United States
     Environmental Protection
July 2008
    Office of Water (4607M) EPA 815-R-07-023 July 2008 www.epa.gov/safewater



       The purpose of this guidance manual is solely to provide technical information for water
systems and States to assist them in complying with the Ground Water Rule (GWR). The
statutory provisions and EPA regulations described in this document contain legally binding
requirements. This guidance is not a substitute for applicable legal requirements, nor is it a
regulation itself. Thus, it does not impose legally-binding requirements on any party, including
EPA, States, or the regulated community. While EPA has made every effort to ensure the
accuracy of the discussion in this guidance, the obligations of the regulated community are
determined by statutes, regulations, or other legally binding requirements. In the event of a
conflict between the discussion in this document and any statute or regulation, this document
would not be controlling.

       Interested parties are free to raise questions and objections to the guidance and the
appropriateness of using it in a particular situation.

       Although this manual describes suggestions for complying with GWR requirements, the
guidance presented here may not be appropriate  for all situations, and alternative approaches
may provide satisfactory performance. The mention of trade names or commercial products does
not constitute endorsement or recommendation for use.


       This manual was developed under the direction of EPA's Office of Ground Water and
Drinking Water and was prepared by EPA  and the Cadmus Group, Inc. Questions concerning
this document should be addressed to:

Michael Finn
U.S.EPA Office of Ground Water and Drinking Water
Standards  and Risk Management Division
1200 Pennsylvania Ave, N.W. 4607M
Washington D.C. 20460-0001


American  Water Works Association
Association of Metropolitan Water Agencies
Association of State Drinking Water Administrators
Michael Focazio-WRD, US Geological Survey (Peer reviewer)
Kate Miller-Public Water Supply and Subdivision Bureau, Montana Department of
Environmental Quality (Peer reviewer)
Steven J. Roy, P.G.-Drinking Water & Groundwater Bureau, New Hampshire Department of
Environmental Services (Peer reviewer)
Source Assessment Guidance Manual


1.  Introduction	1-1
    1.1 Background	1-1
    1.2 Purpose and Scope	1-2
    1.3 Ground Water Rule Summary	1-3
    1.4 Public Health Risk Factors	1-6
    1.5 Hydrogeologic Data Sources for Assessment Source Water Monitoring Decisions	1-15
       1.5.1  State and Federal Hydrogeologic Investigations	1-15
       Wellhead Protection and Source Water Assessment Studies	1-15
       State Geologic Survey, USGS, and Other Hydrogeologic Investigations	1-16
       1.5.2  Hydrogeologic and Geologic Maps	1-16
       1.5.3  Soil Maps	1-18
       1.5.4  Topographic Data	1-19
       1.5.5  Stereoscopic Aerial Photography	1-21
       1.5.6  Other Data Sources for Desktop Analyses	1-22
2.  Sensitive Hydrogeologic Environments	2-1
    2.1 Aquifer Sensitivity	2-1
    2.2 Karst Aquifers	2-2
    2.3 Fractured Bedrock Aquifers	2-6
    2.4 Gravel Aquifers	2-7
    2.5 Hydrogeologic Barriers	2-8
3.  Hydrogeologic Sensitivity Assessments	3-1
    3.1 Identifying Aquifer Types	3-2
    3.2 Karst Regions and Aquifers of the United States	3-2
       3.2.1  Diagnostic Characteristics	3-3
       3.2.2  Desktop Approaches	3-5
    3.3 Fractured Bedrock Regions and Aquifers	3-7
       3.3.1  Diagnostic Characteristics	3-8
       3.3.2  Desktop Approaches	3-9
    3.4 Gravel Aquifer Hydrogeologic Settings	3-10
       3.4.1  Diagnostic Characteristics	3-11
       3.4.2  Desktop Approaches	3-11
    3.5 Hydrogeologic Barriers	3-12
       3.5.1  Data Sources for Hydrogeologic Determinations	3-14
       3.5.2  Desktop Approaches	3-15
4.  Assessment Source Water Monitoring; Number (and Frequency) of Samples	4-1
    4.1 Introduction	4-1
    4.2 Connection to  Hydrogeologic Sensitivity Assessment	4-1
    4.3 Assessment Monitoring Basis and Triggers	4-2
    4.4 Assessment Monitoring; Number (and Frequency) of Samples	4-2
    4.5 Sample Location	4-4
    4.6 Representative Wells	4-5
    4.7 Indicator Selection	4-5
    4.8 Analytical Methods	4-6
5.  References	5-1
Source Assessment Guidance Manual

Appendix A:  Field Methods for Determining the Presence of a Hydrogeologic Barrier

Appendix B:  Ground Water Travel Time

Appendix C:  Microbial Inactivation Rates

Appendix D:  Additional Reference Sources
Source Assessment Guidance Manual              in


Exhibit 1.1   Summary of GWR Requirements	1-4
Exhibit 1.2   Summary of Risk Factors for Targeting Susceptible Systems for Assessment
             Source Water Monitoring	1-7
Exhibit 1.3   Outbreak Examples by Risk Scenario	1-8
Exhibit 1.4   Strike and Dip	1-17
Exhibit 1.5   The Importance of Map Scale for Determining Aquifer Type	1-20

Exhibit 2.1   Example Sensitive Aquifers	2-2
Exhibit 2.2   Examples of Hydrogeologic Barriers and their Properties	2-2
Exhibit 2.3   Map of Sinkholes (closed circular contours with tick marks) in
             Orleans, Indiana	2-4
Exhibit 2.4   Aquifers of the United States	2-13

Exhibit 4.1   Likelihood of Identifying E. coli Occurrence by Assessment Source Water
             Monitoring in a Population of Wells Randomly Selected and Sampled
             (Data from USEPA 2006d and USEPA 2006e)	4-4
Exhibit 4.2   E. coli Methods Approved for Use under the GWR	4-6
Exhibit 4.3   Enterococci Methods Approved for Use under the GWR	4-7
Exhibit 4.4   Coliphage Methods Approved for Use under the GWR	4-7
Source Assessment Guidance Manual

                                List of Acronyms


American Society for Testing and Materials
Composite Analytical-Numerical Model for Viral and Solute Transport
Community Water System
Centers for Disease Control and Prevention (U.S. Department of Health
and Human Services)
Digital Elevation Model
Department of Environmental Quality
Digital Line Graph
Defense Mapping Agency
United States Environmental Protection Agency
Earth Resources Observation Systems
Earth Science Information Centers
Finite Element Contaminant Transport - Unsaturated Zone
Geographic Information System
Global Positioning System
General Particle Tracking
Ground Water Rule
Ground Water System
Ground Water Travel Time
Ground Water Under the Direct Influence of Surface Water
Hepatitis A Virus
Hydrogeologic Sensitivity Assessment
Idaho Department of Environmental Quality
State of Illinois Environmental Protection Agency
Kentucky Department of Environmental Protection
Methylene blue active substances
Methyl tertiary butyl ether
Map Unit Interpretations Record
Multiple Well Capture Zone
National Aerial Photography Program
Non-Community Water System
National Cooperative Geologic Mapping Program
New Hampshire Department of Environmental Services
National Research Council
Natural Resources Conservation Service
National Soil Survey Center
Public Water System
Regional Aquifer-System Analysis
Source Assessment Report
Soil Conservation Service
Safe Drinking Water Act
Soil Survey Geographic (SSURGO) Data Base
State Soil Geographic (STATSGO) Data Base
Source Water Assessment Program
Total Coliform Rule
Source Assessment Guidance Manual

USDA              United States Department of Agriculture
USGS               United States Geological Survey
VIRALT            Virus Analytical Transport
WHAEM            Wellhead Analytical Element Model
WHPA              Wellhead Protection Area
WHPP              Wellhead Protection Program
WIDNR             Wisconsin Department of Natural Resources
Source Assessment Guidance Manual

                                    1. Introduction
1.1    Background

       The 1996 Amendments to the Safe Drinking Water Act (SDWA) required the United
States Environmental Protection Agency (EPA) to develop national primary drinking water
standards requiring disinfection as a treatment technique for all public water systems, including
surface water systems, and, as necessary, ground water systems. EPA promulgated the final
Ground Water Rule (GWR) on November 8, 2006 as one step in addressing these requirements.
(USEPA, 2006a)

       The GWR establishes a risk-targeted approach to identify Ground Water Systems
(GWSs) susceptible to fecal contamination and requires corrective action to correct significant
deficiencies and source water fecal contamination in public GWSs. A central objective of the
GWR is to identify the subset of ground water sources that are at higher risk of fecal
contamination among the large number of existing GWSs (approximately  147,000), and then
further target those systems that must take corrective action to protect public health. This risk-
targeted approach includes the following:

       •   Periodic sanitary surveys of GWSs requiring the evaluation of eight critical elements
          and the identification of significant deficiencies;

       •   Triggered source water monitoring of systems that do not achieve 4-log inactivation
          or removal of viruses;

       •   Corrective actions to eliminate significant deficiencies and fecal contamination;  and

       •   Compliance monitoring to ensure that disinfection treatment for drinking water is
          reliably operated where it is used and achieves a 4-log inactivation or removal of

       The risk-targeted approach also includes assessment source water monitoring and
hydrogeologic sensitivity assessments (HSAs) as tools that States may use  at their discretion to
evaluate ground water sources that may be at risk for fecal contamination. Assessment source
water monitoring involves the collection of source water samples at regular intervals from
ground water sources or wells located in  sensitive aquifers, or from wells that are vulnerable to
contamination due to other factors determined by the State, and analysis of those samples for
fecal indicators. The GWR specifies E. coli, enterococci, or coliphage (male-specific or somatic)
as suitable fecal indicator organisms. HSAs, described later in this guidance, are determinations
of whether GWSs obtain water from hydrogeologically sensitive settings.

       Ground water-supplied public water systems (PWSs) are at greater risk of causing
waterborne disease if the source ground water is  fecally-contaminated and if the finished water
does not receive 4-log treatment of viruses (i.e., a 99.99% reduction). The Sanitary Survey
Guidance for Ground Water Systems document (USEPA, 2006b) provides  additional
information regarding the 4-log inactivation requirements.

Source Assessment Guidance Manual               1-1

       Public water supplies may transmit fecal contamination if their sources are subject to one
or more of the following risk factors:  1) sensitive aquifers; 2) aquifers in which viruses may
travel faster and farther than bacteria (e.g., alluvial or coastal plain sand aquifers; 3) shallow
unconfined aquifers; 4) aquifers with thin or absent soil cover; 5) wells previously identified as
having been fecally-contaminated; and 6) high population density combined with on-site
wastewater treatment systems, particularly those in aquifers with restricted geographic extent,
such as barrier island sand aquifers. Other risk factors also may allow or facilitate fecal
contamination transmission to PWS wells. Assessment source water monitoring should be
conducted at PWS wells if any risk factor is identified.
1.2    Purpose and Scope

       The purposes of this guidance document are to describe: 1) scenarios when assessment
source water monitoring might be advantageous in protecting public health; 2) sensitive aquifers;
3) data sources and methods suitable for use in a HSA and 4) implementation of assessment
source water monitoring.

        This document does not rank or prioritize the various risk factors that govern source
water fecal contamination likelihood. Each of the risk factors listed above,  as well as other
factors, may be important nationally or in certain locales. This document emphasizes ways to
identify readily available information suitable for office, rather than field, determination of risk
at an individual PWS well.

       The main objective of this guidance document is to identify the most significant risk
factors, useful information sources, and simple analytical approaches that may aid State technical
staff when considering the need and locations for assessment source water  monitoring.  Pathogen
survival in the water environment and transport through ground water are discussed with
emphasis on localities where there is increased likelihood that infectious pathogens may arrive at
a PWS well in sufficient number to cause illness. States may elect to perform these assessments
based on data availability and existing knowledge about fecal contamination risk in each State.

       This guidance emphasizes desktop analytical approaches and associated data sources
because a  PWS well's aquifer type can be determined without field investigation in most cases.
Furthermore, EPA is encouraging States to build upon their source water assessment efforts (i.e.,
through their Wellhead Protection Programs (WHPPs) and Source Water Assessment Programs
(SWAPs)) if they choose to conduct HSAs, and to coordinate efforts among these programs
whenever  possible.

       This document does not address issues related to well construction or sanitary setback
distance encroachment. Proper well construction and encroachments into any sanitary setback
exclusion  zones for existing wells are addressed in the Sanitary Survey Guidance Manual for
Ground Water Systems (USEPA, 2006b). Although proper well construction is an important
barrier to contamination for wells situated in all aquifer types, proper well construction does not
preclude contamination of water drawn  into wells.  Conversely, poor well construction may put
wells  at risk of contamination regardless of the aquifer type in which the well is  situated. EPA
has included information on appropriate well construction and survey of land use around the well
Source Assessment Guidance Manual              1 -2

head in the Sanitary Survey Guidance Manual for Ground Water Systems (USEPA 2006b).
Although EPA recognizes that land use outside any sanitary setback boundary is often not under
the control of the utility, States should consider land use in making decisions about the need for
assessment source water monitoring.

       This document does not address ground water sources that are under the direct influence
of surface water (GWUDI). The GWR is not applicable to GWUDI sources. Sources that have
been identified as GWUDI are subject to the requirements for surface water supplies (CFR 40,
Part 141, Subparts H, P, T and W). If assessment source water monitoring or other
investigations,  as part of GWR implementation,  identify GWUDI sources that have not been
previously identified, then evaluations and regulatory determinations based on surface water
supplies should be implemented.
  As used in this guidance, "State" refers to the agency of the State or Tribal government
  that has jurisdiction over public water systems. During any period when a State or
  Tribal government does not have primacy enforcement responsibility pursuant to
  section 1413 of the Safe Drinking Water Act, the term "State" means the Regional
  Administrator, US Environmental Protection Agency.
1.3    Ground Water Rule Summary

       The GWR applies to all PWSs that use ground water, except PWSs that combine all of
their ground water with surface water or with GWUDI prior to treatment. The GWR also applies
to consecutive systems receiving finished ground water. Ground water systems (GWSs) must
comply with the GWR beginning December 1, 2009. Key components of the GWR are:

       1.  Sanitary surveys,

       2.  Triggered source water monitoring,

       3.  Corrective actions, and

       4.  Compliance monitoring.

       Each of these components is discussed further below and Exhibit 1.1 provides a summary
flowchart of the final GWR requirements.
Source Assessment Guidance Manual               1 -3

                                    Exhibit 1.1  Summary of GWR  Requirements
                 All GWSs"
        Initial and periodic sanitary surveys performed by the State
       •  Community water systems (CWSs): every 3-5 years
       i  Non-community water systems (NCWSs): every 5 years
            Conduct routine sampling
          under the Total Coliform Rule
                  Was TCR
                 sample total
                              Did the
                         State identify any
                                                        Consult State within 30 days of notification regarding
                                                            appropriate corrective action, if necessary
                                                 Does the
                                                State require
                                                source water
                                                     Implement State-approved or -specified corrective actions.
                                                                       Options include:

                                                    • Eliminate source of contamination
                                                    • Correct significant deficiency
                                                    • Provide an alternate water source
                                                    • Provide treatment to achieve 4-log reduction of viruses
                     Complete correction or be in
                      accordance with a State-
                    approved schedule within 120
                    days of notice of fecal indicator
                    positive or significant deficiency
Conduct triggered
source water

                                                                  Continue GWR compliance:   \
                                                                   sanitary surveys, triggered
                                                               j monitoring, TCR compliance, and (
                                                                    assessment monitoring
Perform public notification
and consult State within 24
                    Compliance monitoring - options include:
      Per State
direction, take corrective
      action or
      5 additional
                                                                               Perform public notification
                                                                               and consult State within 24
                                                                                    Consult State
                                                                                    within 30 days
                                                                                 regarding appropriate
                                                                                  corrective action, if
                                          Perform public notification
                                          and consult State within 24
Were any of the 5
 repeat samples
. '

Monitor the
process in
with State-



daily via grab
sample at
peak flow


Monitor the
process in
with State-
                                                                         (1) The GWR applies to all public water systems (PWSs) that use ground
                                                                            water, except public water systems that combine all of their ground water
                                                                            with surface water or with ground water under the direct influence of surface
                                                                            water prior to treatment.

                                                                         (2) Treatment using inactivation, removal, or State-approved combination to
                                                                            achieve a 4-log reduction of viruses before or at the first customer. Compliance
                                                                            monitoring required.

                                                                         (3) If the State determines that the distribution system is deficient or causes total
                                                                            coliform-positive samples, the system may be exempted from triggered source
                                                                            water monitoring.

                                                                         (4) The State must provide the GWS with written notice describing any significant
                                                                            deficiencies within 30 days of identifying the significant deficiency.
Source Assessment Guidance Manual

Sanitary Surveys

       The final GWR requires regular (every three years for community water systems (CWSs)
and every five years for non-community water systems (NCWSs)) comprehensive sanitary
surveys of 8 critical components: (1) source; (2) treatment; (3) distribution system; (4) finished
water storage; (5) pumps, pump facilities, and controls; (6) monitoring and reporting, and data
verification; (7) system management and operation; and (8) operator compliance with State
requirements.  If a significant deficiency is identified, corrective action is required or a
treatment technique violation is incurred.

Source Water Monitoring

       In the final GWR, systems not achieving, or not performing compliance monitoring for,
4-log treatment of viruses (using inactivation, removal, or a State-approved combination of these
technologies) must conduct triggered source water monitoring for the presence of at least one of
the following fecal indicators: E. coli, enterococci, or somatic coliphage.  The triggered
monitoring requirements apply to systems that are notified that a Total Coliform Rule (TCR)
routine sample is total coliform-positive. Within 24 hours of receiving the total coliform-
positive notice, GWSs must collect a source water sample and test it for the presence of a fecal

       If the State does not require corrective action (see Corrective Action section below) for
an initial fecal indicator-positive source water sample, the system must collect five additional
source water samples within 24 hours of being notified of the initial fecal indicator-positive
source water sample. The GWR requires systems to take corrective action if any of the five
additional source water samples are fecal-indicator positive.

       The GWR provides  States with the option to require systems to conduct assessment
source water monitoring for ground water sources as needed. States may specify the number of
samples, the duration of sampling and the analytical method used for analysis. States may use
HSAs as a tool to determine if GWSs are obtaining water from hydrogeologically sensitive
settings. The GWR requires GWSs to provide the State with any existing information that will
enable the State to perform  the HSA. States also have the option to require corrective action for
any fecal indicator positive  sample found during assessment source water monitoring.

Corrective Action

       The GWR requires that systems implement corrective action for;

       1.  Significant deficiencies,

       2.  Fecal-indicator positive samples, if directed by the  State, after the initial fecal
          indicator-positive in triggered monitoring, or for a fecal indicator-positive found
          during assessment monitoring, or

       3.  A fecal indicator-positive sample in any of the five additional source water samples
          collected after the initial fecal indicator-positive source water sample during triggered
Source Assessment Guidance Manual              1 -5

       The system must implement at least one of the following corrective actions: correct all
significant deficiencies; provide an alternate source of water; eliminate the source of
contamination; or provide treatment that reliably achieves at least 4-1 og treatment of viruses.
Furthermore, the system is required to notify the public served by the water system of any
uncorrected significant deficiencies and/or source water contamination. (The State may also
require notification of corrected significant deficiencies.)

Compliance Monitoring

       Compliance monitoring requirements are the final defense against microbial
contaminants provided by the final GWR. All GWSs that provide 4-log treatment of viruses,
either as a corrective action or in lieu of GWR triggered source water monitoring, must conduct
compliance monitoring to demonstrate continual treatment effectiveness.
1.4    Public Health Risk Factors

       In the GWR preamble, EPA identified several risk factors that may indicate the need for
assessment source water monitoring. A PWS well may be at risk if there is an increased
likelihood for pathogenic bacteria or viruses to arrive at the well in an infectious state. Because
most pathogens are not native to ground water, they are unable to reproduce in the ground water,
and their survival is limited. As the subsurface residence time increases, the proportion of
infectious pathogens (with that residence time) decreases. Thus, subsurface residence time is
often used as a surrogate measure of pathogen risk. An important corollary is that the larger the
number of introduced pathogens, the greater the likelihood that some pathogens will  remain
infectious at any particular residence time. In general, at shallow ground water temperatures,
viral pathogens likely remain infectious for a significantly longer time as compared with the
bacterial pathogens (Appendix C).

       To illustrate this concept, consider a visitor to a highway rest area served by a septic
tank. The visitor is recently ill and is shedding pathogens in his stool, which are flushed into a
poorly sited and performing septic tank. The risk that others will become ill  from drinking water
from a nearby PWS well that produces water from a shallow unconfmed aquifer and  whose
water is not receiving 4-log treatment of viruses depends  on the time that the pathogens spend in
the subsurface. If the average ground water travel time from the aquifer fed by the septic tank is
10 years, then it is likely that all pathogens will be inactivated during their residence  in the
subsurface, and it is likely that the PWS well is not at risk. If the average ground water travel
time is 2 years, then some ground water will take a fast path and arrive in 1 year or less, and
other ground  water will take  a slower path and arrive in 3 years or more. Because pathogens
remain infectious in the subsurface for a maximum of about one year (See Appendix  C), the
health risk depends on the proportion of ground water that arrives most  rapidly at the well. That
risk increases if the number of pathogens introduced into  the groundwater is large rather than
small  because there is a greater likelihood that the fast arriving ground water will entrain
pathogens if there are more pathogens entering the ground water. In any case, the number of
infectious pathogens must be sufficiently large (an infectious dose) so as to overcome the innate
defense systems in the human host. Viruses have significantly lower infectious doses than do
bacteria (McBride et al., 2002).
Source Assessment Guidance Manual              1-6

    Exhibit 1.2 Summary of Risk Factors for Targeting Susceptible Systems for
                         Assessment Source Water Monitoring
Risk Factor
Example sources
of Information
Sensitive Aquifers

Karst, fractured
bedrock, or gravel
Aquifers in which viruses may  Alluvial or coastal
travel faster and further than    plain sand aquifers
Shallow unconfined aquifers    Any
Aquifers with thin or absent     Any
soil cover
Wells previously identified as
having been fecal ly-
High population density
combined with on-site
wastewater treatment system
Other Risk Factors1
Barrier island sand
£. co//, Enterococci, or

£. co//, Enterococci, or

Based on historical

                      £. co//, Enterococci,
See section 1.5 and
Chapter 2

Research literature.
See also shallow
unconfined aquifers
Well logs, well
construction reports,
or well permits
Soil maps. See
section 1.5.3
Sanitary survey

Sanitary survey
                      Sanitary survey,
                      Source Water
                      Assessment, and
                      field visit records

 Including but not limited to: well near a source of fecal contamination; well in a flood zone; improperly constructed
well (e.g., improper surface or subsurface seal); well of unknown construction (e.g., no driller's log or other record of
construction); other non-microbial indicators of potential for fecal contamination (e.g., Methylene blue active
substances (MBAS), high chloride or nitrate levels from baseline or historic trends).
Source Assessment Guidance Manual

                    Exhibit 1.3 Outbreak Examples by Risk Scenario
   Risk Factor
                             Example Outbreak
Aquifers in
which viruses
may travel faster
and farther than
Aquifers with
thin or absent
soil cover
Wells previously
identified as
having been

High population
combined with
A norovirus outbreak in 2001 was associated with visitors to a snowmobile lodge in
Wyoming (Anderson, et al, 2003). A detailed investigation by Gelting et al (2005)
concluded that the bedrock underlying the shallow (coarse-textured) soils at the lodge
consisted of fractured granite. When the wastewater reached the fractured granite
bedrock, it traveled within the fractures which served as conduits to the PWS wells.

Passengers traveling by bus through Alaska and the Yukon Territory became ill after
consuming water at a restaurant in the Yukon Territory (Beller, et al, 1997). Water
was supplied to the restaurant by two shallow wells. Wastewater was piped to a
septic pit located about 15 m from one well. Well #1 had total and fecal coliform
counts of 10-50 per 100 ml and  2-18 per 100 ml. Norovirus was identified in the well
water from well #1 and matched to the norovirus recovered from ill individuals. Dye,
introduced into the septic pit, arrived at well #1 in about 24 hours.
A norovirus outbreak is associated with well water at a new resort in Arizona
(Lawson, 1991). Although the latest technology was used to design the resort's water
and sewage treatment plants, well water was sufficiently contaminated so as to cause
illness in about 900 individuals. The waste water passed through about 10 m of sandy
alluvium, 70 m of sandstone and contaminated the underlying aquifer. The aquifer is
unconfined and is shallow compared with other aquifers in the arid western States.

At an island in northern Michigan, 39 people became ill from drinking tap water at a
resort (Ground Water Education in Michigan, 1992; Chippewa County Health
Department, unpublished report, 1992). The septic tank was believed to be the
contamination source. Dye introduced into the septic tank was found in the PWS after
two days. As described in the outbreak report, much of the island is covered with a
thin layer of soil, sometimes insufficient for proper filtration of surface water and
sewage effluent.

An £. co// O157:H7 outbreak in Walkerton, Ontario resulted in six deaths and 2300
illnesses (Hrudy and Hrudy, 2004) in 2000. The source water from the wells and fror
2 to 6 locations in the distribution system were typically sampled weekly. The
available data are summarized in a report by B.M. Ross and Associates Ltd. (2000).
In a 1992 report, 3% of 125 samples showed adverse results. In contrast,  no adverse
results were reported in 1989 and 1991. In 1996, 3% of source water samples and
6% of distribution system samples were adverse. In 1998,16 adverse samples were
identified, and £. co//were identified in low numbers (one to four) in treated water and
distribution system samples. In 1999, total coliform bacteria were identified in 7 of 151
source water samples and in 3 of 146 distributions system samples. No samples
were positive for £. co//. In early 2000, prior to the contamination event and outbreak,
total coliform bacteria were identified in source water, treated water and the
distribution system samples.
Tourists visiting a resort island in Ohio (and drinking water from PWS wells rather
than the centralized water treatment plant) became ill from Campylobacter,
Arcobacter, Salmonella and perhaps other pathogens (O'Reilly et al, 2007; Fong et al,
in press). The island is served by a centralized water treatment plant but 13 PWS
wells and numerous domestic wells were also in use.  The island hosts about 25,000
visitors per day during the tourist season. O'Reilly et al (2007) conclude that there
was widespread contamination of the aquifer because many wells from various
locations on the island  showed evidence of contamination.
Source Assessment Guidance Manual

       The following text identifies risk scenarios that may merit additional assessment source
water monitoring. These risk scenarios are identified based on the lessons learned from past
waterborne disease outbreaks. One common characteristic of the risk scenarios is that all are
based on a short ground water residence time (months rather than years) either because the
ground water flow is naturally fast or because well pumping combined with short ground water
flow paths precludes longer residence times. For example, when fecal indicators are found in
PWS well water, that finding is evidence of a short ground water residence time because fecal
indicators probably cannot survive in ground water longer than one year outside of a mammalian
host (see Appendix C). A short ground water residence time can include both the time to
infiltrate from the surface to the aquifer as well as the time exclusively within the aquifer. The
purpose of this document is to emphasize issues related to the time within the aquifer as opposed
to the infiltration time. Short infiltration times are a feature of GWUDI of surface water wells
and are considered in detail in the Surface Water Treatment Rule guidance.

       Sensitive aquifers - A sensitive aquifer is herein defined as any karst, fractured bedrock,
or gravel aquifer. In these aquifers, ground water flow velocities are typically very high, and
flow typically takes the shortest and most direct path. This rapid transport allows fecal
contaminants to  travel without significant reduction in numbers through inactivation or removal.
 Furthermore,  ground water velocities in these aquifers are even higher in the vicinity of a
pumping well  than they are under natural flow conditions, increasing the potential for pathogens
to migrate to a well.

