5524
905R84118
Interim status ground-water monitoring
implementation study
Phase III
U.S. Environmental Protection Agency
Office of Solid Waste
1984
U.S. Environmental Protection Agency,
R;gion V, Library
2>;0 South Dearborn Street
Chicago, Illinois 6Gj3Q4.
-------
U,S. Environmental Protection Agency
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
IATE X P ^"^ i
|ECT . Draft of The RCRA Phase III Ground-water Study
ROM John H. Skinner, Directo
Office of Solid Waste
T0 Addressees
J
Attached please find a draft of the RCRA Phase III
Ground-water Study prepared by Bill Meyers of OSW. The
final report will probably be widely distributed and quoted.
Please review and send me your comments by January 15, 1985.
Thank you for your assistance.
Attachment /
Addressees \' .n^ ' t *\ Q • >
•.i. —i. .—. i- , / \ !,_ fr* ") J*
. . . . T^ >? -A ^ ' < '
HW Division Directors, Regions I-X /' I'rtf"^ ! ' i"1
Gene Lucero I/* ' ' Ji -l ^
Bill Hedeman _ y \. \<1.<1
Cora Bebee / ^/o Z>
Dick Morgenstern .
OSW Senior Staff f)
cc: Lee Thomas (with copies)
Jack McGraw (with copies)
rm 1320-4 (R.v. 3-76)
-------
EXECUTIVE SUMMARY
Background
This is the third and final phase of the Interim Status Ground-Water
Monitoring Implementation Study requested of EPA in early 1982
by the Office of Management and Budget. Phase I addressed initial
implementation issues as perceived by State agencies and the EPA
Regional Offices (report dated September 17, 1982). Phase II
addressed "first year" ground-water monitoring issues and included
compliance evaluations at 189 non-randoraly selected hazardous waste
management facilities (report dated March 10, 1983). The Phase
III Study was conducted between October 1983, and May 1984.
The purpose of this part of the study was to critique the 40 CFR 265
Subpart F program in its latter stages of implementation.
The report consists of 4 subparts. Part one addresses the
implementation of the program by the States and EPA Regional Offices.
Specific issues addressed were: the number of facilities in each stage
of the ground-water monitoring program (detection, assessment,
or waiver); organizational problems being encountered by the States
and Regions; available resources; State authorization problems;
and enforcement. Information for this part of the study was collected
during headquarters Regional reviews, in discussions with State
and Regional personnel during site visits for the assessment part
of this study, by follow-up telephone calls and from the Major
Facility Status Sheet file.
Part two addresses laboratory capacity for and cost of analyzing
the 40 CFR 265.92 parameters. Also included was whether the laboratory
had gas chromatography or gas chromatography/mass spectometry
capabilities which are necesssary for the more complex Appendix
VII and VIII organic analytes. The source for this information was
a non-random telephone survey of 88 laboratories located across
the country.
Part three is a brief discussion on the Agency's requirement that
the Student's t statistical test be used to ascertain whether or not
a regulated unit is discharging to the ground water. Sources for
this portion of the study included an ongoing study on the t-test's
validity by Lockheed Engineering and Management Services, Inc.,
a report on alternatives by JRB Associates, and various EPA workgroup
analyses that have taken place over the past year.
Part four is a detailed analysis of the ground-water systems of 22
RCRA sites located in nine EPA Regions that have been reported to
be in an assessment mode. Each site was visited by EPA headquarters
staff. Split samples were obtained and sent to an EPA contract lab
for comparison with site laboratory results. The sampling and analysis
procedures that were used by the site during the visit were critiqued
and an evaluation made of well construction and placement as weL% as
the site's implementation of its assessment plan. The individjaarl sites
-------
were chosen by the EPA Regions in conjunction with the States. The
only firm criteria used was that they be in the assessment mode.
While this sample was not randomly chosen, EPA has no reason to
believe it is not representative of the entire universe.
Major Findings
Implementation
o The Agency, at this point in time, still does not have an accurate
list of the facilities in the nation that are subject to
ground-water monitoring requirements.
o Determining whether a facility is subject to Subpart F regulations
is often difficult because of the complexity of the regulations
that specify which waste constituents or processes are covered.
o Data compiled as of February 1984, show that approximately 60%
of the sites reported are in the detection program; 21.5% are
in one of the assessment programs; 6% have claimed waivers;
and about 12% are undetermined.
o The Facility Status Sheets showed 15% of the sites in the
detection program and 18% in the assessment program to be
in compliance with Subpart F standards.
0 o Several large States have a bifurcated organizational structure
-lilt that has resulted in very poor implementation.
o Many EPA Regions and most States have severe technical staffing
problems due to insufficient fiscal resources, poor salaries,
and an inability to attract experienced personnel.
o EPA's policy of turning the RCRA program over to the State's
as soon as possible has resulted in authorizing many States
that were ill prepared to implement the program. This, in part,
accounts for the program's poor compliance record.
o A number of authorized States have a very lax enforcement policy
towards Subpart F violatons.
o EPA's accountability system encourages larger numbers of cursory
inspections and discourages attention to time consuming cases
involving Part 265 Subpart F regulations.
-------
Laboratory Capability and Capacity
o There appears to be sufficient commercial laboratory capability
and capacity in the country to perform analytic testing for
Interim Status parameters.
o The per sample cost to a facility of having all the required Part
265.92 parameters tested for ranges between $200 and $2000
depending upon the laboratory chosen and the test methods used.
Statistical Comparisons
o The t-test, as it is currently being applied, is ill equipped
to deal with the very small data sets being generated by the
facilities, nor can it effectively handle the wide and largely
unknown variabilities due to spatial, temporal, sampling,
and analytic problems.
o The indicator parameters do not function efficiently or effectively
detect the possible impacts a regulated unit may be having on
the ground-water regime.
Assessment Study
o 20% of the sites examined did not have sampling and analysis
plans as required by 265.92.
o 56% of the plans lacked sufficient specificity and depth to
yield reproducible results.
o Only one site followed their written sampling and analysis plan.
o 60% of the sites did not use equipment appropriate for the
parameters sampled.
o 79% of the sites that sample for volatile organic constituents
used methods that would strip them from the samples before
they were shipped to a lab for analysis.
o Equipment cleaning between wells is often inadequate or not done.
o At least 25% of the sites have poorly constructed wells.
o About 75% of the sites do not have an adequate number of wells
to detect waste migration from their units.
-------
o About 50% of the sites have incorrectly screened their wells
so that they may miss contaminant plumes even though they
are in the correct place to intercept the plume.
o 30% of the sites have incorrectly placed at least some of their
wells.
o Most of the facilities that started out in an alternate assessment
mode have not complied with 265.90 to assess the rate and
extent of contaminant migration on site.
o Of the sites that were "triggered" into the program, none had
performed the rate and extent study required by 265.93.
o 10% of the sites had not determined the direction of ground-water
flow.
o 45% of the sites did not start their program on time.
-------
INTRODUCTION
This is the third and final phase of the Interim Status Ground-
Water Monitoring Implementation Study requested of EPA in early
1982 by the Office of Management and Budget. Phase I addressed
initial implementation issues as perceived by State agencies and
the EPA Regional Offices (report dated September 17, 1982).
Phase II addressed "first year" ground-water monitoring issues
and included compliance evaluations at 189 non-randomly selected
hazardous waste management facilities (report dated March 10, 1983).
Phase III is concerned with State and Regional implementation
issues; the cost and availability of analytic lab services to
perform RCRA 40 CFR 265.92 parameters; the success of the t-test
as a means of indicating a statistically significant increase
in contaminant concentrations in downgradient wells when compared
to upgradient background water quality; and how well sites that
have indicated they may have affected the ground-water quality
are implementing the Subpart F 265 Interim Status requirements.
-------
PART I INSTITUTIONAL
-------
REGIONAL AND STATE IMPLEMENTATION
This part of the Phase III study evaluates the success up through
April 1984, of the State and Regional Offices in implementing
the ground-water monitoring program with regards to the status
of the regulated universe, State and Regional program office
organization, available resources, state authorization, and
government enforcement efforts. It uses as major sources Regional
Review documents, facility status tracking sheets, State annual
reports, and interviews conducted during the course of the study
with State and Regional Officials. The quality, quantity, and
reliability of the above information is necessarily uneven so
that it is difficult to make quantitative comparisons or
draw broad conclusions that would apply across all Regions
and States. However, trends are apparent, and these will be
pointed out during the course of the report.
Status of the Regulated Universe.
In FY-84 the Agency initiated a specific reporting requirement
for the Regions and States referred to as the Major Facility
Status Sheet (MFSS) . This status sheet was designed to track
individual facility performance/compliance with interim status £
ground-water monitoring, closure/post closure, and financial
responsibility regulations. The importance of this device to
the Phase III study is that it reports the State and Regional
evaluations of the adequacy of well systems that facilities are
using to comply with the Subpart F requirements. Regions and
authorized States were to complete facility status sheets for
all major facilities, which were to include ^11 eii-ot CHKJQQ«-
Subpart F. While some facilities suoject to Subpart F may
nor oe included in the MFSS universe, these reports do provide a
picture of the status of the ground-water program nationwide at
the same point in time-- February 1984. '0
In addition to the above definitional problems, there are a number
of caveats that need to be interjected before discussing the data
available from the MFSS. First, the data presented in this study
do not include Ohio. Ohio was the only State to follow EPA
headquarters instructions to input their data directly into the
Hazardous Waste Data Management System (HWDMS) . The FY-84 guidance
had "encouraged States to adopt HWDMS as it should considerably
ease their reporting burden. The Agency intend(ed) that States...
enter data through the Regional mini computers and obtain HWDMS
reports... by communicating directly with the Agency's IBM main
frame in North Carolina." The guidance also stated that the States'
costs for adopting HWDMS were grant eligible. However, the
Agency has had severe technical problems in getting HWDMS to
function and these have limited the system's usefulness. In fact,
these technical problems have made it impossible for EPA to
retrieve the MFSS information that has been entered into the system.
-------
In other words, the only State to follow EPA headquarters direction
can not have its data accessed because the system does not work.
