
Quality Assurance (QA) Project Plans developed under the Uniform Federal Policy for
Quality Assurance Project Plans (UFP), (EPA 2007a and 2007b). A QA Project Plan is a
formal document describing in comprehensive detail the necessary quality assurance,
quality control, and other technical activities that must be implemented to ensure that the
results of the work performed will satisfy the stated performance criteria. Nationally, the
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 8 ^ij^Environmenta! Protection
-------
UFP was developed as a joint initiative between the EPA, the Department of Defense
(DOD), and the Department of Energy (DOE), to consistently implement quality system
requirements. The QA Project Plans for this study present the overall project description,
project organization, staff responsibilities, and QA objectives associated with each phase
of sampling. The two plans comply with all QA requirements and underwent peer-
review.
Plans for addressing study Goals 1 and 2 are included in the QA Project Plan dated
January 31, 2007 (EPA 2007a). During this first phase of study, the QA Project Plan
specifies that 10 Sparta aquifer monitoring wells and 10 alluvial aquifer monitoring wells
will be installed and sampled, and that geological information (core samples and logs)
will be obtained during the drilling. Plans for study Goals 3 and 4 are contained in the
QA Project Plan dated December 18, 2007 (EPA 2007b). This plan states that EPA will
collect additional groundwater samples from all 20 EPA groundwater monitoring wells,
collect water level data, conduct reconnaissance level surface water sampling, and
conduct specialized sampling for perchlorate isotopes.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 9 ^ij^Environmenta! Protection
-------
3. GOAL 1: GEOLOGIC INVESTIGATION
(Evaluate the vulnerability of the Sparta aquifer with respect to: (i) the
presence of confining clay overlying the Sparta aquifer, and (ii) by making
observations about the occurrence of clay and sand within the Sparta
aquifer.)
A. Geologic Setting
The purpose of Section A is to briefly introduce the geologic history, depositional
environments, and stratigraphy needed to understand the basis for the geologic
investigation and related complexities under Goal 1. Information in this section was
compiled from published sources.
Shumaker NAD lies within the Gulf Coastal Plain physiographic province. This
province is characterized by low relief and heavily timbered lands and hills characteristic
of many parts of the southern United States. The Gulf Coastal Plain physiographic
province extends from the Florida Panhandle to southern Texas, and geologically
includes sedimentary rock and loose sediments deposited through cyclic marine and
nonmarine depositional events. These events took place during a geologic time known as
the Tertiary Period. The Tertiary Period occurred from 2 to 65 million years ago, and is
subdivided into individual time units (series) based on geological events that occurred
during the earth's history at those times. The deposition of sediments most relevant to
this study occurred within the Eocene and Holocene series resulting in the Sparta
Formation and the Cook Mountain Formation (both Eocene Series), and the overlying
Holocene Series alluvial deposits (Table 1).
Eocene sediments are divided into separate geologic units called the Wilcox,
Claiborne, and Jackson Groups. The Wilcox is the oldest group and contains sediments
from continental depositional environments including fluvial, lacustrine, lagoonal, and
deltaic environments which produced complexly interbedded sands, slits, clays, and
lignite (Hosman, 1988). Claiborne sediments were deposited during alternating marine
and nonmarine depositional cycles which produced distinctive lithologies; thereby
allowing Claiborne sediments to be differentiated into individual Formations. Sediments
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 10 ^ij^Environmenta! Protection
-------
in the Claiborne Group, whether deposited under subaerial or submarine conditions or
both, are composed of near shore deposits (Albin, 1964). These formations are called the
Carrizo, Cane River, Sparta, Cook Mountain, and Cockfield Formations. The Eocene
Period ended with deposition of the Jackson Group which is composed of marine
sediments.
Era
Cenozoic
Mesozoic
System
Tertiary
Cretaceous
Series
Holocene
Pleistocene
D
Gulf
Group
Jackson
Wilcox
Formation Or
Subdivision
Alluvium
Terrace Deposits
Cockfield Formation
Cook Mountain
Formation
Sparta Sand
Cane River Formation
Carrizo Sand
Porters Creek Clay
Clayton Formation
Arkadelphia Marl
Approximate Number
Years Ago
11,000
500,000 to 2,000,000
58 000 000
Re nnn nnn
135,000,000
Table 1: Generalized Geologic Column
Column shows sediments relevant to this investigation in yellow. They are the Eocene Sparta, Cook
Mountain, and Cockfield Formations, and Pleistocene-Holocene terrace and alluvial deposits. Table
modified from Albin, 1964.
The geology and hydrogeology of the three county areas surrounding Shumaker,
including the counties of Bradley, Calhoun, and Ouachita, were described by Albin
(1964). He reported that rocks of these counties were deposited in the shallow
Mississippi embayment part of the Coastal Plain physiographic province as the sea
alternately advanced and retreated over the land. During early and mid-Eocene time,
including the time when the Sparta Formation was deposited, the main depositional
environment was deltaic deposition of sand, silt, clay, and lignite. During late Eocene,
which may have included deposition of the Cook Mountain Formation, moderately deep-
water clay and marl were deposited. These sediments are reported to dip to the east and
southeast at approximately 25-50 feet per mile towards the axis of the Mississippi
Embayment (Payne, 1968, and Albin, 1964).
Sparta Aquifer
Vulnerability Assessment
Page 11
xvEPA
United States
Environmental Protection
Agency
-------
Additional detailed information on depositional environments is necessary to help
understand the differences between Sparta and Cook Mountain deposits. The following
descriptions illustrate that the Sparta Formation was deposited in a near-shore marine/non
marine environment, but a range of specific near-shore environments may exist as noted
by the author's referenced below.
The Sparta Formation is described by Albin (1964) as having been deposited as
the beach of an advancing sea. Shallow-water clay and back beach lignitic clay and
lignite, indicate that the shoreline fluctuated in the Calhoun and Ouachita County area.
Another perspective, however, is provided by Payne (1968), who reported on the
percentage of sand within the Sparta Formation.
Payne reports that based on sand percentage (total thickness of sand divided by
total formation thickness), the Sparta could be divided into two areas having different
depositional environments. One area covers Louisiana, Mississippi, southern Arkansas,
and eastern Texas, and the other area extends from Grimes to Webb Counties, Texas. For
the area including southern Arkansas, the distribution of sand represents a system of
braided stream channels and interlacing lakes, swamps, and marshes as would be
developed on a large deltaic-fluvial plain; the delta represents an ancestral Mississippi
River system that existed during Claiborne deposition. Payne reported that areas which
contain at least 50% sand represent areas of channel "flow-ways" similar to channel
development along areas of present courses of the Ouachita and Mississippi Rivers, and
areas containing less than 50% sand represent interchannel swamp, marsh, and lake areas
where finer detritus and vegetation accumulated. For the Shumaker area, Payne's
regional mapping indicates that approximately the eastern half of Shumaker contains
from 30 to 50% sand, with a maximum sand unit thickness of 150 feet. Regional data are
not presented for the western half of Shumaker, which is indicated as being a Sparta
outcrop area.
Literature contains only limited descriptions of the Cook Mountain Formation but
it is briefly described in several articles on regional and local geology and hydrogeology,
including Albin (1964), Fitzpatrick, et. al. (1990), and Joseph (2000). According to
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 12 ^ij^Environmenta! Protection
-------
Fitzpatrick, the Cook Mountain Formation ranges between 100 and 150 feet thick and is
composed chiefly of carbonaceous clay, with some lignite and lenticular beds of sand less
than a few feet thick. He reports that the formation is a significant confining bed
covering the Sparta aquifer in east-central Arkansas. Similarly, Joseph (2000) reports
that the Cook Mountain Formation occurs as a massive clay serving as an upper
confining unit to the Sparta aquifer. The article by Albin (1964) reports that the Cook
Mountain Formation occurs as a moderately deep water marine clay in most of the
Mississippi embayment. However, an exception occurs in Bradley, Calhoun, and
Ouachita Counties where the formation consists of near-shore shallow-water dark-grey to
dark-brown silty clay. Albin states that the thickness of the Cook Mountain Formation is
approximately 150 feet, and includes silt, sand, and lignite clay deposited in a back-beach
environment.
Lying above the Cook Mountain Formation is the Cockfield Formation. The
Cockfield Formation consists of sands, slits, and clays deposited primarily under
subaerial conditions Albin (1964). The Cockfield Formation is less significant than the
Cook Mountain Formation in limiting vulnerability of the Sparta aquifer, because the
Cockfield occurs only over approximately the eastern 1/5 of Shumaker and does not
underlie industrialized areas to the west.
The youngest sediments at Shumaker are Quaternary (Pleistocene) terrace
deposits and Holocene alluvium. Quaternary deposits unconformably overly the
Claiborne Group, and the Jackson Group is absent at Shumaker. Terrace deposits occur
as a relatively thin (approximately 45 feet) blanket of sediments covering both the Cook
Mountain and Sparta Formations. These deposits consist mainly of gravel, sand, silt, and
clay which coarsen downward. Where the Cook Mountain Formation is absent,
Quaternary sediments (terrace deposits) rest directly on the Sparta Formation. Gravel at
the bottom of the Quaternary sediments forms an aquifer capable of producing significant
quantities of groundwater. The gravel layer averages about 10-20 feet thick and is the
most transmissive shallow aquifer in the Shumaker area. Holocene alluvium occurs as
deposits along stream channels.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 13 ^ij^Environmenta! Protection
-------
B. Subsurface Investigation
The geologic investigation filled important data gaps with new information
previously unavailable from either regional geologic studies or local industrial facility
reports. The geologic investigation included collecting subsurface information about the
stratigraphy and sediments comprising the Sparta Formation, Cook Mountain Formation,
and Quaternary deposits. The investigation determined whether the Sparta Formation is
overlain by Quaternary sediments or the Cook Mountain Formation, or both, and also
provided information on sand and clay sequences occurring within the Sparta aquifer
itself.