       In addition to karst, fractured bedrock, and gravel aquifers, States may designate
additional aquifer settings (e.g., sand and gravel aquifers) as sensitive and require assessment
source water monitoring in them if they believe this designation is necessary to protect public
health. Sensitive aquifers are discussed in more detail in  Chapter 2, including specific methods
for identifying sensitive aquifers and case study examples of the process.

       As mentioned above, fractured bedrock is defined in this document  as a sensitive aquifer.
 Granite is recognized as an example of fractured bedrock.  Wells in granitic and/or similar rocks
are able to produce water because there are fractures present and should be considered for
assessment source water monitoring to protect public health. This is because unfractured granite
is among the most dense and compact rock types and has insufficient connected void space to
provide sustained water yield to a well. Thus, to serve as an aquifer, granite must have
numerous large,  open (unmineralized) and connected fractures which can serve as conduits for
contamination. Section 1.5 identifies the hydrogeologic  data sources suitable for identifying
sensitive aquifers.

       Because  ground water flow is fast and direct through the relatively large voids within
these aquifers, there are few opportunities for entrained fecal indicator microorganisms to
interact with the solid aquifer materials. Organism, size,  charge or other factors governing
transport in ground water are not likely to be significant if there is little interaction with the
aquifer solid material. Thus, these organisms will most likely pass through the subsurface at or
near the average ground water velocity. All three indicator microorganisms will likely pass
through sensitive aquifers at similar rates. Based on source water risk factors, there is no reason
to favor one organism over another in selecting a recommended fecal indicator organism.
Source Assessment Guidance Manual               1 -9

       Aquifers in which viruses may travel faster and farther than bacteria - Aquifers are
broadly classified into two categories; porous and non-porous media. A sand aquifer is an
example of a porous media aquifer. In a sand or other porous media aquifer, the ground water
takes a relatively slow and indirect path around the myriad sand grains, thereby providing ample
opportunity for bacteria or viruses, entrained with the water, to come in contact with a sand
grain. In a non-porous media aquifer (e.g., fractured bedrock aquifer), the ground water flow can
be idealized as fast (laminar) flow between parallel plates. In this flow regime, bacteria and
viruses tend to remain within the high velocity zone in the middle and less commonly approach
or transition through the boundary layer to attach to the fracture wall.

       In this document, non-porous media aquifers and one porous media aquifer, gravel, are
defined as sensitive aquifers. Among the remaining porous media aquifers (e.g. sand or sand and
gravel aquifers), sand and gravel aquifers (that is, aquifers containing a combination of both sand
and gravel) transmit fecal contaminants more efficiently than sand aquifers because average
ground water velocity is higher. Assessment source water monitoring in sand and gravel aquifers
may be  conducted using any of the GWR fecal indicators.  Some sand aquifers may also
efficiently transmit fecal contamination but, as discussed next, appear to more efficiently
transmit viruses as compared with bacteria.  Thus, sand aquifers, if targeted for assessment
source water monitoring may be monitored using coliphage rather than E. coli or enterococci.

       Because viruses may travel faster and farther in some aquifers, monitoring in these
aquifers using coliphage (virus) may be a better choice than monitoring using bacteria. This
issue is  also discussed in the GWR Source Water Monitoring Methods Guidance document
(USEPA 2006c). Microbial transport in porous media aquifers is an active research area and
consensus is difficult. It is generally agreed that microbe size is an important element in
determining subsurface transport in porous media (mobility), although many other factors, such
as surface charge may also have significant influence. Assuming that size is important, the
significant (one-thousand fold) size difference between viruses (measured in nanometers) and
bacteria (measured in micrometers) increases the likelihood that an infectious virus rather than
an infectious bacterium will reach a PWS well in porous media.

       All subsurface particles, including microbes,  may be transported by flowing ground
water. Particles may permanently (i.e., removal) or temporarily (i.e., retardation) become
associated with the solid aquifer materials (both porous and non-porous media). The thousand-
fold size difference between viruses and  bacteria may be significant in sand aquifers for two
reasons: (1) viruses are less likely to be subject to removal or retardation at pore margins by
straining, wedging, or micro-straining, and (2) viruses may be more likely to be excluded from
the smaller pores where ground water velocities are slower.  As a result of this pore-size
exclusion (which is due indirectly to size because charge effects predominate for smaller
particles), viruses may be favored over bacteria because the viruses remain in faster flowing
ground water for longer periods. As a result of straining and pore-size exclusion, sand aquifers
may facilitate virus, as compared with bacterial transport.

       In other aquifers, such as non-porous media and gravel aquifers where average ground
water velocities are exceptionally fast, straining and pore-size exclusion are much less
significant and bacteria and viruses are assumed to travel at equal rates. In general, this guidance
assumes that straining and pore-size exclusion effects are more significant in sand aquifers
where ground water velocity is moderate because mean grain size is moderate (sand aquifers as
Source Assessment Guidance Manual              1-10

compared with sand and gravel). As ground water velocities increase because of increasing
gravel content or increasing proximity to a pumping well, the differences between virus and
bacterial transport efficiency become less important. On the other hand, the finest grained porous
media, such as shale and clay beds, are not considered to be important aquifers because water
and entrained pathogens are not transmitted efficiently in all directions, so they are not further
considered, despite the much greater significance of straining and pore-size exclusion.

       Although the presence of viruses in water drawn from a well in porous media implies
fecal contamination at or near the surface, some microbial pathogens such as Legionella
pneumophila (Costa et al., 2005; Riffard et al., 2004), Helicobacterpylori (Hegarty et al, 1999;
Rolle-Kampcyk et al, 2004), Naegleriafowlerii (Blair and Gerba, 2006) and perhaps
Toxoplasma gondii  (Sroka et al, 2006) are not associated with fecal contamination and, instead,
may be resident members of aquifer ecosystems.  For these microbes, transport from the surface
or near surface is not an important risk element because the microbes can colonize the well
gravel pack or the aquifer immediately surrounding the gravel pack. In these instances, the
bacterial versus viral size difference and associated subsurface mobility become much less
important. Municipal waste water that is injected into deep brine formations or aquifer storage
and recovery operations are not typically shallow fecal contamination sources and are regulated
under the Underground Injection Control provisions of the SDWA. Because they are specifically
regulated, they are not further considered here.

       In the Alaska outbreak of Exhibit 1.3, norovirus was sufficiently long-lived and mobile in
the subsurface to contaminate a well in large numbers sufficient to provide an infectious dose
that made visitors ill. In contrast, fecal coliform were probably less mobile and long-lived and
arrived at the well only in small numbers. No travelers became ill from bacterial pathogens and
no bacterial pathogens were recovered from the well water. The short flow path between the
septic pit and the well provided for a relatively short ground water residence time. Wells with
fecal contamination sources that are near or at the State-mandated or recommended setback
distance  should be considered for assessment source water monitoring. Sanitary surveys and
source water assessments identify fecal contamination sources present in or near State mandated
or recommend setback distances.

       There are no simple, desk-top methods to identify PWS wells in which viruses may be
more likely than bacteria to arrive at a well. This is an active research area using water and
bioparticle tracers and is usually conducted at a field research site or in a laboratory column.
However, it is sometimes suggested that PWS wells producing from shallow, unconfmed
aquifers  are examples of PWS wells in which viruses are more likely than bacteria to arrive at a
well. This risk factor is discussed next.

       Shallow unconfined aquifers - As discussed  above, it is commonly assumed that fecal
contaminant sources originate at or near the surface. For  aquifers that are relatively close to the
surface and unprotected by a hydrogeologic barrier such  as a confining layer, the transport path
is relatively short and unimpeded. Thus, there is a greater likelihood that infectious fecal
contamination will reach a PWS well that produces water from a shallow, unconfined aquifer as
compared with a deeper or confined aquifer. Shallow aquifers are often,  but not always, sand,
sand and gravel, or sensitive aquifers.  A PWS well producing water from a shallow, unconfined
sandstone aquifer has an increased likelihood of infectious fecal contamination as compared with
a deep sandstone aquifer. In this instance, it is relatively easy to determine shallow versus deep
Source Assessment Guidance Manual              1-11

aquifers because well depth data are always available on the well log, well construction report,
or final well permit. If well construction reports are used, it is important to get the well depth
from the final report rather than from the well construction permit because many wells are not
constructed as designed.

       This guidance document does not define which unconfmed aquifers are shallow, may be
at risk, and therefore are suited for assessment source water monitoring. Aquifer depth is highly
variable due to pumping, recharge, geologic structure, and climate, and no single definition
applies nationwide. States  should consider the range of PWS well depths within their State and
geologic provinces and conduct source water monitoring assessment monitoring at those wells
producing water from unconfmed  aquifers that are among the shallowest in the State.

       Fecal indicator microorganisms may pass through shallow unconfmed aquifers at vastly
differing transport rates, depending on the aquifer type. Because groundwater flow and indicator
transport is so variable, no fecal indicator is favored and any of the three recommended
indicators is appropriate in these hydrogeologic settings.

       See section 2.5 for  further information on hydrogeologic barriers.

       Aquifers with thin or absent soil cover - Because fecal contamination sources are
assumed to be located at or near the surface, the presence and thickness of soil may be important
to attenuating infectious pathogen risk for drinking water wells. Soils are defined herein as
unconsolidated material formed in place by natural processes from geologic parent material.
Thick soils have enhanced capability to remove pathogens as compared with thin soils. Soil
typically has high natural organic matter content (as opposed to material added by septage input)
which is relatively efficient at retarding pathogen transport by favoring attachment to soil
particles. Where soils are red due to the presence of iron oxides or other favorable soil  conditions
occur, pathogens may also be efficiently removed.  In general, septic tanks are permitted in soils
that exhibit variable saturation conditions because  such conditions enhance removal of
pathogens and other septage contaminants. On the other hand, septic tanks are generally
prohibited from continuously saturated soils which do not  exhibit such conditions. Where soils
are thin or absent, pathogen removal from septage  and sewage is minimized and infectious
microorganisms are more likely to reach a PWS well tapping an aquifer unprotected by thick
soils. However, thick soils can have macropores, root casts, and other openings that mimic the
conduits and fractures typical of sensitive aquifers. Where present, macropores can  allow direct
and efficient transport of fecal contamination through soil, thereby by-passing some of the
protective properties of thick soils

       Soil formation occurs continuously, but soils may be thin or absent if erosion by wind,
water, anthropogenic activity, or glaciers predominates over soil formation processes. In general,
humid climates have thicker soils than arid climates. Unglaciated terrain has thicker soil than
glaciated terrain. States located in the western US are more likely to be arid, and States located
in the northern US are more likely to have had glacial removal of soil. More detailed data on soil
presence or thickness are typically available from US Department of Agriculture (USD A)
County Soil Surveys (maps and reports) as discussed in Section 1.5.3.
Source Assessment Guidance Manual              1-12

       Where soils are thin or absent, the soil capacity to remove or inactivate pathogens is
minimal. Assessment source water monitoring may be appropriate for PWS wells located in
areas with thin or absent soil (thick soils with soil macropores may also perform poorly).

       Because some fecal contamination sources such as septic tanks release fecal indicator
microorganisms into the shallow subsurface, these indicators may be insufficiently attenuated
when the soil is thin. The variably or completely saturated bedrock may allow passage of
indicator microorganisms at greatly differing rates, depending on the bedrock type. Since no
fecal indicator type has a greater likelihood  of passaging to the wellscreen, any of the three fecal
indicator organisms would be appropriate.

       Wells previously identified as having been fecally-contaminated - All PWSs must
conduct monitoring under the Total Coliform Rule (TCR) and that monitoring may identify fecal
contamination of ground water sources. The TCR-approved method identifies many but not all
E. coli serotypes. For example, the E. coli O157:H7 serotype is not identified using TCR-
approved methods. Although E. coli is used as a fecal indicator organism, some serotypes such
as E. coli O157:H7 are frank pathogens.  In particular, E. coli O157:H7 may cause kidney failure
in children and the elderly. Wells with a  history of E. coli contamination, where identified, are
more likely to experience additional fecal contamination and can be identified as wells with
greater likelihood of infectious pathogen contamination. As part of the sanitary survey, the fecal
contamination history of a PWS well would be reviewed. The Sanitary Survey Guidance Manual
for Ground Water Systems provides additional discussion about identifying PWS well fecal
contamination history.

       Wells and systems with a previous history of fecal contamination such as E. coli
occurrence should be considered to be at high risk from fecal contamination and should be
considered for assessment source water monitoring.

       The choice of fecal indicator microorganism should be governed by the available data.
All indicator microorganism groups (and probably types within each group) are transported at
differing rates in the same aquifer materials. If the well has a history ofE. coli occurrence, then
E. coli should be selected as the recommended indicator microorganism. It is less likely that
wells will have a history of enterococci or coliphage occurrence but if a well has previous
occurrence of one of these organisms, then,  based on that fecal indicator occurrence history, that
fecal indicator should be selected because the aquifer materials appear to permit efficient
transport of that indicator to the well.

       High population density combined with on-site wastewater treatment systems - Any
aquifer may be at risk of fecal contamination if the aquifer's natural attenuation capabilities are
overwhelmed. For example, the high-density fecal contamination discharged into the subsurface
by septic tank drainfields and other on-site wastewater treatment systems can pose such a risk.
Greater population density combined with restricted areal extent of an aquifer is an especially
risky combination because aquifer recharge by septage discharge is significant compared with
infiltrating precipitation. Some aquifers,  such as barrier island or marine island aquifers,  are
capable of supplying only limited yield because over-pumping will result in seawater intrusion,
permanently damaging the aquifer. Where population density is high and yield is limited,
dilution and other natural attenuation processes are also limited, and fecal contamination is more

Source Assessment Guidance Manual              1-13

likely. PWS wells located in resort island communities should be targeted for additional

       Large numbers of visitors to an island, combined with on-site wastewater treatment
facilities, are important risk factors that have contributed to aquifer contamination. Localities
with large transient populations and on-site wastewater treatment may be considered for
assessment source water monitoring to provide additional public health protection.

       As discussed in the Sanitary Survey Guidance Manual for Ground Water Systems, land
use and sanitary setback distances around PWS wells are important elements of a sanitary
survey. It is not the purpose of this document to address land use. However, resort communities
with large seasonal population changes are relatively easy to recognize. These locations are
likely to have high population density and restricted aquifer yields, and some have on-site
wastewater treatment. In such locations, assessment source water monitoring will provide
additional public health protection.

       Resort communities may be located on, for example, barrier island sand aquifers or karst
limestone islands. Because resort island aquifers types  are varied, the ground water flow and the
subsurface passage of fecal indicator microorganisms is also highly variable. Thus, no fecal
indicator is favored and any of the three recommended indicator organisms may be appropriate.
However, as discussed previously, sand aquifers may be an aquifer type in which viruses are
favored for transport as compared with bacteria and thus coliphage might be a more appropriate
fecal indicator organism.

       Other Risk Factors - States may have information collected during sanitary surveys,
Source Water Assessments, and field visits that indicate a ground water source may be subject to
fecal contamination. States may require PWSs with these sources to conduct source assessment
monitoring based on this information. Examples of deficiencies or information that may indicate
ground water sources are at risk from fecal contamination and therefore indicate that a need to
conduct assessment source water monitoring include:

       •  Well near a  source of fecal contamination;

       •  Well in flood zone;

       •  Improperly constructed well (e.g., improper surface or subsurface seal);

       •  Well that does not meet State codes or standards (e.g., depth or setback

       •  Well of unknown construction (e.g., no driller's log or other record of construction);

       •  Other non-microbial indicators potential for fecal contamination (e.g., MB AS, high
          chloride or nitrate levels from baseline, or historical trends).
Source Assessment Guidance Manual              1-14

       Due to the great variability among other possible risk factors, no specific fecal indicator
organism would be favored and thus any of the three GWR fecal indicator organisms is
1.5    Hydrogeologic Data Sources for Assessment Source Water Monitoring

       A number of EPA publications provide detailed discussions of hydrogeologic data
sources. An EPA workgroup was convened in 1993 to develop a guidance document on ground
water resource assessment. The guidance describes sources of hydrogeologic data and how these
data may be used to evaluate aquifer sensitivity (USEPA 1993a). EPA also published the
Ground Water Information Systems Roadmap, A Directory of EPA Systems Containing Ground
Water Data (USEPA  1994a). An EPA Handbook titled  Ground Water and Wellhead Protection
(USEPA 1994b) also  summarizes hydrogeologic data sources.

       This section augments the discussions contained in these earlier documents, emphasizing
desktop analyses. In cases where desktop analysis is not possible because the needed data are
not available or do not have sufficient resolution, field investigations may be necessary.
1.5.1   State and Federal Hydrogeologic Investigations

       The data sources described in this section are electronic or hard copy reports and/or data
produced through previous desktop analyses or field investigations.  Such information may have
been generated to meet the requirements of SWAPs or through water quality and/or water supply
investigations initiated at the local, State, or Federal level.

       Existing data for a given PWS well may be used.  For example, if an existing report or
appropriate scale map indicates whether or not a PWS well is screened in a sensitive (i.e., karst,
fractured bedrock, or gravel) aquifer, then that information can be used to satisfy the HSA
requirement. Generally, spatial data at the scale of 1:100,000 or larger (e.g., 1:24,000) are
sufficiently detailed for most purposes [Note: large scale maps provide detailed information of
small geographic areas.] Wellhead Protection and Source Water Assessment Studies

       The SDWA, as amended in 1986, created the Wellhead Protection Program (WHPP).
Each State is required to adopt a program to protect wellhead areas within its jurisdiction from
contaminants that may have adverse health effects and to submit the program plan to the EPA
Administrator. Currently, 49 States and two territories have WHPPs in place.  In their WHPPs,
States address all program elements including how to delineate wellhead protection areas
(WHPAs) and how to identify and inventory all potential sources of contamination.

       Section 1453 of the 1996 SDWA Amendments required  all States to establish SWAPs
and to submit plans to EPA for approval by February 6,  1999. These SWAPs address both
surface water and ground water protection, and their SWAP plans detail how States will: (1)
Source Assessment Guidance Manual              1-15

delineate source water protection areas; (2) inventory significant contaminants in these areas;
and (3) determine the susceptibility of each public water supply to contamination.  States may
use any available information to carry out the SWAP, including data generated through the
WHPP.  After plan approval, the States must have completed susceptibility determinations for all
PWSs by November 6, 2001, unless the State was granted an 18-month extension until May 6,

       EPA encourages States to build upon previous SWAP or WHPP efforts to determine
hydrogeologic sensitivity. A review of selected, approved State SWAP plans across EPA regions
indicates that many  States intend to evaluate hydrogeologic information that may enable them to
determine a PWS well's aquifer type.

       At least one  State addresses the three sensitive aquifer types (i.e., karst, fractured
bedrock,  and gravel) and the presence of a hydrogeologic barrier as part of a susceptibility
determination process in its approved SWAP plan (WIDNR 1999). Other approaches to
fulfilling SWAP requirements are also likely to result in data that will be useful for source water
assessment. Case studies # 2 and # 4, presented in sections 3.2.2 and 3.3.2, respectively,
illustrate just two ways in which data can be extracted from SWAP investigations. State Geologic Survey, USGS, and Other Hydrogeologic Investigations

       Many State geologic surveys and/or agencies of natural resources have significant
experience studying local and regional aquifer systems and investigating ground water quality
and quantity issues. Although many of these studies may have directly supported, or continue to
support, SWAP or WHPP work, many more studies have been conducted independently of these
efforts.  In addition to State geologic surveys, the United States Geological Survey (USGS) has
district offices that perform similar work in each State, sometimes in cooperation with State
agencies.  Universities, local governments, and non-governmental organizations also conduct
pertinent hydrogeologic research.
1.5.2   Hydrogeologic and Geologic Maps

       Hydrogeologic or aquifer maps generally show the location, spatial extent, and depth of
aquifers in a region.  Such maps typically include information on aquifer type as well.
Hydrogeologic maps will often be the most direct means of evaluating important risk factors
such as aquifer type.

       Geologic maps may depict a region's surficial geology, which would include the
locations and extent of distinct unconsolidated deposits and bedrock units exposed at the earth's
surface, or, alternatively, the bedrock geology of an area. Surficial geologic maps are available
for many areas from the USGS and often include a key to interpret the results of various test
holes shown on the map. Using geologic maps is a less direct means of identifying aquifer type
than using hydrogeologic maps. But analytical techniques such as projection (described below)
and information such as well depth can help in determining aquifer type.
Source Assessment Guidance Manual               1-16

       The availability of hydrogeologic maps at an appropriate scale varies among States and
among regions. The following sources may be useful to States in obtaining appropriate maps for
use in preparing HSAs. As part of its Regional Aquifer-System Analysis (RASA) program, the
USGS has produced a large variety of hydrogeologic maps at various scales. Some of these
maps are at scales that may be useful for an HSA (Sun et al. 1997). The RASA program
completed studies of 25 major US aquifer systems in 1995.  The Ground Water Atlas of the
United States was developed as part of the RASA program and provides small-scale (i.e.,
numerically large, less detailed coverage of large geographic areas) hydrogeologic data for the
country both as a printed atlas and as a digital dataset (available on the Internet at:
http://capp.water.usgs.gov/gwa/). The printed atlas has 13 individual chapters that cover specific
US regions.  The Ground Water Atlas data, however, are compiled at scales that may not be
suitable for public water system-specific HSAs (e.g., at 1:5,000,000 and 1:2,500,000 scales).

       In areas where hydrogeologic maps are not available, a geologic map along with the
projection method may be used to determine the aquifer type for a well  of a given depth.
Projection is a structural geologic technique which can be used to determine aquifer depth, or
the depth of any local geologic unit at a well, using the strike and dip of the aquifer as measured
at nearby outcrops. Typically, bedding (layering) can be described in terms of its strike and dip.
Bedding occurs (but may be indistinct) in some sedimentary rocks, in metamorphosed
sedimentary rocks (metasediments), and in some igneous rocks such as  volcanic flows (e.g.,
basalts).  Outcrops of the bedrock are shown on many geologic maps along with the values of the
strike and dip of the bedding.  The strike is the compass direction or azimuth of the line formed
by the intersection of the bed with its horizontal (planar)  surface. The dip is the angle in degrees
between the bedding and a horizontal surface, measured at right angle to the strike (Exhibit 1.4).
If the bedrock is a known aquifer, the depth  to that aquifer can be determined by projecting the
dip over the distance to the well location. Using simple trigonometry, the depth to the aquifer is
then equal to the tangent of the angle multiplied by the distance.  This method can be used in
areas of simple geology.
                              Exhibit 1.4 Strike and Dip
Source Assessment Guidance Manual

       More detailed hydrogeologic and geologic maps are available from a variety of public
and private entities. The USGS, as well as State geologic surveys or natural resources agencies,
are the most prolific sources. However, coverage is highly variable from State to State. The
National Research Council (NRC) estimated in 1988 that less than 20 percent of the United
States has been geologically mapped at a scale of 1:24,000 or larger (NRC 1993).

       In response to this situation, Congress enacted the National Geologic Mapping Act of
1992. This act established the National Cooperative Geologic Mapping Program (NCGMP) to
implement expanded geologic mapping efforts through a consortium of geologic mappers. As
part of this program, the USGS conducts Federal mapping projects through its FEDMAP
program; STATEMAP, run by State geological surveys, is a matching-funds grant program; and
universities participate in another matching-funds program - EDMAP.  The USGS coordinates
the NCGMP, which has a long term goal of producing l:24,000-scale geologic maps for
high-priority areas of the States and national coverage at the l:100,000-scale.

       The NCGMP also maintains an exceptionally useful database for locating existing
geologic maps produced by a wide variety of entities, and it includes mapping currently in
progress through the consortium and is searchable by location, scale, and other parameters.  The
database, as well as general information on the program, is available on the Internet at
http://ngmdb.usgs.gov/.  A geologic map index is also available for many States showing
boundaries for compiled map projects and references.
1.5.3   Soil Maps

       The Natural Resources Conservation Service (NRCS), a division of the United States
Department of Agriculture (USDA), is responsible for soil mapping in the United States. Soil
survey maps have been completed for most of the United States, as of December 2002 (NRCS-
NSSC 2002). The mapping is done using aerial photographs at scales that depend upon needs,
but scales of 1:24,000, 1:15,840, and 1:12,000 are common.  Soil maps are published as Soil
Survey Reports, usually on a county by county basis.

       Although they do not provide direct information regarding aquifer type, careful
interpretation of county soil surveys can yield useful information on other risk factors such as
presence or absence of thick protective soil.  Soil maps are accompanied by detailed descriptions
of the mapped soil series which indicate the earth materials from which they were derived (i.e.,
the "parent materials").

       In some cases, the parent materials may be the underlying bedrock.  Soils formed in place
by the weathering of bedrock are called residual soils. A soil may also form in sediments
transported to the site (e.g., by stream or glacial deposition).  Therefore, even though the soil
series descriptions always indicate a soil's parent material, only in the case of residual soils will
that information directly indicate the type of underlying bedrock, and possibly the aquifer type
for the PWS of interest. Nonetheless, the underlying bedrock type may be noted in the soil
series description even if it is not the parent material. For example, in the Soil Survey for Essex
County, Massachusetts, Southern Part, the Chatfield  series profile description notes that the
parent material is glacial till, but granitic bedrock is found 34" below the  land surface. Soil

Source Assessment Guidance Manual              1-18

series profile descriptions are based on a profile from the survey area that is considered typical of
the given series (USDA-SCS 1984).

       The availability of soil survey maps and suggestions for how to obtain maps not available
on the Internet can be checked on the Internet at http://soils.usda.gov/survey/. Maps in the
"published" category can be considered up-to-date. The "initial mapping complete" category
refers to maps that can be obtained either on CD from NRCS  or as a hard copy, per request by
the State.  Similarly, areas listed as "Update Field Work Complete" may be available on the
Internet, and areas listed as "Update Field Work in Progress"  are likely to have maps available
upon request to the NRCS. Maps in the categories of "Maintenance Needed" or "Maintenance"
are likely to require some minor changes (e.g., due to regional floods changing the character of a
floodplain, as occurred with the Mississippi River floodplain  in 1993). Areas on this map that
are considered "non-project" by  the NRCS may still have soil survey maps that can be obtained
from the Federal agency that has responsibility for the land in that area.  Alternatively, local
NRCS field offices may also have data for "non-project" sites.  Areas that are listed as "initial
mapping in progress" (only 5 percent of the United States) do not have soil survey maps
currently available.

       Digital soil survey data are also available. The Soil Survey Geographic (SSURGO) Data
Base is comprised of digitized county-level soil survey maps  (compiled at scales ranging from
1:12,000 to  1:63,360), and are of much higher resolution than STATSGO data. Therefore,
SSURGO data are more suitable for performing HSAs.  Furthermore, SSURGO data are
designed for use with GIS because they are linked to a Map Unit Interpretations Record (MUIR)
attribute data base. SSURGO data are not yet available for every  county and area, however the
database continues to grow. States can easily check to see if a soil survey is available for a
particular county by accessing the SSURGO list. All soil surveys  on this list can be obtained
either on the Internet, or are available on CD by requesting the information from the NRCS.  The
SSURGO data available on the Internet are updated monthly.
1.5.4   Topographic Data

       Well coordinates, depth to the screened interval of a well, and topographic maps
(described below) can be used to determine whether or not a well is drawing water from a given
aquifer. Imprecise plotting of a well's location could lead to an erroneous assessment of the
aquifer type from which the well is drawing water (and thus possibly an incorrect evaluation of
whether or not the well is drawing from a sensitive aquifer).