Second, there are several different estimates of the number of
ground-water monitoring facilities nationwide. At the time EPA
required the MFSS data to be submitted, the Part A file in HWDMS
contained over 2,000 ground-water facilities. Projections from the
RIA survey estimated just over 1,000, and quarterly surveys
conducted prior to the MFSS reporting requirement indicated
approximately 1,300. To date there have been status sheets
submmitted for 972 ground-water monitoring facilities (excluding
those in Ohio). The considerable variance in these estimates
indicates that, even at this stage of the program, with all sites
having been inspected at least once, there is still a significant
uncertainty concerning how many facilities are subject to Subpart F.
This uncertainty about the regulated community was manifested during
interviews undertaken as part of this study. One State official,
for example, remarked that he thought his State had over 100 subject
facilities while the EPA Regional Office maintained that the number
was closer to 40. Part of the difficulty in determining the
regulated community is caused by protective Part A filings. When
the RCRA program began, many facilities were unsure whether they
were covered under the regulations. In order to protect themselves
from EPA enforcement actions, they sent in Part A permit notifications,
A large percentage of these facilities need not have notified EPA.
Another major cause for uncertainty is the highly complex regulations
regarding exemptions from the ground-water monitoring requirements.
These exemptions most often involve whether the activity is
regulated or whether the waste stream is considered hazardous
under 40 CFR 261. For example one site visited as part of the
assessment section of this study had five impoundments. Three of
these were considered regulated units because they contained wastes
of a listed process stream (40 CFR 261.32). The other two contained
carbon tetrachloride wastes and were not considered to be regulated
because the process that produced them was not listed. This in
spite of the fact that carbon tetrachloride appears both as a listed
commercial product which is a hazardous waste when discarded and
as a hazardous waste when recovered in degreasing operations.
Because of these types of nuances in the regulations, one State
official estimated that it would take years to sort the facilities
out; that in many cases this may involve court proceedings; and that
the result may be a greater than 100% increase in his State of
regulated facilities over what is currently reported.
-------
Table A presents the status of the ground-water program
(i.e., detection monitoring, assessment monitoring, or waiver)
for all facilities subject to the Subpart F ground-water
monitoring requirements for which the Agency has received status
sheets in the Spring of 1984. This table also displays the
number of facilities with wells installed, whether evaluations
have been conducted, and the adequacy of the well systems evaluated.
Authorized States and Regions have submitted MFSS to Headquarters
for 972 facilities. Of the 972, 858 sites (88%) have been determined
to be either in the detection, assessment, or waiver category.
It was not possible to classify the remaining 114 facilities because
of major discrepancies in the sheets.
The MFSS data indicate that there are 586 facilities in the detection
monitoring program. Of these, 501 (85%) have well systems present,
while 85 (15%) have yet to put any wells in. (The date by which
well systems had to be installed was November 1981). As of the
end of February, the Regions and States had evaluated 65% of the
well systems present for adequacy. Approximately half (175) of the
328 evaluated were found to be adequate. The figures are similar for
the evaluation of sampling and analysis (S&A) plans and ground-water
records. Less than 50% of the detection-monitoring facilities (276)
had had their S&A plans or records evaluated; and of these
approximately 60% were found to be adequate.
The compliance rate for meeting all of the ground-water monitoring
requirements remains very low. Data from the facility status sheets
show that a total of 87 facilities (15%) are complying with all the
evaluated Subpart F standards.
The compliance data for the assessment monitoring program do not
present any better picture. Although all facilities in assessment
(210) have well systems, the data show that the Regions and States
have examined 60% of the sites and found only 33% (70) to be adequate.
For the S&A plans and ground-water records, the regulatory authorities
have determined that 32% and 39% respectively are known to be adequate,
The overall compliance status for meeting all requirements is
very similar to the detection program. The MFSS data show that
approximately 37 assessment monitoring facilities (18%) are in
full compliance with the Subpart F regulations. Furthermore,
half of the EPA Regions did not have any of their assessment sites
in full compliance.
In a separate part of the Phase III study, which is described later
in this report, Headquarters staff conducted thorough evaluations
of 22 sites with assessment monitoring programs. The staff then
compared data gathered during these evaluations with the information
provided on the Facility Status Sheets for these facilities.
-------
o
z
3
y
s
..
»* r«»
fl —
u
41 CO
•J
goo
•o
3 2
s » 2
^rf l_fl *
z « """
Si «•• "
if 3
i j
eri en 3
•
^^ «*
•^ o
41 U-)
JS
p^
«H
41
30
«J
i
•• 2
ta
_ i
-4 O
4» —
3 (N
£
-------
The results of the comparison indicate that the data submitted on
the sheets is suspect. For example, Headquarters staff found
major problems with two sites that had been judged adequate. One
had mistakenly determined its ground-water flow direction and
hence had its wells incorrectly placed. The other had submitted
an assessment report on its activities that was technically
incorrect. In addition the State had agreed to a proposal in
the report that permitted the site to stop monitoring its designated
upgradient well and eliminate from its quarterly sampling routine
most of its downgradient wells (cross gradient ones were still
included). These two examples call into question the adequacy
figures cited above and may indicate that they are too optimistic.
It should be noted that not all of the 22 sites, which had been
recommended for the study by the States and Regions, had status
sheets-- a further indication that the 972 figure is soft.
Data from the waiver program follow the same trend of low levels
of evaluation and adequacy as illustrated in the detection and
assessment monitoring programs. Table A shows that 62 facilities
have been reported to have claimed waivers for the Subpart F
requirements. The Regions and States have evaluated waiver
demonstrations for only 39 sites (63%). They found that a little
more than half of these waivers were legitimate. All sites that
are claiming waivers may not be identified since the Interim
Status regulations do not require facilities to submit waiver #
demonstrations but to have them available for inspection at the
facility. The regulatory authorities can identify waivers
only by inspection and questioning.
State and Regional Program Office Organization
In many States responsibility for evaluating the adequacy of
ground-water monitoring systems is located in an agency that is
different from the agency that is responsible for RCRA permitting
and overall RCRA implementation. The States often do this because
other agencies have been historically responsible for water problems
in the State and already have the needed technical expertise to
perform ground-water evaluations. However, this decentralization
can make it difficult for the lead agency to direct activities
in a way that achieves program goals. In many cases the roles
and responsibilities of the various agencies involved are not
clearly defined, which allows many critical activities to
"slip through the cracks".
The situation in California provides an example of the types
of problems that can result from this bifurcated approach. The
Department of Health Services (DOHS) is the lead agency in the State.
The State's 10 Regional Water Quality Control Boards (RWQCB) are
responsible for ground-water problems. However, because the
roles of these two agencies have not been clearly defined, neither
has been performing RCRA ground-water monitoring evaluations.
In fact, in many cases the Water Boards have been granting exemptions
to facilities from RCRA standards which they by and large were not
authorized to do. This has resulted in a situation where in
-------
February 1984, neither the State nor the Region could provide
to EPA headquarters a status list (detection, assessment, waiver)
of their estimated 134 facilities. This was because many had not
been performing the required RCRA analytical testing and their
status was hence unknown. By the end of the investigation period
of this report (April 1984) the Region had reported that a
Memorandum of Agreement between the two agencies had been signed
and they were in the process of rescinding their exemptions and
bringing the facilities back into the RCRA program. In the mean
time the Region was assuming a more active enforcement role.
Other States where the organizational structure has been an
impediment to effective implementation of the Subpart F requirements
would include New Jersey, Pennsylvania, Massachusetts, and
Washington State. New Jersey has two separate ground-water units.
One deals with offsite commercial facilities, the other with all
other facilities. In addition they have a separate enforcement
office that handles all regulatory inspections. The RCRA programs
in Pennsylvania, Massachusetts, and Washington State are administered
through regional or district offices, which report to a central
office. However, the quality of the regional programs and staff
varies widely among regions. The EPA Regional Offices find that it
is necessary to work with each district office individually rathe-r
than work through the central office, which makes coordination *
more difficult.
Organization problems are not confined to the States. Several EPA
Regions rely on the expertise of the Environmental Services Division
(ESD) to support RCRA ground-water monitoring requirements. In
Region II ESD is committed to several sampling inspections. The
Region II review indicated that there was some confusion within
the Region regarding which facilities ESD should inspect and
ESD's role in inspections versus enforcement. Similarly, the
Region IV RCRA program relies on ESD to perform a portion of its
comprehensive ground-water monitoring inspections and sampling
inspections. The headquarters review of Region IV found ESD's
RCRA workplan grossly insufficient to justify the total of 16
RCRA workyears allocated to it. Regions V and VI also provide a
significant number of workyears to ESD so that that division can
perform a portion of the RCRA oversight, sampling, and compliance
inspections. Just as in the other Regions, there are coordination
problems between the program offices and ESD which have adversely
affected Subpart F implementation.
Responsibilities for Subpart F are not always clearly defined
between permitting and compliance and enforcement. Often only
the permitting sections have the expertise to perform the in
depth analyses required to evaluate "adequacy" for purposes of
determining compliance and must make these resources available
to the compliance sections at the expense of permit writing.
-------
Available Resources.
While State and Regional resources to support the ground-water
monitoring program have improved since the first round of regional
program reviews (FY-83), staffing levels continue to remain
inadequate, especially at the State level. All Regions interviewed
for this study or reviewed by headquarters this fiscal year
indicated that resource problems are impairing their ability to
implement the ground-water program. The major dilemma appears
to be a shortage of qualified technical personnel to conduct
inspections and assess ground-water monitoring systems whether
for enforcement or permits. For instance Region I, Massachusetts,
which has about 20% of the Region's ground-water monitoring
facilities, has only one staff person with ground-water expertise
and that person is responsible for both CERCLA and RCRA work in the
State. The State has relied heavily on Region I for expertise in
this area. Both EPA Region I and headquarters have been unsuccessful
in their efforts to get the State to increase staffing for
Subpart F implementation. The State of Louisiana, which has
considerably more sites subject to the ground-water monitoring
regulations than Massachusetts, faces a similar problem in that they
too have only one ground-water specialist to handle all their sites.