From February through April 2007, EPA drilled 7 exploratory stratigraphic test
holes at selected locations across Shumaker (fig. 4). Test hole locations were chosen to
address the geologic unknowns previously described, and they were chosen as locations
for Sparta aquifer monitoring wells. Monitoring wells served multiple purposes
including monitoring the Sparta near the Shumaker Public Service Company wells,
monitoring the Sparta near known contaminated areas and industrial areas, monitoring
groundwater at Locust Bayou, and providing data on groundwater flow directions.
Drilling depths ranged from 200-300 feet per location and approximately 1800
feet of sediment core were collected and examined from the 7 locations. Coring began at
the contact of the terrace deposits and the Sparta aquifer, and continued to approximately
300 feet per location. The percentage of core recovery, lithology, and related geological
characteristics was recorded in field notes. The depth of 300 feet was determined to be
adequate based on published information on regional geologic structure. Field
information is available in TechLaw, Inc. (2007, 2008).
C. Extent of the Cook Mountain Formation
The most important geological factor affecting the vulnerability of the Sparta
aquifer to contamination is the presence or absence of the Cook Mountain Formation.
Where the Cook Mountain Formation is present in relative thickness, it provides a
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 14 ^ij^Environmenta! Protection
-------
significant layer of clay over the Sparta aquifer, restricting downward groundwater
movement, and thus decreases Sparta aquifer vulnerability. The formation was identified
as discussed below.
Figure 4: Exploratory Test Holes DW1 - DW7
Because the depositional environments were similar during the time the Sparta
and Cook Mountain Formations were deposited, their respective lithologies are also
similar and make them difficult to differentiate based strictly on sediments. However, the
amounts of sand each formation contains, and their stratigraphic sequences of sand and
clay, were found to be different based on core samples. Cores from each stratigraphic
test hole were examined for stratigraphy, lithology, texture, uniformity, bedding,
laminations, fractures, sedimentary particles, core recovery, and other related
information. The most distinguishing characteristic of the Cook Mountain Formation
was found to be its overall stratigraphic composition which is thick massive clay
interspersed with thin beds of sand, thin layers of siltstone, sand and clay laminations,
and minor amounts of coal. Sparta stratigraphy is significantly different, and exhibits
much thicker deposits of sand and thinner layers of clay. Comparing cores to
depositional environments also assisted in differentiating formations. Albin (1964)
reported the Cook Mountain Formation consists of silt, sand, and lignite clay, deposited
in a near-shore back-beach environment. The sediments described above are consistent
Sparta Aquifer
Vulnerability Assessment
Page 15
United States
Environmental Protection
Agency
-------
with this type of system. The Sparta, by comparison, was deposited in more of a beach
front environment and thus contains thicker sand units along with thinner clays associated
with other lower energy fluvial-deltaic processes.
Drilling indicates the eastern part of Shumaker, near the eastern end of the former
rocket test range, was found to contain 108 feet of dense clay. At location DW6, coring
began at approximately 36.6 feet below ground surface (bgs), continued through the
Cook Mountain Formation to the Sparta aquifer contact at approximately 144 feet bgs,
and advanced deeper into the Sparta aquifer to a depth of 203 feet bgs. The transition
between formations is shown in core sample photographs A-D (fig. 5). No sediments
representing the Cockfield Formation were identified. Published maps by Albin (1964)
indicate the Cockfield probably overlies the Cook Mountain Formation further to the east
than DW6. If so, the Cockfield would provide an additional layer of protection over the
Sparta, although no industries exist east of DW6.
A. Core sample from DW6. Core transitions from B. Core sample from Cook Mountain Formation.
Quaternary sediment (brown iron stained sediment at Clay appears generally uniform in bulk composition
36.6') to Cook Mountain Formation. and density. Sand and coal laminations were
encountered in some samples.
C. Core showing transitions from Cook Mountain
to Sparta aquifer. Formation contact is
approximately 144'. At 144', sand content begins
to increase forming broken core wedges consisting
of sand and clay.
D. Core from Sparta aquifer. Material consists of
wedges and laminations of sand and clay.
Figure 5: Core Sample Photographs A-D
Sparta Aquifer
Vulnerability Assessment
Page 16
United States
Environmental Protection
Agency
-------
The Cook Mountain Formation provides a significant protective layer over the
Sparta aquifer where it exists. Cores indicate the Cook Mountain Formation probably
pinches out in a westerly direction between DW5 and DW6, and is absent over
approximately 40,000 of the 73,000 acre study area. Where the Cook Mountain
Formation is absent, the Sparta aquifer occurs beneath approximately 30-40 feet of
Quaternary terrace deposits and alluvium including the shallow gravel aquifer.
D. Sedimentary Particle Analysis
The drilling program provided an opportunity to closely examine material making
up the Cook Mountain and Sparta Formations by microscopically comparing sediments
and mineral compositions from each formation. The intention of this comparison was to
see whether sediment and mineral characteristics could be used to help differentiate
between Cook Mountain and Sparta sediments.
The analysis involved performing evaluations of discrete sediment samples taken
from cores in 5 foot intervals. Samples were examined for mineral types and
composition, grain size and shape including the degree of roundness and sphericity, and
vertical and interwell consistency and variability. The analysis shows that sand within
the Sparta is about 95% quartz [SiCy. Most quartz grains are rounded, but overall
shapes range from being well rounded to subangular. Quartz grains become rounded
when they contact and abrade against each other during transport. The degree of
roundness and sphericity indicates the amount of energy in the transport process and the
nature and type of the depositional environment. Besides quartz, sand grains also include
small percentages of biotite, muscovite, reworked coal, opaque minerals, and possibly
trace amounts of glauconite (fig. 6). This finding is consistent with Payne (1968) who
reported that Sparta sand is composed almost entirely of rounded to subrounded, fine to
medium quartz grains and is generally well sorted. The analysis indicated that there is
not any significant difference in mineral composition between the Sparta and Cook
Mountain Formations, and that sediments alone cannot be used to differentiate these
Formations. This is probably due to the fact that the alternating marine and non-marine
depositional environments were similar for each Formation as previously noted.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 17 ^ij^Environmenta! Protection
-------
Therefore, the most reliable way to identify these Formations is by making careful
observations on sequences, cycles, and thicknesses of sand and clay.
Based on findings about the area's geological framework alone, impacts to the
Sparta aquifer would be most likely to occur in areas where overlying regional confining
clay is absent, where the aquifer is close to the surface, and where existing contamination
is present in the overlying alluvial aquifer. This is mainly the case for saturated sands
near the top of the Sparta aquifer. As depth increases, however, geologic, hydrologic and
chemical data indicate that vulnerability decreases as will be pointed out in subsequent
sections.
Figure 6: Photomicrographs of Sand and Clay within the Sparta Aquifer
Samples from well DW2 at depths of 282 feet bgs (left) and 157 feet bgs (right). View on left shows well
rounded to rounded quartz grains indicative of high-energy depositional environment typical of near shore
marine sediments. View on right, taken under same magnification, shows clay fraction of Sparta with
cohesive clay-sized particles.
E. Deeper Sparta Sands
Geologic vulnerability was also assessed by examining the occurrence and
sequences of clay, fine sand, and rock (sandstone) deeper within the Sparta aquifer.
Sparta core samples indicate that Sparta stratigraphy is complex and consists mainly of
discontinuous layers of saturated fine sand and clay in thicknesses ranging from tens of
feet to only a fraction of an inch. Vertically, there is frequent alternation between sand
and clay strata, and individual sand and clay units do not appear continuous between test
Sparta Aquifer
Vulnerability Assessment
Page 18
United States
Environmental Protection
Agency
-------
holes. The occurrence of discontinuous sand and clay strata in various thicknesses is
consistent with the types of deposits occurring in a flu vial-deltaic depositional system.
Sand and clay also occur as laminations, seams, and lenses, with deposits of organic
matter, lignitic clay and sand, and lignite seams typically less than 1 foot thick.
Almost all of the Sparta sediments, like most sediments on the Gulf Coastal Plain,
were found to consist of lose sediments. A few thin layers of rock (sandstone) were
found, but only in thicknesses of several inches and which probably have no more effect
on groundwater flow than layers of clay. When sediments are changed to rock, it occurs
through a process known as lithification, involving the compaction and cementation of
sediments. Compaction reduces pore space within a body of sediments, and cementation
is a chemical process by which particles are held together by cements such as calcium
carbonate. The ability of the Sparta to produce large quantities of water and function as a
regional aquifer, is a reflection of an overall lack of compaction and cementation
associated with the Sparta deposition.
Each core interval percent recovery was estimated for every 10 feet of core.
Percent recovery is the amount of core actually obtained divided by the total possible
core length (10 foot core barrel length). Percent recovery ranged from 100% in
sediments consisting of mostly clay, to 0% in sediments which was practically all sand. A
recovery of 0% was frequently encountered in deeper sections of the Sparta, because
thicknesses of saturated, loose, fine sand would wash out of the core barrel as the barrel
was withdrawn from the ground. Where sand was not recoverable by core, its presence
was verified by geophysical logging which increased confidence in the ability of core
recovery to indicate the relative amounts of sand and clay.