       Accurate determinations of well locations are critical for making sensitivity
determinations using a desktop analysis; thus, it is important to use large scale topographic maps
(e.g., 1:24,000 topographic quadrangles) for plotting the well's location (see Exhibit 1.5). In the
absence of a detailed topographic map (e.g., 1:24,000), abase map of comparable scale is needed
to accurately locate the well. Such a map might be available from the local  community (e.g.,
Assessor's Office, Engineering Department, Department of Public Works, Water Board, Board
of Health, Planning Board, and Conservation Commission) or from State, Federal, or regional
natural resource agencies and planning departments.
Source Assessment Guidance Manual              1-19

       Accurate well coordinates may be sought first from the PWS. Well registration
information collected by Federal, State, and local regulatory programs also usually include
coordinates.  They may also be available from the well drilling company records. If necessary,
well coordinates can also be obtained in the field using Global Positioning System (GPS)
technology. The City of Tallahassee (1996) has described the process of locating PWS wells
using GPS receivers and discussed the important issue of receiver accuracy.
       Exhibit 1.5 The Importance of Map Scale for Determining Aquifer Type
                               A A1
Lake Wobegon e
                                      incorrect  correct
In Exhibit 1.5, X indicates the location of a well with known areal coordinates and depth. On the smaller scale map,
(a), precise plotting of the well's location is impossible.  The larger scale map, (b), shows the location of X with much
greater precision.  Cross-section (c) shows a correct identification, based on map (b), of the well's aquifer as gravel (a
sensitive aquifer) and an incorrect identification, based  on map (a), of the well's aquifer as sand (a non-sensitive
       Topography can be represented in two dimensions with contours, continuous lines that
join points of equal value (equal elevation in this case). The contour interval, which is the
change in elevation between each successive contour line (e.g., 20 feet), is chosen depending
upon the scale of the map and the topographic relief. The USGS and the Defense Mapping
Agency (DMA) have produced most of the topographic maps for the United States (NRC  1993).
 The USGS produces maps at a variety of scales, but the most common scales for topographic
maps are 1:24,000/1:25,000,  1:100,000, and 1:250,000. The 1:250,000 scale maps are available
for the entire United States. The much more detailed topographic quadrangles (1:24,000 or
1:25,000) are available for most of the country.  Index maps showing available topographic maps
for each State are provided by the USGS without charge.  Each 1:24,000 topographic map covers
approximately 58 square miles, where 1 inch corresponds to 2,000 feet.
Source Assessment Guidance Manual

       Digital topographic data for the United States are also available from the USGS as
Digital Line Graphs (DLGs) and Digital Elevation Models (OEMs). DLGs are vector data files
that represent linear and areal features commonly found on topographic maps, including contour
lines. OEMs are data files that store point elevations spaced at regular intervals in a matrix.
Detailed OEMs have 10- and 30- meter resolutions. Because national coverage is incomplete for
both DLGs and OEMs, and State-wide coverage varies considerably by State, the remainder of
this section will focus on paper topographic quadrangles.

       Topographic maps are based on aerial photos, and skilled topographic map interpretation
may reveal landform features  such as sinkholes and "losing streams" (in this context, a stream
that disappears and loses its water to ground water via an underground route). Such features are
indicative of the underlying bedrock type and/or structure (in cases where the structure controls
the topography). Discontinuous drainage networks are also revealed on detailed topographic
maps and indicate a karst environment.  Drainage may follow the underground joint pattern in
the rock, which is expressed on the topographic map. Contour lines representing elevation may
also reveal distinct features of the local bedrock structure such as folds and faults; such
structures are almost invariably  associated with fracturing.  The topographic map will also show
the orientations of folds and/or faults that have a surface expression, helping to establish the
orientations of regional fracture networks.

       The surficial or geomorphic features associated with a particular soil type may be
represented on a topographic map.  Deposits likely to consist of coarse gravel can be readily
identified  on a topographic map by their surface expression. The geometries or drainage patterns
of streams can provide  clues to the underlying geology.  A dendritic drainage pattern will most
likely be found in horizontal sedimentary rocks or massive igneous rocks, but can also be seen in
folded or complex metamorphic terrain. Trellised and rectangular drainage patterns indicate
faulted and jointed rock. Centripetal patterns, with or without trellised drainage, can indicate the
presence of sinkholes.  In areas covered by overburden, the lack of surface streams is an
indicator that well-draining granular soils underlie the area.
1.5.5   Stereoscopic Aerial Photography

       Aerial photographs taken with approximately 30 percent overlap allow three dimensional
imaging of land surface features with the aid of stereoscopes.  In regions with limited geologic or
topographic data, stereoscopic air photos may help locate wells. In most cases, however, such
photos will be most useful for determining aquifer types when used in conjunction with other
data sources. For example, if low resolution geologic maps or well log  data indicate that a given
PWS well may be screened in a karst aquifer, stereoscopic aerial photos could be used to
determine the presence or absence of sinkholes and/or other characteristic karst landform

       Aerial photographs are available from several entities within the USDA and from the
USGS. The NRCS and the Forest Service, both under the USDA, have  extensive U.S. coverage
at scales appropriate for HSAs. As noted above, the NRCS uses high resolution aerial
photography to compile their county level soil surveys at scales ranging from 1:12,000 to
1:63,360 (see section 1.5.3).  The USDA Aerial Photography Field Office, Farm Service Agency
acts as the clearinghouse for all USDA aerial imagery, archiving over 10,000,000 images dating
Source Assessment Guidance Manual               1-21

to 1955.  USD A aerial photo coverage, availability, and ordering information are available
through their website at: http://www.apfo.usda.gov/.

       The USGS National Mapping Division administers the National Aerial Photography
Program  (NAPP). The NAPP coordinates the collection of cloud-free coverage of the
conterminous United States and Hawaii at a uniform scale (approximately 1:40,000) about every
five years. NAPP photographs are available in black-and-white, and in many cases, color
infrared.  The imagery is available from the USGS's Earth Resources Observation Systems
(EROS) data center (http://edc.usgs.gov/) or Earth Science Information Centers (ESIC).  NAPP
photos are also available from the USDA Aerial Photography Field Office, Farm Service Agency
(see link  above).
1.5.6   Other Data Sources for Desktop Analyses

       Well registration information and well logs collected by local, State, and Federal
regulatory programs may be very useful for determining aquifer type. Well registrations usually
indicate well locations and information necessary to conduct an HSA. A sufficiently detailed
driller's log for a PWS well could itself, or in combination with other data sources, adequately
characterize the subsurface  stratigraphy and aquifer type. For example, based upon a regional
bedrock geology map that is of moderately low resolution (e.g., 1:700,000), a State may identify
that a PWS well is located in an area underlain primarily by limestone.  The State may review
the driller's log (if available) to confirm that,  in fact, the well is screened in a limestone aquifer.
Certain States such as New  Jersey and New Hampshire require drillers to file a log for each well
with the appropriate State agency, such as a water well board or the State environmental
protection agency.

       A driller's log typically records changes in lithology with depth, although local
terminology may be used and may need deciphering. For example, in much of the United States
the term "artesian well" is used by drillers as  a lay term to indicate a producing bedrock well.
This contrasts with the hydrogeologist's definition - a confined aquifer where the water in a well
rises above the top of the aquifer, sometimes flowing to the land surface. Another example is the
use of the term "hardpan" by drillers to describe what may be a dense glacial till,  a cemented
soil, or a hard clay.  A driller's log may also include information on the drilling method
employed, which may give  clues to the type of materials the drillers encountered.

       Additional desktop sources include consultant reports and database searches for property
site assessments conducted  by private search  companies. These searches of Federal,  State, and
local agency databases are conducted as part of due diligence investigations for property site
assessments and are usually in accordance with the standards of the American Society for
Testing and Materials (ASTM).  These database searches include a description of the bedrock
and surficial geology, a well inventory, and usually aerial photo coverage for the area in
question. The well inventory summarizes well locations, construction, soil and bedrock type,
water quality, and other pertinent data.
Source Assessment Guidance Manual               1-22

                     2. Sensitive Hydrogeologic Environments
2.1    Aquifer Sensitivity

       Aquifers in which pathogens can move quickly to public supply wells, allowing little
filtration or time for inactivation, are considered sensitive.  These aquifers generally have rapid
ground water flow velocities (that increase in the vicinity of pumping wells) and short, direct
flow paths. Three aquifer types are most likely to have these properties: karst, fractured bedrock,
and gravel aquifers. These aquifer types are described below in sections 2.2, 2.3, and 2.4,
respectively. Exhibit 2.1 provides examples of sensitive aquifers and their common names as
well as examples and sources of USGS maps of aquifer types. Exhibit 2.2 provides examples of
hydrogeologic barriers and their properties.

       Aquifer type, which is usually well-correlated with lithology, is important and should be
identified to determine if a specific aquifer is sensitive.  This document includes all limestone
aquifers, igneous and metamorphic aquifers, and gravel aquifers as sensitive. This is because
limestone is the lithology most likely to be karst; igneous and metamorphic aquifers are likely to
be highly fractured; and gravel aquifers are likely to have direct flow paths and rapid ground
water velocities due to the shape and large size of their pores. Pumping wells increase the
natural flow velocities in a sensitive aquifer to a greater degree  than they would in a fine-
grained, unconsolidated aquifer, for example.

       States may designate additional aquifer types as sensitive if they believe it is necessary to
do so to protect public health (e.g., a State may designate sand and gravel aquifers as sensitive).
This guidance will not cover other potentially sensitive aquifers because States have the
flexibility to set their own criteria regarding other aquifer types that may or may not be sensitive.

       The following sections - 2.2 through 2.4 - describe the rationale for including or
identifying karst, fractured bedrock, and gravel aquifers as sensitive. Brief summaries of
bacterial contamination research in each of these aquifer types are included,  as well as
summaries of known disease outbreaks resulting from such contamination.  Because karst
aquifers make up such  a large proportion (40 percent) of all productive aquifers in the United
States (USGS, 2002), section 2.2 describes the surface and subsurface hydrologic characteristics
of karst regions, and the potential for bacterial contamination of karst aquifers, in some detail.
Nevertheless, fractured bedrock and gravel aquifers are considered equally sensitive, and when
encountered,  are worthy of the same time and consideration as karst aquifers.
Source Assessment Guidance Manual              2-1

                         Exhibit 2.1 Example Sensitive Aquifers
 Aquifer Type     Common     Example   USGS Classification and Map of Aquifer Types in US
 Carbonate       Karst        Limestone  http://capp.water.usgs.gov/aquiferBasics/carbrock.htmWlist
 Igneous         Fractured     Granite     http://capp.water.usgs.gov/aquiferBasics/volcan.html
 Metamorphic     Fractured     Gneiss     http://capp.water.usgs.gov/aquiferBasics/volcan.html
 Unconsolidated  Gravel       Glacial     http://capp.water.usgs.gov/aquiferBasics/uncon.html
         Exhibit 2.2 Examples of Hydrogeologic Barriers and their Properties
Glacial Till
Porosity1 (%)
Range1 (cm/s)
Specific Yield (%)
(1 - 20)2
(5 - 20)3
(0.5 - 5)4
(1 - 35)5
Conductivity6 (cm/s)
1. 4x10* to 1. 4x1 rj9
1. 4x1 Q-6 to 9.4x1 0'10
  Brown etal., 1983
 2Depends on source. Heath, 1983; Morris and Johnson, 1967, as compiled by McWhorter and Sunada, 1977; Sevee,
 1991; Devinny et al, 1990
 3Devinny et al, 1990
 4Sevee, 1991
 5Depends on source. Morris and Johnson, 1967, as compiled by McWhorter and Sunada; Sevee, 1991; Devinny et
 al, 1991
 6Compiled from Morris and Johnson, 1967, by Barton et al, 1985
 2.2    Karst Aquifers

      Karst is defined as a type of geologic terrain within which flowing ground water has
 dissolved significant portions of the area's soluble (usually carbonate) rocks (Fetter 2001).
 Where karst regions occur, infiltrating precipitation and ground water create a permeability
 structure characterized by  numerous and often large, interconnected conduits.  Through time,
 these conduits continue to  enlarge, creating unique surface and subsurface drainage networks
 and characteristic surface landforms. Ground water velocities are usually rapid and flow paths
 are very direct in karst environments, especially in the vicinity of pumping wells.  All limestone
 aquifers are designated as  sensitive aquifers in this document, due to the likelihood that they are
 karst environments. Microbial pathogens released into karst aquifers that intersect the ground
 surface from subsurface sources such as septic systems or surface sources such as livestock
 feedlots are likely to reach drinking water consumers in an infective state. For example, the
 Walkerton, Ontario E. coli outbreak in May, 2000 is believed to have been caused by fecal
 pollution of a karst aquifer system (Worthington et al. 2001, 2002). Two outbreaks in  Braun
 Station, TX (D'Antonio et al., 1985), and outbreaks in Georgetown, TX (Hejkal et al., 1982) and
 in Brushy Creek, TX (Bergmire-Sweat et al., 1999; Lee et al., 2001) all resulted from
 contaminated wells located in the Edwards Aquifer, a sensitive karst limestone aquifer.

 Source Assessment Guidance Manual               2-2

Similarly, outbreaks in South Bass Island, OH (Ohio EPA, 2005; USCDC, 2005; Fong et al.,
2007; O'Riley et al., 2007) and Walkerton, Ontario (Golder Associates, 2000; Health Canada,
2000; Hurley and Hurley, 2004) resulted from contaminated wells located in the Upper Silurian
Bass Island Formation, a sensitive karst limestone aquifer.

       Karst features are most commonly found in limestone, but marble, dolomite, evaporites
(e.g., gypsum), and other soluble rocks may also have karst features. Calcium magnesium
carbonate is less soluble than calcium carbonate, which is why limestone (composed primarily of
calcium carbonate) is more likely to be karstic than dolomite (composed of calcium magnesium
carbonate) (Freeze and Cherry 1979). Solution-enlarged fractures and conduits - typical karst
features - tend to provide the dominant ground water flow paths in limestone aquifers. For the
purposes of this guidance manual, it is assumed that all limestones are karstic and are therefore
sensitive. Other potentially soluble rocks  such as dolomite may also be recognized as sensitive
aquifers, especially if the dissolution openings are enhanced by fracturing (e.g., Door Peninsula,
WI).  States may choose to consider all carbonate (potentially soluble) aquifers, and other
aquifers formed by soluble rocks (e.g., evaporite aquifers in the Las Vegas Valley; see USGS
Circular  1170, 1998), as possibly karstic and designate them sensitive.

       Karst regions are typically characterized by the following: underground  drainage
networks with solution openings that range in size from enlarged fractures to large caves; closed
surface depressions, known as sinkholes, where the dissolution of the underlying bedrock has
caused the collapse of overlying rock and  sediment; and discontinuous surface water drainage
networks that are related to the unique subsurface hydrology (Winter et al. 1998). In areas such
as Orange County, Indiana, there are so many sinkholes (over 1000  per square mile, Exhibit 2.3)
that they coalesce into compound sinkholes (Thornbury 1954).  In other mature karst landscapes,
characterized by relatively pure limestone in areas of high precipitation, caves and caverns are
formed in the subsurface.  Conduits in carbonates and gypsum can be quite large with some
exceeding 100 feet in diameter (i.e., caves) and several  miles in length. Mammoth Cave,
Kentucky has a mapped length of more than 340 miles of interconnected  conduits distributed
over five horizontal levels. Ground water velocities have been measured there at more than
1,000 feet per hour (USEPA 1997).
Source Assessment Guidance Manual               2-3

     Exhibit 2.3 Map of Sinkholes (closed circular contours with tick marks) in
                                   Orleans, Indiana
       Indeed, it is the rapid ground water velocities in karst aquifers that necessitate their
characterization as sensitive aquifers.  In the karst region of Slovenia, bacteriophage injected into
a karst aquifer reportedly traveled approximately 24 miles in less than 4 months (Bricelj 1999).
Using conservative ground water tracers, scientists have measured ground water velocities in
karst aquifers to be as high as approximately 0.3 miles per hour (USEPA  1997). In Florida,
ground water velocities surrounding a well have been measured at several hundred feet per hour
(USEPA 1997). In a confined karst aquifer in Germany which was breached by monitoring
wells, ground water traveled approximately 650 feet in less than 4 days (Orth et al. 1997). In the
Edwards Aquifer, Texas, Slade et al. (1986) reported that dye traveled 200 feet in 10 minutes.
This data all indicates that ground water flows extremely rapidly through  karst aquifers.
Source Assessment Guidance Manual

       Well-developed karst systems may have underground streams because of the large size of
interconnected openings in the rock. Underground streams can have flow rates as great as those
of surface streams. It is also not unusual in karst terrains for surface streams of considerable size
to disappear into solution cavities (swallow holes) intersecting a streambed, creating a
discontinuous surface drainage system. These same streams may reappear at the surface at other
locations (Winter et al. 1998).  Seeps and springs are thus common in karst regions.

       Sinkholes in karst regions can play a particularly devastating role in the microbial
contamination of ground water supplies. For example, sewage treatment lagoons have been
known to leak and eventually collapse over sinkholes.  This phenomenon has been documented
in West Plain, Missouri in 1978 (Craun 1984); in Lewiston, Minnesota in 1991; and in Altura,
Minnesota in 1974 and 1976 (Jannik et al.  1991).  In Missouri, 759 illnesses resulted from the
contamination of domestic wells due to this 1978 sinkhole collapse (Craun 1984).

       Even in the absence of sinkhole collapse, the potential for rapid infiltration of fecal
contamination through overlying soils into karst aquifers is great when karst aquifers intersect
the ground surface. Residual soils, formed by bedrock dissolution, are characteristic of well-
developed karst regions. These soils are typically clay-rich, but can have great variation in
thickness and hydraulic conductivity (the capability to transmit water).  Soil macropores transmit
water rapidly, and are caused by channels formed by decayed  roots, insect and animal burrows,
dessication cracks, soil failure surfaces, and soil piping (USEPA 1997). Rapid flow in the
overlying soil may also occur via vertical fissures, even when there is substantial residual soil
cover (Smart and Frederich 1986, cited in USEPA 1997).  Where the mantle of glacial till or
outwash deposits is thin, infiltration velocities may also be high (Crowther 1989, cited in
USEPA 1997).

       The actual transport of fecal bacteria within karst aquifers has been studied at a variety of
localities (Malard et al. 1994; Orth et al. 1997; Tranter et al. 1997; Gunn et al. 1991).  Malard et
al. (1994)  suggested that both fractures (discussed in section 2.3) and karstification contribute to
rapid bacterial transport in limestone.  For this reason, Malard et al. (1994) consider the risk of
bacterial contamination greater in limestone than in any other type of aquifer.

       It is important to note that concentrations of bacteria within karst environments often
vary significantly with rainfall.  Personne et al. (1998) found that high aquifer water levels,
induced by high rainfall, correlated with high bacteria levels in the aquifer. The water level in
one Edwards Aquifer well (582 feet deep with a water table 240 feet deep) began rising within 1
hour after a rainfall event (Slade et al. 1986).  Mahler et al. (2000) studied fecal coliform and
enterococci bacteria near a wastewater irrigation site, and found the presence of bacteria in
ground water directly followed rainfall events. Mahler's data suggests that small sampling
intervals of 3 to 4 hours are necessary to describe the breakthrough of bacteria at a monitoring
well screened in a karst aquifer.

       The potential for rapid transport of bacteria and viruses through karst aquifers
necessitates that they be monitored carefully for contamination.  Bacteria can rapidly percolate
into the unsaturated zone of karst aquifers, as well as be farther transported to the saturated zone
during periods of intensive rainfall.  In fact, Malard et al. (1994) found high occurrence rates for
bacteria in a karst aquifer as long as a year after surface pollution had essentially ceased.  This
data demonstrates that sensitive aquifers can be  contaminated even when surface pollution
Source Assessment Guidance Manual              2-5

sources are difficult to identify.  Furthermore, research shows that surface water and ground
water drainage divides generally do not coincide in karst regions due to complex patterns of
surface water and ground water  flow. For example, a stream may disappear in one surface water
basin and reappear in another basin.  This situation makes it even more difficult to successfully
inventory sources of fecal contamination in the recharge area of a karst well (Winter et al. 1998).
Such situations are part of the motivation behind the GWR's focus on monitoring sensitive
aquifers,  rather than merely looking for potential sources of bacterial contamination. In
summary, bacterial contamination of karst aquifers is both fairly likely and highly unpredictable,
although correlations with rainfall events are common.
2.3    Fractured Bedrock Aquifers

       This document considers all igneous and metamorphic aquifers to be fractured bedrock
aquifers and designated sensitive.  They are considered hydrogeologically sensitive due to the
rapid velocities and direct flow paths through the fractures. Under the influence of pumping,
already naturally high flow velocities increase to even higher rates in fractured rock aquifers.  In
general, fractures have a role in ground water movement through any consolidated aquifer,
however fractured bedrock aquifers are those in which fractures provide the dominant flow
paths. Other aquifer types that may also be fractured (e.g., sandstone aquifers) are not considered
sensitive in this document. Nevertheless, States may choose to investigate the degree to which
these other aquifers are fractured, and decide if these aquifers should be monitored for bacterial

       Any solid block of igneous or metamorphic rock (i.e., the matrix) that is surrounded by
fractures is considered essentially impermeable (Domenico and  Schwartz 1990). Thus, all flow
is forced to take place within the fractures. A detailed understanding of flow in a fractured
bedrock aquifer requires knowledge  of fracture widths, orientations, the degree to which
individual fractures are mineral-filled,  and the degree of fracture interconnection and spacing.
Most fracture widths are smaller than one millimeter (mm), and a fracture's capability to
transmit ground water (i.e., hydraulic conductivity) is roughly proportional to the cube of the
fracture width (NRC  1990). Thus, small changes in fracture width result in very large changes in
hydraulic conductivity. For example, a 1 mm fracture can transmit 1000 times more water than a
0.1 mm fracture, provided that other factors are constant (e.g., hydraulic gradient).

       Freeze and Cherry (1979) report void space as high as 10 percent of total volume in
igneous and metamorphic rock.  Other data presented in Freeze and Cherry (1979) suggest that
the first 200 feet beneath the ground surface produces the highest water yields to wells because
fractures at shallow depths are wider, more numerous, and more interconnected.  Nevertheless,
municipalities sometimes derive high volumes of water from wells located in fault zones that
extend to depths measured in miles.

       EPA (1991a)  discusses several  basic differences between fractured bedrock aquifers and
unconsolidated, granular aquifers (e.g., sand aquifers).  In unconsolidated, granular aquifers,
flow tends to be slow, laminar, and predictable using Darcy's Law. Such aquifers can more
easily be assumed to  be homogeneous  and isotropic.  In contrast, flow through fractured bedrock
aquifers may be fast (most commonly 3 to 330 ft/yr; and sometimes over 3,000 ft/yr; USEPA
1987a), and Darcy's Law will often not apply. Flow takes place through fractures rather than
Source Assessment Guidance Manual               2-6

through pores between individual grains. Furthermore, fractured bedrock aquifers tend to be
more heterogenous (i.e., flow properties vary with location in the aquifer) and anisotropic (i.e.,
flow properties vary with the direction of flow) than granular aquifers. EPA (1991a) provides
more detailed descriptions of each these properties.

       Tracer tests have been used in several studies to estimate ground water flow rates in
fractured bedrock. Malard et al. (1994) report that dye traveled approximately 140 feet in a
fractured bedrock aquifer in 2 hours. Becker et al. (1998) report that water traveled
approximately 118 feet in about 30 minutes. Ground water velocities in fractured bedrock
aquifers are comparable to velocities in karst aquifers. Thus, fractured bedrock aquifers are
vulnerable to contamination by waterborne pathogens.

       As with other sensitive aquifers, the rapid ground water velocities in fractured bedrock
aquifers provide a means by which pathogenic bacteria or viruses can travel quickly from
contaminant sources to PWS wells.  It is important to note that the pumping of a PWS well
causes a greater increase in ground water velocities in  fractured bedrock than it would in a non-
sensitive aquifer. Recent cases of waterborne disease outbreaks due to contamination of wells
screened in fractured bedrock aquifers occurred in Couer d'Alene, ID (Rice et al.,  1999), Big
Horn Lodge, WY (Anderson et al., 2003; Getting et al., 2005) and northern Arizona (Anderson
et al., 2003; Getting et al., 2005).
2.4    Gravel Aquifers

       Gravel aquifers, as defined here, are unconsolidated water-bearing deposits of well-sorted
pebbles, cobbles, and boulders. Gravel aquifers consist primarily of coarse grains larger than
approximately 4 mm or approximately 0.16 inches in diameter, although they may have minor
amounts of smaller diameter material as well. Gravel aquifers are often limited in area and are
generally produced by high energy events such as catastrophic glacial outburst floods or
flash-floods at the periphery of mountainous terrain.  They can also sometimes be found at
fault-basin boundaries or in glacio-fluvial deposits such as crevasse fillings, eskers, kame
terraces, and outwash/valley trains.  Typically, these are small, relatively localized aquifers.

       Gravel aquifers are not particularly numerous, as compared with sand and gravel
aquifers, karst aquifers, and fractured rock aquifers.  Very few PWSs use this type of aquifer and
reports of outbreaks are correspondingly limited. No data are available that specifically
implicates a gravel aquifer as a contributing factor in a published waterborne disease outbreak.
Nevertheless, ground water velocities  in gravel aquifers can be quite rapid due to the aquifers'
large interconnected pore spaces,  offering little resistance to flow.  Such velocities increase
significantly under the influence of pumping wells. Gravel aquifers thus have the potential to
become contaminated with microbial pathogens and are therefore designated as sensitive.

       The following paragraphs  discuss the formation of gravel aquifers due to catastrophic
glacial outburst floods. This information may be useful to States in helping to identify potential
locations of sensitive gravel aquifers.  The discussion below focuses on the western United
States. Additional information on possible locations of sensitive gravel aquifers throughout the
country can be found in Chapter 3.

Source Assessment Guidance Manual               2- 7

       Repeated catastrophic floods, resulting from the breaching of large ice-dammed lakes
during glacial periods that ended about 12,000 years ago, are believed to be responsible for the
formation of the larger pebble, cobble, and boulder aquifers that are more widely distributed in
the western United States (Bretz 1925; Baker et al., 1987). Glacial Lake Missoula, one of the
largest glacial lakes of the Wisconsin Glaciation (the most recent glacial period), was estimated
to have a maximum water depth at the ice dam of about 2,100 feet (Pardee 1942, cited in Baker
et al. 1987). The area inundated by Missoula flooding as a result of the breaking of the ice dam
is hypothesized to include large parts of western Washington, Idaho, and Oregon (O'Connor and
Baker 1992).

       The Missoula floods exhibited exceptional sediment transport capability as evidenced by
the size of boulders entrained by the flood waters.  Baker et al. (1987) conducted field
measurements and performed calculations that estimate the peak Missoula flood discharges at
three or four orders of magnitude greater than the modern flood discharges of major rivers such
as the Amazon or the Mississippi. The floods produced very large pebble, cobble, and boulder
deposits including an approximately 15 square mile area of coarse gravel dunes near Spirit Lake,
Idaho.  The gravel dunes near Marlin, Washington are approximately 6.5 feet high and 200 feet
apart, and probably formed in response to 200 foot deep flood waters.  The largest Glacial Lake
Missoula discharges likely occurred through the Spokane Valley-Rathdrum Prairie area,
resulting in as much as 500 feet of flood deposits over an area of about 350  square miles
(O'Connor and Baker 1992).  Peak discharge through the valley is estimated at approximately
600 million cubic feet per second (O'Connor and Baker 1992).