State staff shortages are often aggravated by salary structures that
limit the State's ability to pay competitive wages. The State of
Indiana, for example, is plagued by severe resource problems. The
State's salary scale is so low it has difficulty attracting and
retaining qualified technical personnel. Last year, the State
tried to return the program to EPA because repeated efforts to
get the legislature to raise the salaries had been unsuccessful.
The State knew that it would not be able to operate a program
without a significant increase in resources. At the time of
this report the State had gotten authority to hire an additional
29 staff but the salary issue was still under consideration by
the legislature.
State hiring freezes present another resource obstacle. Despite
increased State grant funding, many States are prevented from
hiring additional personnel and are unable to meet their grant
commitments. For example at the time of this report both California
and Pennsylvania were under hiring freezes. Another problem
in this area is when a State agency has its personnel ceilings
set by a legislature that meets biannually and has very little
flexibility between sessions. This is the situation in Texas.
The Regions are also facing staffing shortages. Despite a substantial
increase in resources in FY-84 from FY-83 (144 workyears), some
Regions found it difficult to find competent, experienced,
hydrogeologists and other technical personnel to fill their vacancies.
While the Regional salary structure is generally much more competitive
than the States' it still is not as competitive as the private sector.
-------
Another difficulty in RCRA staffing is competition from the
Superfund program. In FY-84, the Superfund program also received
a dramatic increase in workyears. Some Regions gave priority to
hiring Superfund positions over RCRA. For instance, Region IVs
RCRA hiring plan in FY-84 did not allow the Air and Waste Management
Division to hire personnel above the RCRA workyear ceiling.
In other words, one person was being equated to one full workyear
regardless of the date of hire. In.effect, this decision guaranteed
unused workyear ceiling. However, The Region's hiring plan for
Superfund gave authority to the Air and Waste Management Division
to hire on-boards greater than the Superfund workyear allocation.
The Region planned on converting unused RCRA workyears into Superfund
workyears at the end of the fiscal year to cover the Superfund
overhires.
Region IV was not alone in its approach to the management of RCRA
resources. Region V also implemented a lapse management strategy.
Permanent workyears that are not utilized throughout the year are
reallocated by the Management division to other priority Regional
areas. These programs include: Great Lakes, Regional Counsel, and
Regional Management. Region III also did not give overhire authority
to the RCRA program but it did for the Superfund program. Moreover,
Region Ill's resource problem was exacerbated by the decision to *
impose a 7 percent tap on all workyears except Superfund in order
to establish a Regional pool of resources for later distribution.
Resource taps exist in almost all Regions and have contributed
significantly to the inability of RCRA managers to meet their
goals and program objectives.
State Authorization.
It has always been the intent of Congress to transfer program
responsibility from EPA to the States as quickly as possible.
EPA carried out this objective by placing priority emphasis on
State authorization early in the program. EPA has been successful
in terras of the number of States it has given interim authorization,
but recent Regional Reviews have shown that the Agency has been
negligent in ensuring that these authorized States have sound
RCRA programs.
Headquarters decided, based upon the RCRA statute, that it could not
look at past State performance in making interim authorization
decisions because Congress intended the interim authorization period
to be a time for States to develop their capabilities. Therefore,
during the past few years, the Agency has approved and authorized
State programs regardless of the State's capabilities to implement
an effective and credible hazardous waste program. Headquarters
developed an interim authorization process that looked at State
programs in terms of the "substantial equivalency" of their
statutes and regulations. That is to say, States were authorized
on the basis of their written regulations. They were not examined
for staffing levels or program experience. The Agency failed to
-------
react to the slowness of some States' enforcement process, or the
unwillingness of several to even file RCRA cases.
Oregon presents a good example of the failure of the interim
authorization process. Oregon's basic philosophy is that ground-
water monitoring is not particularly useful in the State. Its
program includes licensing procedures and other rules that allow
them to impose ground-water monitoring, however, they are not
required to force ground-water monitoring under State law. In fact,
until recently, none of the land-based facilities in Oregon were
monitoring their ground water. Oregon had determined, with little
documentation, that there is little risk to the ground water as
a result of facility location, or nature of wastes, etc. This
approach to hazardous waste management blatantly contradicts the
thrust of EPA's hazardous waste program and land disposal regulations.
In addition to States with weak regulations and questionable
attitudes, the Agency also authorized States that lacked the basic
staff and expertise to carry out a successful ground-water program,
along with States that lacked strong enforcement authorities to
assess penalties against major violators. These substantive program
deficiencies are major contributors to the low level of
compliance with the ground-water monitoring requirements in the f
regulated community.
Despite these inadequate hazardous waste programs, EPA has granted
interim authorization for Phase I to 48 States. These States are
responsible for implementing and enforcing the ground-water monitoring
regulations. This confined the EPA Regions' role to providing
technical assistance and oversight. However, EPA's ability to take
appropriate enforcement followup action if States failed to take
action themselves was also limited by the Agency's philosophy with
respect to authorized States. This philosophy essentially prohibited
the Agency from playing a strong oversight role and taking action
when necessary. Upon receiving interim authorization, States were
free to run their own programs with very little interference from
EPA. This attitude is reflected by the extremely limited amount of
information States were required to provide EPA on all aspects of
Phase I implementation. _ The FY-84 guidance represented the first
time the Agency attempted to collect data on implementation from
authorized States and marked a reversal of prior Agency policy.
Enforcement.
As mentioned earlier in this part of the study, because the States
have primary responsibility for implementing Subpart F, EPA's
ability to take enforcement actions has been limited. Until
recently, States in general relied on low level actions to achieve
compliance with little followup or escalation of activity if
facilities did not comply. States in Regions IV, VI, and IX,
in particular, have been reluctant to take strong enforcement action
in response to Class I violations. These "voluntary", approaches
-------
have not been very successful in returning facilities to compliance,
The situation in California typifies this approach. California
has the authority to issue compliance orders but generally does
not use it. The State prefers to talk to facilities regarding the
technical nature of the violations to encourage compliance. If this
approach fails, the State refers the case to the Attorney General's
Office where it may or may not be pursued.
Texas has a situation similar to California. The State can issue
compliance orders but has no penalty authority. Facilities that
do not respond to initial "jawboning", warning letters, or
compliance orders aren't penalized for their behavior. The only
recourse the State agencies have is to refer the case to the
State AG.
As will be explained in the 22 site study section of this report,
the quality of the ground-water systems being installed for
RCRA purposes has been found to be very poor. A major reason for
this is the reluctance on the part of the Agency to enforce its
265 performance standards. There are several reasons for this
but the two most cited ones are the difficulty in preparing a
compliance order for a performance standard and the difficulty in
winning the case when it is appealed to an administrative law
judge. Also the Region can spend a great deal of time, using »-
scarce technical resources, to prepare these cases, and the
"bean" counting system used to determine Regional "success" or
"failure", at least at the time of this study, is not set up
to account for level of effort per action. Hence there is a
natural disincentive for the Regions to pursue these cases.
-------
PART II ANALYTIC LABORATORY SURVEY
-------
BACKGROUND
There has beat some discussion over the past year concerning the
availabilityof commercial analytic laboratories that are capable
of performing RCRA required analyses. EPA's Office of Drinking
Water has recently estimated that for eight of EPA's 10 Regions
there are approximately 2000 analytic laboratories that are
capable of analyzing at least some of the parameters called for
under Subpart F of 40 CFR 265. This estimate includes: 1300
microbiological labs capable of testing for coliforms; 52
radiological laboratories capable of testing for radium, gross
alpha, and gross beta; and 700 general chemistry labs that can
test for most inorganics. These latter labs may or may not be
able to perform organic testing.
EPA had a consultant conduct a telephone survey of a sample
of independent commercial laboratories to address several
issues. (1) What RCRA parameters were these labs capable of
performing? (2) What did they charge for this work? (3) Did
they have gas chromatography or gas chromatography/mass
spectrometry equipment capable of performing complex organic
analysis? (4) Were they working at or near capacity i.e., could
they use more work? Quality assurance and quality control at
analytic labs was also addressed in this study.
r
METHODOLOGY
The laboratories chosen for the study were selected from "yellow
pages" directories nationwide. The laboratories chosen were
limited to those listed in the telephone directories that advertised
ground-water testing capabilities. The original list of 113
was reduced to 88 to account for eight laboratories which are
no longer in business, eight that could not be reached for
information, and nine which, despite their advertisement, do
not perform ground-water analyses. Although the final sample
of 88 laboratories does include facilities in all ten EPA
Regions, there was a high concentration of analytic laboratories
in certain areas of the country and limited capabilities in others.
Each of the 88 laboratories was contacted by telephone and
asked a series of questions about its capability and charges
to test for .30 ground-water monitoring parameters required under
the Interim Status Subpart F standards.1 These parameters address:
(1) characterization of the suitability of ground-water as a drink-
ing water supply (40 CFR 265.92(b)(1)); (2) establishment of
ground-water quality (40 CFR 265.92(b)(2)); and (3) indicators
of potential ground-water contamination (40 CFR 265.91(b)(3)).
^Information on testing for chromium, which is required by the
Subpart F regulations, was inadvertently omitted from the set
of questions asked of some of the laboratories. Based on those
labs that did provide information, it is assumed that the
availability of testing for chromium will be similar for other
metals-- over 90 percent. The price is assumed to be in the
$15 range.
-------
In addition to the parameters currently required under the Subpart F
regulations, each laboratory was asked about (1) its capability to
perform a volatile organic scan and an analysis for base/neutral
extractables, plus their costs, and (2) whether it had gas
chromatography (GC) or gas chromatography/mass spectrometry
(GC/MS) capability. Finally it was determined whether the lab
was currently operating at full capacity.
In collecting price data on the particular parameters, information
was requested on a single sample basis, in order to allow price
comparisons among the laboratories sampled. In some cases, how-
ever, price data were only available on an aggregate basis. For
those laboratories that could only provide aggregated data, the
total cost estimate was divided evenly among the parameters
included in the estimate. Where possible, the estimates of
prices for the different laboratories reflect the same analytical
testing method. Laboratories were also asked whether an additional
fee is charged for replicate results.