F. Summary of Findings
In summary, the geological risk to deeper sands of the Sparta aquifer is mainly
associated with stratigraphy, including the presence/absence of clay within the formation,
continuity of clay strata, thickness of clay, and the interconnection of saturated fine
sands. Interconnected fine sands occur relatively close to the surface at industrial sites as
the Sparta aquifer becomes closer to the surface in a westerly direction. Based on
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 19 ^ij^Environmenta! Protection
-------
analyses of core sample material and percent recovery, the likelihood of downward
contaminant migration into deeper saturated fine sands of the Sparta aquifer (i.e., beyond
the first 100' of the aquifer), is greatest at EPA well location DW1, followed by DW2,
DW4, DW3, and DW7. At each of these locations, saturated fine sands alternate with
clay layers, and the percentages of fine sand layers in the first 100' are approximately
55% (DW1), 49% (DW2), 51% (DW4), 51% (DW3), 35% (DW5), and 28% (DW7). The
position of sand layers in DW2, being at the top of the section, seem to increase the
likelihood of downward migration at DW2. Estimations were not made for DW6 where
the Sparta is overlain by Cook Mountain.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 20 ^ij^Environmenta! Protection
-------
4. GOAL 2: SPARTA AQUIFER MONITORING
(Examine the vulnerability of the Sparta aquifer by: (i) installing
permanent groundwater monitoring wells, (ii) by monitoring for impacts
to the aquifer, and (Hi) by examining groundwater flow directions.)
A. Procedures Summary
The second major component of the vulnerability assessment was performing
groundwater monitoring. EPA installed a total of 10 wells in the Sparta aquifer. These
wells were constructed from the 7 exploratory test holes used for Goal 1, plus 3
additional monitoring wells with deeper screens to help understand vertical flow (fig 7).
Well construction information and field sampling procedures are described in detail in
field activity reports by Tech Law Inc. (2007 and 2008). Quality assurance criteria are
contained in the two QA Project Plans previously mentioned in Section II. C (EPA 2007a
and 2007b).
Figure 7: Location Of The 10 Sparta Aquifer Monitoring Wells
Wells DW2L, DW3L, and DW4L are screened lower in the Sparta than adjacent wells.
Aquifer vulnerability was assessed by performing monitoring and determining
flow directions and gradients. Monitoring was conducted for volatile organic compounds
(VOCs), semi-volatile organic compounds (SVOCs), explosives, perchlorate, and RCRA
Sparta Aquifer
Vulnerability Assessment
Page 21
United States
Environmental Protection
Agency
-------
metals. Analytes are listed in EPA 2007a and 2007b. There were two groundwater
sampling events: (1) March-April 2007, and (2) January 2008. Analytical results were
compared to EPA Regional Screening Levels (RSLs). EPA uses screening levels when a
site is initially investigated to determine if potentially significant levels of contamination
are present to warrant further investigation. Screening levels represent relatively
protective environmental concentrations. The significance of detections to vulnerability
is discussed under Key Observations and Uncertainties (Section 7).
B. Analytical Results Summary
In 2007, arsenic, lead, and bis(2-ethylhexyl)phthalate exceeded RSLs for tap
water. Table 2 contains a summary of detections versus RSLs. Samples from wells DW-
1, DW-2, DW-2L, DW-2LD, DW-3L, DW-4 and DW-4L exceed RSLs for arsenic with a
maximum concentration of 18.30 ug/1. Samples denoted with a "D" are QA duplicate
samples. Samples from DW-2, DW-2L, DW-2LD, DW-4, DW-4L exceeded RSLs for
lead with a maximum concentration of 65.9 ug/1. Samples collected from DW-1, DW-
2LD, DW-3, DW-4L, and DW-6 exceeded RSLs for bis(2-ethylhexyl)phthalate with a
maximum concentration of 9.20 ug/1. Perchlorate was not detected in any of the Sparta
aquifer samples. The complete results are in Tech Law (2007 and 2008).
Contaminant
Arsenic
Lead
bis(2-
ethylhexyl)phthalate
Frequency of
Detection
7/9
11/12
8/12
Screening Concentration
(RSL)
0.045
15.0
4.80
Well
ID
DW1
DW2
DW2L
DW2LD
DW3L
DW4
DW4L
DW2
DW2L
DW2LD
DW4
DW4L
DW1
DW2LD
DW3
DW4L
DW6
Concentration
(M9/D
10.50J
9.80J
18.30J
16.50J
4.00J
1 1 .40J
5.50J
43.40
65.9
56.7
20.50
18.80
9.20J
8.60J
8.30J
5.20J
7.20J
Table 2: Screening Levels vs. Sample Concentrations 2007
Laboratory qualifier "J" stands for estimated results. RSL for lead is i
technology action level.
• treatment
Sparta Aquifer
Vulnerability Assessment
Page 22
vvEPA
United States
Environmental Protection
Agency
-------
In 2008, samples were collected and analyzed for the identical analytical suite as
in 2007 (Table 3). Arsenic and lead exceeded their constituent-specific RSLs. The
sample collected from deep well DW3L showed an arsenic concentration of 17 (j.g/1. The
samples collected from deep well DW2L exceeding the RSL for lead with a maximum
concentration of 20
Contaminant
Arsenic
Lead
Frequency of
Detection
1/12
2/12
Screening
Concentration (RSL)
0.045
15.0
Well
DW3
DW2L
DW2L
Concentration
(M9/D
17.0
18.0
20.0
Table 3: Screening Levels vs. Sample Concentrations 2008
C. Sparta Aquifer Flow Directions
Measurement data was collected for both horizontal and vertical groundwater
flow directions. Water levels were measured on two separate occasions, on May 9, 2007,
and then approximately 1-year later on May 19, 2008. To reduce any temporal effects on
measurements, all water levels were measured within a 24-hour period on each of the two
occasions. Wells DW1, DW2L, DW3L, DW4L, DW5, DW6, and DW7 were used to
determine horizontal flow directions. Vertical gradients were determined by using well
clusters. These clusters are: (1) DW2, DW2L, and SW9; (2) DW3, DW3L, and SW8;
and (3) DW4, DW4L, and SW5. Wells SW9, SW8, and SW5 are shallow alluvial wells
installed as part of alluvial aquifer monitoring discussed under Goal 4. Water level
measurements for 2007 and 2008 are listed below in Table 4.
Well
DW1
DW2
DW2L
DW3
DW3L
DW4
DW4L
DW5
DW6
DW7
May 9, 2007
110.42'
138.68'
114.67'
132.88'
110.13'
119.23'
97.56'
190.53'
210.51'
179.70'
May 19, 2008
111.23'
138.60'
115.12'
132.16'
110.48'
119.49'
98.02'
190.19'
210.71'
179.94'
Table 4: Water Level Measurements (mean sea level)
Sparta Aquifer
Vulnerability Assessment
Page 23
vvEPA
United States
Environmental Protection
Agency
-------
For both years, the highest water levels measured were in the eastern part of the
study area at well DW6. From that location, water levels were found to decrease in all
wells towards the west and southwest. The lowest water level measured was in well
DW4L located approximately 12 miles from DW6. Horizontal Sparta flow directions are
provided in figure 8. Based on the gradients in figure 8, the total change in Sparta
hydraulic head across Shumaker is approximately 145' (over 15 miles), as measured from
an approximate 230'contour line on the east, to an approximate 85' line on the west. This
results in an average approximate groundwater gradient of 9.6' per mile, or 0.18%.
Figure 8: Groundwater Flow Direction In Sparta Aquifer 2008
Wells DW2L, DW3L, and DW4L were used for developing figure 8, rather than wells DW2, DW3, and
DW4. Wells DW2L, DW3L, and DW4L have well screen depths more consistent with the remaining
Sparta wells (i.e., wells DW1, DW5, DW6, and DW7). A comparison of well screen depths and water
levels by aquifer is provided in Table 5.
Figure 8 only presents flow directions for 2008, because 2007 water levels and
flow directions are practically the same. Contoured data are most accurate where
contours are nearest well locations/control points that constrain contour lines. As one
moves away from well locations and reduced data control, more careful interpretation of
flow direction is needed because contour lines are not constrained by actual measurement
data. This is particularly important near the outside edge of the Shumaker NAD
Sparta Aquifer
Vulnerability Assessment
Page 24
United States
Environmental Protection
Agency
-------
boundary and even further around the perimeter of flow direction maps. Flow gradients
appear more uniform and consistent in the central part of Shumaker than the outer part,
and this is probably a function of the spacing of wells.
EPA data from well clusters indicate a substantial downward groundwater flow
gradient. The downward gradient exists from the alluvial aquifer to the upper sands of
the Sparta aquifer, and then again from the upper Sparta sands to where the lower Sparta
wells are screened. Water level elevations decrease in each successive deeper aquifer
from a high of 143.38 ft msl to a low of 98.02 ft msl (Table 5).
Well
Cluster
SW8
DW3
DW3L
Water Level
(ft msl)
135.97
132.16
110.48
Screen Interval
(ft msl)
122.78-112.78
99.56-89.56
21.36-1.36
Aquifer
Alluvial
Sparta
Sparta
Gradient
Magnitude
(unitless)
0.16(SW8toDW3)
0.26 (DW3 to DW3L)
SW5
DW4
DW4L
123.22
119.49
98.02
111.50-101.50
75.86-65.86
10.89- (-9.11)
Alluvial
Sparta
Sparta
0.10(SW5toDW4)
0.30 (DW4 to DW4L)
SW9
DW2
DW2L
143.38
138.60
115.12
134.52-124.52
104.82-94.82
25.96-5.96
Alluvial
Sparta
Sparta
0.16(SW9to DW2)
0.28 (DW2 to DW2L)
Table 5: Water Level Data For Well Clusters, 2008
Magnitude is change in water level divided by change in well screen
midpoint (dh/dl). More data on alluvial water levels is contained in
Table 8.