       Glacial lake outburst flooding on a variety of scales occurred in other areas of the United
States that were at the ice margin during the Wisconsin Glaciation. For example, a gravel
aquifer is associated with glacial flooding along the Umatilla River in Milton-Freewater,  Oregon.
 Some of these regions are further discussed in Chapter 3.
2.5    Hydrogeologic Barriers

       A hydrogeologic barrier is defined as the physical, biological, and chemical factors,
singularly or in combination, that prevent the movement of viable pathogens from a contaminant
source to a water supply well. Where present, a hydrogeologic barrier may protect wells
screened in sensitive aquifers (i.e., those in karst, fractured bedrock, and gravel aquifers) from
microbial contamination. States have the option to investigate and verify the presence of an
adequate hydrogeologic barrier for sources located in hydrogeologically sensitive settings.  If a
thorough investigation proves that the proposed barrier is protective of the well, then the State
may use that information in making further decisions such as evaluating the need for assessment
source water monitoring.

       A confining unit, a common example of a hydrogeologic barrier, is a low permeability
subsurface stratigraphic layer that overlies an aquifer and acts to prevent significant infiltration
to the aquifer.  Low permeability strata often consist of unconsolidated clay or silt, or their
consolidated counterparts, shale and siltstone, but they may also consist  of other lithologies
(USEPA 1991b). For example, certain glaciated areas of the United States are underlain by
cemented till, which is relatively impervious and acts as a hydrogeologic barrier. In order to
prove that a given well is protected by a confining unit or other  hydrogeologic barrier, it will
Source Assessment Guidance Manual               2-8

generally be necessary to conduct a thorough investigation of the surrounding area. A successful
investigation will usually need to include evidence that the hydrogeologic barrier for a specific
well protects at a minimum the aquifer over an area that includes the zone of influence when the
well is pumping.

       Confining layers may be discontinuous.  They may also be breached by natural processes
(e.g., fractures) or anthropogenic activities (e.g., improperly constructed or abandoned wells).
Confining layers are rarely absolutely impermeable.  Instead, a confining layer is simply
characterized by a hydraulic conductivity that is orders of magnitude lower than that of an
adjacent aquifer.  Confining layers may, in fact, be leaky in that they slowly transmit water from
one aquifer to another.

       Other environmental conditions that may act to prevent the movement of viable
pathogens to an aquifer include the following:

       1.  Sufficiently long subsurface horizontal or vertical ground water travel times
          (especially due to vertical flow through a thick unsaturated zone) so that pathogens
          become inactivated as they travel from a source to a public water supply well.

       2.  Site-specific physical and chemical (and perhaps even biological) properties of the
          aquifer and ground water which may serve to decrease the longevity of particular
          microbes, increase their adsorption to aquifer material, or otherwise decrease the rate
          at which they are transported to a public water supply well.

       It is important to emphasize, however, that the geochemical factors which affect virus
and bacteria fate and transport are complex, poorly understood, and vary significantly with virus
and bacteria type. Furthermore, in sensitive hydrogeologic settings, pathogens have fewer
opportunities to interact with aquifer material  as compared with other hydrogeologic
environments, such as those consisting of fine-grained, unconsolidated sediments.  The potential
for biological predation to provide a hydrogeological barrier to pathogen contamination is
discussed further in section 3.5; however, there is little documentation of such predation on
waterborne pathogens in the scientific literature. Thus, the use of geochemical or biological
conditions as a basis for determining an adequate hydrogeologic barrier is present is likely to
require detailed, site-specific field investigations. Continuous, non-leaky confining units  or very
thick unsaturated zones are more likely to be proven adequate hydrogeologic barriers.

       The presence or absence of a confining layer is sometimes difficult to determine. The
office procedures for determining presence or absence of a confining layer protecting a sensitive
aquifer apply equally to determining the presence or absence of a confining layer protecting a
shallow aquifer of any kind. The following text provides information to differentiate an
unconfmed from a confined aquifer.

       There is no established permeability (or hydraulic conductivity) range for confining
strata. Low permeability rocks typically have permeability values below 10"3 cm/sec (10"5 ft/sec).
Permeability of a confining unit is typically three orders of magnitude lower than the
permeability of the producing aquifer (Freeze  and Cherry, 1979). Confining layers can be
extremely variable in composition so that permeability (and confining performance) can vary

Source Assessment Guidance Manual               2-9

significantly in all directions. In general, confining layers that are formed by deposition in open
marine environments are the most homogeneous type of confining bed (USEPA 1991b).

       Rates of vertical leakage are an important consideration in differentiating highly confined
from semi-confined aquifers (USEPA 1991b). Rates of vertical leakage can be calculated using
Darcy's Law. Information needed to perform the calculation include: water level for the
confined aquifer; water level at the water table; vertical  hydraulic conductivity; and confining
bed thickness. The equation is presented in EPA (1991b) as equation (2). Unlike horizontal
hydraulic conductivity data for most Darcy's Law applications, vertical hydraulic conductivity
data are often very difficult or expensive to obtain.

       Vertical travel time calculations may be used to differentiate semi-confined from highly
confined aquifers (USEPA 1991b). However, the method is not appropriate for identifying
confining beds and is difficult to implement.  Implementation problems arise because little data
is generally available on porosity and vertical permeability of confining beds. Other data, less
difficult to obtain, is also necessary, such as confining bed thickness and hydraulic gradient
across the confining strata. Because these data are not typically available, the travel time method
(EPA, 1991b)(see Appendix B) should be conducted together with other verification methods,
such as age-dating with tritium analyses (see Appendix A), to reduce overall uncertainty.

       EPA (1994) identifies 14 indicators of confinement and the characteristics used to
identify the presence of a confining layer (measured in the confining layer or in the aquifer, as
specified). More detailed discussion about each of the 14 indicators is presented in EPA
(1991b). Some of the methods described in EPA (1991b) may not be appropriate for confined
karst or fractured bedrock aquifers. More detailed information is necessary to apply these
methods to a particular hydrogeologic setting. The fourteen indicators are:

       1. Geologic maps and cross-sections showing the presence of a continuous, suitable
          confining layer such as clay  above the aquifer.

       2. Water level elevation above  the top of the aquifer, as measured in a single well
          screened in the aquifer.

       3. Hydraulic head differences between two (with one being karst) aquifers as measured
          in wells cased to and open in the differing aquifers.

       4. Water level fluctuations in the aquifer as the result of barometric or tidal effects but
          with no response to infiltrating precipitation  and recharge.

       5. No changes in water level in the aquifer in response to large pumping stress and
          diurnal water level fluctuations.

       6. Pump test with storativity value for the aquifer calculated to be less than 0.001.

       Note: Storativity values for confined aquifers may range from 0.005 to 0.00005; much
lower than values for unconfmed aquifers (which range  from 0.01 to 0.30) (Freeze and Cherry,
1979). The concept of storativity was originally developed for analyzing well hydraulics in
confined aquifers. It is defined as the volume of water that an  aquifer releases from storage per
Source Assessment Guidance Manual              2-10

unit surface area per unit decline in hydraulic head due to pumping. Karst limestone or fractured
bedrock aquifers do not generally release much water from storage, even when unconfined.
Because these aquifers may have very low storativity values even when unconfined, storativity
may not be useful for the purposes of identifying confining layers in these aquifers.

       7.  Leakage pump test (in the aquifer) plotted as drawdown versus time matches
          analytical solutions (calculated leakage less than 10-3 gal/day/ft2).  However,
          calculation of leakage values for well fields is not routine and there may be limited
          information available.

       8.  Three-dimensional numerical ground water flow model with good parameter input
          data (and appropriate number of nodes) showing low leakage across confining layer.

       9.  Water chemistry in the ground water indicative of long distance from recharge.

       10. No anthropogenic atmospheric tracers such as detectable tritium or fluorocarbons in
          the ground water.  EPA (1991b) notes that tritium analyses may be inappropriate for
          the case of a confined limestone aquifer, where horizontal flow may be fast enough
          that ground water contains tritium from lateral recharge and not from vertical leakage
          through the confining bed.

       11. Isotope chemistry showing carbon-14 dates greater than 500 years in the ground

       12. No detectable contaminants in the ground water identified based on inventory of
          contaminant sources.

       13. Long-term head decline due to pumping does not cause accompanying water
          chemistry changes in the ground water indicative of vertical leakage.

       14. Time of travel through confining strata that exceeds 40 years, where the travel time
          calculations are based upon hydraulic head gradient, porosity, and hydraulic
          conductivity measurements or estimates for the confining layer.

       EPA (1991b) provides recommendations for the methods  that are most appropriate for
evaluating confinement.  The most important recommendation is that the determination be based
on an integration of geologic, hydrologic, and hydrochemical approaches. The geologic
approach is necessary to determine whether there is a confining bed and whether there are
pathways through that bed. The hydrologic and hydrochemical approaches document whether
there is actually leakage through the confining bed. Collecting both hydrologic and
hydrochemical data allows for a comparison of the results from one approach with the results
from another. Of the available hydrologic methods, those based upon water level data (including
continuous recorder data) and potentiometric surface data are most useful as such data are the
easiest and least expensive to obtain (USEPA 1991b).  Of the hydrochemical data, tritium
analyses are the most useful (USEPA  1991b).

       The Groundwater Section of the Illinois Environmental Protection Agency has developed
a list of five diagnostic properties to determine if a well is protected by a confined aquifer
Source Assessment Guidance Manual             2-11

(ILEPA 1995). According to the list, wells are most likely to be protected if they satisfy the
following criteria:

        1.   At least one contiguous unit of impermeable geologic materials greater than 10 feet
           thick overlies the aquifer (excluding the top 10 feet of soil materials).

       2.   The top of the uppermost aquifer is greater than 50 feet from the surface.

       3.   The static and pumping water levels of the PWS are above the top of the aquifer
           (using the most recent data).

       4.   The well  is located in an upland (i.e., non-alluvial/outwash) geologic setting.

       5.   The storativity value for the aquifer is less than or equal to 0.001.

       The Illinois Environmental Protection Agency uses the above criteria in a weighted
scoring process in determining whether or not an aquifer is confined. Out of a total possible
score of 10 points, criteria  1 and 2 are weighted with 3  points each; criterion 3 is worth 1 point if
just the static water level is above the top of the aquifer and 2 points if both the static and
pumping water levels are above the top of the aquifer; criteria 4 and 5 are worth 1 point each. If
a public water supply well  meets criteria 1 and 2, and also amasses at least 6 points in the scoring
process, it is determined to be protected by a confining unit. Exhibit 2.4 provides a national
perspective of aquifers and their locations.
Source Assessment Guidance Manual               2-12

                      Exhibit 2.4 Aquifers of the United States
Source Assessment Guidance Manual

                                        UrhCOK-iaolij  Liiisi 1-1- H II ^icji-iiti=Ttt

                                                 Sti-Hk.?- River Pi-am toa^-ir-i-ft II -aciuilers
                                                 Willamette Lowt-and txasin-IMI e» q LJ 1 11 er

                                                 hJorthmrn R-ockv MTJ'-irn.-»h-.!:
                                                    I rii«=irrT-ir«i-i|-:irn=" Gtasi n^. B
                                                       •  -il : .«t:) -il-ir.n- ;:  ;;. |  i  |.  r

                                                       "fc''jimooK« ;>ci n if-nr

                                                         hldfiOJioiL: ki^sin  a •_)!_! i ft? rti
                                                 •C;*rntif-inn OrrJn^-jnj
                                              3   Upper Tertter-y aQLiilece

                                          ii r* d & t cn-^e" tind -Ci*rl>inni*t^- ro na
                                                 Art>u chl

                                                 Sil urian—

                                                 I'.T) li-i'^-llTI. ll'l  I.  I I".  -",

                                                 LJpprsr ci*rkH-n>rtnrtin m^uifiir

                                                 FlcimJu ii  anrt>orvai@-ro

                                                 Piedmont  amd  B-lue- Ri-.1-a-« c-artt^on^it-e-ricichi

                                                 Orasll K- Hay-ii^; n t_| LJ i f «=• r
                                                  -..».-.  Rive-r Pl^in t>-3solttc-rocK -gcii.n Core

                                                 OnKimhi^ Plntnnn tios.-alt tr.-rrn-k  nr|i li-f^rn

                                                 M^iwnii:«n wolnn n in-rr>r:k .-( <-- 1 ..= i -f •-- r r.
                                        Other   RdkK cH.-vt iirc^ mininrall-y n^rmf^-atilci but msry
                                                  onlEam loc-ally p-roc^uctiv« .=1 n n 1 1 •= r v
                                      Source:  National  Atlas of the United States, March 5,  2003,
Source Assessment  Guidance Manual

                     3. Hydrogeologic Sensitivity Assessments
       As discussed in section 1.1, States may use hydrogeologic sensitivity assessments
(HSAs) as a tool to determine if GWSs are obtaining water from hydrgeologically sensitive
settings. Desktop HSA approaches are emphasized in this guidance document because, in most
cases, States can determine a PWS's aquifer type without conducting a field investigation. Field
investigations can be conducted if desktop analyses provide insufficient information to complete
an HSA. Because a desktop analysis depends on the availability of reliable data and
information, hydrogeologic data sources are discussed first.  The desktop HSA approaches are
then presented, and case studies are included to demonstrate how data generated through Source
Water Assessment Program (SWAP) efforts and/or other water resources investigations can be
used to support HSAs. In particular,  section 3.3.2 demonstrates how in the City of Twin Falls,
Idaho, a Source Water Assessment Report for the Blue Lakes Well Field, a community drinking
water system, provides all of the necessary information for completing an HSA.  Similarly, the
Source Water Assessment Program provides documents regarding the community water system
in Trenton, Kentucky (discussed in section 3.2.2), which also contain all of the information
necessary to complete an HSA.

       This document includes three aquifer types as sensitive: karst, fractured bedrock, and
gravel.  Such designations  allow States to simply identify the aquifer type of a ground water
source for sensitivity assessments, instead of conducting detailed hydrogeologic investigations to
characterize each source's  ability to rapidly transmit potential contaminants. A State can
designate other sensitive hydrogeologic settings if it chooses. States can also choose to
demonstrate that a hydrogeologic barrier is protective of a public water system (PWS) well by
showing that the sensitive aquifer is adequately protected by the barrier over an appropriate area
surrounding the well.

       Among the many risk factors applicable to assessing the likelihood for fecal
contamination of a PWS well, one risk factor, sensitive aquifers, is perhaps most important. To
illustrate this importance, consider a geologic map of the Michigan basin. Such a map can be
created at http://nationalatlas.gov. Click "Map Maker." Click "Geology."  Click "Geologic
Map."  Then click on the area of the  map to be zoomed (Michigan, in this case).  The map shows
a large geographic area extending from southern Ohio to northern Michigan and northern
Ontario. The boundary between rocks of Silurian and Devonian age is marked as an arc that
extends from Ohio to Ontario to Michigan. As one traces the boundary with a finger, it passes
over the locations of three PWS waterborne disease outbreaks. Despite being widely separated,
these three outbreaks have one common feature; they all are located in wells that produce from
the Upper Silurian Bass Island Formation, a karst limestone aquifer. The outbreaks in South Bass
Island,  Ohio, Walkerton, Ontario and Drummond Island, Michigan are not random occurrences.
Rather, they are associated with a specific aquifer type. Another example is found in Texas,
where outbreaks in New Braun, Georgetown, and Brushy Creek, Texas are all associated with
the karst limestone Edwards Plateau  aquifer. Because karst limestone and other sensitive
aquifers represent a significant risk factor, this document provides additional information to
identify sensitive aquifers.
Source Assessment Guidance Manual               3-1

3.1    Identifying Aquifer Types

       The geologic and hydrogeologic characteristics of karst, fractured bedrock, and gravel
aquifers are reasonably well known. Such knowledge allows the identification of sensitive
settings using high resolution data and a desktop analysis. This section outlines some diagnostic
characteristics and general approaches for identifying each sensitive aquifer type without
conducting field investigations. If a desktop analysis is not possible due to limited data
availability, the diagnostic characteristics noted below for each aquifer type can help guide the
types of field investigations that are needed to make a sensitivity determination. Field
investigations are likely to be needed only on rare occasions. Karst aquifers comprise a large
percentage of all productive aquifers in the United States. When considering hydrogeologically
sensitive aquifers only, karst aquifers make  up an even larger percentage of the category.  Thus,
this guidance manual provides more detail on the complex flow regimes and diagnostic
characteristics of karst regions than it does for the other two sensitive aquifer types (fractured
bedrock and coarse gravel deposits). Karst  aquifers are the sensitive aquifer type most likely to
be encountered by the largest number of water systems. Nevertheless, karst, fractured bedrock,
and gravel aquifers are considered equally sensitive. Each of these aquifer types may pose a risk
of pathogen contamination to PWS wells located within them.
3.2    Karst Regions and Aquifers of the United States

       This section discusses U.S. karst regions and aquifers in general, although only those
developed in limestone bedrock are considered sensitive. On the other hand, potentially soluble
rocks such as dolomite and evaporites are not explicitly considered sensitive.  Again, States may
choose to consider all carbonate and other soluble rock aquifers as possibly karstic and designate
them sensitive.

       A map of the United States showing karst regions was presented by Davies et al. (1984)
at a scale of 1:7,500,000. Detailed descriptions of the major U.S.  karst regions accompany the
map. A list of significant karst aquifers was also published by EPA (1997).

       About  20 percent of the United States is underlain by karst aquifers of various types.
Karst aquifers underlie almost 40 percent of the United States east of the Mississippi River
(Quinlan 1989).  Where soluble rocks such as limestone and dolomite are present at  or near the
surface, solution openings are common.  A region's topographic relief, soluble rock  thickness,
and hydrogeology determine the depth to which solution openings occur. For example, the
vertical extent of solution openings is known to be as great as 1,100 feet in mountainous areas of
the western United States. In the eastern United States, where topographic relief is lesser, depths
are generally less than 400 feet, with  a maximum of 650 feet. Beneath many broad river valleys
throughout the country, solution features in carbonate rocks extend to depths of approximately
100 feet (Davies et al. 1984).

       In the region of the United States that was formerly covered by Pleistocene ice sheets
(glaciers), karst-related caves and fissure openings may have been partially filled by glacial
debris. The southernmost advance of the ice sheets covered New  England, New York, northern
New Jersey, northeastern and northwestern Pennsylvania, most of the States bordering the Great
Lakes, and much of the land north of the Missouri River (Davies et al. 1984).  Because of glacial
Source Assessment Guidance Manual                3-2

erosion or filling, caves in glaciated terrain typically have lengths of less than 1,000 feet. South
of the formerly glaciated areas, caves and other solution features are more identifiable, and in
general, the number and size of solution features increases with decreasing latitude.
Furthermore, the number of solution openings varies according to the age and structure of the
soluble rocks.  More deformed or folded rocks (deformed as part of mountain building
processes) may have fewer solution openings than younger, undeformed soluble rocks.  Solution
openings are most highly developed in Mississippian or younger limestones (Davies et al. 1984).

       The carbonate bedrock may be present at (or directly beneath the soil cover of) the land
surface in many karst regions (Winter et al. 1998). The Edwards Aquifer in south-central Texas
is an example of such a region. Across this area, enlarged fractures, solution cavities along
fractures, and sinkholes are present at (or very near) the land surface and extend down into the
bedrock. Precipitation in such karst regions (particularly when the precipitation falls on
outcropping bedrock) tends to infiltrate the land surface rapidly, seeping downward and acting as
a source of recharge to the underlying aquifer. A considerable amount of recharge to karst
aquifers can also be provided by streams that flow across local karst features.  Water in karst
region streams can flow out of the base of the stream as it encounters the karst features, rapidly
infiltrate downward, and thereby be lost from the surface as it recharges the underlying aquifer
(i.e., a "losing stream").  Even the largest  streams that originate outside a karst region outcrop
area can be dry within the outcrop belt for most of the year (Winter et al. 1998).

       In other karst regions, the karstic limestones or dolomites are covered with a veneer of
material that can obscure some of the karst features. This is referred to as a "mantled karst." An
example is northwest Arkansas, a region characterized by many springs, seeps, losing streams,
and a regolith mantle of variable thickness (Brahana J.V. et al. 2002). As karst solution features
enlarge in subsurface limestones or dolomites, the underlying bedrock support is removed, and
the overlying cover or mantle material can slump into  the karst solution features. In these cases,
even if the mantle material may act as a confining unit, the resulting slumpage breaches the
confining unit and water can readily infiltrate the underlying karst aquifer.

       Mantled karsts are quite susceptible to slumpage. In Florida, most land-surface
depressions containing lakes formed by the slumpage  of unconsolidated surficial deposits into
sinkholes that are caused by dissolution of the underlying limestone (Winter et al. 1998).
Shallow ground water can quickly flow through such a lake and into the underlying karst aquifer.
 In this way, ground water flowing through mantled karst aquifers may transport fecal
contamination from surficial water bodies to wells.  In northwest Arkansas, samples taken from
springs flowing from mantled karst were found to be contaminated with fecal coliform (Whitsett
etal. 2001).
3.2.1   Diagnostic Characteristics

       Although lithology cannot be used to definitively identify whether or not an aquifer is
karstic, the presence of carbonate rocks, especially when subareally exposed in a humid
environment, generally indicates the existence of a karst aquifer (USEPA 1997). For the purpose
of simplifying the HSA process, this document presumes that all limestone aquifers are karstic,
and therefore sensitive, while presuming that all other soluble aquifers are not karstic and not
sensitive.  The information in this section may be useful to States because they may choose to
Source Assessment Guidance Manual              3-3

designate karst aquifers developed in other carbonate or soluble rocks (e.g., dolomite and
gypsum) as sensitive. Furthermore, the characteristics diagnostic of karst regions and aquifers
may be useful for HSA determinations in cases when a State may not have definitive information
on a given PWS's lithology (i.e., whether it is limestone).

       This section presents geomorphic, geologic, and hydrogeologic characteristics of karst
regions and/or aquifers.  This information may be useful for conducting desktop hydrogeologic
sensitivity assessments.  Information that can be gleaned from maps, aerial photos, and other
spatial data sources are discussed first, followed by a brief discussion of how karst hydrology
may manifest itself in a variety of existing hydrogeologic data sources and reports.  Distinctive
karst landforms produced by the dissolution of underlying carbonate bedrock may be the first
clue that a locality's PWS is drawing water from a karst aquifer.  Nevertheless, site-specific
hydrogeologic data is the most direct evidence.

       EPA (USEPA 1997) suggests that karst regions can be identified by the presence of one
or more of the following land surface features: sinkholes; springs; caves; sinking or losing
streams;  discontinuous drainage networks;  and  dry valleys in humid climates.  Field surveys are
the  most reliable way to identify karst features, but many are also recognizable on sufficiently
detailed aerial photos or topographic maps.  Stereoscopes (optical instruments used to create
three dimensional images from stereoscopic aerial photographs, see section 1.5.5) can be used to
facilitate interpretations of stereoscopic aerial photographs. Identifying karst features using
topographic maps may be more challenging.  Nevertheless, contour lines representing closed
surface depressions (sinkholes) and discontinuous drainage networks (including disappearing
and reappearing streams) are indicative of a karst landscape (see Exhibit 2.2).  Local names for
natural features may also be noted on a map and can provide clues that an area is influenced by
karst hydrology. For example, the name "Lost River" is often a good indication that the stream
in question is located in a karst landscape, and disappears into the subsurface.

       Geologic and geomorphic characteristics diagnostic of karst regions and/or karst aquifers
may also be described in existing data and reports. For example, the following features are
indicative of karst terrain and may be noted in reports describing the field reconnaissance of a
particular area: dissolution-enlarged fractures or bedding surfaces; karren (dissolutional,
subaerial, water-carved grooves in rock, commonly subparallel); and grikes (soil-filled,
dissolutionally-enlarged fractures or grooves; also known as cutters or soil karren) (USEPA
1997). These features may  be seen in outcrops or road cuts, or encountered through drilling.

       Hydrogeologic data compiled and described in existing reports may also reveal
characteristics indicative of karst hydrology.  For example,  a karst aquifer may be identified
using aquifer pump test results for a well penetrating, and open in, the aquifer.  Continuous
pumping data showing stepped drawdown in the pumped well, or one or more nearby
observation wells, suggests that there is flow through conduits or fractures.  However, the
converse (a smooth drawdown curve) does not always indicate a non-karstic aquifer because
karst aquifers are highly variable (USEPA  1997). Stepped drawdown may not be observed in a
well that does not intersect sufficient numbers of solution-enlarged fractures or conduits,
although drawdown data collected for another well nearby may show entirely different results.
Source Assessment Guidance Manual              3-4

       EPA (1997) suggests that the following hydrogeologic characteristics, possibly
documented in existing data and/or reports, can be used to identify karst (or fractured) aquifers:

       •  Stepped drawdown during continuous pumping.

       •  Irregular cone of depression, or no cone of depression, as defined by multiple
          observation wells around a pumped well.

       •  Drawdown plotted against discharge indicates non-linearity.

       •  Stair-stepped, irregular configuration of potentiometric surface.

       •  Bimodal hydraulic conductivity data (when plotted as the logarithm of hydraulic
          conductivity) for a suite of wells completed in the same formation. Note that
          unimodal hydraulic conductivity is also possible in a karst aquifer.

       •  Differing water levels in closely adjacent wells.

       •  Bimodal or polymodal distribution of temporal specific conductance data from a
          given well.

       •  Significant spatial variation in specific conductance and hydraulic conductivity as
          interpreted from well logging devices.

       •  Significant variation in well discharge distribution during constant rate pumping.

       •  Significant variations in the distribution of flow with depth, for a pumped or
          unpumped well, measured using electromagnetic or thermal flow meters.

       •  Significant differences in tracer breakthrough depending on the location of the
          injection well (given recovery at the same well).
3.2.2   Desktop Approaches

       It is feasible to determine a PWS well's aquifer type without field investigation in most
cases.  A variety of hydrogeologic data sources (discussed in section 1.5) may provide the
necessary information.  In some cases, one data source may suffice (e.g., an existing study of the
PWS well of interest may identify the well's aquifer type).  Nevertheless, determining aquifer
type in a desktop analysis will often require multiple data sources used together.  This guidance
document recommends a simple step-like approach to determining aquifer type where the most
directly relevant information is used first (e.g., primary data sources such as driller's logs,
aquifer maps, geologic maps, and existing hydrogeologic investigation reports). If such data
sources are insufficient or unavailable, an HSA can progress to using less direct information
(e.g., soil surveys, topographic maps, and aerial photos) to facilitate an aquifer type
Source Assessment Guidance Manual              3-5

       This section illustrates the step-like HSA approach using two case studies.  The first case
study shows how primary information gathered in a wellhead protection study (i.e., a driller's
log) identifies the aquifer type for an undisinfected well.  Similarly, the second case study
illustrates how information available from State SWAP reports can be used for an HSA. EPA is
encouraging States to use information from their source water assessment efforts (i.e., through
their WHPPs and SWAPs) as they conduct their HSAs, and to coordinate efforts among these
programs wherever possible.

Case Study #1 -Fincastle, Virginia

       Fincastle, Virginia is located in Botetourt County, a mostly rural area of approximately
30,500 residents near Roanoke, Virginia. The Town of Fincastle, Botetourt County's historic
county seat, owns and operates a small PWS that served 975 people using two wells at the time
of the study described below. Rapid growth in this region is likely to have placed additional
demands on the PWS. The PWS wells were constructed in the mid-1970's after rural private
wells that pumped from the region's shallow aquifer experienced hydrocarbon and fecal
contamination.  The two PWS wells tap a deeper aquifer that was not affected by the shallow
aquifer's contamination. Available data for these wells include design and construction
information (e.g., well depth and diameter), a driller's log, pump tests, and well yields.  The
PWS's service area includes residences, businesses, government offices, churches, a nursing
home, and three  schools (Virginia DEQ 1993).