Not all of the 88 laboratories contacted were willing to divulge
price information for the tests they perform. These laboratories
were still included however, in the estimate of the percentage of
laboratories sampled that offer particular tests. Similarly, £
several of the 88 laboratories do not perform certain tests in-house
but rather send them out for analysis. While these laboratories
were not included as having the capability to perform these
particular analyses, their price estimates were included in the
analyses.
FINDINGS
Exhibit 1 presents data on prices of ground-water analyses charged
by the laboratories in the sample as well as the share of laboratories
with the specific testing capability. The prices and availability
listed in this Exhibit refer to those parameters required under
40 CFR 265.92(b)(l), (b)(2), and (b)(3) of Subpart F.
As shown in Exhibit 1, the costs to an owner or operator of sending
a single monitoring well sample for tests required by Subpart F
regulations to a laboratory can vary significantly. The wide
range in the price of tests was found consistently for all the
parameters examined. An especially large range was found for
the costs of analysis for certain herbicides, radioactivity, and
other organic tests. For example, the cost of testing for TOX
ranged from nine dollars to two hundred per sample.
The prices shown in Exhibit 1 include the costs for one sample
per well. The tests for indicator parameters required in
Section 265.92(b)(3), however, must be performed in quadruplicate.
Several of the laboratories that provide this service offer a
special discount for four replicates of indicator parameters
which is not reflected in the unit costs presented in the
Exhibit. Where the specified price included four samples,
-------
EXHIBIT I
PRICES AND AVAILABILITY OF GROUND-WATER ANALYSES
UNDER THE INTERIM STATUS SUBPART F REGULATION
Parameters
Price Rang*
Mean
Price
($)
Median
Price
($)
Percentage of
Surveyed Labs1
Offering Service
» T , - , . XT , V,,y
Required by
§265
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
.92(b)(l)
Arsenic
Barium
Cadmium
Lead
Mercury
Selenium
Silver
Fluoride
Nitrate
Turbidity
Coliforms
Radium
Gross Alpha
Gross Beta
Endrin
Lindane
Methoxychlor
Toxaphene
2,4 - D
2,4,5-T P Silvex
8
5
5
5
8
8
7
5
7
1
6
13
8
8
11
11
11
11
13
13
- 50
- 35
- 43
- 43
- 55
- 66
- 40
- 39
- 50
- 35
- 75
- 187
- 66
- 66
- 69
- 69
- 69
- 69
- 200
- 200
22
14
14
14
24
23
14
15
15
8
18
70
31
33
29
29
29
29
49
49
20
14
12
13
22
20
14
14
14
8
15
65
27
29
25
27
25
29
43
43
95
95
95
97
95
95
97
97
98
89
86
14
15
15
74
74
74
74
74
74
Required by
c *%/*e
§265
1.
2.
3.
4.
5.
6.
•92(b)(2)
Iron
Manganese
Phenols
Chloride
Sodium
Sulfate
5
5
13
5
5
4
- 35
- 35
- 130
- 35
- 35
- 35
12
12
27
10
12
13
10
10
25
9
10
12
97
97
90
97
97
97
Required by
c ^^c
§265
1.
2.
3.
4.
•92(b)(3)
pH
Specific Conductance
TOX
TOC
^,
0
2
9
8
- 15'
- 25*
- 200*
- 80*
4
6
67
26
.»
^
60
*> •
100
94
40
40
1 The total number of laboratories only includes chose 88 facilities
that perform ground-water analyses.
2 These prices account for only one sample; the regulations require that
four samples be analyzed.
-------
the total was divided by four to derive a per sample price.
As a result,, the per sample price presented may be slightly
under-estimated. However, because the regulations require
that these tests be performed in quadruplicate, costs to the
owner/operator for indicator parameters will usually be nearly
four time the estimates shown in Exhibit 1.
Exhibit 1 also presents both the mean and median prices charged
for each test to show the affect of the distribution of costs
on the mean estimates. Although the mean and median prices
were similar, both the mean and median estimates may still be
underestimates of the actual prices charged for individual
tests. This may be especially true in the case of the organics.
The mean and median prices may in fact be underestimates in
these instances because most laboratories offering these tests
quoted aggregate price packages for testing several organics
(e.g., costs of tests for multiple pesticides). To the extent
that discounts offered for packages were incorporated into the
prices quoted, the estimates in the Exhibit may be lower than
if the laboratories had provided prices on a per-parameter basis.
The prices presented in Exhibit 1 do not account for possible
discounts offered for large numbers of samples or to frequent
customers and thus may overestimate the actual cost incurred
by facility owners or operators. Most lab managers contacted
during the course of the survey indicated that discounts are
available for a large volume of samples (generally for quantities
greater than 10) and for contract work.
The prices shown in Exhibit 1 also do not include the costs of
replicate analysis. As mentioned previously, the regulations
require that four replicates be performed on the indicators
parameters. An owner or operator may wish to have replicate
samples for other parameters as well for purposes of quality
assurance. Most of the laboratories in the sample indicated
that, regardless of the number of samples involved, there
would be an additional charge for replicates above the number
necessary to satisfy their own quality control requirements.
Most laboratories did state, however, that replicates would
usually be offered at a discounted rate which would likely
vary depending on the number of samples involved. An estimate
of a fifty percent discount for replicates was often cited.
Exhibit 1 also shows the percentage of the 88 laboratories
surveyed that perform analyses for each parameter required
under the Subpart F regulations. Of these 88 labs, only seven
have the capability to perform analyses for all the parameters
listed under 40 CFR 265.92. Most of the laboratories in the
sample specialize in either inorganic or organic chemistry,
biological, or radioactive testing. As shown in the Exhibit,
most of the laboratories can test for inorganic salts and
-------
EXHIBIT 2
•
AVAILABILITY AND PRICES OF GROUND-WATER
REQUIRED UNDER THE INTERIM STATUS SUE
Percentage of
Procedure
EXHIBIT 3
AVAILABILITY OF GC AND GC/MS EQUIPMENT
Percentage of Surveyed
Equipment Labs Offering Service
1. GC only 24
2. GC/MS 35
1. Volatile Organic Scan 50 - 1,500 208 195 55
2. Extractables-(base/neutral) 40 - 1,500 307 275 47
-------
metals (95-97 percent) and a large share offer analyses of
pesticides and herbicides (74 percent). On the other hand,
only about 15 percent have the capability to perform radioactive
testing. The number of laboratories with the ability to test
for TOX and TOC was also limited (about 40 percent of the
facilities sampled).
Exhibit 2 presents information on the share of laboratories in
the sample capable of performing two analyses not currently
required under Interim Status — volatile organic and base/
neutral GC or GS/MS scans-- and the prices charged. As with
the parameters required by the Subpart F regulations, there is
an extremely wide range in prices charged for these two tests.
Because the prices presented in Exhibit 2 do not differentiate
between methods of analysis, some of the wide variance is due
to the fact that tests performed using GC/MS equipment are
significantly more expensive than those using GC alone. In
most cases, where laboratories have both GC and GC/MS
capabilities, GS/MS is used because of the more reliable
identification of specific compounds.
Exhibit 3 shows the percentage of ground-water laboratories £
surveyed that have the GC or GC/MS capability that is necessary
for performing volatile organic scans and base/neutral extractables
Of the sample of 88 laboratories, 24 percent have GC capability
only, and an additional 35 percent have GC/MS capabilities.
The current absence of a laboratory certification program and
standard laboratory methods for ground-water analyses could
call into question the accuracy and adequacy of laboratory
results under Subpart F regulations. To address part of this
problem, EPA is in the process of adding ground-water methodology
to the current technical guidance manual titled "Methods for
Evaluating Solid Waste" (SW-846). In addition, EPA is currently
drafting a regulation to require facility owners or operators
to use only those methods prescribed in SW-846.
Requiring an owner or operator to use only EPA prescribed methods
should help to improve the quality of test results by precluding
the use of unacceptable methods. Such a requirement may not
guarantee that the data produced by these laboratories will be
accurate. While establishing a laboratory certification program
would further increase the confidence in test results by allowing
EPA to specify that all tests be performed only by a certified
laboratory, the resources necessary to administer such a laboratory
certification program are not currently available.
Rather than develop a certification program, the Office of Solid
Waste is planning to establish a "lab evaluation program" where
EPA Regional, State, and EPA contract laboratories would be
required to participate, and independent laboratories could
-------
voluntarily participate if they so choose. Participating
laboratorie^.-iiould periodically receive evaluation samples
and if the llraoratory satisfies criteria specified by EPA
for the test sample it would receive a letter from Headquarters
stating that is has successfully completed the sample evaluation.
These letters could then be made available by laboratory managers
to all prospective clients. Evaluations would occur several times
a year. OSW began to phase in this quality control program
in 1984.
CONCLUSIONS
Based on the data provided by the 88 laboratories in the sample,
the costs to a facility owner or operator of using commercial
testing laboratories for all of its ground-water analyses required
under Subpart F could range from slightly over $200 to over $2000
per sample if a client chose the laboratory offerring the lowest
or highest prices for each of the tests required. The total costs
for all tests incurred by owners or operators will not likely
vary by a factor of ten because individual laboratories generally
did not consistently charge the lowest or highest prices for all
the tests offerred.
An important factor to consider when evaluating the results is *
the extent to which wide variance in prices charged is a reflection
on the quality of the laboratory. This issue was outside the
scope of this analysis. The establishment of a voluntary
laboratory evaluation program may, however, help to determine
the extent to which quality assurance and quality control affect
the prices charged for testing. Other possible explanations for
the wide range in prices may be (1) the laboratory's lack of
experience with a particular test and consequent underestimation
of costs; (2) the availability of package prices which resulted
in lower per parameter prices; and (3) different analytic methods
used.
All the laboratories contacted indicated that they were not
operating at full capacity and could expand to satisfy an increase
in demand for ground-water testing.