The alluvial water levels appear distinct from upper Sparta water levels, which are
about 4 feet lower than the alluvial levels. As depth increases, separation in water levels
also increases and the difference between the upper Sparta and lower Sparta levels
averages about 22 feet. Water levels in the upper Sparta more closely resemble alluvial
water levels than lower Sparta water levels, suggesting increased vulnerability in the
upper Sparta aquifer. Separation between upper and lower Sparta levels does not
mean the lower Sparta cannot become contaminated, but it is less likely. If
Sparta Aquifer
Vulnerability Assessment
Page 25
vvEPA
United States
Environmental Protection
Agency
-------
contamination were to find a pathway downward, it would probably move quickly to
impact deeper sections of the Sparta.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 26 ^Erwironmenta! Protection
-------
5. GOAL 3: SURFACE WATER INVESTIGATION
(Examine aquifer vulnerability by conducting surface water sampling
including streams, lakes, and ponds over the extent of Shumaker, and the
Locust Bayou area.)
A. Streams and Lakes
Considering the possibility of interaction between surface water and groundwater
in the Shumaker vicinity, EPA conducted surface water sampling and analysis to
determine whether surface water contains contamination that may impact the Sparta
aquifer. The natural down-cutting action of streams of the Shumaker drainage network
has caused a dissection of terrace deposits overlying the Sparta aquifer. The presence of
terrace deposits incised by streams, and the occurrence of relatively high and fluctuating
water table levels, increases the possibility of surface water and groundwater interaction.
Topographic data from U.S. Geological Survey quadrangle maps indicates stream level
elevations are close to water table elevations.
There are approximately 172 miles of stream courses at Shumaker. Most streams
are probably intermittent and flow only after significant rainfall events. Other streams
obtain water from groundwater seepage (i.e., baseflow), and land surface runoff. The
area's larger streams are Two Bayou, Locust Bayou, and Caney Creek, which are all
probably perennial streams. These and other relatively large streams connect with
numerous smaller creeks and tributaries forming a dendritic drainage pattern. This
pattern suggests that stream courses are controlled by slope, as opposed to other factors
such as geologic structure, and that sediments have a relatively uniform resistance to
erosion. Stream channel dimensions are highly variable and range from small tributaries
which are approximately 10-15 feet wide and several feet deep, to large streams which
are up to 30-40 feet wide and possibly 5 or more feet deep at high water stage.
The direction of stream flow is generally northeast to southwest (fig. 9). A few
streams have different flow directions including Taylor Creek, located at the eastern end
of Shumaker which flows to the south, and an unnamed stream just north of Taylor Creek
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 27 ^ij^Environmenta! Protection
-------
which flows to the southeast. Other exceptions to southwest flow occur at Two Bayou
and in channelized drainage (flowing to the south) near the western end of Shumaker.
Surface water also occurs in 4 relatively large (5-10 acre) lakes/ponds, and also
within a small number of man-made channels constructed near former and present
operating industrial areas. The most significant lakes/ponds are Covington Pond, North
Pond, Middle Pond, and South Pond. Water from streams and lakes/ponds were sampled
as described below.
Figure 9: Surface Water Sampling Locations
All locations were streams except for locations 26, 28, 29, and 30 which were lakes.
B. Sampling Results
During February 2008, EPA conducted reconnaissance surface water sampling
and analysis for perchlorate, explosives, and metals. QA plans and procedures were
established in the QA Project Plan dated December 18, 2007 (EPA 2007b). Detailed
analytical information is presented in the field activity report by Tech Law 2008.
Sparta Aquifer
Vulnerability Assessment
Page 28
United States
Environmental Protection
Agency
-------
EPA collected surface water from a total of 34 locations, as shown in figure 9.
Locations were determined by selecting sampling points which captured surface water
runoff representing as much of the Shumaker land area as possible. As part of the
planning process, a field reconnaissance trip was made prior to sampling to plan and
inspect each possible location. The purpose was to check for water availability, site
accessibility, and locations of stream courses relative to industrial areas and potential
land-based sources of contamination. Prior to collecting samples, global positioning
system (GPS) coordinate data, pH, and water flow characteristics were collected and
recorded at each sampling site.
RSLs were not exceeded in surface water samples. Detections still occurred for
metals (arsenic, barium, cadmium, chromium, lead, and selenium), explosives (2-
nitrotoluene and 3-nitrotoluene), and perchlorate, although these were below screening
levels. The full list of detections and other field observations by location is available in
Tech Law (2008). Although results for explosives are reported, difficulties with
laboratory analyses for explosives indicate a level of uncertainty for those results.
Perchlorate was detected at many surface water sampling locations at concentrations
ranging from less than 1.0 |^g/l to 5.2 jig/1. The origins of perchlorate are discussed in the
next Section under Goal 4, and uncertainties about explosives are discussed in the chapter
on Key Observations and Uncertainties.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 29 ^ij^Environmenta! Protection
-------
6. GOAL 4: PERCHLORATE AT LOCUST BAYOU
(Determine the source of per chlorate detected in groundwater monitoring
wells and drinking water at Locust Bayou.)
A. Background
Locust Bayou and vicinity consists of a small community located several miles
southeast of East Camden extending along State Highway 278 and various county roads
(fig. 10). Perchlorate was initially detected in residential tap water samples in 2005, and
since then has been detected in follow up sampling and in monitoring wells used for this
study. To accomplish Goal 4, EPA performed monitoring of groundwater in the
Pleistocene terrace deposit gravel, considered the presence and location of known
existing perchlorate sources, evaluated groundwater velocity relative to perchlorate
releases, tested for perchlorate in surface water, and conducted fingerprinting of
perchlorate by using chlorine and oxygen isotopes.
Easi
Camden
r
Lakeside
Locust
Bayou
Figure 10: Locust Bayou and Vicinity
Sparta Aquifer
Vulnerability Assessment
Page 30
United States
Environmental Protection
Agency
-------
EPA targeted the gravel aquifer because residential wells at Locust Bayou are
probably screened in that aquifer. The gravel aquifer is capable of producing large
quantities of groundwater and is relatively shallow, thus reducing drilling and pumping
costs, making the aquifer an economical water source. The gravel is about 5-20 feet thick
and is significant because it is the most transmissive water bearing zone above the Sparta.
It can transport relatively large amounts of groundwater fairly rapidly, and provide a
pathway for perchlorate to spread out and contact the top of the Sparta aquifer. The size
of the area selected for monitoring the gravel aquifer encompasses approximately 16,000
acres (25 square miles) including the area between Locust Bayou and highway 274 to the
north (fig 11). The area includes land both within and outside the Shumaker NAD
footprint.
Figure 11: Location of Gravel Aquifer Monitoring Wells
Because fundamental groundwater information needed to be collected, EPA
believed it was best to place gravel aquifer monitoring wells over a wide area to collect
fundamental information on flow direction and perchlorate concentrations. EPA
installed 10 monitoring wells (SW1 - SW10) which range from 26 to 55 feet deep.
Individual well locations were determined by considering site characteristics including
the locations of existing industrial areas, site accessibility, topography, possible
Sparta Aquifer
Vulnerability Assessment
Page 31
United States
Environmental Protection
Agency
-------
groundwater discharge areas, and monitoring well distribution. Although perchlorate was
the main constituent of interest, sampling was also conducted for VOCs, SVOCs,
explosives, and metals. The purpose for these constituents was to help determine
whether perchlorate was occurring by itself, or was occurring as one of a number of
contaminants in a chemically complex plume. Analytical results, except for perchlorate,
were compared to RSLs. Perchlorate results are discussed separately in Section D below.
Field procedures are described in detail in field activity reports by Tech Law (2007 and
2008). Quality assurance criteria are contained in the two QA Project Plans previously
mentioned in Section II. C (EPA 2007a and 2007b).
B. Monitoring Results
In 2007, RSLs were exceeded for arsenic, lead, and bis(2-ethylhexyl)phthalate.
Samples from wells SW-1, SW-1D, SW-5, SW-6, SW-7, SW-8 and SW-10 showed
arsenic concentrations exceeding the RSL with a maximum concentration of 29.60 ug/1.
Lead concentrations exceeded the RSL with a maximum concentration of 65.90 ug/1, and
were found in samples from wells SW-5, SW-6, SW-7 and SW-10. Samples collected
from wells SW-1, SW-2, SW-5, SW-7 and SW-10 exceeded the RSL for bis(2-
ethylhexyl)phthalate with a maximum concentration of 72.00 ug/1. Table 6 provides a
summary of exceedances from the 2007 shallow groundwater samples. The full set of
analytical results may be found in Tech Law (2007 and 2008).