       A wellhead protection study for the two Fincastle  wells identified farm animal wastes and
residential septic systems as potential fecal contaminant sources.  The study also noted that State
regulations allow septic systems to be located as close as  75 feet from a PWS wellhead (Virginia
DEQ, 1993). One of the Fincastle PWS wells is located 100 yards from a commercial gas station
and across the street from a farm that  obtained a permit for a septic system in the early 1990s
(Virginia DEQ 1993). Clearly, potential sources exist for fecal contamination of the Fincastle
PWS wells.

       The first  well, Fincastle #1, is drilled 400 feet into dolomite and limestone. Fincastle #2
penetrates to a depth of 475 feet in the same rock types. Both wells intersect solution cavities
and water-bearing fractures according to the wellhead protection study (Virginia DEQ, 1993):

          The Driller's Log indicates two caves or openings with muddy water were
          encountered from a depth of 125 feet to 130 feet and from 190 feet to 195
          feet, and both of these openings were cased off. The material from a depth of
          195 feet to 470 feet was described simply as limestone  with "broken
          limestone" zones from 405 feet to 407 feet and from 428 to 431 feet.  A cave
          opening was reported at the bottom of the well, for the  depth interval 470-475
          feet (Virginia DEQ 1993).

       Based on the driller's log, the aquifer is determined to be karst.  Additional information
from other sources, such as the Virginia geologic map and the Botetourt County Soil Survey,
help confirm the determination of this well's hydrogeologic setting.

       According to the Virginia Department of Health Environmental Engineering Field Office
in Lexington, Virginia, Fincastle well #1 is not chlorinated. Fincastle well #2 is chlorinated. An
Source Assessment Guidance Manual               3-6

undisinfected well located in a sensitive hydrogeologic setting, such as the limestone and
dolomite karst from which both Fincastle wells are pumping, should be considered for source
water assessment monitoring. If the chlorination system at Fincastle well #2 does provide a four
log virus inactivation, it should also be considered for fecal indicator analysis.

Case Study #2 - Trenton,  Kentucky

       The City of Trenton, Kentucky operates a community water system (CWS) that serves
868 people (KYDEP 1999).  The CWS has an average daily withdrawal of 110,000 gallons and
is supplied by three ground water production wells.  Two of the wells were constructed in 1900
and 1936, respectively, and were probably the wells for Trenton's first public water supply.  The
third well was completed  in 1992 (KYDEP 1999).

       The State of Kentucky has completed a Phase IWHPP for the City of Trenton (KYDEP
1999), as part of its Source Water Assessment Program. The WHPP notes that the CWS's three
wells are drilled in a karst aquifer, and more specifically, a hydraulic conduit  (solution opening),
and includes a general discussion of the regional geology and hydrology. The report provides
farther hydrogeologic details, most  notably a well log for well #3 (drilled in 1992), indicating
that the city's wells are drawing from a limestone karst aquifer.  The well #3 driller's log
indicates that limestone bedrock is encountered at 51 feet. At 84.5 feet, a void is recorded that
continues to a depth of 85.5 feet. Limestone is encountered below the void to a depth of 90 feet,
which is the well depth, and the well log records "good water" for this interval (KYDEP 1999).

       Drillers logs are not available in the Phase IWFIPP for the City of Trenton's two other
wells.  However, Attachment 3 of the report includes water well inspection records for wells #1
and #2. These records indicate that all three of the City's wells are in very close proximity and
that wells #1 and #2 are both drilled to a depth of 85 feet. Comparing these depths to the well #3
driller's log indicates that these two wells are indeed drawing from a void (solution opening) in
limestone karst, supporting the narrative information found in Attachment 2.

       Although this CWS is discussed here as a hypothetical case for preparing an HSA, the
desktop analysis would (hypothetically) designate the aquifer from which this system's three
wells pump as a sensitive aquifer.
3.3    Fractured Bedrock Regions and Aquifers

       All igneous and metamorphic aquifers are considered fractured bedrock aquifers, and
therefore sensitive aquifers for the purposes of an HSA.  Unlike karst hydrogeologic settings,
regions underlain by fractured bedrock generally do not have characteristic topographic
expressions or landforms. An exception may be domal highlands with thin soils surrounded by
flat topography, examples of which include the Adirondacks in New York, the Llano uplift in
Texas, South Dakota's Black Hills, and the St. Francis Mountains of Missouri. Fractured
bedrock regions may have mountainous (e.g., the Sierra Nevada) or gently rolling topography
such as the metamorphic Piedmont region of the eastern United States. Therefore, topography is
not necessarily a diagnostic tool for identifying fractured bedrock regions. Instead, States are
encouraged to  use geologic maps to identify their igneous and metamorphic fractured bedrock
regions. These rocks may be exposed at the surface or, in the previously glaciated regions of the
Source Assessment Guidance Manual               3-7

United States, covered by a relatively thin veneer of sediment. In a few regions of the United
States, fractured bedrock is known to be overlain by sedimentary rocks, but is shallow enough to
be an aquifer. Sedimentary strata of Cambrian or Ordovician age, identifiable on geologic maps,
may thinly cover igneous or metamorphic bedrock in a few places in the mid-continent.
3.3.1   Diagnostic Characteristics

       This document presumes that any igneous or metamorphic aquifer is a fractured bedrock
aquifer and therefore sensitive. Therefore, the characterization of an aquifer that feeds a given
PWS well as igneous or metamorphic is sufficient to complete an HSA. Available data and
reports, however, may reveal (in cases where lithology is uncertain) additional aquifer
characteristics that are diagnostic of fractured bedrock aquifers and could support an HSA
aquifer type determination.

       EPA (199 la) suggests that the following data may be useful as part of the
characterization of an aquifer as a fractured versus a porous medium. Although this data can be
collected in the field when necessary, in many cases it will already be available to States through
the desktop data  sources discussed in sections 1.5 and 3.2.2.

       1.  Discharge-drawdown plots using aquifer pump test data.

       Hickey (1984, cited in USEPA 199la) suggests using an aquifer pump test with
incrementally greater discharge rates in the pumping well, accompanied by drawdown
measurements in observation wells measured at one-hour increments, to test whether an aquifer
is fractured or porous media. Plotting discharge versus drawdown on an arithmetic scale, a
non-linear fit to the data may suggest that fracture flow is occurring.

       2.  Time-drawdown plots using aquifer pump test data.

       Time-drawdown curves for observation wells located in two or more different directions
from the pumped well that have different shapes or sharp inflections may be indicative of
fractured bedrock aquifers.

       3.  Contour map showing points of equal drawdown.

       Maps showing lines of equal drawdown (drawdown contours) are compiled using
drawdown values measured at multiple observation wells. Linear or irregular drawdown
contours, rather than circular or elliptical contours,  are indicative of fractured bedrock aquifers.
Furthermore, Risser and Barton (1995) suggest that if water levels in multiple observation wells
decline, but the response is greatest at some distant well, then the aquifer may be a fractured
bedrock aquifer.  Also, if water levels in some observation wells do not decline in response to
pumping, while levels in other nearby wells decline, a poorly connected fracture network may be
Source Assessment Guidance Manual

       4.  Water table surface configuration.

       A "stair-step" water table configuration could be indicative of a fractured bedrock
aquifer. Such configurations occur in sparsely fractured rocks when there are large contrasts in
hydrogeologic properties between massive blocks and the fracture zones that bound them.

       5.  Hydraulic conductivity distribution.

       Aquifers with strongly bimodal hydraulic conductivity distributions may be fractured
bedrock aquifers.

       Black (1989), as cited in Risser and Barton (1995), cautions that tests to determine
whether an aquifer is fractured can provide misleading results because fractured bedrock aquifer
characteristics can change with time. For example, a non-linear response may occur at the
beginning of a well pump test, indicating flow from a single, planar, vertical fracture (i.e.,
one-dimensional flow). As the pump test continues, horizontal fractures begin to contribute to
the flow. The flow pattern becomes more two-dimensional, and a radial response in the aquifer
may result. Thus, a pump test analysis relies on evaluation of the early pumping results to
determine if the aquifer is fractured.
3.3.2   Desktop Approaches

       As noted above, this guidance document recommends a simple step-like approach to
determining aquifer type using the most directly relevant information first (e.g., primary data
sources such as driller's logs, aquifer maps, geologic maps, and existing hydrogeologic
investigation reports). If primary data sources are unavailable or do not definitively identify
whether an aquifer is composed of igneous or metamorphic rocks, less direct information may
help to confirm an identification and complete an HSA. This section illustrates the simple step-
like approach for completing an HSA for PWSs located in a fractured bedrock setting.  Case
Study #3, below, demonstrates how State source water assessment program data can be
integrated with other existing data, in this case a State-wide bedrock geologic map, to determine
aquifer type.  Case Study #4, on the other hand,  demonstrates that in some cases, a source water
assessment report alone can be used to complete an HSA.

Case Study #3 - Enfield, New Hampshire

       The State of New Hampshire is underlain almost entirely by igneous and metamorphic
bedrock.  In some places, the bedrock is at or very near the surface.  However, glacial
overburden of variable thickness is widely distributed throughout the State.  Ground water PWSs
in New Hampshire often draw water from glacial overburden aquifers, but some draw from
igneous or metamorphic bedrock aquifers (i.e., fractured bedrock aquifers).

       Enfield is a small, rural town in western New Hampshire located approximately 10 miles
east of the Connecticut River, which forms the border with Vermont. The Enfield Water
Department operates four active wells serving 1,125 people. The State of New Hampshire has
completed a source assessment report (SAR) for the Enfield Water Department (NHDES 200la).
 The SAR indicates that all four of the PWS's wells are bedrock wells, and that three of them
Source Assessment Guidance Manual              3-9

have well depths of at least 425 feet. Well depth is not reported for one of the wells. A GIS map
accompanying the SAR locates three of the wells in the northern part of Enfield, and one well is
shown farther northwest and just across the town line in nearby Canaan (NHDES 200la).
Consulting the State's 1:250,000 scale bedrock geology map indicates that the area in the
vicinity of the wells (and the whole region) is underlain by igneous and metamorphic rocks
(Bothner and Boudette 1997).  Therefore, given the fact that the wells are bedrock wells, and the
GWR's presumption that igneous and metamorphic aquifers are fractured bedrock aquifers,
Enfield's public water supply wells are drawing water from a fractured bedrock (and thus
sensitive) aquifer.

Case Study #4 - Blue Lakes Well Field, City of Twin Falls, Idaho

       The City of Twin Falls, Idaho has a community drinking water system consisting of 10
ground water source wells, four of which comprise the Blue Lakes Well Field, north of the
Snake River in Jerome County. These four wells are joined together in a manifold. Although
total coliform bacteria were detected in the water distribution system on four occasions between
1994 and 1998, no microbial contaminants have ever been detected in the samples that were
collected from the well manifold according to the Idaho Department of Environmental Quality
(IDEQ) (IDEQ 2002a).

       According to IDEQ (IDEQ 2002a), the Blue Lakes wells  draw from a regional aquifer
consisting of highly fractured, layered basalts of the  Snake River Group.   These basalts host one
of the most productive aquifers in the United States,  often yielding up to 3,000 gal/min for wells
screened in only 100 ft of the sometimes 5000 ft. thick flows. The aquifer is  unconfined over
most of its areal extent, although interbedded clays and small areas of less fractured basalt may
create confined conditions in some localized areas. The aquifer is recharged  by surface water
irrigation, stream losses, direct precipitation, and underflow from tributary basins. According to
well logs, the aquifer is 500 to 1500 feet thick in the region of the Blue Lakes wells, and is
overlain by 1 to 23 feet of sediment. In the area immediately surrounding the wells, the water
table is only about six feet deep. Local area well logs further indicate that the vadose zone is
predominantly fractured basalt. The Source Water Assessment Final Report, prepared by the
Idaho Department of Environmental Quality, contains all of this information  and also indicates
that the Blue Lakes wells are only about 20 feet in depth (IDEQ 2002a).

       Given that the aquifer from which the four Blue Lakes wells draw water is composed of
basalt (i.e., igneous rock), it is considered a fractured bedrock aquifer for the  purposes of an
HSA. Thus, this aquifer is considered sensitive,  assuming the City has not shown that a
hydrogeologic barrier exists for these wells.
3.4    Gravel Aquifer Hydrogeologic Settings

       Glacial Lake Missoula is thought to have produced some of the largest glacial outburst
flood deposits known in North America (see section 2.4).  Glacial lakes on a smaller scale
(although still large) also formed in other ice margin environments of North America during the
Wisconsin period.  The southernmost advance of ice sheets during the Wisconsin glaciation
covered New England, New York, northeastern and northwestern Pennsylvania, and much of the
areas north of the Ohio and Missouri Rivers. Recent research describes coarse gravel deposits in
Source Assessment Guidance Manual              3-10

Wisconsin that may be the result of subglacial-lake outburst flooding from the Laurentide Ice
Sheet (Cutler et al. 2000).

       Large lakes also formed in some of the interior basins of the Intermountain West during
glacial periods.  These lakes were not proximal to the margins of the continental glacier, but their
formation was related to the climatic conditions that caused, and were perpetuated by, the
continental ice sheets. Specifically, increased precipitation, less evaporation, and meltwater
from nearby alpine glaciers raised lake levels in many interior basins because they have no outlet
to the ocean.  An example was Glacial Lake Bonneville, the remnant of which is the modern, and
considerably smaller, Great Salt Lake in Utah.

       In contrast to Glacial Lake Missoula, flooding from Glacial Lake Bonneville is estimated
to have a peak discharge of about 35 million cubic feet per second.  Glacial Lake Bonneville
floods also produced coarse gravel flood deposits (O'Connor 1993). Bonneville flood deposits
may be aquifers along portions of the Snake River in southern Idaho, including areas mapped as
the Melon Gravel (Malde and Powers 1972 cited in O'Connor 1993) and the Michaud Gravel in
the Pocatello and American Falls areas (Trimble and Carr, 196la and  1961b, cited in O'Connor,
3.4.1   Diagnostic Characteristics

       As noted in section 2.4, catastrophic floods can produce coarse gravel deposits.  These
floods are typically associated with the rapid failure of ice-dammed lakes during glacial periods.
 Coarse gravel deposits can be produced by other processes, such as flash flooding in steep
terrain, but these deposits tend to be very small and localized, and are unlikely to form public
water supply aquifers. This section, therefore, will focus on some U.S. regions known to have
coarse gravel deposits produced by glacial lake outburst flooding or related proglacial outwash
processes. States are encouraged to consider the Quaternary  depositional history of a region
when conducting HSAs for PWS wells screened in unconsolidated aquifers. The Quaternary
period is defined by approximately the last 2 million years, during which glacial periods were
common. Quaternary glaciers influenced the modern landscape in large areas of North America.
 If a particular site has a geologic history of glacial lake outburst flooding, it is likely that a
coarse gravel aquifer is present, as opposed to an unconsolidated aquifer with significant
amounts of fine-grained material.
3.4.2   Desktop Approaches

       As noted above, this guidance document recommends a simple step-like approach to
determining aquifer type using the most directly relevant information first (e.g., primary data
sources such as driller's logs, aquifer maps, geologic maps, and existing hydrogeologic
investigation reports). If primary data sources are unavailable or do not definitively identify
whether or not an aquifer is composed of coarse gravel sediment, less direct information such as
county soil surveys may provide the information necessary to confirm an identification and
complete an HSA. This section illustrates the simple step-like HSA approach with a case study
demonstrating how State source water assessment program data can be integrated with other
existing data to determine aquifer type.
Source Assessment Guidance Manual               3-11

       Among unconsolidated (e.g., sand, sand and gravel, gravel) aquifers, only coarse gravel
deposits resulting from glacial outburst floods are considered sensitive aquifers for this guidance
document.  States have discretion to consider any unconsolidated aquifer a sensitive aquifer. As
in the above sections, the following case study illustrates how information gleaned from a
desktop analysis can be used to determine a PWS well's aquifer type.

       In the following case study,  most of the relevant information for completing an HSA can
be gleaned from the system's source water assessment report in combination with accompanying
maps (IDEQ 2000). Determining aquifer type from a desktop analysis may often require the use
of additional data sources.  Again, EPA is encouraging States to use information from their
source water assessment efforts (i.e., through their SWAPs) as they conduct their HSAs, and to
coordinate efforts among these programs wherever possible.

Case Study # 5 - Post Falls, Idaho

       The surficial geology of Post Falls, Idaho consists primarily of Rathdrum Prairie gravels,
deposited by the repeated catastrophic flood releases of Pleistocene glacial Lake Missoula,
discussed in section 2.4. Post Falls  is located 30  miles downstream from the former Clark Fork
ice dam (responsible for forming Glacial Lake Missoula), and 15 miles upstream from the lake
where most of the flood waters were channeled (Breckenridge and Othberg, 1998).

       The Riverbend Water Company is a non-community, non-transient public water system
in western Post Falls, which uses two wells to supply 26 connections with water in a
commercial/industrial area (IDEQ 2002b). It is among 186 water systems that draw their water
from the Rathdrum Prairie Aquifer.  This aquifer is the sole source of water for over 400,000
people (i.e., most of the residents  of Spokane County, Washington, and Kootenai County,

       According to the source water assessment report, the Riverbend Water Company's two
wells are screened at around 180 ft and 160 ft from the surface just north of the Spokane River in
Post Falls (IDEQ 2002b), which is located in the  flood gravel deposits (personal communication,
David Risley, Source Water Assessment Program lead, IDEQ).  An aquifer atlas that has been
sent to all water systems indicates that the entire  Spokane-Valley-Rathdrum Prairie Aquifer is
composed of thick layers of coarse-grained flood gravel (and cobble and boulder) deposits
(IDEQ 2000). The source water assessment report references well logs to show that there is
neither a thick unsaturated zone nor a confining unit in the vicinity of the wells (IDEQ 2002b).
Thus, the Riverbend Water Company's two public water system wells are good examples of
wells screened in a sensitive gravel  aquifer, with  no hydrogeologic barrier to negate the sensitive
aquifer designation.
3.5    Hydrogeologic Barriers

       States may wish to consider the presence of a protective hydrogeologic barrier for public
water systems (PWSs) in their evaluation of wells drawing water from sensitive aquifers and in
any follow-up activities. For example, a limestone aquifer and all wells producing water from
that aquifer may be designated as sensitive, but the State may also identify the presence of a
Source Assessment Guidance Manual              3-12

hydrogeologic barrier (i.e., confining layer) and may able to demonstrate that the barrier is
protective of the aquifer.

       All proposed hydrogeologic barriers should be carefully evaluated. For example, it
cannot be assumed that all confining layers are protective, are continuous over the area of
interest, or have identical properties.  Confining layers may be breached by unplugged water or
injection wells or may be excessively leaky, allowing rapid transport of fecal contaminants from
near-surface environments to the underlying aquifer.  It may be possible at some sites to use
tracer tests and/or pumping tests, described in Appendix A, to evaluate such situations.  The
difficulties associated with determining if site-specific geochemical conditions will provide an
adequate hydrogeologic barrier were reviewed briefly in section 2.5. In sensitive aquifers with
relatively unpredictable ground water flow, the identification of a barrier to pathogen
contamination will generally be a technically-based determination that uses as much site-specific
hydrogeologic information as possible. The following examples demonstrate why it is important
that hydrogeologic barriers be evaluated carefully.

       The Floridan aquifer system is the primary drinking water source for the Orlando, Florida
area.  It is a carbonate aquifer system overlain in the region by a confining unit composed of 150
feet of sandy clay, silt, and shell. The integrity of this protective layer is compromised because
Orlando has over 300 stormwater drainage wells that  are completed in the upper karstic unit of
the aquifer system (i.e., the Upper Floridan aquifer).  The operation of these drainage wells is
considered necessary for the disposal of excess surface water (including urban stormwater
runoff), but the water receives no treatment prior to injection into the aquifer system. Although
most of the Orlando area's public supply wells draw water from the deeper Lower Floridan
aquifer, separated from the Upper Floridan aquifer by a semiconfming layer, some of the water
wells are screened in the Upper Floridan (NRC 1994). Although widespread contamination has
not occurred, the Floridan aquifer system in the Orlando area is clearly a sensitive aquifer
(because it is limestone), and the 150 ft confining unit is not an adequate hydrogeologic barrier
(because it has been breached by the stormwater drainage wells).

       In a paper by Johnson et al. (2000), emphasis was placed on the importance of careful
investigation of site-specific conditions when evaluating a confining layer. Although a confining
unit overlies the Charnock well field in Santa Monica, California, aggressive pumping rates
twice as great as the aquifer's natural inputs dewatered a significant portion of the upper aquifer.
 This, in turn, caused water containing methyl tertiary butyl ether (MTBE) to flow toward the
well field from all directions and contaminate the source water (Johnson et al. 2000). This
situation  highlights the importance of examining pumping rates and other ways that
hydrogeologic barriers can be compromised. States are encouraged to use site-specific
approaches to identify hydrogeologic barriers and ensure that those barriers remain effective
during the expected range of pumping operations.

       Biological factors in an aquifer are considered in the definition of a hydrogeologic
barrier. This refers primarily to the possibility that certain microorganisms may be predators,
serving to reduce the concentrations in an aquifer of pathogens that may cause waterborne
disease.  Such a possibility should be evaluated on a site-specific basis because predation is
highly dependent on temperature, soil type, aquifer mineralogy, and other geochemical
conditions. Protistan grazing (i.e. grazing by protists) on ground water bacteria has been
investigated by Kinner et al.  (1998), who focused on the rates and size-selectivity of such
Source Assessment Guidance Manual              3-13

grazing.  Although the study found that up to 74 percent of the bacterial prey could be consumed
fairly rapidly, this study is not directly applicable to hydrogeologic barrier investigations because
the bacteria preyed upon in the study were not pathogens.  Rather, these bacteria biodegraded
chemical contaminants in ground water.  Nasser et al. (2002) investigated the role of microbial
activity in reducing concentrations of viruses in saturated soil.  The results of this study, though
highly dependent on virus type, are promising in terms of potential future applicability to studies
of potential hydrogeologic barriers.  Nevertheless, current research into the role of biological
factors as potential barriers to the transport of waterborne pathogens is at a very early stage.
Therefore, few, if any, PWS's will choose biological factors as hydrogeologic barriers.

       The following sections present approaches for identifying hydrogeologic barriers using
reliable data which may already have been collected  and through field investigation. The use of
site-specific field data is emphasized because hydrogeologic barriers are local phenomena,  and
the determination of their adequacy for protecting sensitive aquifers will often be difficult.
Additional details and references regarding the types of field investigations that may be
necessary are provided in Appendix A. Finally, travel time calculations may be informative if it
is suspected that there is a hydrogeologic barrier consisting of a very long flow path through the
unsaturated zone or through a confining unit. If the calculated travel times are longer than the
expected lifetime of waterborne pathogens (see Appendix  C), a hydrogeologic barrier may be
present. Methods for calculating travel times, including a  variety of computer models, are
summarized in Appendix B.
3.5.1   Data Sources for Hydrogeologic Determinations

       Many of the data sources discussed above will be useful resources for identifying
potential hydrogeologic barriers. For example, source water assessment reports often indicate
that a PWS well's aquifer is unconfined, in which case there will generally be no need to look
further for possible confining layers.  On the other hand, if documents produced by the Source
Water Assessment Program (SWAP) indicate that a confining unit is present and the State
wishes to pursue a potential hydrogeologic barrier investigation, EPA encourages States to use
the diagnostic criteria, described below in this section, to further evaluate the ability of the unit
to adequately protect the sensitive aquifer.

       Well logs available through the Wellhead Protection Program (WHPP) and SWAP or
directly from the driller may be the first indication that a confining unit is present.  Geologic,
hydrogeologic, and soil maps, if available, can be used to identify the presence, thickness, and
areal extent of a confining bed.  In addition, these maps may identify places and potential
pathways by which fecal contamination may leak into a sensitive aquifer. Under these
circumstances, the hydrogeologic barrier would not be effective in protecting the well and its
associated aquifer from contamination.

       Geologic maps, which depict geologic formations, may be used to determine
confinement. A confining formation is commonly composed of one predominant rock type, such
as shale, or of sediment, such as clay. The dominant rock or sediment type and the formation's
estimated ground water production rate may indicate whether it is an effective confining unit.
Again, the issue of map scale is relevant to the applicability of the information. This chapter
provides information on a variety of means to obtain such maps.
Source Assessment Guidance Manual              3-14

       Unfortunately, maps alone are unlikely to be adequate for determining whether a
hydrogeologic barrier exists and is protective. The most useful data sources for hydrogeologic
barrier determination will be those that describe the PWS wells and surrounding geology and
hydrogeology in as much detail as possible. Again, hydrogeologic barrier determinations will
generally be made on a site-by-site basis, and desktop analyses alone will often be insufficient.
Given that hydrogeologic barrier determinations are not required, the detailed desktop and field
investigations would be conducted at the State's discretion.

       Available information from sources such as existing State, Federal, or academic
hydrogeologic investigations; geologic, environmental, and hydrogeologic maps; or other data
sources may be useful for identifying the presence, thickness, and areal extent of a
hydrogeologic barrier.  They may also identify breaches in the hydrogeologic barrier that are
potential pathways for fecal contaminants to enter a sensitive aquifer.  However, ensuring that a
hydrogeologic barrier is functioning adequately  to prevent leakage of contaminated recharge into
a PWS source will very often require field data because hydrogeologic barriers are local
phenomena.  In most instances, some combination of site-specific field investigation and desktop
analysis based on available data will be necessary.
3.5.2   Desktop Approaches

       At a few sites, State, Federal, or academic institutions may have already conducted
significantly detailed field investigations that show the presence of an adequate confining unit,
sufficiently thick unsaturated zone, or other hydrogeologic barrier. In such instances, it may be
appropriate to conclude that a hydrogeologic barrier exists, without the need to conduct
additional field investigations.  On the other hand, if the study in question may be outdated, it
will be necessary to conduct additional research to ascertain that, for example, the confining unit
has not been breached in the recent past.
Source Assessment Guidance Manual              3-15


     4. Assessment Source Water Monitoring; Number (and Frequency) of
4.1    Introduction

       The GWR provides States with the option to require systems to conduct assessment
source water monitoring of any ground water source at any time, and States may require systems
to take corrective action based on the results of these analyses. Assessment source water
monitoring allows States to initiate a more thorough source water monitoring approach than that
resulting from the GWR triggered monitoring provisions on a case-by-case basis. Assessment
source water monitoring also allows States to address recent development of additional source
vulnerability based on sanitary surveys or other State oversight activities.

    States may identify high risk GWSs and require assessment source water monitoring based

       •  Information from Source Water Assessments and sanitary surveys;

       •  Information from Hydrogeologic Sensitivity Assessments (HSAs);

       •  GWR triggered monitoring results;

       •  TCR monitoring results;

       •  Well construction information (or lack of information);

       •  Historical or water quality data from the system; and

       •  Well located in a sensitive aquifer.

    The purpose of this chapter is to describe additional rationale suitable for  identifying when
optional assessment source water monitoring might be beneficial to protecting public health.
Included in this discussion are available data to guide decisions about monitoring number and
frequency, as well as information regarding sample locations,  indicator choice, and analytical
methods. More information on indicator choice and analytical methods can be found in the
Source Water Monitoring  Methods Guidance document. Corrective actions are described in
more detail in the Corrective Action Guidance document.