-------
PART III STATISTICS
-------
STATISTICS
Section 265.93(b) States that:
"For each indicator parameter specified in 265.92(b)(3), the
owner or operator must calculate the arithmetic mean and
variance, based on at least four replicate measurements
on each sample, for each well monitored in accordance with
265.92(d)(2) , and compare these results with its initial
background arithmetic mean. The comparison must consider
individually each of the wells in the monitoring system,
and must use the Student's t-test at the 0.01 level of
significance to determine statistically significant
increases (and decreases in the case of pH) over initial
background."
The main reason for using a statistical test is to allow the Agency
to specify a relatively neutral method for determining if incremental
changes in the ground-water quality near a regulated unit are being
caused by that unit. The t-test is a fairly powerful yet easily used
statistical method for doing this.
However, experience in the program has shown that t-test is not
performing efficiently due to a variety of problems that have little^
to do with statistics. Some of the more major of these problems
would include: 1) too few data points to compensate for naturally
occurring temporal changes in ground water quality; 2) too few
wells both up and down gradient to. compensate for naturally occurring
spatial ground-water quality differences; 3) insensitivity of the
designated indicator parameters to certain chemical species
of interest or at concentrations of interest; and 4) sampling and
analysis practices of facilities that introduce a tremendous
amount of variability into the data.
1. Too few data points. This problem relates to the specification
that the site sample its ground water quarterly for the first year
and semi-annually thereafter unless a "significant" problem is found.
Since the comparison is to a single base year the assumption is
made that that year will not be statistically different from other
years. However, a "wet" base year may in fact see recharge water
flowing through a different set of substrates than a dry year and
hence produce a different ground water chemistry— whether the
site is in the area or not. The converse with a "dry" base year
and wet comparison year is also true. In addition there may be
serious year to year quarterly variations depending on the weather.
For example in the base year a sample taken in early April may
be taken in the middle of the spring thaw while in the subsequent
year a late spring may mean that the snow cover is still there.
The difference in sampling at a point when considerable recharge
is occurring versus one where it is not may be significant. The
obvious answer to this problem would be to sample weekly over many
years to establish a true base and then start making comparisons.
This approach would be both costly and time consuming. •
-------
2. Too few wells. It would appear from the available evidence that
most sites that are in detection are opting to put in as few wells
as possible. This essentially means that the assumption has to be
made that the substrate that the single up-gradient well is in is
representative of the site substrate as a whole. While in some
cases this may be true it is not generally so. It also forces the
assumption that "background" water coming onto a site is uniform in
quality and hence the small window that is intercepted by the
upgradient well is representative of all water entering the down-
gradient wells' windows. In a nonindustrial area this would be a
dubious assumption. It is even more questionable in an area where many
sites are located upgradient from each other. The Agency has actively
encouraged sites to put in more than the minimum number of wells
specified in the regulations— both up and down gradient— but has
not been particularly successful in geting them to do so.
3. Insensitivity of the parameters. This problem will be discussed
in more depth in the next section. However, it relates to the
ability of the indicator parameter to respond to particular species
at concentrations of concern. For example, the most prevalently
used method for TOC measurement is sensitive only to about 1 ppm
and in general will not pick up volatile organics such as benzene.
This limits its effectiveness yet EPA at the present does not have
available another more effective single parameter that can be used *"
as a surrogate for these "missed" organics.
4. Sampling and analysis practices. By requiring replicates to be
run on each sample taken the Agency has attempted to account for
laboratory variability at a single point in time. It is not known
at this time whether this approach has been successful in achieving
its goal. However, the assessment part of this study has revealed
that many facilities also change labs over time. Replicates cannot
compensate for between laboratory variability. In addition, the
assessment part of this study has discovered that many sites are
not taking their samples properly and in many cases are using methods
that will eliminate constituents of concern or give highly varying
results between sampling events. No statistical test can be made to
account for these problems.
EPA has recognized these problems and is currently taking steps to
develop methods to compensate for directly related to the statistics.
The other problems will have to be solved by educating the regulated
community and through enforcement actions.
-------
PART IV SITE ASSESSMENT STUDY
-------
ASSESSMENT STUDY
Background*
Regulatory-- Subpart F of 40 CFR 265 was designed to be self
implementing. It consists of four major performance standards.
If a site operates a surface impoundment, landfill, or land treatment
unit, it must install a well system that is (1) Capable of determining
the facility's impact on the quality of ground water in the uppermost
aquifer underlying the facility (265.90); and (2) Capable of yielding
ground-water samples for analysis (265.91). It must (3) Develop
and follow a ground-water sampling and analysis plan that has to
include procedures and techniques for sample collection, sample
preservation and shipment, analytical procedures for required
parameters, and chain of custody control (265.92). Finally if it
is determined through statistical or other technique that the facility
is affecting the ground water then it must, as soon as is technically
feasible, implement an assessment plan that at a minimum determines
the concentrations of the hazardous waste or waste constituents, the
rate at which they are moving through the subsurface, and the depth
and areal distance they have reached to date (265.93).
The well system that is to be installed has certain requirements dhat
must be met. At a minimum it must consist of:
(1) Monitoring wells (at least one) installed hydraulically
upgradient (i.e., in the direction of increasing static head) from the
limit of the waste management area. Their number, location, and depths
must be sufficient to yield ground-water samples that are:
(i) Representative of background ground-water quality
in the uppermost aquifer near the facility; and
(ii) Not affected by the facility; and
(2) Monitoring wells (at least three) installed hydraulically
downgradient (i.e., in the direction of decreasing static head)
at the limit of the waste management area. Their number, locations,
and depths must ensure that they immediately~detect any statistically
significant amounts of hazardous waste or hazardous waste constituents
that migrate from the waste management area to the uppermost aquifer.
In the preamble to the regulation the Agency extensively discussed
these requirements and in the case of the minimum number or wells
(1 up and 3 down) made it clear that this minimum would probably
not be sufficient for any but the most special of cases.
A unit that would normally be required to install a complete ground-
water system could avoid doing so in full or in part if it could
demonstrate a low potential for migration of hazardous waste or
hazardous constituents from the facility via the uppermost aquifer
to water supply wells or to surface water. This waiver need not
be submitted to the Agency but must be kept on file at the facility
for review if requested.
-------
At the beginning of the program a facility was given the option of
entering into a detection monitoring phase or an alternate assessment
phase. The detection system makes an a priori assumption that the
site is probably not affecting the ground water. It calls for the
collection of four quarters of data for all RCRA wells which will then
be statistically compared to all subsequent semi annual samplings.
The parameters to be tested for in the semi annual events are far
fewer than those required in the background collection. The statistical
comparisons to be made are for four indicator parameters only (pH,
specific conductance, total organic carbon (TOG), and total organic
halogen (TOX)). If the comparisons show a statistically significant
increase (or decrease for pH) then the site must implement its
assessment plan and make the determinations required in 265.93.
If a site decided at the outset to enter into an alternate assessment
phase it was making an a priori assumption that it was affecting
the ground water. It would not in this phase be required to
collect all the chemical data required of a detection program
but would have to initiate, not later than one year after the
start of the program, the same determinations required under
265.93. Therefore one would expect the first rate and extent
determination findings to be submitted by the site sometime in 1982.
Study-- The site portion of Phase III, to which this section is
devoted, was designed to closely examine for adequacy the ground-water
systems of 20-30 sites across the country that were in the assessment
or final stages of the 265 program.
The initial plan was to request each of the 10 EPA regions to nominate
for study consideration 3 sites that were in some stage of assessment.
The preference expressed by headquarters for these sites was that one
have started in the alternate assessment mode and that two have been
"triggered" into assessment by a statistically significant change
finding. If possible these sites were to be in different States.
Due to various problems that included site visit timing, Regional
and State cooperation (Region VIII declined to nominate any sites) ,
and actual existence of sites in assessment (Region X had only two
and the California program was in such disarray that Region IX could
only offer one in Nevada) only 22 were finally selected.
The Regional break-out of these sites is as follows:
Region I-- 3 in Connecticut
Region II-- 1 in New York and 1 in New Jersey
Region III-- 1 in Pennsylvania and 3 in West Virginia
Region IV-- 3 in Alabama
Region V-- 1 in Ohio
Region VI-- 2 in Texas and 3 in Louisiana
Region VII-- 1 in Kansas and 1 in Missouri
Region VIII-- 0
Region IX-- 1 in Nevada
Region X-- 1 in Washington
-------
Of the 22 sites chosen, 11 started out in the alternate assessment
mode and 11 were triggered. The visits, which occurred between
October 1983 and April 1984, were timed to allow an EPA headquarters
person to be at the site when it was actually doing a regularly
scheduled sampling. Both EPA regional and State staff were invited
to participate at these visits for information exchange and critiquing
purposes. By being at the facilities during an actual sampling
round it was possible to evaluate their sampling procedures, check to
see if the sampling and analysis plan required by the regulations
was being followed, and by splitting samples at the end, check their
laboratory results against those of an EPA contract lab. To accomplish
this final end, the sites were asked to use their own equipment and
to furnish EPA with a set of sampling containers that had received
the same handling as the ones the site was going to use.
In addition to the sampling concerns, management at each site
was asked a series of questions about their ground-water monitoring
systems. These questions revolved around several major topics which
included the nature of the waste the facility handled, the chemistry
of the ground water, subsurface geology, well construction, well
evacuation procedures, and well placement. Documents requested
generally included well boring logs, the assessment plan, any
results from that plan's implementation, the sampling and analysis^,
plan, and any chemistry results. A copy of the questionnaire, *
which was developed after the first site visit, is attached as
Appendix A.
The information collected was then examined on a site by site basis
to determine how well the system met the performance standards.
-------
SAMPLING AND ANALYSIS
Section 265.92 of Subpart F requires that a facility subject to the
ground-water monitoring regulations develop a sampling and analysis
plan. The plan has to include procedures and techniques for sample
collection, analytical procedures for any required parameters, a
description of their chain of custody control, and sample preservation
and shipment procedures.