In 2008, samples were collected and analyzed for the identical analytical suite as
in 2007 (Table 7). Arsenic and lead exceeded their constituent-specific RSLs. Samples
collected from shallow wells SW2, SW7, and SW8 showed arsenic concentrations
exceeding the RSL with a maximum concentration of 29 ^ig/1 at well SW7. Samples
collected from shallow wells SW2, SW4, SW7, SW8, and SW9 showed lead
concentrations exceeding the RSL with a maximum concentration of 41 ^ig/1 at shallow
well SW7. Table 7 provides a summary of RSL exceedances for 2008 sampling results.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 32 ^ij^Environmenta! Protection
-------
Contaminant
Arsenic
Lead
bis(2-
ethylhexyl)phthalate
Frequency of
Detection
7/12
11/12
8/9
Screening Concentration
(RSL)
0.045
15.0
4.80
Well
ID
SW1
SW1D
SW5
SW6
SW7
SW8
SW10
SW5
SW6
SW7
SW10
SW1
SW2
SW5
SW7
SW10
Concentration
(M9/D
2.30J
3.20J
7.0J
29.60
12.10J
4.30J
2.0
18.60
48.90E
24.20E
22.6
72.07
5.71J
6.70J
4.90J
6.60J
Table 6: Screening Levels vs. Sample Concentrations, 2007
Laboratory qualifier "J" stands for estimated result "E" means results did not meet serial dilution
acceptance criteria.
Contaminant
Arsenic
Lead
Frequency of
Detection
3/10
5/10
Screening
Concentration
0.045
15.0
Lab ID/Well
SW2
SW7
SW8
SW2
SW4
SW7
SW8
SW9
Concentration
(M9/D
16
29
21
29
21
41
37. 7J
22
Table 7: Screening Levels vs. Sample Concentrations, 2008.
Laboratory qualifier "J" stands for estimated result.
C. Flow Directions
Water level measurements were collected within a 24 hour period on May 9, 2007
and then on May 19, 2008. Water levels are similar for each year as indicated on Table
8. The highest water levels were measured east and north of Locust Bayou at wells SW 9
and SW10. From SW10, the approximate flow direction is to the west-southwest, and
flow from SW9 is towards the south. The lowest water level measured was at SW4
located just south of highway 278.
Sparta Aquifer
Vulnerability Assessment
Page 33
vvEPA
United States
Environmental Protection
Agency
-------
Well
SW1
SW2
SW3
SW4
SW5
SW6
SW7
SW8
SW9
SW10
May 9, 2007
135.08'
134.28'
126.29'
117.99'
123.67'
122.30'
129.37'
133.49'
143.47'
158.04'
May 19, 2008
133.84'
133.53'
125.28'
117.17'
123.22'
121.84'
128.82'
135.97'
143.38'
158.40'
Table 8: Water Level Elevations for Alluvial Wells 2007 and 2008.
Groundwater flow directions are shown on Figure 12. Because water levels,
contoured gradients, and flow directions are nearly the same for 2007 and 2008, only the
2008 mapped data are presented. Similar to data for the Sparta aquifer, care must be used
when interpreting contoured data. Contours should only be considered reliable in areas
where there are actual data points, as is the case for the area within and north of Locust
Bayou between highways 278 and 274. Where contours extend beyond this area,
additional care must be used in making interpretations about flow directions and
gradients.
The approximate total change in head from SW8 to SW6 is 14.13 ft over 2.2
miles, which gives an approximate hydraulic gradient of 0.0012. A notable feature is a
flattened gradient just south of SW1 and extending to near SW2. This anomalous feature
is probably the result of SW1 being screened lower than the other alluvial wells. SW1
was screened lower because it was the first well installed during the alluvial aquifer
drilling program, and consistency on screen elevations had not yet been developed. If
SW1 had been screened higher, the gradient would probably be more uniform showing a
more consistent south-south west flow direction.
Sparta Aquifer
Vulnerability Assessment
Page 34
vvEPA
United States
Environmental Protection
Agency
-------
Figure 12: Groundwater Flow Directions in Alluvial Aquifer Near Locust Bayou
D. Perchlorate Fingerprinting
L Introduction
The determination of the source of perchlorate required using a special analytical
technique, fingerprinting, which can differentiate between types of natural and man-made
perchlorate. Perchlorate fingerprinting relies on an analysis of chlorine and oxygen
stable isotopes. Perchlorate occurs as a natural and man-made chemical consisting of one
chlorine atom bonded to four oxygen atoms. An isotope is one of two or more atoms
whose nuclei have the same number of protons, but different number of neutrons. A
stable isotope is a non-radioactive isotope that does not decay. The rationale for using
stable isotope analysis is that it would provide isotopic signatures which would
significantly narrow down the field of possible sources.
Man-made perchlorate, also referred to as synthetic and anthropogenic, is
commonly used as an oxidizer in explosives, road flares, fireworks, rocket motors, and
other uses (EPA, 2008). Man-made perchlorate is also found in some hypochlorite
products and disinfectants (Massachusetts DEQ, 2006). EPA (2008b) reports that about
90% of all domestically produced perchlorate is used by the defense and aerospace
industries in the form of ammonium perchlorate. The most well known natural
Sparta Aquifer
Vulnerability Assessment
Page 35
United States
Environmental Protection
Agency
-------
perchlorate is associated with nitrate ore deposits from the Atacama Desert of Chile.
These ore deposits are used to make fertilizer which has long been imported into the U.S.
and is still imported today. One metric ton of imported fertilizer could yield as much as
40 to 340 million gallons of water with 6 [j.g/1 of perchlorate (Bohlke, 2005). Perchlorate
may also occur naturally in other areas such as arid areas of the western U.S., and
possibly in west Texas where perchlorate has been detected in groundwater over wide
areas.
Both natural and synthetic perchlorate are possible at Locust Bayou, and were
both given earful consideration during this investigation. Since Shumaker was used for
the manufacture, testing, distribution, destruction, and storage of naval rockets, it was
reasonable to consider that historical operations might have been a source of perchlorate
in groundwater. Existing perchlorate contamination at nearby industries was also
considered as possible sources. The most well known perchlorate plume in the Shumaker
area is at the Aerojet facility north of Locust Bayou. Aerojet purchased the facility from
Atlantic Research in 2003, which had existing perchlorate groundwater contamination.
Imported fertilizer was also considered a possible source since Shumaker was used for
farming prior to the existence of Shumaker, and perchlorate has the ability to persist in
groundwater for decades. Household bleach was also considered. An analysis of the
perchlorate content of 4 household bleach products was shown to contain from 89 to
8000 [ig/l (Massachusetts DEQ, 2006). Still other possible sources were thought to be
road flares and fireworks. The stable isotope analysis provided a means to confidently
rule out many of these possibilities and provide positive identification of source types.
ii. Field Sampling
Field sampling had to be carefully designed to ensure that enough perchlorate was
collected for analysis. Each sample had to contain 10 mg of perchlorate. The laboratory
analysis was performed by the Environmental Isotope Geochemistry Laboratory at the
University of Illinois at Chicago. Field sampling took place from May 19, 2008 through
May 24, 2008.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 36 ^ij^Environmenta! Protection
-------
Perchlorate was identified in 8 of 10 alluvial monitoring wells with
concentrations ranging from 0.57 jig/1 to non-detect (fig 13). Because it was difficult
to collect 10 mg perchlorate per sample from water containing low concentrations, the
two wells showing the highest concentrations were chosen for isotopic sampling (i.e.,
wells SW5 and SW7). Well SW5 contained 0.57 ^ig/1 perchlorate, and well SW7
contained 0.25 ^ig/1 perchlorate. For comparison purposes, a third well was also chosen
for isotopic analysis, which was located at the Aerojet facility a few miles to the north.
Sampling was completed by concentrating perchlorate in a highly perchlorate-
selective bifunctional ion exchange resin (Purolite A-53E), in columns designed for low
perchlorate concentrations. Pumping through the ion exchange columns at well SW5
took 42 hours at a rate of 8 liters/minute, and pumping at well SW7 took 55 hours at a
rate of 12 liters/minute. Water pressure was set not to exceed 30 pounds per square inch
to allow the proper residence time in the filter. A filter (with 5-15 micron sand/sediment
filter insert) was used between the pump and ion-exchange column to reduce the
sand/sediment particulates in the groundwater entering the ion-exchange column. The
sampling team observed the sampling 24 hours a day. No additional QC samples were
required. At the laboratory, perchlorate was converted to a form that could be isotopically
analyzed using a gas-source isotope ratio mass spectrometer.
»SW 10 10.03 5j|
Locust Bayou
•SW-H0.14)
Figure 13: Perchlorate Detections In Monitoring Wells Near Locust Bayou
Sparta Aquifer
Vulnerability Assessment
Page 37
United States
Environmental Protection
Agency
-------
The sample collected from Aerojet was collected from a known perchlorate
source area (Building 52) where perchlorate had been released into groundwater during
previous operations by Atlantic Research (1983-1985). This sample did not require ion
exchange columns and extended pumping times because the concentration was
sufficiently high to provide 10 mg perchlorate by direct sampling. The well sampled was
Aerojet monitoring well 25s, identified as AJ-01 for this study. Five gallons of water
were collected from well AJ-01 and sent to the University of Illinois for analysis.
Hi. Isotopic Results
This section begins by explaining how isotopic results are reported. Analyses are
reported by comparing isotopic ratios for 537C1, 518O, A17O and by comparing those ratios
to specific reference standards. The ratios are determined from the following
relationships:
Chlorine Isotope Ratio Analysis
S37Cl (%o) = [(37Cl/5Cl)sample/(37Cl/5Cl)smoc* -1] x 1000
Oxygen Isotope Ratio Analysis
818O (%c) = [(18O/6O)sample/(18O/6O)vsmow* - 1] x 1000
A17O (%o) = [(1+S17O/1000)/(1 + S18O/1000)a525] -1] x 1000
*smoc is standard mean ocean chloride; vsmow is Vienna standard mean ocean water.