4.2    Connection to Hydrogeologic Sensitivity Assessment

     A Hydrogeologic Sensitivity Assessment (HSA) can be  an effective screening tool in
identifying GWSs with ground water sources that are susceptible to fecal contamination and for
which assessment source water monitoring would be appropriate and beneficial. The HSA
identifies wells located in  sensitive aquifers that should be targeted for assessment monitoring.
Chapter 3 describes the details of performing an HSA.  States  have the option to investigate and

Source Assessment Guidance Manual              4-1

verify the presence of an adequate hydrogeologic barrier for sources located in
hydrogeologically sensitive settings as part of a determination that assessment source water
monitoring is appropriate. Section 2.5 provides more information about evaluating
hydrogeologic barriers.
4.3    Assessment Monitoring Basis and Triggers

       As discussed in Chapter 1, wells may be identified as candidates for assessment source
water monitoring due to a variety of hydrogeologic and microbial monitoring factors. For
example, assessment source water monitoring may be appropriate for wells in sensitive aquifers
and wells with a history of total coliform occurrence (based on TCR monitoring). However,
other factors, not hydrogeologic or microbial in nature, may also indicate wells for which
assessment source water monitoring is appropriate. For example, wells located adjacent to
concentrated animal feeding operations, land spreading of manure or biosolids, or near municipal
landfills might be suited for such  monitoring. Indicators other than microorganisms, such as
MB AS, chloride, or nitrate/nitrite levels could indicate that anthropogenic activities are affecting
well water quality. Where high concentrations (above normal background levels  or based on
historic trends) of conservative tracers (contaminants that travel at the same velocity as ground
water) such as nitrate and chloride are found in well water, this indicates that there exists a
relatively efficient ground water recharge pathway that permits only minor contaminant dilution
and attenuation. Microbial pathogens  may be more likely to arrive as infectious agents when
such an efficient pathway is present. Thus, these wells could be subject to assessment source
water monitoring.
4.4    Assessment Monitoring; Number (and Frequency) of Samples

       Assessment source water monitoring is one of several barriers (e.g. State sanitary setback
distances, GWR sanitary surveys) designed to protect the public health of drinking water
consumers using ground water sources. Ground water travel paths are complicated, and it is
often difficult to establish that all ground water reaching a well emanates from a protected
aquifer and has resided in the subsurface for years, decades, centuries or longer. In general, a
small amount of ground water often takes the fastest path.  Therefore, even in the most protected
aquifer, there may be a small but significant component of recent ground water capable of
carrying pathogens.  As a result of this uncertainty, public health protection principles suggest
that source water samples be collected as necessary. In general, public health is best protected if
frequent source water samples are collected and assayed because each sample increases the
probability that infrequent source water contamination, if present, is identified.

       The added public health protection value due to collection of one or more source water
samples is an important issue. In particular, one key aspect is identifying the additional
information value of each or several source water samples. EPA evaluated this aspect using
available E. coli data presented in the Occurrence and Monitoring background document of the
Ground Water Rule  (USEPA, 2006d) and in the Ground  Water Rule Economic Analysis
(USEPA 2006e).
Source Assessment Guidance Manual

       Exhibit 4.1 evaluates the population of all wells randomly sampled for E. coli( using data
from the documents described above). In Exhibit 4.1, the horizontal line at 0.029 represents the
total E. coli proportion in untreated ground water. If all water produced by all wells could be
analyzed for E. coli  (100 ml at a time), about 2.9% of samples would be positive for E. coli.
Because it isn't feasible to analyze all water all of the time, Exhibit 4.1 displays the proportion
that would be identified if all wells were to assay the same number of samples (1, 2, ...24). The
exhibit shows that less than one-half percent of the total occurrence would be identified if all
wells were assayed only one sample each, but that nearly all of the occurrence would  be
identified if every well was assayed using 24 samples each. If each well were assayed using 5
samples, then nearly half of the wells would be identified. The exhibit shows that wells having
about 80% of the E.  coli occurrence should be identified by a positive E. coli result if 12 samples
are assayed for each well.

       Based on this analysis, twelve or more source water monitoring samples should be
collected from wells identified for assessment source water monitoring, including wells in
sensitive aquifers. This more frequent sampling will identify more than 80% of fecal
contamination occurrences. Assessment source water monitoring should be representative of the
system's typical operations. Using a minimum of 12 samples ensures sampling for each month
that most systems are in operation, and addresses the impact that seasonal events can  have on
contamination (e.g., heavy rain events). Sampling for seasonal systems should be equally
distributed (12 samples per season) or conducted during consecutive years and States may set
other sampling frequencies based on local conditions.
Source Assessment Guidance Manual

  Exhibit 4.1 Likelihood of Identifying E. co// Occurrence by Assessment Source
Water Monitoring in a Population of Wells Randomly Selected and Sampled (Data
                      from USEPA 2006d and USEPA 2006e)



           E. Coli Occurrence (overall and elminated thru assessment monitoring)
  '-g    0.01

                           Number of Assessment Monitoring Samples
4.5   Sample Location

      Ground water samples used for triggered or assessment monitoring must be collected at a
location prior to any treatment of the ground water source unless the State approves a sampling
location after treatment. If the system's configuration does not allow for sampling at the well
itself, the system may collect a sample at a State-approved location, provided the sample is
representative  of the water quality  of that well.

      Assessment source water monitoring sampling locations should be located as close to the
wellhead as possible. For the smallest systems with the shortest length distribution systems, it
might be expected that tap water and source water quality are similar; a sample at the tap might
be reasonable.
  Systems without sample taps at or near the well should install a sample tap
Source Assessment Guidance Manual

4.6    Representative Wells

       The Ground Water Rule allows systems, with State approval, to identify one or more
wells that are representative of multiple wells in the system for the purposes of GWR assessment
source water monitoring. The representative wells are wells that draw water from the same
hydrogeologic setting and that have the same, or greater, vulnerability to source water
contamination as the wells that are represented (e.g., distance to source of contamination,
pumping rates, etc.). Information that supports the use of representative wells includes drillers
logs, well construction details, and water quality data (e.g., pH, temperature and other physical
parameters, mineral analyses).
4.7    Indicator Selection

          Aquifers are broadly classified into two categories; porous and non-porous.  In this
document, non-porous aquifers (e.g., fractured igneous or metamorphic rock aquifers) as well as
gravel aquifers are defined as sensitive aquifers. Among the remaining porous aquifers (e.g.,
sand, or sand and gravel, aquifers), sand and gravel aquifers more efficiently transmit fecal
contaminants than sand aquifers because average ground water velocity is higher.

       All subsurface particles, including microbes, may be transported by flowing ground
water. Particles may be removed from flow or be retarded.  That is, they  may permanently or
temporarily become associated with the solid aquifer materials in either porous or non-porous
aquifers. Microbial transport in porous media aquifers is an active research area and  consensus
is difficult in many issues in this field. It is generally agreed that microbe size is an important
element in determining mobility in porous media, although many other factors, such as  surface
charge, may also have significant influence. The significant (one-thousand fold) size  difference
between viruses (measured in nanometers) and bacteria (measured in micrometers) increases the
likelihood, other mobility factors being equal, that an infectious virus rather than an infectious
bacterium, will reach a GWS well in a porous aquifer.

       In other aquifers, such as non-porous aquifers (e.g., fractured igneous or metamorphic
rock aquifers) and gravel aquifers, average ground water velocities are exceptionally  fast, and
straining and pore-size exclusion are much less significant and bacteria and viruses are  assumed
to travel at equal rates. In general, straining and pore-size exclusion effects are more significant
in sand aquifers than in sand and gravel aquifers. In sand aquifers, ground water velocity is
moderate because mean grain size is moderate. As ground water velocities increase because of
increasing gravel content or increasing proximity to a pumping well, the differences between
virus and bacterial transport efficiency become less important, and either a viral or bacterial
indicator may be recommended (See Exhibit 1.1).

       On the other hand, the finest grained porous aquifers, such as shale and clay beds, are not
considered to be aquifers because ground water velocities through them are generally very slow.
 Thus, despite the great significance of straining and pore-size exclusion in such environments,
entrained pathogens are not transmitted efficiently through shale or clay,  therefore such
subsurface formations are not considered further in this guidance.
Source Assessment Guidance Manual

4.8    Analytical Methods

       The Ground Water Rule Source Water Monitoring Methods Guidance document provides
a detailed explanation of EP A-approved analytical methods. A compilation of these methods is
listed in Exhibits 4.2, 4.3, and 4.4. Details about each method can be found in the Source Water
Monitoring Methods Guidance document. The last column in each table identifies the section of
the Source Water Monitoring Methods Guidance document where additional explanatory text
resides. The tables list approved methods for GWR monitoring as of the date of this guidance
document. Additional methods may have been approved for GWR monitoring since this date.
Additional information, including additional approved analytical methods for GWR monitoring,
can be  found at the following URL: http://www.epa.gov/safewater/methods/methods.html.
           Exhibit 4.2 E. coli Methods Approved for Use under the GWR
mEndo or
LES Endo6
Ml Medium
SM1 9223
SM1 9223
SM1 9221 F
SM1 9222B6
SM1 9222G
EPA Method 1604
Approved Formats
Membrane Filtration
Membrane Filtration
Membrane Filtration
Description of
Positive Result
Yellow, fluorescent
Red/magenta, fluorescent
Blue/green, fluorescent
Growth and the presence
of acid and/or gas in LTB,
fluorescent in EC-MUG
Pink to red colonies with
metallic (golden-green)
sheen that fluoresce after
transfer to NA-MUG
Blue colonies
Blue colonies
 Standard Methods for the Examination of Water and Wastewater, 20 edition.
Source Assessment Guidance Manual

        Exhibit 4.3 Enterococci Methods Approved for Use under the GWR
Azide Dextrose /
SM1 9230B
SM1 9230C
Description of Positive Results
Growth at 45EC in BHI and growth
in BHI with 6.5% NaCI at 35EC
Pink to red colonies that form black
or reddish-brown precipitate on
underside of filter
All colonies with a blue halo
Presence of blue-white
Standard Methods for the Examination of Water and Wastewater, 20m edition.
 Annual Book of ASTM Standards Water and Environmental Technology, Volume 11.02, 2000.
         Exhibit 4.4 Coliphage Methods Approved for Use under the GWR
Two-Step Enrichment
Single Agar Layer
EPA Method 1601
EPA Method 1602
Description of
Positive Result
Presence of plaques
(circular lysis zones)
Presence of plaques
(circular lysis zones)
Source Assessment Guidance Manual


                                    5. References
Aley, T., and M. W. Fletcher. 1976. Water tracers cookbook. Missouri Speleology, Journal of the
      Missouri Spelelogical Survey. 16(3):  1-32.

Anderson, A. D., A. G. Heryford, J. P. Sarisky, C. Higgins, S.  S. Monroe, R. S. Beard, C. M.
      Newport, J. L. Cashdollar, G.  S. Fout, D. E. Robbins, S. A. Seys, K. J. Musgrave, C.
      Medus, J. Vinje, J. S. Bresee, H. M. Mainzer, andR. I. Glass. 2003. A waterborne
      outbreak of Norwalk-like virus among snowmobilers-Wyoming, 2001. Journal of
      Infectious Diseases. 187: 303-306.

B. M. Ross and Associates, Ltd, and the Municipality of Brockton. 2000. Municipality of
      Brockton: Contamination of Walkerton Water Supply,  May 2000 - Report on Cause.
      Unpublished report, 59 p.

Baker, V. R., R. Greeley, P. D. Komar, D.A. Swanson, and R.  B. Waitt, Jr. 1987. Chapter 11:
      Columbia and Snake River Plains. In: Geomorphic systems of North America, W.L. Graf
      (Ed.), Centennial Special Volume 2, p. 403. Boulder, CO: Geological Society of America.

Barton, A. R., Jr., and J. C. Redwine. 1985.  Groundwater Manual for the Electric Utility
      Industry, Volume 1: Geological Formations and Groundwater Aquifers, Final Report (1st
      ed). EPRI CS-3901-Vol. 1. Palo Alto, CA: Electric Power Research Institute.

Becker, M. W., P. W. Reimus, and P. Vilks. 1998. Transport and attenuation of
      carboxylate-modified latex microspheres in fractured rock laboratory and field tracer
      tests. GroundWater. 37 (3): 387-395.

Bedient, P. B., H. S. Rifai, and C. J. Newell. 1999. Ground Water Contamination:  Transport and
      Remediation. New Jersey: Prentice Hall, Inc.

Bergeisen, G. H., M. W. Hinds, and J. W. Skaggs. 1985. A waterborne  outbreak of hepatitis A in
      Meade County, Kentucky. American Journalof Public Health.  75: 161-164.

Bergmire-Sweat, D., J. Morgan, K. Wilson, K. Von Alt, L. Marengo, T. Bennett, Y. M. Lee, V.
      C. Tsang, W. R. MacKenzie, andB. Furness.  1999. Cryptosporidiosis at Brushy Creek:
      Describing the epidemiology and causes of a large outbreak in Texas, 1998. In:
      Proceedings, International Conference on Emerging Infectious Diseases, Milwaukee, WI.
      Denver, CO: American Water Works Association.

Beyerle, U., W. Aeschbach-Hertig, M. Hofer, D. M. Imboden, H. Baur, and R. Kipfer. 1999.
      Infiltration of river water to a  shallow aquifer investigated with  3H/3He, noble gases and
      CFCCs. Journal of'Hydrology.,220: 169-185.

Bitton, G., S. R. Farrah, R. H. Ruskin, J. Butner, and Y. J. Chou. 1983.  Survival of pathogenic
      and indicator organisms in ground water. Ground Water. 21 (4): 405-409.
Source Assessment Guidance Manual              5-1

Biziagos, E., J. Passagot, J. Crance, and R. Deloince. 1988. Long-term survival of Hepatitis A
       virus and Poliovirus Type 1 in Mineral Water. Applied Environmental Microbiology. 54
       (11): 2705-2710.

Blair, B. A., and C. P. Gerba. 2006. Occurrence of Naegleria fowleri in well water in Arizona.
       University of Arizona National Science Foundation Water Quality Center. Unpublished

Bothner, W. A., and E. L. Boudette. 1997. Generalized and Simplified Bedrock Geologic Maps
       of New Hampshire. Adapted from Bedrock Geologic Map of New Hampshire by Lyons,
       L. B., W. A. Bothner, R. H. Moench, and J. B. Thompson, Jr.. Reston, VA: United States
       Geological Survey.  Available on the Internet at:
       http://www.des. state.nh.us/pdf/gen_br.pdf

Bowen G. S., and M. A. McCarthy. 1983. Hepatitis A associated with a hardware store water
       fountain and a contaminated well in Lancaster County, Pennsylvania, 1980. American
       Journal of Epidemiology. Ill: 695-705.

Brahana, J. V., R. K. Davis, T. J. Davis, T. E. Tin, S. Al-Rashidy, K. Whitsett, and M. Al-Qinna.
       2002. Recurring, multicomponent ground water tracing experiments at a well-
       characterized basin  in mantled-karst: Lessons from the Savoy Experimental Watershed.
       In: Proceedings - GSA Annual Meeting, October 27-30, Denver, CO.

Breckenridge, R. M. and K. L. Othberg. 1998. Surficial Geologic Map of the Post Falls
       Quadrangle and Part of the Liberty Lake Quadrangle, Kootenai County, Idaho. Idaho
       Geological Survey.  Available on the Internet at:
       http://www.idahogeology.org/PDF/Maps (MVSurficial  Geologic Maps (SGMVPDF/S

Bretz, J. H. 1925. The Spokane flood beyond the Channeled Scabland. Journal of Geology.  33:

Bricelj.1999. Parti culate Biological Tracers. Abstract from the Proceedings of the Workshop on
       Groundwater Pollution in Karst-Preserving Quality in Karst Systems ^November 4-6,

Brown, R. H., A. A. Konoplyantsev, J. Ineson, and V. S. Kovalensky. 1983. Ground-Water
       Studies: An International Guide for Research and Practice. Studies and Reports in
       Hydrology No. 7. Paris: UNESCO.

Chippewa County Health Department. 1992. A preliminary report on the water-borne disease
       outbreak on Drummond Island in late August 1991. Sault Ste. Marie, Michigan.
       Unpublished report.

City of Tallahassee.  1996. A Model Approach to Aquifer and Wellhead Protection in Florida,
       Final Report. City of Tallahassee, Water Quality Division, Aquifer Protection Section,
       FDEP Contract GW075..
Source Assessment Guidance Manual              5-2

Costa, J., I. Tiago, M. S. da Costa, and A. Verissimo. 2005. Presence and persistence of
      Legionella spp. in groundwater. Applied Environmental Microbiology.  71 (3): 663-671.

Craun, G. F. 1984. Health aspects of ground water pollution. In:Groundwater Pollution
      Microbiology, Bitton, G., and C. P. Gerba (Eds.), pp. 135-179. New York: John Wiley
      and Sons, Inc.,

Crowther, J. 1989. Karst geomorphology of South Wales. In: Limestones and Caves of Wales,
      T.D. Ford (Ed.), pp. 20-39. Cambridge University Press. Cited in EPA  (1997).

Cutler, P. M., L. Clayton, D. M. Mickelson, P. M. Colgan, and J.W. Attig. 2000. Tunnel
      Channels and Associated Fan Deposits in Wisconsin, U.S.A.: Insights into the Plumbing
      of the Southern Laurentide Ice Sheet. In: International Symposium on Modern and
      Ancient Ice-Marginal Landsystems, Russell, A. J., and P. M. Marren (Eds.), pp. 19-20.
      England: Keele University.

D'Antonio, R. G., R. E. Winn, J. P. Taylor, T. L. Gustafson, W. L.  Current, M. M. Rhodes, W.
      Gary, and R. A. Zajac. 1985. A waterborne outbreak of Cryptosporidiosis in normal
      hosts. Annals of Internal Medicine.  103: 886-888.

Davies, W. E., J. H.  Simpson, G. C. Ohlmacher, W. S. Kirk, and E. G. Newton. 1984.
      Engineering Aspects of Karst in National Atlas of the United States of America. U.S.
      Geological Survey. Map #38077-AW-NA-07M-00.

Devinny, J. S., L. R. Everett, J. C. S. Lu, and R. L. Stollar. 1990. Sub-surface Migration of
      Hazardous Wastes. New York: Van Nostrand Reinhold.

Domenico, P. A., and F. W.  Schwartz.  1990. Physical and ChemicalHydrogeology. New York:
      John Wiley and Sons, Inc.

Driscoll, F. G. 1986. Groundwater and Wells. St. Paul, MN: Johnson Filtration Systems.

Faybishenko, B. A., I. Javendel, and P. A. Witherspoon.  1995. Hydrodynamics of the capture
      zone of a partially penetrating well in a confined aquifer. Water Resources Research. 31
      (4): 859-866.

Fetter, C. W. 2001. Applied Hydrogeology. New Jersey: Prentice Hall, Inc.

Field, M. S. 1999. The QTRACER Program for Tracer-Breakthrough Curve Analysis for Karst
      andFractured-Rock Aquifers. EPA Report 600/R-98/156a.  Washington, DC: U.S.
      Environmental Protection Agency.

Fong, T., L. S. Mansfield, D. L. Wilson, D. J. Schwab, S. L. Malloy, and J.B. Rose. 2007.
      Massive microbiological groundwater contamination associated with a waterborne
      outbreak in Lake Erie, South Bass Island, OH. Environmental Health Perspectives. 115
      (6): 856-864.

Freeze, R. A.  and J. A. Cherry. 1979. Groundwater. Englewood Cliffs, NJ: Prentice Hall.
Source Assessment Guidance Manual              5-3

Gelting, R., J. Sarisky, C. Selman, C. Otto, C. Higgins, P. O. Bohan, S. B. Buchanan, and P. J.
       Meehan. 2005. Use of a systems-based approach to an environmental health assessment
       for a waterborne disease outbreak investigation at a snowmobile lodge in Wyoming.
       International Journal of Hygeine and Environmental Health. 208 (1-2): 67-73.

Gerba, C. P., S. M. Kutz, A.  W. Bradford, and R. DeLeon. Undated. Rotavirus Survival and
       Transport in the Subsurface, Cooperative Agreement Final Report to Steve Hutchson,
       EPA, Ada, OK, CR-813566-0101.

Golder Associates. 2000. A report on hydrogeologic assessment, bacteriological impacts,
       Walker ton Town Wells, Municipality of Brockton, County of Bruce, Ontario.

Grondin, G. H. 1987. Transport ofMS-2 andf2 bacteriophage through saturated Tanque Verde
       Wash Soil. Unpublished M. A. thesis, University of Arizona, Department of Hydrology
       and Water Resources Administration.

Groundwater Education in Michigan (GEM) Program. 1992. A Vacation That's Enough to Make
       You Sick? GEMNotes. 4 (1). February, 1992.

Gunn, J., J. Tranter, J. Perkins, and C. Hunter.  1991. Sanitary bacterial dynamics in a mixed
       karst aquifer. In Karst Hydrology ^Proceedings of Workshop W2, Rabat Morrocco. IAHS
       Pub. 247, pp. 61-70.

Haitjema, H. M. 1995. Analytic Element Modeling of Groundwater Flow. Academic Press.

Hansen, C. V. 1991. Description and Evaluation of Selected Methods  Used to Delineate
       Wellhead-Protection Areas Around Public-Supply Wells Near Mt. Hope, Kansas. U.S.
       Geological Survey Water-Resources Investigations Report 90-4102.

Health Canada. 2000. Waterborne outbreak of gastroenteritis associated with a contaminated
       municipal water supply, Walkerton, Ontario, May-June 2000. In: Communicable Disease
       Report, 26-20 (October 15, 2000).

Heath, R. C. 1983. Basic Ground-Water Hydrology. U.S. Geological Survey Water-Supply
       Paper 2220. Republished in a 1984 edition by National Water Well Association, Dublin

Hegarty, J. P., M.  T. Dowd, and K. H. Baker. 1999. Occurrence of Helicobacterpylori in surface
       water in the United States. Journal of Applied Microbiology. 87: 697-701.

Hejkal, T.  W., B. Keswick, R. L. LaBelle, C. P. Gerba, Y. Sanchez, G. Dressman, B. Hafkin, and
       J. L. Melnick. 1982. Viruses in a community water supply associated with an outbreak of
       gastroenteritis and infectious hepatitis.  Journal of the American Water Works
       Association. 74: 318-321.

Hickey, J.  J. 1984. Field Testing the hypothesis of Darcian flow through a carbonate aquifer.
       Ground Water. 22 (5): 544-547.
Source Assessment Guidance Manual              5-4

Hrudey, S. E. and E. J. Hrudey. 2004. Safe Drinking Water, Lessons from Recent Outbreaks in
       Affluent Nations. London: IWA Publishing

Hurst, C. J., J. J. Moster, and M. V. Yates. 1997. Modeling the compatibility of viral data sets.
       1997 Water Quality Technology Conference, AWWA (abstract), November 9-12, 1997,
       Denver, CO.

Hydrogeologic, Inc. 1988. FECTUZ, Finite Element and Semi-Analytical code for Simulating
       One-Dimensional Flow and Solute Transport in the UnsaturatedZone.

Hydrogeolgic, Inc. 1994. VIRALT, A Modular Semi-Analytical and Numerical Model for
       Simulating Viral Transport in Ground Water, Version 3.0. Available from CoSMoS,
       EPA, Ada, OK or the International Ground Water Modeling Center, Colorado School of
       Mines, Golden, CO.

Hydrogeologic, Inc. 1995. CANVAS, A Composite Analytical-Numerical Model for Viral and
       Solute Transport Simulation, Version 2.1. Available from CoSMoS, EPA, Ada, OK or
       the International  Ground Water Modeling Center, Colorado School of Mines, Golden,

Idaho Department of Environmental Quality (IDEQ). 2002a. City of Twin Falls - Blue Lakes
       Well Field (PWS 5420058) Source Water Assessment, Final Report.

IDEQ.  2002b. Riverbend Water Company (PWSNO 1280230) Source Water Assessment Report.

IDEQ.  2000. The Spokane Valley - Rathdrum Prairie Aquifer Atlas. August, 2000.

ILEPA - State of Illinois Environmental Protection Agency (ILEPA). 1995. Methodology for
       Evaluating the Vulnerability of Community Water Supply Wells. March, 1995.
       Unpublished memorandum.

Jannik, N. O., E. C. Alexander, Jr., and L. J. Landherr. 1991. The sinkhole  collapse of the
       Lewiston, Minnesota waste water treatment facility lagoon. In Proceedings of the Third
       Conference on Hydrogeology, Ecology, Monitoring, and Management of Ground Water
       inKarst Terrains, Dublin, Ohio, 715-724. National Ground Water Association.

Jansons, J., S. W.,  Edmonds, B. Speight, and M. Bucens. 1989.  Survival of viruses in
       groundwater. Water Research. 23  (3): 301-306.

Johnson, R., J. Pankow,  D. Bender, C. Price, and J. Zogorski. 2000. MTBE: To what extent will
       past releases contaminate community water supply wells? Environmental Science &
       Technology. 34 (7):  210A-216A.

Kentucky Department of Environmental Protection (KYDEP). 1999.  WHPP for the City of
Source Assessment Guidance Manual              5-5

Keswick, B. H., C. P. Gerba, S. L. Secor, and I. Cech. 1982. Survival of enteric viruses and
       indicator bacteria in groundwater. Journal of Environmental Science and Health. A17
       (6): 903-912.

Kinner, N. E., R. W. Harvey, K. Blakeslee, G. Novarino, and L. D. Meeker. 1998. Size-selective
       predation on groundwater bacteria by nanoflagellates in an organic-contaminated aquifer.
       Applied and Environmental Microbiology. 64 (2): 618-625.

Kaemer, S.R., V.A. Kelson, andH. M. Haitjema. 2007. WhAEM2000. Model available online:
       http ://www. epa. gov/athens/software/whaem/index.html

Kraemer, S. R., H. M. Haitjema and V. A. Kelson. 2007. Working with WhAEM2000: Capture
       Zone Delineation for a City  Wellfield in a Valley Fill Glacial Outwash Aquifer
       Supporting Wellhead Protection. EPA Office of Research and Development. EPA/600/R-
       05/151.  Available online: http://www.epa.gov/ceampubl/gwater/whaem/index.htm

Lawson, H. W., M. M. Braun, R. I. M. Glass, S. Stine, S. Monroe, H. K. Atrash, L. E. Lee, and
       S. J. Engelender.  1991. Waterborne outbreak of Norwalk virus gastroenteritis at a
       southwest U.S. resort: role of geological formations in contamination of well water. The
       Lancet. 337: 1200-1204.

Lee, Y.M., P. W. Johnson, J. L. Call, M. J. Arrowood, B. W. Furness, S. C. Pichette, K. K.
       Grady, P. Reeh, L. Mitchell, D. Bergmire-Sweat, W. R. Mackenzie, and V. C.  W. Tsang.
       2001. Development and application of a quantitative, specific assay for Cryptosporidium
       parvum oocyst detection in high-turbidity environmental water samples. American
       Journal of Tropical Medicine and Hygeine. 65 (1): 1-9.

Lerner, D. N. 1992. Well Catchments and Time-of-Travel Zones in Aquifers with Recharge.
       Water Resources Research. 28 (10): 2621-2628.

Mahler, B. J., J. C. Personne, G. F. Lods, and C. Drogue. 2000. Transport of free and particulate-
       associated bacteria in karst. Journal of Hydrology. 238: 179-193.

Malard, F., J. Reygrobellet, and M. Soulie. 1994. Transport and retention of fecal bacteria at
       sewage-polluted fractured rock sites. Journal of Environmental Quality. 23: 1352-1363.