The criteria for judging the adequacy of the plan itself were that
it be written in such a manner that if two trained technicians were
to be given the plan separately and instructed to follow it the
results would be nearly, if not exactly, identical. Thus, if a plan
were to say the well is to have 3 well casing volumes evacuated
before a sample is taken without specifying how, it is deemed
inadequate because there are a wide variety of methods for doing this
which could be chosen. Similarly, if the plan merely refers the
reader to an analytical methods reference book such as Standard
Methods, without calling out the specific method, it would be considered
inadequate. This is because several methods are generally given per
analyte. Samples taken in sequential quarters may be tested differently
if the commercial or site laboratory is left to choose among these^
methods.
The above insistence on standardization of a site's procedures is
very important to the success or failure of the program. Both the
265 and 264 regulations are heavily dependent on variability analysis
conducted between upgradient and downgradient wells to determine if
a site is having an impact on the underlying aquifer. Ideally this
variability measurement should occur between two "pure" determinations
of chemical constituent concentrations in the water upgradient and
downgradient from a regulated unit. Variability introduced through
extraneous means such as evacuation, sampling, and testing methods
will tend to partially if not completely obscure what is actually
happening in the subsurface. This problem becomes especially
acute in 264 when alternate concentration limits (ACL's) are set
for a leaking unit. If the ACL is exceeded an expensive corrective
action may ensue.
Using these simple criteria, which do not judge whether a chosen
method is appropriate, but only that it is spelled out, 17%
of the plans present were adequate. About 20% of the sites did
not have a plan although they should have been inspected for this
at least once over the past three years. For each of the specific
plan requirements the results were:
(1) Procedures for sampling
adequately laid out 44%
(2) Sample preservation method
specifically named 67%
-------
(3) Analytic procedures/method
specifically named 50%
(4) Chain of custody procedures
spelled out including form
example 72%
Oddly enough, of the 8 sites whose plans provided specific and
adequate sampling procedures, they were followed by the staff
performing the sampling only twice.
This points up two very serious enforcement issues. First the
regulatory community (States and EPA Regions) does not
appear to be checking on the plan's presence. Second, a careful
records check would not appear to do the job since the site
personnel (contractors or in-house) seem to ignore the plan--
some are not even aware of its existence. The only sure way for
the government to guarantee that sites follow their plans would
be to perform an unannounced, spot, split sampling inspection,
that occurs at the same time as the site is scheduled to do its
normal sampling. In conversations with inspection personnel
during this study the view was expressed that this would require
a considerable amount of extra effort on their part since there
is currently no requirement that they be given advance notification
of a facility's sampling schedule. Indeed, it is not clear at
this point whether the government, under its current set of
regulations, has the power to require a site to share its sampling
schedule, if it has one.
The presence or absence of a plan, and its adequacy in spelling out
procedures to be followed is, however, only a very minor problem
compared with the actual procedures and equipment being used for
the sampling, changing of these procedures over time and changing
laboratory methods over time. These problems and a critique of
the 265 indicator parameters (pH, specific conductance, TOG, and
TOX) will be discussed below under 3 major headings: evacuation
procedures, sampling procedures, and laboratory analysis.
Evacuation Procedures
The science of evacuation and sampling is still in a state of flux.
For evacuation, the generally accepted procedure to prepare a well
for the taking of a water sample representative of that found in
the formation is to remove 3-6 casing volumes from the well if it
is a relatively productive formation; or to evacuate to dryness
if it is not. The actual sample taking should generally occur
within several hours of the evacuation procedure with a slightly longer
period being allowed for wells placed in very tight (low permeability)
soils. The accepted rules for evacuation are based upon the
assumption that water passing from a soil environment into a well
environment may undergo chemical transformations. The longer the
water is exposed to the well environment the more likelihood of
-------
chemical changes occurring. Since we are interested in the
chemical make-up of the water in a soil environment, it is
necessary to first remove the water that has been exposed to the
atmosphere. This is because over time gases will escape from
the water and exposure to air in the casing may result in
oxidation reduction reactions of the chemicals of interest.
In addition, it is critical to consider the evacuation procedure
in combination with the procedures for taking the actual sample.
In many cases these may not be compatible. For example, in a
highly productive formation water can enter the screen of the well
as fast as it is being evacuated. If a pump is used and its intake
placed at the bottom of the screen it is possible to withdraw
3-6 casing volumes of water without drawing down the stagnant water
above the pump intake. If a normal bailer with a bottom check valve
is then used for taking the sample (a bailer of this design takes
water from the top of the column) it may take in the stagnant water.
Other areas of concern were: whether the materials of construction
of the evacuation device would cause contamination of the water
through desorption/adsorption/absorption; whether the same procedure
has been maintained throughout the program; and how evacuated
water was disposed of. With few exceptions adequacy in this
area is very difficult to judge since there are not presently
any quantitative standards on the use of non-compatible materials £
or the number of casing volumes.
In summary:
1. Three of the sites used a stainless steel bailer for evacuation.
For the materials being tested for this was deemed adequate.
2. Five of the sites used PVC bailers for evacuation. Since all five
were looking for either specific organics or were testing for
TOG and TOX the use of PVC may cause interferences. It is not
possible to judge the significance of these interferences especially
since this is the evacuation step not the actual sampling.
3. Two sites used submersible gas bladder pumps. One of these was
equipped with tygon tubing and was used to test for organics.
Tygon tubing is generally unacceptable for organics. Whether
using this equipment for the evacuation stage of the sampling
will produce unacceptable interferences is unknown.
4. Six of the sites used submersible impeller driven pumps. Several of
these were dedicated. It was not possible to determine whether
the materials of construction would be a problem for the evacuation
procedure.
-------
5. One sice employed a gas displacement submersible pump. This is
not recommended for sampling but is generally acceptable for
evacuation.
6. One site used a dedicated forced air system which relied on
pumping air into the casing which had been made airtight. The
dedicated return tubing was tygon which may cause interferences.
7. Six of the sites used a peristaltic pump with tygon tubing. This
is generally acceptable for evacuation purposes, although the tubing
may cause interferences.
8. One site used a surface gasoline driven pump with dedicated
tygon tubing. The pump itself is accceptable for evacuation
purposes. The tubing may cause interferences.
9. Ten of the devices used above were lowered into the wells using
a synthetic cord. Some of these cords were dedicated, some new,
and some washed between wells. Cords of this sort have
some potential for releasing materials of construction, for
adsorbing other constituents of interest and they are practically
impossible to clean. This is not a good practice. To date
no one has quantified what affects these problems may have
on the sample itself. I
10. Eight of the sites had changed evacuation procedures at least
once since they began sampling. The affects of these changes
can not at this time be determined. However, it can be assumed
that any change in method will increase variability in the
sampling results.
11. Four of the sites either did not calculate the amount of water
they evacuated or did not evacuate a sufficient amount (at least
3 casing volumes or to dryness).
12. Only one site did not dispose of the evacuated water to the ground.
13. Eleven sites sampled the same day as evacuation; 5 waited 24 hours;
and 3 waited 48 or more hours before returning to the well for
sample taking. 5
Sampling Procedures
Once a well has been properly evacuated it remains to remove the
sample and transfer it to the appropriate containers for shipment
and analysis. There are several pitfalls in doing this, and extreme
care should be taken in choosing the method since each one has the
potential to yield different results. Over the past couple of years
a number of articles have appeared in the open literature on
sampling, and EPA has invested a good deal of effort in putting out
various guidance documents that are concerned in part with the proper
method to take a ground water sample. The main points to be concerned
with, and which this study looked for, are the compatibility of the
materials of construction with the anticipated chemicals to be found
-------
in the well; whether the physical method of recovery and transfer
will affect the chemicals in the water; whether, if the same equipment
is to be used for more than one well, proper cleaning procedures
are being employed; at what depth in the well the sample is being
taken; and whether the same sampling procedure is maintained
through out.
Materials of Construction. There is a wide variety of commercially
available sampling equipment from which a site may choose. Commercial
bailers are offerred primarily in teflon, stainless steel, or
polyvinylchloride (PVC). The standard bottom ball valve PVC bailer
is the cheapest and was chosen by more sites in this study than
any other. PVC bailers have the disadvantage that they are interactive
with organic compounds and can cause various interferences with the
analysis. As an example, one site had just purchased new PVC bailers.
At the split sampling performed at the site, a GC/MS base/neutral
scan was done. About the only chemicals picked up were phthalate
compounds which were probably plasticizers that had leached from the
bailer itself. Because of these interference problems (absorption,
adsorption, and desorption), the Agency considers a PVC bailer to
be inappropriate if the sample is to be used for organic analysis.
This would include two of the 265 indicator parameters-- TOC and TOX.
The next most used bailer material was stainless steel. In general
these are acceptable for most sampling work. Care should be taken,,,
however, if the water is suspected to be highly acidic and the *
constituents of interest include chromium or other alloy metals.
The four sites using stainless steel bailers in this study did not
have this problem and their sampling equipment was judged to be
adequate. Two sites in the study used teflon which to date has been
found to be acceptable for all hazardous constituents. There were
also two other sites that used home made equipment which consisted
of a quart jar placed in a metal cage of sorts with a line attached
to a cork in the jar's mouth. While crude, this was acceptable at
one of the sites but not at the other.
Pumps of various design can also be used to draw water from a well
and they generally take less time to do so than bailers. If the
water table is shallow (depth to water 25 feet or less) then
surface stationed pumps relying upon a vacuum may be used. The
most common of these pumps is the peristaltic. The tubing associated
with this type of pump must be very flexible to work. Historically
the material of choice has been tygon. For metals sampling this
is not known to cause severe analytic problems. However for
organics the tubing is a virtual sponge and should be avoided.
Research work performed under contract to EPA has shown even 6-inch
sections (which are threaded through the pump) to be detrimental
to the sample. Five sites were using this type of pump with
tygon tubing. All were deemed to be using inappropriate equipment.
At least two of the sites had this equipment recommended to them by a
major ground-water consulting firm.