Isotopic ranges for natural and synthetic perchlorate are reported by Bohlke, et.
al., (2005) and Sturchio et. al., (2006). Samples from the Atacama Desert and derivative
fertilizer products have 537C1 values range from -14.5 to -11.8%o. The range of 537C1 for
synthetic perchlorate is -3.1 to+1.6%o. The 618O range is -9.3 to -4.2 %0 for natural
perchlorate, and the synthetic range is -24.8 to -12.5 %o. The range of A17O for synthetic
perchlorate is 0.0 + 0.1 %o, and +8.93 to +9.57 %o for natural perchlorate. Sturchio et.al,
(2006) reports the most diagnostic isotopic characteristic of natural perchlorate is its
positive A17O value compared to synthetic perchlorate. An example plot of isotopic
ranges can be viewed graphically in figure 14, using values for 537 and 518O, to illustrate
how data from natural and man-made perchlorate may group. Although EPA data fall
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 38 ^ij^Environmenta! Protection
-------
reasonably close to this idealized graph, actual data results, however, should fall close to
the numerical ranges indicated above.
Range of
Anthropogenic
(Synthetic)CiO4-
West Texas CiO4-
in Groundwater
Range of
Geogenic
(Atacama) CIO4'
Figure 14: Idealized Graph Showing Natural and Man-Made Perchlorate Data.
From Motzer, 2006
Results for the 2 samples from Locust Bayou and sample from Aerojet are
presented in Table 9. Results indicate the 3 groundwater samples are isotopically distinct
and fall within ranges for synthetic perchlorate (sample AJ01) and natural perchlorate
(sample SW5), and include a possible mixture of the two types showing microbial
reduction (SW7). Sample SW7 appears to contain a 2:1 mixture of AJ01 and SW5 where
the mixture experienced a minor extent of biodegradation after being mixed.
Sample ID
AJ01
SW5
SW7
A1S0
-15.8
-1.5
-2.7
A1V0
0.0
8.0
2.7
A37C1
0.8
-10.6
0.5
Table 9: Fingerprinting results for Locust Bayou and Aerojet
E. Most Likely Sources of Perchlorate
Multiple lines of evidence, along with uncertainties, were evaluated to determine
the most likely sources of perchlorate. Major indications of perchlorate sources are the
following: stable chlorine and oxygen isotopes, groundwater flow directions, geologic
Sparta Aquifer
Vulnerability Assessment
Page 39
x>EPA
United States
Environmental Protection
Agency
-------
conditions, travel time estimations, land use, facility operations, history of perchlorate
releases, perchlorate age constraints, and the respective locations of existing facility
perchlorate contamination. The most significant uncertainties involve possible
perchlorate contributions from past naval operations, and differentiating
perchlorate between individual existing industrial facilities.
/. Imported Chilean Nitrate Fertilizer
Chilean nitrate fertilizer was the primary nitrate fertilizer used in the U.S. during
the early 1900s. This fertilizer is still imported today and 68,000 metric tons are used in
the U.S. annually. No historical records were identified that indicate how much imported
nitrate fertilizer may have been used at Shumaker in the past, but the analysis of stable
chlorine and oxygen isotopes clearly indicates imported Chilean nitrate fertilizer is a
source of naturally occurring perchlorate in groundwater at Locust Bayou. It is also
reported that before Shumaker existed, the area was used primarily for farming.
Different populations of perchlorate concentration data may possibly be used to
differentiate between natural and man-made perchlorate at Shumaker. There appear to be
two populations of data based on concentration levels. These two data populations
appear most evident in surface water results, where concentrations of <0.1 ^ig/1 were
found in areas without any indication of past or present industrial activity. These
concentrations appear to represent surface or near surface perchlorate contamination
indicative of the land application of Chilean nitrate fertilizer. As streams pass near
certain industrial areas, perchlorate concentrations increase indicating perchlorate loading
of stream water with synthetic perchlorate. Although more research would be required to
determine specific relationships, there may be similar indications in groundwater. For
example, concentrations of fertilizer based perchlorate may only be a fraction of a part
per billion, whereas synthetic concentrations or mixtures of synthetic and fertilizer based
perchlorate, may generally be higher.
ii. Synthetic Perchlorate
Stable chlorine and oxygen isotopes indicate synthetic perchlorate is occurring as
a mixture with perchlorate from Chilean nitrate fertilizer. However, synthetic
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 40 ^ij^Environmenta! Protection
-------
perchlorate, and perchlorate from Chilean nitrate fertilizer, may also occur separately.
Perchlorate from Chilean nitrate fertilizer was detected, in fact, without synthetic
perchlorate in well SW5.
There are two likely possibilities for how synthetic perchlorate is being
transported to the Locust Bayou vicinity. The first possibility is through the gravel
aquifer. Geologic information, obtained by drilling during this investigation and from
published information previously discussed, indicates the gravel aquifer is a continuous
blanket beneath the entire Shumaker area including the Locust Bayou area. The gravel
provides a physical pathway for transport from upgradient sources.
A second possible transport mechanism is through surface water. Surface water
sampling data indicates perchlorate loading is occurring where streams pass near
industrialized areas associated with the aerospace industry. Perchlorate loading is
occurring in Dogwood Creek as it passes from Section 5 into Section 6 (Township 13
Range 15) where perchlorate increases from 0.027 ^ig/1 to 5.2 ^ig/1. The stream segment
in Dogwood Creek is near the perchlorate groundwater plume at Aerojet. Increases are
also seen in Two Bayou from Section 10 to Section 24 (Township 13 Range 16) where
perchlorate increases from 0.026 ^ig/1 to 1.3 ^ig/1. The increase in Two Bayou is near a
sewage treatment plant outfall owned by Highland Industrial Park, and perchlorate was
discovered near the outfall in 2007. Up to 0.944 mg/1 perchlorate has been detected in
a stream just south of the outfall. Perchlorate is believed to have been transported
to the sewage plant through a pipeline extending approximately 2 miles west from a
building with a basement pump at the Aerojet facility. Since this release was
discovered, Aerojet has ceased discharge from the building sump.
Since a perchlorate plume exists at Aerojet, and the Locust Bayou vicinity
and other residences are downgradient, it is reasonable to consider the existing
perchlorate plume at Aerojet a possible source. It should be noted, however, that any
other perchlorate releases from other past or present facilities or units in the vicinity,
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 41 ^ij^Environmenta! Protection
-------
should they be identified as having released perchlorate to groundwater in significant
quantities and concentrations, could just as well be considered potential sources.
For the Aerojet area, Building 52 was used for washing out rocket motor casings
from 1983 to 1985 (Atlantic Research Corporation, 2001), and was investigated for the
2001 RCRA facility investigation report. Based on perchlorate concentration maps,
Building 52 is the primary source area for the plume at Aerojet, with concentrations in
shallow groundwater reaching several thousand parts per million (mg/1). Groundwater
near Building 52 is currently undergoing remediation, and concentrations have decreased
according to site personnel.
In order to evaluate whether contaminated groundwater could reach the Locust
Bayou vicinity since the 1983-1985 time frame, a basic travel time estimate was made for
seepage velocity in the alluvial gravel layer based on a form of Darcy's law:
n Ax
Where: K = hydraulic conductivity
N = porosity
Ah
— = gradient
Ax
Using a hydraulic gradient of 0.0012, and text book values for gravel hydraulic
conductivity (0.1 cm/s), and a porosity (0.25), the groundwater seepage velocity in the
gravel layer is 1.36 ft/day (496.5 ft/yr). Based on this rough estimate, there is enough
time for the front of the perchlorate plume to have extended approximately 13,405
feet since the release at Building 52 occurred. Other factors affecting the southward
extent of perchlorate contamination are transport of perchlorate through surface water,
and any perchlorate that may have been released from source areas located further south
than Building 52, thus potentially resulting in contamination still further to the south.
Atlantic Research (2001) reported travel times for shallow sands and clays
overlying the gravel layer as being 21 to 49 ft/yr, and reported other velocities as being
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 42 ^ij^Environmenta! Protection
-------
29 to 68 ft/yr to 14 to 28 ft/yr for intermediate and deeper sediments. However, Atlantic
Research noted that the actual rate of water movement may be higher through zones of
coarser aquifer matrix, and indicated a generally westward flow.
Hi. Other Possible Sources and Uncertainties
Although fingerprinting can differentiate between plumes of perchlorate from
Chilean nitrate fertilizer and synthetic perchlorate, the technique cannot tell the difference
between individual sources of synthetic perchlorate, and it cannot determine the
difference between old and new synthetic perchlorate. Therefore, other sources of
information were used to try and further evaluate the scope of these possibilities.
The history of potassium and ammonium perchlorate for use in solid rocket
propellant was described by J.C. Schumacher (1999) in a paper from the American
Institute of Aeronautics and Astronautics. Schumacher presents a timeline for domestic
perchlorate production and describes the evolution of production facilities and
companies, including contracting agreements with the U.S. Navy. The timeline suggests
it is unlikely for perchlorate to have been used in naval solid rocket motors during the
time Shumaker was in operation by the Navy, but does not completely rule out the
possibility either.
During WWII, the Western Electrochemical Company (WECCO) was the
preeminent company which designed, constructed, and operated perchlorate plants from
1940 until its merger with American Potash and Chemical Company in 1955. In 1942,
WEECO designed and constructed a small pilot plant to produce experimental quantities
of potassium and ammonium perchlorate for the Air Corps Jet Propulsion Research
Project in Los Angeles, California, known as the GALCIT Project No 1 (after the
Guggenheim Aeronautics Laboratory at the California Institute of Technology). The
GALCIT project is described as the prelude to the formation of the Aerojet Engineering
Corporation and large-scale commercial development of composite solid rocket motors.