Malde, H. E. and H. A. Powers. 1972.  Geologic map of the Glenns Ferry-Hagerman area,  west-
       central Snake River Plain, Idaho. Miscellaneous Geologic Investigations Map 1-696 (2
       sheets). U.S. Geological Survey. Cited in O'Connor (1993).

McBride, G., D. Till, T. Ryan, A. Ball, G. Lewis,  S. Palmer, P. Weinstein. 2002. Pathogen
       Occurrence and Human Health Risk Assessment Analysis. New Zealand Ministry of
       Health and Ministry of Environment Freshwater Microbiology Research Programme
       Report, November 2002.

McFeters, G. A., G. K. Bissonnette, J.  Jezeski, C.  A. Thomson, andD. G.  Stuart. 1974.
       Comparative survival of indicator bacteria and enteric pathogens in well water. Applied
       Microbiology. 27 (5): 823-829.
Source Assessment Guidance Manual              5-6

McWhorter, D. B. and D. K. Sunada. 1977. Ground-Water Hydrology and Hydraulics. Littleton,
       CO: Water Resources Publications.[Later edition published 1981].

Missouri Department of Health. 1992. Summary of investigation, hepatitis A outbreak, Apostolic
       church and school, Racine, MO. Unpublished report, 5 pp.

Molz, F. J., O. Guven, and J. G. Melville. 1990. A New Approach and Methodologies for
       Characterizing the Hydrogeologic Properties of Aquifers. EPA Report 600/2-90/002.

Moore, A. C., B. L. Herwaldt, G. F. Craun, R. L. Calderon, A. K. Highsmith, and D. D. Juranek.
       1993. Surveillance for water borne disease outbreaks - United States, 1991-1992.
      Morbidity and Mortality Weekly Report. Surveillance Summary SS-5, US Centers for
       Disease Control and Prevention. 42 (SS-05): 1-22.

Morris, D. A. and A. I. Johnson. 1967. Summary of Hydraulic and Physical Properties of Rock
       and Soil Materials as Analyzed by the Hydraulic Laboratory of the U.S. Geological
       Survey. U.S. Geological Survey Water Supply Paper  1839-D, pp. D1-D42.

Nasser, A. M. and S. D. Oman. 1999. Quantitative assessment of the  inactivation of pathogenic
       and indicator viruses in natural water sources. Water Research. 33 (7): 1748-1752.

Nasser, A. M., R. Glozman, and Y. Nitzan. 2002. Contribution of microbial activity to virus
       reduction in saturated soil. Water Research. 36: 2589-2595.

National Research Council (NRC). 1990. Ground Water Models. Washington, D.C.: National
       Academy Press.

NRC. 1993. Ground Water Vulnerability Assessment: Predicting Relative Contamination
      Potential under Conditions of Uncertainty. Washington, D.C.: National Academy Press.

NRC. 1994. Ground Water Recharge Using Waters of Impaired Quality. Washington D.C.:
      National Academy Press. Available on the Internet at:
       http://www.nap.edu/books/0309051428/html/index.html. Accessed December 16, 2002.

Natural Resources Conservation Service - National Soil Survey Center (NRCS-NSSC). 2002.
       Soil Surveys and Maps

New Hampshire Department of Environmental Services (NHDES). 200 la.. Drinking Water
       Source Assessment Report for Enfield Water Department.

NHDES. 2001b. Drinking Water Source Assessment Report for Merrimack Village District.

O'Connor, J. E. 1993. Hydrology, Hydraulics and Geomorphology of the Bonneville Flood.
       Geological Society of America Special Paper, 274.

O'Connor, I.E. and V.R. Baker.  1992. Magnitudes and implications of peak discharges from
       glacial Lake Missoula. Geological Society of America Bulletin. 104: 267-279.
Source Assessment Guidance Manual               5- 7

O'Reilly, C. E., A. B. Bowen, N. E. Perez, J. P. Sarisky, C. A. Shepherd, M. D. Miller, B. C.
       Hubbard, M. Herring, S. D. Buchanan, C. C. Fitzgerald, V. Hill, M. J. Arrowood, L. X.
       Xiao, R. M. Hoekstra, E. D. Mintz, M. F. Lynch, and the Outbreak Working Group..
       2007. A waterborne outbreak of gastroenteritis with multiple etiologies among resort
       island visitors and residents: Ohio, 2004. Clinical Infectious Diseases. 44 (4):  506-512.

Ohio Environmental Protection Agency (Ohio EPA). 2005. South Bass Island, Ottawa County
       Gastrointestinal Illness Summer 2004. Ohio Environmental Protection Agency
       Investigation and Actions. February 22, 2005, amended February 16, 2006. Available on
       the Internet at: http://www.epa.state.oh.us/ddagw/SBIweb/Reports/SBI EPA.pdf. 48pp.

Orth, J. P., R. Netter, and G. Merkl. 1997. Bacterial and Chemical Contaminant Transport Tests
       in a Confined Karst Aquifer (Danube Valley,  Swabian Jura, Germany). InKarst Waters
       and Environmental Impacts, Gunay, G. and A. I. Johnson, p. 173. Rotterdam: Balkema.

Pancorbo, O. C., B. G. Evanshen, W. F. Campbell, S. Lambert, S. K. Curtis,  and T. W. Woolley.
       1987. Infectivity and antigenicity reduction rates of human Rotavirus strain Wa in Fresh
       Waters. Applied and Environmental Microbiology. 3 (8):  1803-1811.

Pardee, J. T. 1942. Unusual currents in glacial Lake Missoula, Montana. Geological Society of
       America Bulletin. 53:  1569-1600.  Cited in Baker et al. (1987).

Parshionikar, S.U., S. Willian-True, G. S. Fout, D. E. Robbins, S. A. Seys, J. D. Cassady, and R.
       Harris. 2003. Waterborne outbreak of gastroenteritis associated with a norovirus. Applied
       and Environmental Microbiology. 69 (9): 5263-5268.

Parsonnet J., S. Trock, C. A. Bopp, D. J. Wood, D. G. Addiss, F.  Alai, L. Gorelkin, N. Hargrett-
       Bean, R. Gunn, and R. V. Tauxe. 1989. Chronic diarrhea associated with drinking
       untreated water. Annals of Internal Medicine. 110:  985-991.

Personne, J. C., F. Poty, L. Vaute, and C. Drogue.  1998. Survival, transport and dissemination of
       Escherichia coli and enterococci in a fissured environment. Study of a flood in a karstic
       aquifer. Journal of Applied Microbiology. 84: 431-438.

Quinlan, J. F. 1989. Ground Water Monitoring in Karst Terranes: Recommended Protocols and
       Implicit Assumptions. U.S. Environmental Protection Agency, Environmental  Monitoring
       Systems Laboratory, Las Vegas, Nev. EPA Report 600/X-89/050. 79pp.

Reilly, T. E., O. L. Franke, and G. D. Bennett. 1987.  The principle of superposition andits
       application in ground-water hydraulics: U.S.  Geological Survey Techniques of
       Water-Resources Investigations, Book 3, Chapter B6., 28 p.

Rice, E. W., M. R. Rodgers, I. V. Wesley, C. H. Johnson, and S. A. Tanner. 1999. Isolation of
       Arcobacter butzleri from ground water. Letters in Applied Microbiology. 28: 21-35.
Source Assessment Guidance Manual

Rice, E. W., C. H. Johnson, D. K. Wild, and D. J. Reasoner. 1992. Survival of Escherichia coli
       O157:H7 in drinking water associated with a waterborne disease outbreak of
       hemorrhagic colitis. Letters in Applied Microbiology. 13: 38-40.

Riffard, S., S. Springthorpe, L. Filion and S. A. Sattar. Occurrence ofLegionellain
       Groundwater. American Water Works Association Research Foundation Report 90985F,
       Denver, Colorado. 164 pp.

Risser, D. W., and T. M. Madden, Jr. 1994. Evaluation of Methods for Delineating Areas That
       Contribute Water to Wells Completed in Valley-Fill Aquifers in Pennsylvania. U.S.
       Geological Survey Open-File Report 92-635.

Risser, D. W. and G. J. Barton. 1995. A Strategy for Delineating the Area of Ground-Water
       Contribution to Wells Completed in Fractured Bedrock Aquifers in Pennsylvania. U.S.
       Geological Survey Water-Resources Investigations Report 95-4033. 30 pp.

Rolle-Kampcyk,  U. E., G. J. Fritz, U. Diez, I. Lehmann, M. Richter, and O. Herbarth. 2004. Well
       water-One Source of Helicobacter pylori colonization. InternationalJournal of Hygiene
       and Environmental Health. 207: 363-368.

Sevee, J. 1991. Methods and Procedures for Defining Aquifer Parameters.  In: Practical
       Handbook of Ground-Water Monitoring, D.M. Nielsen (Ed.), pp. 397-447. Chelsea, MI:
       Lewis Publishers.

Slade, R. M., M.  E. Dorsey, and S. L. Stewart. 1986. Hydrology and Water Quality of the
       Edwards Aquifer Associated With Barton Springs in the Austin Area Texas. U. S.
       Geological Survey. Water Resources Investigations Report 86-4036.

Smart, P. L. and H. Frederich. 1986. Water movement in the unsaturated zone of a maturely
       karstified carbonate aquifer, Mendip Hills, England. In: Environmental Problems in
       Karst Terranes and Their Solutions Conference, held in 1st Bowling Green, Kentucky,
       J.F. Quinlan (Ed.), pp. 1-14. Dublin, Ohio: National Water Well Association. Cited in
       EPA (1997).

Sobsey, M. D., P. A. Shields, F. H. Hauchman, R. L. Hazard, and W. Caton. 1986. Survival and
       transport  of hepatitis A virus in soils, groundwater and wastewater. Water Science  and
       Technology. 18: 97-106.

Solomon, D. K., R. J. Poreda, S. L. Schiff, and J. A. Cherry. 1992. Tritium and Helium 3 as
       groundwater age tracers in the Borden Aquifer. Water Resources Research. 28 (3):

Sroka, J., A. Wojcik-Fatla, and J. Dutkiewicz. 2006. Occurrence of Toxoplasmagondii in water
       from wells located on farms. Annals of Agricultural and Environmental Medicine.  13:

Strack, O. D. L. 1989. Groundwater Mechanics. Prentice-Hall.

Source Assessment Guidance Manual              5-9

Stute, M., J. Deak, K. Revesz, J. K. Bohlke, E. Deseo, R. Weppernig, and P. Schlosser. 1997.
       Tritium/3 He dating of river infiltration: an example from the Danube in the Szigetkoz
       area, Hungary. Ground Water. 35 (5): 905-911.

Sun, R. J., J. B. Weeks, and H. F. Grubb. 1997. Bibliography of Regional Aquifer-System
       Analysis Program of the U.S.  Geological Survey, 1978-96. U.S. Geological Survey
       Water-Resources Investigations Report 97-4074. Available on the Internet at:
       http://water.usgs.gov/ogw/rasa/html/introduction.html. Accessed Dec. 14, 2002, last
       updated February 20,  1999.

Swerdlow, D. L., B. A. Woodruff, R. C. Brady, P. M. Griffin, S. Tippen, H. Donnel, Jr., E.
       Geldreich, B. J. Payne, A. Meyer, Jr., J. G. Wells, K. D. Greene, M. Bright, N. H. Bean,
       and P. A. Blake. 1992. A waterborne outbreak in Missouri of Escherichia coli O157:H7
       associated with bloody diarrhea and death. Annals of Internal Medicine. 117(10): 812-

Thornbury, W. D. 1954. Principles of Geomorphology. New York: John Wiley and Sons, Inc.

Tranter, J., J. Gunn, C. Hunter, and J. Perkins. 1997. Bacteria in the Castleton Karst, Derbyshire,
       England. Quarterly Journal of Engineering Geology.  30: 171-178.

Trimble, D. E. and W. J.  Carr. 196la. TheMichauddelta andBonneville River near Pocatello,
       Idaho. U.S. Geological Survey Professional Paper 424-B. U.S. Geological Survey, pp.
       B164-B166. Cited in O'Connor (1993).

Trimble, D. E. and W. J.  Carr. 1961b. Late Quaternary history of the Snake River in the
       American Falls region, Idaho. Geological Society of America Bulletin 72: 1739-1748.
       Cited in O'Connor (1993).

U.S. Center for Disease Control and Prevention (USCDC). 1996. Shigella sonnei outbreak
       associated with contaminated drinking water - Island Park, Idaho, August 1995.
       Morbidity and Mortality Weekly Report. 45 (11): 229-231.

USCDC. 2005. U.S. Center for Disease Control and Prevention. Environmental Health
       Assessment for Epi-Aid 2004-076: Outbreak of gastroenteritis with multiple etiologies
       among resort island visitors and residents of Ohio, 2004. National Center for
       Environmental Health, Division of Emergency and Environmental Health Services,
       Environmental Health Services Branch, Technical Assistance Team, August, 2005. 74
       pp. Unpublished report

U.S. Department of Agriculture - Soil Conservation Service (USDA-SCS).  1984. Soil Survey for
       Essex County, Massachusetts, Southern Part. In cooperation with the Massachusetts
       Agricultural Experiment Station. Washington, D.C.: Government Printing Office. 167 pp.

United States Environmental  Protection Agency (USEPA). 1987a. Agricultural Chemicals in
       Groundwater - Proposed Pesticide Strategy. Office of Pesticides and Toxic Substances.

Source Assessment Guidance Manual             5-10

USEPA. 1987b. Guidelines for Delineation of Wellhead Protection Areas. EPA Report

USEPA. 199 la. Delineation of Wellhead Protection Areas in Fractured Rocks. Environmental
      Protection Agency, Washington, D.C.

USEPA. 1991b. Wellhead Protection Strategies for Confined-Aquifer Settings. EPA Report

USEPA. 1993 a. Ground Water Resource Assessment. Office of Water, EPA. EPA Report 813/R-
      93-003 166 pp and 4 appendices.

USEPA. 1993b. WHPA, A Modular Semi-Analytical Model for the Delineation of Wellhead
      Protection Areas Version 2.2. Available from CoSMoS, EPA, Ada, OK or the
      International Ground Water Modeling Center, Colorado School of Mines, Golden, CO.

USEPA. 1994a. Ground Water Information Systems Roadmap, A Directory of EPA  Systems
      Containing Ground Water Data. EPA Report 813-B-94-001.

USEPA. 1994b. Handbook, Ground Water and Wellhead Protection. EPA Report 625/R-94/001.
      269 pp.

USEPA. 1994b. WhAEM: Program Documentation for the Wellhead Analytic Element Model.
      EPA Report 600/R-94/210. 120 pp.

USEPA. 1997. Guidelines for Wellhead and Springhead Protection Area Delineation in
      Carbonate Rocks. EPA Report 904-B-97-0003.

USEPA. 2002. Delineation of Source-water Protection Areas in Karst Aquifers of the Ridge
      and Valley and Appalachian Plateaus Physiographic Provinces: Rules of Thumb for
      Estimating the Capture Zones of Spring and Wells.  EP A-816-R-02-015.

USEPA. 2006a. National Primary Drinking Water Regulations: Ground Water Rule; Final
      Rule. Federal Register 71(224):67427-65660.

USEPA. 2006b. Sanitary Survey Guidance for Ground Water Systems. Office of Water, U.S.
      Environmental Protection Agency, Washington, D.C.  EPA-815-D-07-006.

USEPA. 2006c. GWR Source Water Monitoring Methods Guidance. Office of Water, U.S.
      Environmental Protection Agency, Washington, D.C. EPA-815-R-07-019.

USEPA. 2006d. Occurrence and Monitoring Document for the Final Ground Water Rule. Office
      of Water, U.S. Environmental Protection Agency, Washington, D.C. EPA-815-R-06-012.

USEPA. 2006e. GWR Economic Analysis. Office of Water, U.S. Environmental Protection
      Agency, Washington, D.C. EPA-815-R-06-014.
Source Assessment Guidance Manual             5-11

USEPA. 2006f. GWR Corrective Action Guidance. Office of Water, U.S. Environmental
       Protection Agency, Washington, D.C. EPA-815-R-06-xxx..

USGS. 1999. Groundwater Atlas of the United States. Available on the Internet at:
       http://capp.water.usgs.gov/gwa. Accessed December 27, 2002.

USGS. 2002. Office of Groundwater website. Available on the Internet at:

USGS. 1998. Water Quality in the Las Vegas Valley Area and the Carson and Truckee River
       Basins,  Nevada and California, 1992-1996. USGS Circular 1170.

Virginia Department of Environmental Quality (DEQ). 1993. Wellhead Protection: Case Studies
       of Six Local Governments in Virginia. Prepared by the Virginia Ground Water Protection
       Steering Committee.

Walton, W. C.  1989. Groundwater Pumping Tests, Design and Analysis. Chelsea, MI: Lewis
       Publishers, Inc.

Weissman, J. B.,  G. F. Craun, D. N. Lawrence, R. A. Pollard, M. S. Saslaw, and E. J. Gangarosa.
       1976. An  epidemic of gastroenteritis traced to a contaminated public water supply.
       Journal of Epidemiology. 103  (4): 391-398.

Whitsett, K. S., R. K.  Davis, T. Ting,  G. Thoma, and J. V. Brahana. 2001. Sediment and bacteria
       tracing  in mantled karst of northwest Arkansas. In: Proceedings: GSA Annual Meeting,
       November 5-8, 2001, Boston, Massachusetts.

Winter, T. C., J. W. Harvey,  O. L. Franke, and W. M. Alley. 1998. Ground Water and Surface
       Water:  A Single Resource. U.S. Geological Survey Circular 1139. Denver, CO: U.S.
       Government Printing Office. 79 pp.

Wisconsin Department of Natural Resources (WIDNR). 1999. Wisconsin's Source Water
       Assessment Program Plan. Available on the Internet at:
       http ://www. dnr. state, wi .us/org/water/dwg/swap/index. htm

Witten, J., and  S. Horsley. 1995. A Guide to Wellhead Protection. American Planning
       Association, Planning Advisory Service Report Number 457/458, 102 pp.

Worthington, S. R. H., C. C. Smart, and W. Ruland. 2001. Karst hydrogeological investigations
       at Walkerton. Prepared for Concerned Walkerton Citizens. Available on the Internet at:

Worthington, S. R. H., C. C. Smart, and W. W. Ruland. 2002. Assessment of groundwater
       velocities to the municipal wells at Walkerton. Ground and Water: Theory to Practice. In
       Proceedings of the 55th Canadian Geotechnical and 3rd Joint IAH-CNC and CGS
       Groundwater Specialty Conferences, held in Niagara Falls, Ontario, October 20-23,
       2002, D. Stolle, A. R. Piggott, and J. J. Crowder (Eds.), pp. 1081-1086. Southern Ontario
       Section of the Canadian Geotechnical Society.
Source Assessment Guidance Manual             5-12

Yahya, M. T., L. Galsomies, C. P. Gerba, and R. C. Bales. 1993. Survival of bacteriophages MS-
       2 and PRD-1 in ground water. Water Science and Technology. 27 (3-4): 409-412.

Yates, M. V. 1992. Unpublished, untitled spread sheet compilation of measured parameters for
       viruses in saturated and unsaturated materials based on 34 references (saturated) and 10
       references (unsaturated).

Yates, M. V., and Yates, S.R.  1988. "Modeling microbial fate in the subsurface environment."
       CRC Critical Reviews in Environmental Control. 17: 307-344.
Source Assessment Guidance Manual              5-13


           Appendix A:  Field Methods for Determining the Presence
                             of a Hydrogeologic Barrier
Pump and Slug Tests

       Pump tests can be used to estimate aquifer properties on local and regional scales.  A
pump test requires pumping water at a known rate from a test well or a production well, and
measuring water-level drawdown over time in the pumping well and/or nearby monitoring wells
(Witten and Horsley, 1995).

       Careful test design is critical to obtaining accurate pump test data. A monitoring well can
be sited by estimating the pumping rate of the well and the expected drawdown. It is important
to locate monitoring well screens in the same hydrostratigraphic unit as the screen of the pumped
well (Witten and Horsley, 1995). Driscoll (1986) and Walton (1989) provide further information
on pump test design.  Pump test analytical methods are available for unconfined, confined, and
semi-confined (leaky) aquifers.  Pump tests are usually analyzed using the drawdown data
collected during pumping, but tests may also use the hydraulic head recovery measurements
taken as the potentiometric surface returns to its initial level after pumping is ceased (Witten and
Horsley, 1995).

       Slug tests may also be useful to determine aquifer properties such as hydraulic
conductivity and transmissivity. When a slug ( a solid rod, or, alternatively,  a given volume of
additional water) is dropped into a well, the water level in the well is forced to suddenly rise, and
the rate at which as the water level returns to its initial level can be measured. The initial test
results can be verified by running the test in reverse (i.e., by removing the slug from the well).
To achieve reliable test results, it is important that a well be properly constructed (Witten and
Horsley, 1995).

Geophysical Methods

       Borehole geophysical data can be used to help interpret which fractured bedrock intervals
have low permeability and may act as confining layers. These well logging devices are usually
run in combination and can include caliper, resistivity, spontaneous potential, neutron,  gamma,
and television logs. Risser and Barton (1995) report that caliper, single-point resistance, and
gamma logs are commonly used for identifying fractures and fracture zones.  Television logging
is particularly useful for  identifying vertical fracturing. Comparatively expensive,
unconventional logs,  such as full-waveform acoustic, acoustic televiewer, vertical seismic
profiling, borehole radar, and resistivity tomography may provide good data  on fractured
bedrock aquifers.

       Borehole flowmeters are designed to measure the flow into or within a well bore. Heat
pulse, electromagnetic, and impeller flow meters are designed to locate productive fracture zones
(Molz et al. 1990). If no productive fracture zones are identified in the upper part of the aquifer,
the interval may act as a  confining layer.
Source Assessment Guidance Manual              A-l

Ground Water Age Dating

       Ground water age dating of samples collected from a public water supply well can help
quantify ground water travel time from the surface to the well. At locations where ground water
travel times are short, ground water velocities are greater for a given flow path length (well
depth). For the short ground water travel times characteristic of sensitive aquifers, helium-
3/tritium ratios, and oxygen and hydrogen isotope concentrations are best suited for age dating.
Although these methods are being developed in a research setting, their application to sensitive
aquifers has not been routine. Risser and Barton (1995) caution that age dating can lead to
erroneous conclusions if a well is receiving water from two or more distinct intervals with
significantly different ground water ages.  The mixed water will provide ground water ages
intermediate between the two actual values.

       Tritium, the radioactive isotope of hydrogen with a half-life of 12.38 years, was  released
from nuclear bomb tests in the atmosphere and serves as a time marker.  For ground water with
travel times of a few months, the helium-3/tritium ratio can be used (Beyerle et al. 1999; Stute et
al. 1997).  The concentration of helium-3 dissolved in ground water increases  as soon as the
ground water is isolated from the atmosphere because helium-3, which is produced by tritium
decay, can no longer escape. Therefore, the helium-3/tritium ratio is a measure of the time
elapsed since a water parcel was last in contact with the atmosphere (Beyerle et al. 1999).  Other
sources of radiogenic helium may also be present, produced by the decay of uranium and
thorium in mineral grains; care is needed in using this technique (Solomon et al. 1992).

       Stable isotope ratios such as oxygen-18/oxygen-16 and deuterium/hydrogen can also be
used to identify the young ground water (i.e., ground water with short residence times) typical of
sensitive aquifers.  Stable isotope methods are designed to compare the isotopic character of
different waters by using plots of isotope ratios and their deviation from  a recognized standard.
Because isotopic ratios differ by season, seasonal recharge can sometimes be recognized in
ground water.  With a sufficient surface water stable isotope record, short residence periods in
ground water can be determined by their differing isotopic signatures (Beyerle et al.  1999).

Tracer Testing

       The most direct way to determine ground water velocity is by introducing a tracer
substance  at one point in the flow field and observing its arrival at another point in the flow field
(typically  a monitoring well) (Freeze and Cherry, 1979). Tracer tests are the most conclusive
method for evaluating the direction and travel time of ground water flow in bedrock aquifers
(Risser and Barton, 1995). Tracers can consist of organic dyes, inorganic salts, gas, or solid
particles.  Tracer recovery at a given location indicates a hydraulic connection, and the time
required to detect the tracer can be used to calculate ground water velocities. Additional
information on tracer testing can be obtained from Aley and Fletcher (1976) or Field (1999).
Source Assessment Guidance Manual              A-2

                     Appendix B: Ground Water Travel Time
       Ground water travel time (GWTT) is the time that it takes a small amount or "packet" of
ground water to move from one point in the aquifer to an endpoint (e.g., a pumping well). It is
sometimes helpful to use a circle or ellipse on a map to represent the area surrounding a
continuously pumping well that will contribute water to the well after, for example, 2 years of
travel. The enclosed area is sometimes referred to, for simplicity, as the "travel time zone."
Such zones can be delineated and drawn on a map when the ground water travel times are known
from many locations in the aquifer, and  an appropriate travel time of interest is chosen.  In order
for such maps to be meaningful, it is necessary to presume horizontal flow conditions. This can
be a useful method for representing areas on a map that may contribute to the contamination of a
pumping well over the time period of interest.

       Pumping increases the rate of ground water flow and shortens the time it takes for ground
water (and any associated contaminants  and pathogens) to move from one point in the aquifer to
another.  In sand, sandstone, and similar aquifer types, ground water travel time can be
calculated using  previously determined estimates of flow system parameters, including hydraulic
conductivity, induced hydraulic gradient, and porosity.  Alternatively, travel time estimates can
be based on measurements of natural or  artificial  tracer transport. Tracers mimic the behavior of
ground water itself, and ideally have little chemical interaction with aquifer material.  Due to
natural spreading or "dispersion" in the  subsurface, a certain amount of any tracer and the
ground water which carries it has an arrival time  at a given well that is shorter than the average
arrival time. This is the result of some water (and tracer) molecules taking a more direct path
rather than the more typical tortuous path through the granular aquifer.  In general, viruses and
bacteria transported through ground water typically arrive later than the average water "packet."
 However, due in part to their small size, some viruses may take the fastest path from the source
to a well and arrive before the average ground water travel time.  Thus, the United States
Environmental Protection Agency  (EPA) recommends that estimates of ground water travel
times be interpreted carefully in order to ensure that proposed  hydrogeologic barriers are truly

       Because viruses are acute contaminants, capable of causing infection at very low doses,
one possible strategy to protect the public is to focus on those viruses that arrive at wells in an
infective state. If a particular virus may survive for four to six months in the subsurface, it may
be advisable to add another six-month "safety factor" to the determined ground water travel time
to account for uncertainties such as dispersion in  the calculation of travel times.

       Ground water travel time calculations often necessitate the use of computer models. All
methods for calculating GWTT require simplifying assumptions.  Even the most complex ground
water calculations (three-dimensional numerical ground water flow models) are simplified
representations of the aquifer/well  system. EPA  (1987b) has grouped the GWTT calculation
methods into four groups.  These groups are (in order of increasing computational complexity):
(1) uniform flow, (2) analytical, (3) semi-analytical, and (4) numerical methods. Under certain
conditions, described below, an even simpler one-dimensional method may be used.

Source Assessment Guidance Manual              B-l

       All methods considered in this document are based on the simplifying assumption of
steady state ground water flow.  Under the steady state assumption, the well pumping effects do
not change with time and ground water flow has achieved a new equilibrium (different than the
natural equilibrium prior to the start of pumping). When the actual well pumping rate is variable
over the time period of interest (several months or 1 or 2 years), for the purposes of evaluating a
possible hydrogeologic barrier, the well is assumed to be uniformly pumped at the maximum
sustainable rate for the entire time period of interest.