-------
In locations where the depth to ground water is greater than 25 feet,
submersible pumps must be used. Two types were used by the sites
in this study-- gas bladder and impeller driven. Bladder pumps are
superior to airlift pumps in that they do not allow the sample water
to come into contact with the gas. However they can be made from
different materials, the most common being silicon rubber for the
bladder and tygon tubing for the return hose. Both of these materials
are inappropriate for organics. For this reason, all three of the sites
using them were considered to be using inappropriate equipment. This
is not to say that gas bladder pumps are not desirable for RCRA
sampling. However, the site should specify that the tubing and
bladder be made of chemically inert materials. Two sites used
dedicated submersible pumps that were impeller driven. The composition
of the internal materials was unknown, although neoprene was
suspected in one. Both pump systems had return tubing that was
constructed of a black polyethylene or polypropylene. Whether
the polymer or the material selected to make them black could
cause interferences is not known.
Also of interest in the materials of construction are the lines that
are used to lower the equipment into the wells. Eleven sites used
some form of wrapped or braided synthetic fiber (nylon for example).
What type of interferences, if any, these may cause by their presence
in the well water is not known. However, it is known that if they
are used for sampling more than one well, which was the case at £
many of the sites using them, one can expect some cross contamination
between wells because they are practically impossible to clean.
In addition, where bottom filling bailers are used, the method
of transferring the sample is generally to pour from the top. This
practice generally puts the sample water over and through the attached
cord. What this does to the sample is unknown, but the possibility
for introducing variability is there. Two of the sites used teflon
coated wire. Because teflon coated wire is easy to clean and relatively
inert, it provides the best protection against cross contamination
and possible interferences from the material.
Physical Recovery and Transfer Methods. This refers to how the
equipment moves the sample from the well to the appropriate containers
and whether this will affect the chemical constituents of concern.
For example, an air lift system (which was not used by any of the
study sites) is designed to introduce a gas under pressure into the
water column. As the gas moves up the well casing it carries the
water with it in a very turbulent state. The result is the stripping
away of dissolved gases and volatile organics and the possible
oxidation of dissolved cations to form, in some cases, insoluble
precipitates. This method should never be used to take samples
although it does have its use for well development.
Peristaltic pumps which utilize a vacuum system to draw their
water should never be used to sample for volatile materials.
(Five of the study sites did.) This is also true of impeller driven
pumps because the high speed rotation of the impeller creates a
pressure differential at the back of the blade that can strip dissolved
gasses and volatile organics before they leave the pump as well as
-------
pumping the water in a highly turbulent state. For the two sites in
this study that used dedicated impeller pumps there was the added
problem that they did not have a speed control but operated on an
on/off basis. This resulted in a high pressure stream being emitted
at the well head that guaranteed the water would be aerated.
Studies performed for EPA under contract seem to indicate that the
equipment least disturbing to the sample are either gas bladder
pumps or bailers. While this may be true for the bailers it is only
true for the gas bladders if they are properly maintained and are
equipped with a variable pressure gas control device. Several
problems were encountered by the personnel operating the
equipment at the study sites. All of the pumps used by the study
sites lacked the variable pressure control devices. This resulted in
the sample being ejected from the tubing in a turbulent, hard stream
which essentially aerated it. Second, the top check valve of several
of them became fouled at one point or other during the sampling
because of silt or other foreign matter. Most of the time the
operators would not recognize this problem immediately, but would rather
wait until the pump could no longer successfully pump any water before
making any repairs or adjustments. A close inspection of the pump's
sample stream during this process reveals that it will become very
much more aerated than it would without the partial obstruction of the
check valve. While this process may not strip all materials of £
concern, it will cause sufficient variability in the sample results
over time as to make interpretation of the data impossible.
The use of the bailers is also not trouble free. One site had
6-inch monitoring wells and used a very large stainless steel bailer.
When taking the sample they dropped the bailer, as opposed to carefully
lowering it, into the well. This had the effect of creating a surge
in the water column that agitated it causing possible loss of
dissolved gasses and a stirrring up of well sediments. Another
transferred the water from the bailer into the sample container by
releasing the bottom check valve with a gloved finger. The resultant
water stream was a turbulent spray that had the effect of aerating
the sample as well as putting it across a glove that may or may not
have release and/or other chemical agents on it. All the sites
using large bailers appeared to have problems in sample transfer
because of their bulkiness.
For volatile organic sampling, the less the sample is handled the
better. Several of the sites visited had a policy for at least
some of their wells of compositing the collected water before
transferring it to the appropriate containers. This is not an
appropriate method of sample handling which can be illustrated by
the fact that at at least two sites in the study EPA requested that
they not composite but rather transfer directly to the appropriate
container. At these two sites, the results were an order of
magnitude higher than when they had previously taken their samples.
Another problem with volatile organics, including TOX, is that of
headspace left in the containers before shipping. A much lower
contaminant level generally results as volatiles escap.e into the
air space in the container.
-------
Proper Cleaning of Equipment. For those sites that do not dedicate
their equipment to a particular well (15 of 22) proper cleaning is
essential to ensure that results are indicative of the water in
that well. There is currently no officially recommended guidance
on the proper method of cleaning equipment. In a recent study
sponsored by EPA it was suggested that when organics are the
constituents of concern that the cleaning be accomplished by
steam cleaning, followed by rinsing with a solvent such as hexane
or acetone, followed by rinsing with distilled or deionized
water. None of the sites looked at in this study used steam
cleaning as a procedure. The Agency at this point is not in a
position to say whether one method of cleaning is superior to
another. However, several of the sites in this study did make
use of transfer blanks to check their cleaning procedure. This
practice essentially has the technician clean the equipment
after sampling a contaminated well as would normally be done. He
then pours or pumps distilled water through it and into
the appropriate sample containers for subsequent shipment to the
laboratory as a regular sample. If analysis shows hazardous
constituents in this sample then one can assume that a cross
contamination problem existed during the sampling and the data
becomes suspect. At least two of the sites in the study that used
transfer blanks and had extensive cleaning procedures also had
those blanks turn up contaminated. It would probably be best t«
for a facility to test their cleaning procedures by use of transfeV
blanks on a regular basis and change the procedures if there
appears to be a problem.
Depth of Taking Sample. Depending upon the constituents of interest,
the depth within a well at which a sample is taken can affect what
concentrations of those constituents are seen in the subsequent
laboratory analysis. What appeared to drive the choice of sampling
depth at the sites looked at in this study was their choice of
sampling equipment. Since this choice was generally related to ease
of use or relative expense, not a conscious examination of the
constituents of concern, this could present a definite problem. Of
the 22 sites examined, only one had taken samples within their well
column to definitely ascertain if there would be a difference. After
several quarters of there being no significant difference, they now
only sample at the bottom of the well. On the other hand, another site
was using a bottom valve bailer, which only allows for top of the
column sampling, for a ground-water problem that was in part caused
by a very low pH brine. Since one can expect the brine to sink,they
were probably not showing the full impact of the unit on the water.
This of course cannot be ascertained for sure unless they do, in fact,
draw a sample from the bottom of the well and compare the lab results
to what they get when it comes from the top of the column.
Same Procedure. The above discussion has mentioned a number of the
individual problems to be encountered when using each of the common
sampling methods. Each of these sampling methods will probably impact
the sample in a different way. If a site chooses to mix sampling
methods over time, it is making a conscious (or unconscious) choice
-------
to introduce variability to its laboratory results. This is not a
a good idea, as the basis for decisionmaking in this program is
very dependent on minimizing variability. At least 9 of the 22
sites examined had changed sampling methods once since the start
of the program.
To sum up:
1. Fourteen of the sites used a bailer to perform some part of their
sampling. Of these 5 were PVC; 4 were stainless steel; 2 were
teflon; 2 were glass and steel; and one was brass.
2. Ten of the sites used a pump to perform some part of their sampling.
Of these, 3 were submersible bladder; 5 were surface peristaltic;
and 2 were impeller driven. All were judged to be inappropriate
for varying reasons.
3. Forty-one percent of the sites had changed sampling procedures
during the course of their 265 implementation thus making trend
or comparative analysis of the data impossible.
4. Fifty-five percent of the sites did not employ any kind of blanking
system to test for cross contamination either for sample shipping
or cleaning procedures between wells. This lack of a check makes
the data suspect. *"
5. Thirty-six percent of the sites used dedicated equipment for each
well. If the materials of construction for this equipment were
appropriate, this is the surest way to avoid cross contamination
questions.
6. Four sites studied did not clean their equipment between wells,
and of the eleven that did, it was not possible to judge adequacy.
7. All but one of the sites used the correct preservatives for their
samples.
8. All but one of the sites used the correct containers for their
samples although 4 of those using the correct containers did not
employ appropriate caps. The most frequent problem was the use of
poly-seal caps for TOX samples which should have had either teflon
lined or septum type.
9. Sixty-seven percent of the sites that tested for volatiles or TOX
did not check for headspace in their containers. This may result
in the total loss of constituents' of concern between the taking
of the sample and analysis.
10. Seventy-nine percent of the sites that tested for volatiles or
TOX did not use transfer methods designed to avoid the loss of
the constituents of interest.
-------
Laboratory Analysis.
RCRA sites have the option of using a commercial off-site lab or
utilizing on-site facilities if available. For this relatively small
study sample, the sites split about even between those using one or
the other with several using both as cross checks. Whether a site
chose to use off-site services was generally related to whether or not
it needed extensive laboratory capabilities to support its own
production OA/QC operations. For example, most, if not all, the large
organic chemical operations visited performed their own lab work as
they have full analytic (GC or GC/MS) instrumentation on site. On
the other hand, industrial concerns such as metal finishers or
refineries, which are generally not concerned with the actual chemical
composition of their products, usually sent their work to off-site
laboratories.
The Agency has no policy on these practices other than to strongly
object to the changing of laboratories over time. The split sampling
results obtained during this study amply document that 2 labs will
deliver different concentrationvalues for analyses run on basically
the same water sample. In some instances the contract EPA lab reported
higher values; in others lower. This is to be expected given between
lab variability from extraction efficiencies, methods, and precision
and accuracy. This variability could lead to incorrect decisions.
As a result, either the environment will suffer or the site will be £
forced into an unwarranted, expensive clean-up. Eight of the 22 sites
had switched labs at least twice and four, 3 times. Comparing the
quarterly or semi-annual chemical analyses of ground water at these
sites will not be a particularly fruitful exercise— especially in
light of the fact that most sites haven't kept track of the methods
used by the various labs.