The paper reports that the first application of solid rocket motors was for military aircraft
Jet-Assisted-Take-Off (JATO) devices. In 1945, WEECO modified and operated a new
perchlorate plant in Henderson, Nevada, and then began to produce perchlorate for the
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 43 ^ij^Environmenta! Protection
-------
U.S. Navy in 1953 at a capacity of 50 tons/day. A timeline of critical dates relative to
Shumaker and the history of early perchlorate production in the U.S. is given below:
• 1942: WEECO GALCIT Project No 1 (small scale perchlorate production)
• 1944: Shumaker ammunition depot operations begin
• 1945: WEECO starts plant in Henderson, Nevada
• 1953: WEECO contracts with U.S. Navy to produce perchlorate
• 1954: Shumaker Rocket Test Range Completed
• 1956: U.S. Navy announces plans to close Shumaker NAD
• 1959-1961: Shumaker property sold
• 1961: Shumaker ammunition depot closes
Additional recent timeline information follows (from U.S. Army Corps of
Engineers, 2003b)
• 1961: Brown Engineering formed, later named Highland Resources/Industrial
Park (HIP)
• Mid 1960s: Dozens of private businesses operated in HIP, many ordnance
related businesses
• 1987: 37 companies, 4700 employees at businesses located at HIP
• Early 1990s: downturn caused a number of defense contractors to relocate
from HIP
• 2003: 40 companies operate at HIP
In terms of early perchlorate production history, of particular interest is the time
from 1953 until 1961. This is when the U.S. Navy was known to have utilized
perchlorate, although perchlorate may not necessarily have been used in naval rockets at
Shumaker during this period. An indication of perchlorate not having been used is the
non-detect sampling results for groundwater and surface water from the rocket test range.
There were a few low-level perchlorate detections in surface water outside the rocket test
range near the eastern end of Shumaker, but these concentrations are very low (under 0.1
Hg/1) and are probably due to Chilean nitrate fertilizer.
The article by Schumacher indicates perchlorate manufacturing processes used
today are similar to those used in the past, because WEECO and its successor companies
were instrumental in modern plant design. Because of plant engineering and production
similarities, it is difficult to differentiate between multiple sources of aerospace industry
perchlorate. Sturchio (2009, personal communication), reported that the isotopic
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 44 ^ij^Environmenta! Protection
-------
signatures of synthetic (aerospace) perchlorate produced in the past would probably have
the same isotopic signatures of recently produced perchlorate, and different sources of
modern aerospace perchlorate are difficult to determine as well. At Shumaker, there are a
number of aerospace industry facilities besides the Aerojet facility. However, the types
of operations that take place at Aerojet, and which took place when the facility was
owned by Atlantic Research, are to develop and produce solid rocket motors,
automotive air bags, and perhaps other products that utilize ammonium
perchlorate. The other aerospace facilities are believed to function mainly as assemblers
of rocket and missile components, some utilizing perchlorate containing components
from Aerojet.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 45 ^ij^Environmenta! Protection
-------
7. KEY OBSERVATIONS AND UNCERTAINTIES
The following observations about Sparta vulnerability are presented under project
Goals 1-3. No single observation by itself is conclusive about vulnerability. All
observations should be considered collectively because they are all related in a complex
3-dimensional subsurface environment. Observations about the source of perchlorate at
Locust Bayou are presented under Goal 4.
GOAL 1: Evaluate the vulnerability of the Sparta aquifer with respect to: (i) the
presence of confining clay overlying the Sparta aquifer, and (ii) by making
observations about the occurrence of clay and sand within the Sparta aquifer.
Key Observations
Where the Cook Mountain Formation is relatively thick, it provides
substantial protection to the Sparta aquifer. Where the Cook Mountain
Formation is absent, the Sparta aquifer is closer to the surface and has more
exposure to possible contamination.
Clay within the Sparta aquifer provides protection from movement of
contamination to lower strata.
Where contamination exists in the gravel aquifer, contamination would
spread relatively quickly and affect upper sand units of the Sparta aquifer.
Where contamination exists in surface water, contamination would spread
relatively quickly and affect the gravel aquifer.
The two main geological factors affecting vulnerability are: (a) the
presence/absence of the Cook Mountain Formation, and (b) the amount of clay and
alternating sequences of clay and fine sand within the Sparta aquifer. Clay also exists in
the overlying Quaternary sediment, but is not as extensive or protective as thick clays of
the Cook Mountain Formation. The importance of the Cook Mountain Formation is that
it can be a protective clay layer preventing contamination from reaching the top of the
Sparta aquifer. Core sample examination and regional geologic studies indicate the Cook
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 46 ^ij^Environmenta! Protection
-------
Mountain Formation is thickest at the eastern end of Shumaker and thins to the west. The
Cook Mountain Formation was not identified in core samples taken near the industrial
areas west of well DW5.
Within the Sparta aquifer, clay and alternating sequences of clay and fine sand
provide protection limiting the migration of dissolved contamination (e.g., perchlorate)
downward through the strata. Sandy units positioned at the top of the Sparta aquifer are
much more vulnerable to contamination than deeper sands.
Uncertainties
• Sparta stratigraphy between monitoring wells must be extrapolated from what would
normally be expected in geologic deposits in a flu vial-deltaic system, and by using
observations made from cores taken from each well.
• The precise western limit of the Cook Mountain Formation is approximate. Core
samples indicate it probably pinches out between DW5 and DW6.
The Sparta geologic deposits at Shumaker appear highly variable, both laterally
and vertically, and strata cannot be correlated between EPA wells. While the geology
between wells is not exactly known, reasonable predictions can be made from an
understanding of sediments laid down in a fluvial deltaic depositional environment, and
an understanding of the site-specific variations in wells.
GOAL 2: Examine the vulnerability of the Sparta aquifer by: (i) installing permanent
groundwater monitoring wells, (ii) by monitoring for impacts to the aquifer,
and (in) by examining groundwater flow directions.
Key Observations
• Arsenic and lead were detected in the Sparta aquifer above Regional Screening
Levels. Perchlorate was not detected in Sparta monitoring wells.
• A downward groundwater flow gradient exists, but actual flow to strata lower
than the uppermost saturated sands may be only minor. The Shumaker area
may not be a significant recharge area for lower strata of the Sparta aquifer.
• No influences are seen on Sparta monitoring wells from high-capacity public
water supply production wells.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 47 ^ij^Environmenta! Protection
-------
Detections of arsenic and lead occurred near presently operating industrial
facilities and just north of highway 278 at Locust Bayou. No detections above screening
levels occurred in wells to the east where the Cook Mountain overlies the Sparta. The
origin of these metals is questionable and is discussed further in the uncertainties section
below.
Chemical evidence indicates downward flow may be only minor and that low
concentrations of perchlorate may not penetrate very deeply. Perchlorate was detected in
the gravel aquifer well SW5 at Locust Bayou (0.57 |ig/l), but was not detected in adjacent
clustered wells DW4 or the deeper DW4L. Well DW4 is screened about 35 feet below
SW5 and the amount of clay (72% percent core recovery) above the DW4 well screen is
probably a factor limiting downward migration. Alternatively, much higher
concentrations of perchlorate in shallow groundwater appear more likely to migrate
downward into deeper strata. Historical data at the Aerojet facility (Atlantic Research,
2001) shows perchlorate at Building 52 in alluvial groundwater exceeding 100,000 |^g/l.
At this location, perchlorate was detected just below the top of the Sparta aquifer 13.0
l^g/1. The ability of highly concentrated perchlorate (brine solution) to migrate deeper
may result from site specific geological factors, or chemical effects of highly
concentrated perchlorate on clay strata.
Uncertainties
• The natural occurrence of arsenic and lead in the Sparta aquifer at Shumaker is
unknown. Arsenic and lead may be due to either natural or industrial sources. More
data are needed to determine the source.
• Detections of bis(2-ethylhexyl)phthalate may be from well construction material or
laboratory contaminants.
• Results for explosives are problematic because of laboratory error. Confirmatory
sampling is needed.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 48 ^ij^Environmenta! Protection
-------
The presence of arsenic and lead may simply result from geologic deposits
characteristically containing these metals at elevated concentrations. Additional
investigation would be necessary to make that determination.
Laboratory results for explosives (2-nitrotoluene, 3-nitrotoluene, 4- nitrotoluene,
and nitrobenzene) are not conclusive. Two laboratory analyses were conducted for
explosives. The original analysis did not yield any detections, but had low recoveries
which resulted in estimated results. The re-extraction and re-analysis for explosives
yielded detections below the reporting limit, but were analyzed 1 day outside of the
holding time. Therefore, results are not reliable at the present time and additional
sampling should be conducted to verify results. Sampling could be conducted in only
those wells which had explosives detections, rather than re-sampling all wells at
Shumaker.
GOAL 3: Examine aquifer vulnerability by conducting surface water sampling including
streams, lakes, and ponds over the extent of Shumaker, and the Locust Bayou
area.
Key Observations
• There are many detections of perchlorate which appear to result from two types
of sources: nitrate fertilizer and synthetic perchlorate.
• Relatively high concentrations of perchlorate may be moving downstream in
surface water near Two Bayou and Dogwood Creek.
• Analytical results show no exceedences of RSLs. Arsenic, barium, cadmium,
chromium, lead, and selenium were detected, but were below RSLs.