       The simplifying assumption of steady state flow  is a technical requirement necessary to
perform uniform or analytical calculations. However, it is possible to perform semianalytical or
numerical GWTT calculations without this simplifying assumption. Nevertheless, by applying
the steady state assumption to all methods discussed here, comparisons among the methods can
more easily be performed.

       All GWTT calculation methods require input of the average effective porosity of the
aquifer. Porosity, in a saturated portion of an aquifer, is the proportion of interparticle void
space that is filled with water. Void space varies within an aquifer, so the porosity value is
averaged to simplify GWTT calculations.  Effective porosity refers to that portion of an aquifer's
void space through which water can travel. This definition is necessary because some water is
trapped in pores that are sealed in all directions by mineral growth. Other water is bound in very
small pores in clay minerals and is capable of moving only over time scales longer than those of
concern here. The term, "porosity," as used in the following discussion, refers to  average,
effective porosity.

       Commonly, porosity measurements in the vicinity of a well are unavailable, even within
the largest aquifers. More typically, porosity values are  estimated, based on knowledge  of the
aquifer's hydrogeologic setting (e.g., alluvial) or, for regional extensive aquifers, based on the
aquifer type  known from the name of the hydrogeologic unit (e.g., Dakota sandstone). Porosity
values typical of alluvial or sandstone aquifers are available in the scientific literature - for
example, in Freeze and Cherry (1979). Although porosity values can vary by 10 or 20 percent
even within a given aquifer category, variations in hydraulic conductivity are typically several
orders of magnitude. Thus, heterogeneity of hydraulic conductivity is usually the largest single
factor introducing uncertainty into ground water flow models and GWTT calculations which are
not based on tracer tests.

One Dimensional Method for Horizontal Flow

       The one dimensional (ID) GWTT method is a simple equation for calculating horizontal
time of travel (USEPA 1994a). The method is most appropriate for natural ground water flow
(i.e., when there is no pumping well) in localities where  data are available for three input
parameters that describe the properties of the aquifer: horizontal hydraulic conductivity,
horizontal hydraulic gradient, and porosity. EPA (1991b) suggests a supplementary
approximation method that can be used to simulate GWTT in areas of steep water table slope,
such as in the vicinity of a pumping well.  Nevertheless,  because the 1-D GWTT method does
not explicitly account for the presence and actions of a pumping well, it is generally not
appropriate for source water assessment or hydrogeologic sensitivity assessment purposes.  The
mathematical expression is presented in EPA (1994a, p.  74).

Source Assessment Guidance Manual               B-2

Calculated fixed radius GWTT method

       The calculated fixed radius GWTT method explicitly accounts for the presence and
actions of a pumping well. Thus, it is appropriate for GWTT calculations for Source Water
Assessment Program (SWAP) or Ground Water Rule (GWR) purposes.  The method is based on
an assumption of cylindrical flow to a well with pore volume equal to the pumped volume of
water during the specified period. Everywhere in the cylinder, water flows horizontally to the

       The calculated fixed radius GWTT method is a simple calculation that requires only three
input parameters.  Furthermore, of the three input parameters, only the porosity is a property of
the aquifer; the other two parameters describe the construction or operation of the well.  In
general, it is easier to obtain site-specific values for the latter parameters. As discussed above,
porosity values are typically estimated rather than measured.  The required well parameters are
the well pumping rate and the length of the open interval or well screen. The mathematical
expression of the calculated fixed radius GWTT method is presented in EPA (1994a, p. 70).

       The calculated fixed radius GWTT method is valid if the drawdown from pumping is less
than about 10 percent of the aquifer's pre-pumping saturated thickness (Reilly et al. 1987).
Furthermore, the method requires that the well fully penetrate the water-bearing zone of the
aquifer and be open  or screened throughout the entire interval. The method is most appropriate
for aquifers that most closely approximate a homogeneous, isotropic  aquifer of constant
thickness located in  a region with a flat water table.  Hydrologic boundaries are assumed to be
sufficiently distant so that the ground water flow field in the vicinity  of the well is not
significantly affected by those boundaries. It is assumed that all flow is horizontal and ground
water flow velocity is constant. The method can be modified to simulate constant flux recharge
or leakage (Risser and Madden, 1994). Risser and Madden (1994) also investigate the effects of
violating the flat water table assumption.

       EPA suggests that the method be applied only to shallow coastal plain, wide topographic
basin, or mid-continent aquifers where the aquifer and the topography are relatively flat. Risser
and Madden (1994)  suggest that the method is not appropriate for the valley-fill aquifers in
Pennsylvania.  Similarly, EPA suggests that the method is not appropriate for any dimensionally
restricted aquifer (i.e., those that are long and narrow such as a rift-basin, barrier island, glacial
buried valley, alluvial, or esker aquifer).  The method is not appropriate for wells that are located
near surface water or near a topographic high point or in areas with irrigation wells or other high
demand ground water usage.

Uniform Flow GWTT Method

       The uniform flow method is an analytical solution that can be used to estimate GWTT for
steady flow to a well.  The uniform  flow equations are derived by superposition of the Dupuit
equation for radial flow around a well with the one-dimensional, uniform, pre-pumping flow
field (Risser and Madden 1994).

       The method requires six input parameters. Five parameters describe the properties of the
aquifer; porosity, hydraulic conductivity, aquifer thickness, and hydraulic gradient magnitude
and direction.  The remaining parameter describes the pumping well pumping rate.
Source Assessment Guidance Manual              B-3

       To use the uniform flow method, the aquifer is assumed to be confined, of constant
thickness, homogeneous, and isotropic. The pre-pumping potentiometric surface may be flat or
uniformly sloping and pumping from a fully-penetrating well is assumed to have resulted in a
steady state. According to Risser and Madden (1994), the assumptions of a steady-state flow
and a uniformly sloping potentiometric surface are not theoretically possible in an unbounded
aquifer; but, if boundaries are distant, the assumptions may be valid. The method may be used
for unconfined aquifers if the well drawdown is less than about ten percent of the pre-pumping
saturated thickness of the aquifer. Mathematical expressions of uniform flow are presented in
EPA (1994a, p. 81) and Risser and Madden (1994, p. 32) The most efficient approach to make
use of the uniform flow method is to use the wellhead protection area (WHPA) model (USEPA
1993b), which has the added benefit of being applicable to a wider variety of aquifer boundary
types without requiring more input parameters describing the aquifer or well properties.

       Similar to the calculated fixed radius method, the uniform flow method is applicable to
shallow coastal plain, wide topographic basin or mid-continent aquifers where the aquifer and
the topography are relatively flat. EPA suggests that States find multiple solutions in order to
bound the uncertainty in the estimated parameters used with this method.

Analytical and Semi-Analytical GWTTMethods

       Analytical and semi-analytical methods comprise a group of equations that are
superimposed (equation solutions are added together) in various combinations to simulate
particular aquifer settings  and types. In the more simple combinations, the method does not
require any additional knowledge about aquifer or well properties beyond that required for the
uniform flow equation. Additional parameters needed are used to define the aquifer boundaries
(i.e., the distance to the nearest significant surface water body or rock outcrop that forms a
barrier to ground water flow). Typically, combinations of surface water or rock outcrops can be
used to bound the aquifer on two or  four sides.

       Because of the computational complexity of the analytical methods, semi-analytical
methods are typically used.  The semi-analytical method makes use of more advanced
computational methods with the help of a computer, but does not change the number of
significant input parameters. The WHPA model (USEPA 1993b) is one example of a
semi-analytical method that solves the complex analytical flow equations.  The user-interface for
the WHPA model shields the user from the complexity of the method.

       The WHPA model is divided into two basic modules: one that is suitable for an isolated
well and one that is suitable for a well in a well field where nearby wells have a significant
interfering effect on the flow to each well.  The module MWCAP (Multiple Well Capture Zone)
(or RESSQC) is suitable for the former but not the latter. The module General Particle Tracking
(GPTRAC) is suitable for either case.

       The minimum input parameters are the same for both MWCAP and  GPTRAC.  The input
parameters are the same six  parameters as those required by the uniform flow equation.  Five
parameters describe the properties of the aquifer: porosity, hydraulic conductivity, aquifer
thickness, hydraulic gradient magnitude and direction, and the sixth parameter, well pumping
rate, describes the pumping well.
Source Assessment Guidance Manual              B-4

       The assumptions governing the use of the analytical and semi-analytical methods are
similar to the uniform flow equation in terms of the properties of the aquifer, but they are
dissimilar in terms of the boundaries of the aquifer.  To use the analytical and semi-analytical
methods, the aquifer is assumed to be confined, of constant thickness, homogeneous and
isotropic. The prepumping potentiometric surface may be flat or uniformly sloping, and
pumping from a fully-penetrating well is assumed to have resulted in a steady state condition.
The method may be used for unconfined aquifers if the well drawdown is less than about 10
percent of the pre-pumping saturated thickness of the aquifer.  Unlike the uniform flow equation
assumptions, the analytical method may be used near surface water or near a topographic high.
If GPTRAC is used, as discussed above, the semi-analytical method may be used to calculate
GWTT to a well adjacent to another, interfering, well.

       The GPTRAC module of WHPA may also be used to calculate GWTT in a leaky,
confined aquifer.  However, three additional input parameters that describe the properties of the
aquifer are required; confining bed hydraulic conductivity, confining bed thickness, and areal
aquifer recharge rate. The module GPTRAC may also be used to calculate GWTT in an
unconfined aquifer (if drawdown is less than 10 percent of the aquifer thickness) where areal
aquifer recharge is significant.  In this application, information is needed to specify a boundary
condition. This boundary condition is the distance at which the well pumping effects are
negligible.  Such information is typically not available, thus necessitating that this application be
used with caution. Also, for this application, the areal recharge rate is needed.

       EPA suggests that the analytical and semi-analytical methods are applicable to all
granular, porous aquifers. Hansen (1991) suggests that the analytical and numerical flow
methods could be used for calculating GWTT (for microbial protection areas) around public
water supply wells near Mt. Hope, Kansas. These wells are located in an unconfined,
unconsolidated, Quaternerary terrace and alluvial  deposits of silt, clay, and gravel  (High Plains
aquifer).  Risser and Madden (1994) report that the semi-analytical method is a powerful and
flexible method that may be used as long as the water table surface is not highly irregular in
shape. Lerner (1992) suggests that the semi-analytical ground water flow model ROSE performs
better at calculating GWTT for one type of aquifer setting (i.e., aquifers with significant recharge
and with a well distant from the impermeable boundary). For hydrogeologic settings in which
the well only partially penetrates the full saturated thickness of the aquifer, an analytical solution
is available (Faybishenko et al. 1995).

Analytical Element GWTT method

       The analytic element method is the most recent development of solution techniques for
ground water flow (Strack 1989; Haitjema 1995).  The analytical element method uses both
analytical and numerical methods to perform the ground water travel time  calculation. The EPA
analytical element method, wellhead analytical element model (WHAEM) (USEPA 1994a;
Kelson et al. 1999, Kraemer et al. 1999) is designed to calculate GWTTs.  To perform the
GWTT calculation, WHAEM requires the following parameters:  porosity, hydraulic
conductivity, aquifer thickness, areal recharge rate, pumping rate, and stream water levels.
Because it uses a more sophisticated computational method than WHPA, WHAEM is also
capable of simulating hydrogeologic settings in which streams are not fully incised through the
entire saturated thickness of the aquifer (partially penetrating streams). Both WHPA and
Source Assessment Guidance Manual               B-5

WHAEM can simulate large water bodies that perform as hydrogeologic boundaries because
they fully penetrate the aquifer.  However, unlike WHPA, WHAEM can also simulate partially
penetrating streams that do not perform as hydrogeologic boundaries and that gain and lose
ground water independent of areal recharge and well pumping.

       WHAEM is capable of simulating any area around a well that is shaped as a polygon.
WHAEM can also simulate uniform stream or barrier boundaries on each side of the rectangle or

       WHAEM is applicable to all hydrogeologic settings with porous media (e.g., sand,
gravel). If the hydrogeologic  setting appears complex, such as at the confluence of two large
rivers or at the confluence of a large and small river, then WHAEM may provide a more accurate
GWTT calculation than WHPA.

Numerical GWTT method

       Numerical methods have the capability to simulate nonlinear, nonfully penetrating
boundary conditions, complex patterns of recharge and discharge, and spatial heterogeneity of
hydraulic properties (Risser and Madden, 1994). These methods require, at a minimum, all of
the input parameter data required by semi-analytical methods. More complex simulations
require substantial amounts of input data and are not further considered here.

Calculated GWTT for Vertical Ground Water Flow in Granular Porous Aquifers

       The vertical ground water flow to be discussed here is natural gradient flow directed
vertically downward in saturated porous media.  The flow is assumed to be far from a pumping
well so that the drawdown due to well pumping has no effect on the vertical GWTT. An
unsaturated and saturated glacial till overlies a karst limestone aquifer in Minnesota. The
unsaturated portion and a saturated  portion of glacial till is believed (by Minnesota drinking
water staff) to act as a hydrogeologic barrier that prevents fecal contamination of the karst
limestone aquifer.  The vertical GWTT is the travel time for vertical ground water flow through
the saturated glacial till component of the hydrogeologic barrier.

       The flow through the unsaturated glacial till component is also part of the hydrogeologic
barrier.  However,  unsaturated ground water flow is much more complex than saturated ground
water flow because the hydraulic conductivity cannot be assumed to take the same value
everywhere in the medium.  Rather, in unsaturated soil and aquifer material, the hydraulic
conductivity is very strongly dependent on the porous medium percent saturation.  The percent
saturation is typically low (20-30 percent) near the ground surface and increases to 100 percent
at the water table.  At a particular saturation point, the unsaturated material may have two
hydraulic conductivity values, depending on whether the material is in a wetting phase from
recent infiltrating precipitation or a drying phase as drainage, evapotranspiration, and recharge
remove water from the material. Unsaturated drainage may flow in finger-like wetting fronts
rather than as a uniform drainage front.  Some fingers may drain very quickly compared to the
overall movement of the drainage front. As  a result of these physical processes, unsaturated flow
is very complex, and vertical GWTT calculations are difficult.
Source Assessment Guidance Manual               B-6

       For most aquifers covered by a thin layer of unsaturated materials (a few tens of feet is
typical of the humid eastern United States), including the glacial till in the above example, it
might be appropriate to assume that the entire aquifer thickness is saturated. That is, saturated
conditions begin at the ground surface, as might be true after a heavy rain.  This assumption
avoids the complexity associated with unsaturated flow.

Darcy's Law GWTT Method for Vertical Flow Between a Confined Aquifer and an Unconfined

       The calculation of GWTT using Darcy's Law for vertical flow to or from a confined
aquifer is similar to the Darcy's Law calculation for horizontal GWTT (USEPA 1991b, 1994).
Input parameters needed are: 1) water level (hydraulic head) for the  confined aquifer, 2) water
level at the water table, 3) vertical hydraulic conductivity of the confining layer, 4) confining
layer thickness, and 5) confining layer porosity.

Vertical GWTT Method for Flow from the  Water Table to the Bottom of an Unconfined Aquifer

       The vertical GWTT for flow from the water table to the bottom of an unconfmed aquifer
may be difficult to determine. This is because data from nests of piezometers, which measure
hydraulic head at more than one vertical location in an aquifer, are often unavailable. Vertical
GWTT determinations also require knowledge of vertical hydraulic  conductivity (assuming the
aquifer is anisotropic, as most aquifers are), porosity, and the thickness of the unconfmed
aquifers, which varies with any variation in the depth of the water table.  Seasonal recharge,
barometric pressure, and heavy pumping can have significant effects on the depth of the water
table. Assuming all data inputs could be accurately determined, vertical GWTT's could be
calculated using Darcy's Law.

Vertical GWTT Method for Unsaturated Flow to the Water Table

       One-dimensional vertical downward flow through the unsaturated zone to the water table
is a component of the Virus Analytical Transport (VIRALT) and composite analytical-numerical
model for viral and solute transport simulation (CANVAS) methods (Hydrogeologic, Inc., 1994;
1995).  Both methods use the same GWTT calculation method, which is based on the finite
element and semi-analytical code for simulating one-dimensional flow and solute transport in the
unsaturated zone (FECTUZ) numerical and semi-analytical,  unsaturated GWTT method
(Hydrogeologic, Inc., 1988).

       The VIRALT/CANVAS method includes a built-in database of input parameter values
from 12 typical soils that range from predominantly clay-rich to predominantly sand-rich along
with various combinations of sand, silt, and clay-sized particles. The method allows the user to
select one of the 12 soil types for each unsaturated horizon; when selected, the database provides
most of the needed input parameter values for that soil.  Alternatively, the user can specify each
input parameter separately based on site-specific data.

       The flow of water in the unsaturated zone is assumed to be vertically downward
(one-dimensional). The flow is also considered to be at steady-state, isothermal, and governed
by Darcy's Law.  The aquifer is assumed to be homogeneous, the ground water slightly
compressible, and the effects of wetting and drying cycles on the ground water flow parameters
Source Assessment Guidance Manual              B- 7

are neglected. Recharge rates may vary in time, but the flow field is assumed to adjust
instantaneously from an existing steady-state condition to one reflective of the new recharge rate.
 Up to 10 unsaturated layers with differing properties may be specified, but each layer is
assumed to be a uniform and incompressible porous medium (Hydrogeologic Inc., 1994).

       For each unsaturated interval, the following input parameters are needed if the built-in
soil database is not used:  1) saturated hydraulic conductivity, 2) saturated water content, 3)
residual water content, 4) empirical parameter alpha, 5) empirical parameter beta, and 6)
infiltration rate. The method is applicable to all unsaturated porous media.
Source Assessment Guidance Manual              B-8

                    Appendix C:  Microbial Inactivation Rates
       Microorganisms can enter aquifers due to failed septic systems, faulty construction of
waste disposal injection wells, leaking sewer lines, infiltration from surface water
impoundments, or ground water interaction with contaminated surface water bodies (Bedient et
al. 1999).  Once released, the microorganisms are faced with stresses imposed by the
environment, including competition with other microorganisms. Some released bacteria can
replicate in the natural environment, but parasitic protozoa and viral pathogens are unable to
replicate in natural settings.

       The United States Environmental Protection Agency (EPA) is interested in predicting the
fate of those microorganisms released into the environment as a consequence of human
activities.  The Ground Water Rule (GWR) is based on estimates of the length of time over
which bacterial and viral pathogens might pose a hazard to public water supply (PWS) wells.
After a certain amount of time, a released pathogen will no longer pose a human health hazard
due to inactivation.

       According to Hurst (1997), microorganisms released into the environment become
susceptible to inactivation by a variety of physical, chemical, and biological processes. These
processes include desiccation, denaturation, biochemical antagonism from enzymes, and
predation. The inactivation rates may be accelerated by temperature, pH, interaction with
inorganic and organic dissolved and solid phases,  and solar radiation in surface environments
(Yates and Yates 1988).

       Microbial survival studies are designed to  evaluate the time during which
microorganisms remain viable. Hurst (1997) reviews the variety of methods, procedures and
objectives for performing such studies. Microbes are generally too small to be monitored as
individual organisms.  Rather, population survival is studied. Hurst believes that it is best to
study microbial populations within their natural environment, where they are free to move  and
exchange chemicals with their surroundings. However, such studies for large microbial
populations are unlikely to be practical for the purposes of a hydrogeologic sensitivity
assessment (HSA). As an alternative, microorganism survival can be studied in containers
placed within the natural environment, such as in a well, or in environmental media (e.g., ground
water samples) brought into the laboratory.

       Potential pathogen hazard to PWS wells can be evaluated by identifying the longest
period before inactivation occurs for one or more pathogenic bacteria or viruses.  Given
knowledge of the length of the potential hazard period, States can identify wells that may be
sensitive to pathogen contamination due to the short travel time for ground water recharge
(potentially containing  pathogens) to that well.

       The microbial survival data in this review  are restricted to studies of fecally-derived
bacteria and viruses that are most commonly transmitted via oral ingestion of drinking water.
Pathogenic protozoa, such as Giardia and Cryptosporidium, if found in ground water-supplied
PWS systems, would often result in those PWS's being classified as GWUDI and subject to the
requirements of the Surface Water Treatment Rule rather than the GWR. Thus, the survival of
protozoa will not be discussed here. Virus survival in ground water and surface water were
Source Assessment Guidance Manual               C-l

compared by Hurst (1998), who concluded that statistical models used to predict virus
inactivation in surface water could not be applied to ground water.  The reasons for these
differing inactivation characteristics are not known, but they could be due simply to the ubiquity
of naturally antagonistic microorganisms in surface water.

       Microbial survival studies can be conducted in ground water by using a very large
number of amendments, including nutrients, soil, aquifer materials, wastewater, detergents, and
waste products. Each of the amendments has a potential confounding effect on microbial
survival. In order to minimize the confounding effects, only survival studies that were
conducted in unamended ground water (studied either in situ or in the laboratory) are included
in the table below.  Furthermore, the table lists only studies of unamended ground water
conducted at temperatures within the range of 8 to 25 degrees C. Such temperatures are typical
of ground water in the United States, measured at depths of approximately 30 meters (m).

       A survey of the scientific literature identified numerous bacterial and virus inactivation
studies conducted in unamended ground water at typical ground water temperatures. These
studies are listed in the following table and show the longest survival periods (lowest
inactivation rates) for various bacterial and viral pathogens and fecal indicator organisms. An
inactivation rate of 0.1  can be interpolated to indicate four log microbial inactivation in 40 days.
 Similarly, an inactivation rate of 0.01  indicates 4-log microbial inactivation in 400 days. The
longest survival rate for pathogenic viruses is typically about 0.02, which indicates 4-log
microbial inactivation in approximately 200 days

       The virus and bacterial inactivation rate data below are unlikely to be accurate when
antagonistic microorganisms are present in the ground water sample studied. In some studies,
ground water samples were sterilized or filtered before the test population was seeded into the
sample. In other studies, no sample treatment occurred. Any sample treatment is noted in the
table, and the collection location of the ground water sample is provided if available. In one
study (Biziagos et al. 1988), mineral water (ground water collected from a spring) was sampled
and because the data met the criteria of being unamended ground water at a typical ground water
temperature, the information was included in the table.  The mineral water may, however, be
Source Assessment Guidance Manual              C-2

 Greatest Survival: Laboratory-Measured Virus or Bacterial Inactivation Rates in
       Pure Ground Water at Ground Water Temperatures (8-25 degrees C)
(deg. C)
Rice et al.
Shigella sonnei
Shigella flexeri
E. co/;0157:H7
fecal coliform
£. co/;'
£. co/;'
Vibrio cholerae
0.01 9 (est)

Bitton et al.
Poliovirus 1
Poliovirus 1
Poliovirus 1
Echovirus 1
0.011 (est.)



Puy de Dome

Source Assessment Guidance Manual

coliphage (MS-2)
Hepatitis A Virus
Hepatitis A Virus
Hepatitis A Virus
Simian Rotavirus
0.011 (est.)
0.00 (est.)
0.1 8 (est.)
0.1 (est)
(deg. C)



Puy de Dome

Source Assessment Guidance Manual

                   Appendix D: Additional Reference Sources
       Appendix A of EPA's 1994 Ground Water and Wellhead Protection Handbook contains
an extensive list of additional general reference sources that can be consulted for further
information. Some of the most relevant references from that Handbook are included in the list
below. In addition, several more current general reference sources are listed. The Ground Water
and Wellhead Protection Handbook can be found online at:
Aral, M. M. 1990. Ground Water Modeling in Multilayer Aquifers, Vols. 1 and 2. Florida: CRC

Bedient, P. B., H. S. Rifai,and C. J. Newell. 1999. Ground Water Contamination: Transport and
       Remediation. New Jersey: Prentice Hall, Inc.

Charbeneau, R. J.  1999.  Groundwater Hydraulics andPollutant Transport. New Jersey:
       Prentice Hall, Inc.

Cullimore, D. R.  2007.  Practical Manual of Groundwater Microbiology (2nd edition). Florida:
       CRC Press.

Domenico, P. A. and F. W.  Schwartz. 1997. Physical and Chemical Hydrogeology (2nd edition).
       New Jersey: John Wiley and Sons, Inc.

Drever, J. I.  1988. The Geochemistry of Natural Waters (2nd edition). New Jersey: Prentice
       Hall, Inc.

Driscoll, F.  G. 1986. Groundwater and Wells. St. Paul, MN:  Johnson Filtration Systems.

Fetter,  C. W. 2001.  Applied Hydrogeology (4th edition).  New Jersey: Prentice Hall, Inc.

Freeze, R. A.  and J.A. Cherry. 1979. Groundwater. New Jersey: Prentice Hall, Inc.

Gaudy, A. F.  1980. Microbiology for Environmental Scientists an d Engineers (McGraw-Hill
       series  in water resources and environmental engineering). Ohio: McGraw-Hill.

Han, J., Y. Jin, and C. S. Willson. 2006. Virus Retention and Transport in Chemically
       Heterogenous Porous Media under Saturated and Unsaturated Flow Conditions.
       Environmental Science & Technology. 1. 40 (5):  1547-1555.

Hijnen, W. A., A. J. Brouwer-Hanzens, K. J. Charles, and G. J. Medema. 2005. Transport of
       MS2 Phage, Escherichia coli,  Clostridium perfingens, Cryptosporidiumparvum, and
       Giardia intestinalis in a Gravel and Sandy Soil. Environmental Science & Technology.
       39 (20): 7860-7868.

Source Assessment Guidance Manual              D-l

John, D. E. and J. B. Rose. 2005. Review of Factors Affecting Microbial Survival in
       Groundwater. Environmental Science & Technology. 39 (19): 7345-7356.

Knox, R. C., L. W. Canter, D. A., Sabatini. 1993. Subsurface Transport and Fate Processes.
       Florida: CRC Press.

McWhorter, D. B. andD. K. Sunada. 1981. Ground-Water Hydrology and Hydraulics. Littleton,
       Colorado: Water Resources Publications.

National Research Council (NRC). 1990. Ground Water Models. Washington, D.C.: National
       Academy Press.

Paradis, D., R. Mattel, G. Karanta, R. Lefebvre, Y. Michaud, R. Therrien, and M. Nastev. 2007.
       Comparative Study of Methods for WHPA Delineation. Ground Water. 45 (2): 158-167.

Schmoll, O., G. Howard, J. Chilton, and I. Chorus. 2006. Protecting Ground Water for Health:
       Managing the Water Quality of Drinking-Water Sources. World Health Organization
       (WHO). Available on the Internet at:
       http://www.who.int/water sanitation  health/publications/protecting groundwater/en/ind
       ex. html

Spitz, K., and J. Moreno. 1996. A Practical Guide to Groundwater and Solute Transport
       Modeling.  New Jersey: John Wiley and Sons.

Stanford, J. A., and J. J. Simons (Eds.) 1992. Proceedings of the First International Conference
       on Ground Water Ecology. American Water Resources Association Technical
       Publication Series.

Strack, O. D. L. 1989. Groundwater Mechanics. New Jersey: New Jersey: Prentice-Hall Inc.
       Thornbury, W.D. 1969. Principals of Geomorphology (2nd edition). New Jersey:  John
       Wiley and Sons, Inc.

Ward, C. H., W. Giger, and P. L. McCarty. 1985. Groundwater Quality. New Jersey: John Wiley
       and Sons.

Weight, W. D., and J. L., Sonderegger. 2001. Manual of Applied Field Hydrogeology. Ohio:
       McGraw-Hill, Inc.

Zhen, C., and G. D. Bennett. 2002. Applied Contaminant Transport Modeling (2nd edition). New
       Jersey: John Wiley and Sons.
Source Assessment Guidance Manual              D-2