While this study was particularly concerned with sites in assessment,
there was ample opportunity to examine data on the Part 265 indicator
parameters at many facilities. In addition, the Agency in a separate
effort is in the process of examining under contract how effective these
parameters are in detecting contamination. When one looks at the
preliminary results of the contract study combined with the data
collected under Phase III there are a number of tentative conclusions
that can be drawn.
1. pH has been responsible for many of the t-test findings
of significant increase (or decrease for pH). Yet it appears that
these changes are generally small and more related to natural seasonal
variation than to anything the site has done.
2. Specific conductance has also been responsible for many
of the t-test findings of significant increase. Its track record,
while better than pH, is still not good. It would appear that in most
cases it would be better to pick anions of interest or specific metals
for testing rather than specific conductance. This would probably
result in screening out most natural effects. A further problem that
-------
has been brought up, but not thoroughly documented, is that some
sites may not be correcting specific conductance for temperature.
Specific conductance is temperature dependent (about 2% for every 1
degree change in temperature). If a site laboratory either does not
have self correcting instrumentation, or does not hand correct the
figure then there is a chance that significant increase finding
could be unrelated to what is happening in the ground water. This
is especially true in areas where the water table is high and there
are significant seasonal temperature differences.
3. Data on the effectiveness of TOC is sketchy at this
time. However, there are several problems with how it is currently
being analyzed that would tend to impeach its credibility. The first
is its generally accepted detection limit of about 1 part per million
(ppm). This is too high for many of the chemicals of interest whose
level of concern is at the parts per billion (ppb) level. Second,
the most common method of analysis has the laboratory agitating the
sample for 5 minutes (some labs in this study reported agitation
times of up to 1 hour) with ozone or nitrogen to remove all the
dissolved inorganic carbon (C02 for instance). This has the unintended
affect of also purging the water of most if not all of the volatile
organics that may be in it. Since many industrial solvents are volatile
by nature, this test will not detect them. Hence, we do not presently
have in place an indicator that will address non-halogenated volatJLles
such as benzene or methyl ethyl ketone. One site mentioned another
analytic problem they had encountered with TOC. This site, for whatever
reason, had decided to filter their TOC samples before analysis. They
found, however, that this increased the levels found by about 20 ppm.
The reason given was that the filter itself would initially release
detectable levels or non-volatile organics.
4. The current test for TOX is relatively new. Two major
problems have been found in this study related to TOX. The first
is that many of the halogenated chemicals of concern are volatile,
yet the sites in taking their TOX samples do not treat them as such.
This results in an under reporting or non-detectable finding by the
site as it has purged all the volatiles prior to testing. The second
problem is with the laboratories themselves. An examination of
replicate data on TOX coming back to the facilities reveals a serious
precision and accuracy problem. That is the replicate numbers which
should be closely grouped to indicate valid analytical results are in
fact very widely spaced indicating unreliable data.
-------
WELL CONSTRUCTION AND PLACEMENT
The number of wells, how they are constructed, and their placement
relative both to the regulated unit and within the subsurface are
critical factors in determining whether or not a site meets the
performance standards of 40 CFR 265.90 and 91. To meet those standards
they must be able to give an indication of the impact the facility may
be having on the ground water and be capable of detecting any
subsurface discharge from the unit as soon as it happens.
The sites that were visited in this study were all in an assessment
mode; i.e., in the process of determining the rate and extent of
migration of an indicated discharge. Hence they should have in
place considerably more wells than they would if they weren't
trying to define a plume. Table 1 shows the distribution found
among the 22 sites.
Table 1. Number of Designated RCRA Wells.
Number of wells 4 5 6 7 8 9 10 11 12 13 14 15 and over
Number of sites 52211112020 5
The term designated is used because of the practice of some facility
managements of placing and sampling a large number of wells but only
reporting the results for the "designated" ones. In one case a site
had over 100 wells but only 4 "RCRA" ones and one of these was dry.
For the most part sites tended to stay close to the minimum required
under the regulations-- one up and three down-- and had not done
extensive geotechnical work to delineate a plume. The sites having
over eight wells were generally either industrial facilities with
widely spaced regulated units for which they placed one up and
three down for each, or were landfills with large acreages to
monitor.
Whether or not a site has an adequate number of wells is a matter of
balancing professional judgement and economics. In the opinion of
EPA, approximately 75% of the sites examined did not have an adequate
number. This determination was made based on several criteria and
site specific problems. One facility placed wells on 250 foot
centers and expected them to detect anything flowing between; another
placed wells on 100 foot centers in a highly fractured bedrock subsurface
where pollutant migration can be expected to follow very narrow,
discrete pathways. Another facility determined that there are
several possible subsurface pathways and then placed one well with
with a screening interval that will monitor only one pathway.
-------
These latter examples are concerned as much with well placement
as with their numbers. For the sample population 30% had placed some of
their wells in a position that it was unlikely that they could
intercept any migration from the regulated unit because they were
cross gradient from the ground-water flow direction.
Fully 50% of the sites appear to have placed their screens in such a
manner as to raise serious questions as to whether they can intercept
and give accurate indications of the concentrations of hazardous
constituents that may be leaving the regulated unit. The most serious
technical deficiency may be decisions that were made on where to
screen the well and whether subsurface conditions warranted more
expensive nesting or clustering techniques to monitor different flow
zones at the same well location.
There are a number of possible explanations for this happening.
For one, many sites are interpreting the uppermost aquifer to be the
first formation that is likely to be used as a potable water source.
The appropriate interpretation of uppermost aquifer is the first
formation capable of yielding significant quantities of water-- that
is, capable of transporting any release from the regulated unit off
site. The permeability of the subsurface soils should not be a
determining factor in deciding whether or not an aquifer exists.
The existence of water or saturation should be the determining factor.
A second possibility arises where the aquifer beneath the site has
several alternating soil layers-- generally very permeable sands
interspersed with relatively impermeable clays. Several sites in
this study chose to screen beneath the first clay layer.
Most materials that escape from the unit will not readily penetrate
the clay. The plume strikes the monitoring well at a point on the
casing and sampling does not show a problem.
A third possibility is that little attention is being given by the
regulated community to the physical/chemical properties of the materials
that may escape from their units. If, for example, the material of
interest will float, then the screen should by straddle the water
table and be long enough to account for possible seasonal fluctuations.
If the material is likely to sink, as is the case for many organics
and some brines, then the -screen should probably be placed at the
first relatively impermeable layer. If both these conditions exist,
then one well with one screening interval will in all likelihood be
insufficient.
-------
Finally, there is the question of expense. Cluster or nested wells
are more expensive to put in and require added testing and labor
(to evacuate more casing volumes). If a site believes it can get by
with one well and screening interval, then it is likely to choose that
until the State or EPA Regional Office tells it to do otherwise.
This study also collected information on actual construction and
completion techniques for monitoring wells. The drilling method of
choice was the hollow stem auger; distantly followed by water or air
rotary. A cable tool operation was employed by one site and at least
two reported using mud rotary— a technique which is strongly discouraged
by EPA for drilling hazardous waste monitoring wells because of its
potential for causing analytic interferences. In the majority of the
cases the equipment was not cleaned between borings. This raises the
issue of possible cross contamination problems which may invalidate the
first couple of sampling events conducted by the site to establish
background.
At least a quarter of the sites questioned did not know how or whether
their wells had been properly developed to ensure non-turbid, free
flowing water. Turbidity often causes problems with chemical analyses.
The lack of proper completion and development was evident in that 75% of
the sites had at least some wells yielding turbid samples.
£
PVC was the material of choice for 21 of the 22 sites for both
construction of the casing and the screen. No site used inert materials
such as teflon or 316 stainless steel. This raises some serious
questions since PVC has been shown to cause problems with sample
integrity through adsorption and desorption of various organics and
in some cases metals. There have, in fact, been some reported cases
(not among this study's sites) where the chemicals of concern in
the ground water actually dissolved the PVC screen.
The reason for choosing PVC is obvious— it costs approximately 10
times less than teflon. However, it is still a poor economic choice.
A single sampling of a well for 40 CFR 261. Appendix VII constituents
(much less Appendix VIII) can cost more than the entire casing/screening
costs of putting the well in with teflon. It should also be noted that
both 265 standards, which base decisionmaking on relative values, and
264, which tends, especially with the alternate concentration limit
(ACL) provisions, to look to absolute values, can require expensive
assessments and corrective actions that may actually be caused by well
casing interferences. EPA at this time is in the process of issuing a
guidance memo to its Regions instructing them to require that any part
of the well that will be in contact with the water be constructed of
inert materials.
-------
ASSESSMENTS
One of the major purposes of this part of the study was to determine
how sites were complying with 265.93(d) which requires them, after
(1) a determination of a statistically significant change in an
indicator parameter or (2) a declaration of presumed release
(265.90(d)), to define the lateral and vertical extent of pollutant
travel; how fast the pollutants are moving through the subsurface;
and what they are. In a phrase delineate the plume.
What was found was somewhat dismaying. Of the sites examined
23% had not determined hazardous constituents of concern; 77%
had not determined the rate of transport of these constituents;
and 91% had not determined the extent of the contamination. In
short, all of the sites had detected a problem, but had done
very little work beyond this. In addition, several sites had
problems that would obviously extend beyond their property lines.
In this case other Statutory authorities would have to be relied on.
The poor showing of the sites in assessment in complying with
Subpart F only confirms what EPA Regional permits 'staff have been
saying for quite some time— that most if not all of the permit
applications they receive are critically deficient in information about
ground-water quality and subsurface conditions. Without this information
the Agency can not write a permit. The end result is a delay of up
to several years while the site puts together a basic system that
should have been operational since November 1981, and possibly an
extended assessment system that should have been in place since the
summer of 1983. The malfunctioning of our Interim Status
detection/assessment program as pointed out in this report is going
to have a serious impact on our ability to put in place a program of
permitted facilities.
-------
U S. Environmental Protection Agency
Region V, Library
230 South Dearborn Street
Chicago, Illinois 60604
------- |