Perchlorate was detected at many surface water sampling locations at
concentrations ranging from less than 1.0 |^g/l to 5.2 j^g/1. There appear to be two
populations of analytical data which can be roughly divided by fertilizer and synthetic
source type. Groundwater/surface interaction, including mixing with synthetic
perchlorate along parts of Two Bayou and Dogwood Creek, may affect shallow alluvial
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 49 ^ij^Environmenta! Protection
-------
groundwater. Data does not indicate, however, that surface water contamination
increases the vulnerability of deeper sections of the Sparta aquifer.
Uncertainties
• Low concentrations of perchlorate cannot be readily verified.
• Results for explosives are problematic because of laboratory error. Confirmatory
sampling is needed.
• No information is available on the possible contribution of metals to surface water
from soils.
There are two reasons why it is unlikely that the source of low concentration
perchlorate in surface water can be positively verified with current technology. First,
surface water contains interferences that affect isotopic analysis (Sturchio, personal
communication, 2009). Secondly, even if interferences did not exist, 10 mg of
perchlorate are required for analysis, which would be extremely difficult to obtain from
water containing only a fraction of one part per billion.
Surface water may be receiving some or all metals from surrounding soil
horizons. A soil sampling and analysis program, and/or geologic formation sampling,
would be needed to develop analytical data sets to fill this information gap. Surface
water sampling did not indicate sources of metals such as waste disposal units or types of
spills. Metals in surface water are fairly evenly distributed across Shumaker indicating
soil is a possible source.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 50 ^ij^Environmenta! Protection
-------
GOAL 4: Determine the source of perchlorate detected in groundwater monitoring wells
and drinking water at Locust Bayou.
Key Observations
• Isotopic analysis indicates perchlorate at Loucst Bayou and vicinity is from both
Chilean nitrate fertilizer and from synthetic perchlorate. The vast majority of
synthetic perchlorate is used by the aerospace industry.
• A mixture of synthetic and fertilizer based perchlorate was identified in well
SW7, located upgradient of residences near county road 95.
• Synthetic perchlorate is probably being transported through the gravel aquifer
from an upgradient source to the north, and possibly through surface water.
• Relatively high concentrations of perchlorate have been present in the alluvial
aquifer at the Aerojet facility north of Locust Bayou.
There are two distinctive types of perchlorate in groundwater at Locust Bayou and
vicinity as determined from the analysis of chlorine and oxygen isotopes. These types
may be mixed, or may occur separately, as was noted in one sample. The transport
pathway is probably mainly through the gravel aquifer, with additional contamination
from surface water transport. Perchlorate in surface water may infiltrate the gravel
aquifer.
Uncertainties
• It cannot be completely ruled out that more than one aerospace industry source, or
past Navy operations, contributed to perchlorate contamination.
• Seasonal changes in groundwater flow directions may exist.
• Relationship between southward flow direction in gravel aquifer, to west/south-west
flow in shallower zones (above gravel) at Aerojet.
Even with the current state-of-the-art science, isotopic analysis still cannot
differentiate between specific synthetic perchlorate aerospace sources. There are
multiple aerospace industry facilities at Shumaker, but only one, Aerojet, is known to be
associated with significant perchlorate releases that occurred during previous operations
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 51 ^ij^Environmenta! Protection
-------
by Atlantic Research. While it cannot be ruled out that past Naval rockets contained
perchlorate, it was not detected in the former rocket test range.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 52 ^Erwironmenta! Protection
-------
8. CONCLUSIONS
The vulnerability of the Sparta aquifer to contamination across Shumaker varies
by location and depth. Sparta vulnerability is greatest in the western half of Shumaker
where Sparta sand contacts the gravel aquifer. Sparta vulnerability decreases with depth
because of clay within the Sparta and possible limitations to the downward movement of
groundwater. The least likely part of the Sparta to become contaminated from surface
use is the eastern half of Shumaker where the Sparta is covered by a thick layer of clay.
At this time, the Sparta aquifer has not been widely affected by perchlorate. The
shallow gravel aquifer is more susceptible to contamination than the Sparta and has been
widely affected by perchlorate. Periodic monitoring of the Sparta aquifer should be
conducted as a measure of safety, and routine monitoring of the gravel aquifer should be
conducted to track perchlorate levels and seasonal flow directions.
Groundwater and surface water in the vicinity of Locust Bayou has been affected
by synthetic perchlorate from an aerospace industry source, and by perchlorate derived
from past agricultural uses of imported Chilean nitrate fertilizer products. A mixture of
synthetic and fertilizer based perchlorate was identified in groundwater just upgradient of
a residential area. The most well known source of synthetic perchlorate in the area is the
past release of perchlorate that occurred at Atlantic Research/Aerojet, which has been
transported through the gravel aquifer, and through surface water. Other sources of
synthetic perchlorate may exist, but none appear as likely based on sampling data, flow
directions, locations of known releases and existing contamination, and the history of
perchlorate production and use.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 53 ^ij^Environmenta! Protection
-------
9. REFERENCES
Albin, D.R., 1964, Geology and Ground-Water Resources of Bradley, Calhoun, and
Ouachita Counties, Arkansas, U.S. Geological Survey Water Supply Paper
1779-G.
Arkansas Public Service Commission, 2008, Class A and B Water Utility Annual Report.
Atlantic Research Corporation, 2001, Focused RCRA Facility Investigation Report, Open
Burn Unit and Building 52 Areas, Prepared by Environmental Alliance, Inc.
ATSDR, 2007, Health Consultation, Sparta Aquifer Vulnerability Study, Health
Implications of Heavy Metals Contamination in Locust Bayou Private Wells,
Calhoun County, Arkansas.
Bohlke, J.K., Sturchio, N.C., Gu, B., Horita, J., Brown, G.M., Jackson, W.A., Batista, J.,
and Hatzinger, P.B., 2005, Perchlorate Isotope Forensics, in Analytical
Chemistry, vol. 77, p 7838-7842.
EPA 2007a, Site Specific Quality Assurance Project Plan/Sampling Analysis Work Plan
for "A Phased Assessment of Groundwater and Subsurface Conditions Involving
Protection of the Sparta Aquifer," Shumaker Naval Ammunition Depot, East
Camden, Arkansas, prepared by Tech Law Inc., January 31, 2007.
EPA 2007b, Site Specific Quality Assurance Project Plan/Sampling Analysis Work Plan
for "A Phased Assessment of Groundwater, Surface Water, and Isotopic Analyses
in Support of the Shumaker Naval Ammunition Depot Vulnerability Study,"
Shumaker Naval Ammunition Depot, East Camden, Arkansas, prepared by Tech
Law Inc., December, 18, 2007.
EPA 2008a, Safe Drinking Water Information System (SDWIS) report.
EPA 2008b, Interim Drinking Water Health Advisory for Perchlorate, Prepared by:
Health and Ecological Criteria Division, Office of Science and Technology, EPA
Office of Water.
Fitzpatrick, D.J., Kilpatrick, J.M., McWreath, H., 1990, Geohydrologic Characteristics
and Simulated Response to Pumping Stresses in the Sparta Aquifer in East-
Central Arkansas, U.S.G.S. Water Resources Investigations Report 88-4201.
Hosman, R.L. and Weiss, J.S., 1988, Geohydrologic Units of the Mississippi Embayment
and Texas Coastal Uplands Aquifer Systems, South-Central United States, USGS
Open File Report 88-316.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 54 ^ij^Environmenta! Protection
-------
Joseph, R.L., 2000, Status of Water Levels and Selected Water Quality Conditions in the
Sparta and Memphis Aquifers in Eastern and South-Central Arkansas, 1999,
USGS Water-Resources Investigations Report 00-4009.
Massachusetts Department of Environmental Protection, 2006, Perchlorate in
Hypochlorite Products Report.
Motzer, W.E., Mohr, T.K.G., McCraven, S., and Stanin, P., 2006, Stable and Other
Isotope Techniques for Perchlorate Source Identification, in Environmental
Forensics, 7:89-100.
Payne, J.N., 1968, Hydrologic Significance of the Lithofacies of the Sparta Sand in
Arkansas, Louisiana, Mississippi, and Texas, U.S. Geological Survey Professional
Paper 569-A.
Schumacher, J.C., 1999, History of Establishing a Source of Potassium and Ammonium
Perchlorates for use in Solid Propellant Rockets, American Institute of
Aeronautics and Astronautics, Paper 99-2926.
Sturchio, N.C., Bohlke, J.K, Gu, B., Horita, J., Brown, G.M., Beloso, A.D., Patterson,
L.J., Hatzinger, P.B., Jackson, A.J., Batista, J., 2006, Stable Isotopic Composition
of Chlorine and Oxygen in Synthetic and Natural Perchlorate, in Perchlorate:
Environmental Occurrence, Interactions, and Treatment, B. Gu and J.D. Coates
Editors, Springer Publishers, New York.
Tech Law Inc., 2007, Field Activity Trip Report, Shumaker Naval Ammunition Depot
Facility, East Camden, Arkansas.
Tech Law Inc., 2008, Task Order Report, Shumaker Naval Ammunition Depot Facility,
East Camden, Arkansas.
U.S. Army Corps of Engineers, Little Rock District, 2003a, Fact Sheet, Former Shumaker
Naval Ammunition Depot, Arkansas.
U.S. Army Corps of Engineers, Little Rock District, 2003b, Final Decision Document,
Former Shumaker Naval Ammunition Depot, East Camden, Arkansas.
Sparta Aquifer f\ mA Uni!ed States
Vulnerability Assessment Page 55 ^ij^Environmenta! Protection
------- |