MAY 1982
ENVIRONMENTAL MONITORING
AT
LOVE CANAL
INTERAGENCY REVIEW
Comments by:
U.S. Department of Health and Human Services
National Bureau of Standards
U.S. Environmental Protection Agency
Office of Research and Development
U.S. Environmental Protection Agency
401 M Street, S.W.
Washington, D.C. 20460
-------
FOREWORD
In view of the long-standing involvement of the Depart-
ment of Health and Human Services (HHS) in activities at Love
Canal, that Department prepared the enclosed report "HHS
Evaluation of Results of Environmental Chemical Testing
Performed by EPA in the Vicinity of Love Canal, Implications
for Human Health". Included are individual comments by the
scientists and physicians who participated in the HHS evalua-
tion, and copies of communications between EPA and HHS.
In the fall of 1981 EPA requested the National Bureau of
Standards (NBS) to- review the EPA program directed to organic
chemicals, with particular attention to the analytical methods
and quality assurance procedures which were adopted. In
response, NBS prepared the enlosed report "Review of Material
Provided by EPA in the Analysis for Organic Chemicals in the
EPA Love Canal Monitoring Study," dated May 10, 1982. Copies
of communications between NBS and EPA are also included.
Since the NBS Review raised a number of analytical questions,
EPA subsequently prepared the enclosed response, "EPA Observations
on the Review of the National Bureau of Standards of the Love Canal
Monitoring Program."
Courtney Riordan
Acting Assistant Administrator
for Reserach and Development
-------
CONTENTS
TAB
HHS "Evaluation of Results of Environmental
Chemical Testing Performed by EPA in the
Vicinity of Love Canal"
NBS "Review of iMaterial Provided by EPA on
the Analysis for Organic Chemicals in the
EPA Love Canal Monitoring Study" II
EPA "Observations of the Review of the
National Bureau of Standards of the
Love Canal Monitoring Program" Ill
-------
H
H
S
-------
HHS EVALUATION'OF RESULTS 0? E;fVIRONM£NTAL CS2HICAL
TESTING PERFORMED B1? EPA IN T513 VICINITY OF LOVE CANAL
IMPLICATIONS ?OR HUMAN HEALTH
FURTHER CONSIDERATIONS
CONCERNING UAB1TABILITY
S'or CDC: Clark W. lisath, Jr. t H,D,
T^r.ate D. Ki-.broug-h, M.D.
John A. Liddle, Ph.D.
For NIBHS: David P. Rail, M.D,, Ph.D.
Waltsr J- Rogan, M.D.
JULY
, >
H
-------
On Hay 12, 1982, the National Bureau of Standards (S3S) made available to the
of Health and Human Services (HHS) its assessment of the
tal Protection Agency's (S?A) Love Canal Monitoring ^rojacc, Ws
reviewed this assessment find pointed out (June 14., 1932) th£C our esrU^r
statemeni: about hahitabsiity of tha Love Osnai area was predicated on adeq'jats
ehessical -methodology r "Should the K3S review suggest Efithodologic dif ficul ties
in the E?A data sec, the health evaluations presented hare would require
reassesmnent" (T»HS, October 7. 1981). Since the May 12 HSS asssasnrent iraplied
that SpAfs tech-nical ciethods sight be inadaquata in the ar-ea of method
detection limits, we felt compelled to withhold judgment about habitabilitj? -
Since that tiiae; additional information was provided to HHS by EPA on July 9,
1932., regarding reliability of method dsesction liaits and recovery dats for
organic compounds measured in environmental specimens froai tn& Love Canal
araa, E?A assures1 that levels previously dasigriarsd 33 "trace" or "aor
detecre^" are nsoar unlikely to have sxceedsd val'-ie* in the low paTts p=r
billion (ppb) range and in no case would represent a value grsstsr than 1 pare
per million. "We judge chst levels of organic chemicals in the low ppb range
present: minimal health risk, escape in the case of TCDD for -which levels in
the ppb range are clearly of toxicologic significance and which has been
identified in storm sawars and related outflow straps in the Love Canal
In the uiasntiroe, K3S has £-ontinaad its eKamination of the technical ,-sechodg
used by EPA in developing the data- Based on tha findings of this further
review (in psrt abiding from additional inf or^ation provided by 5?A to Ls3S on
June 23, 1982), NBS has indicated by letter to E?A on July 9, 1932, that irtha
-------
sechods of analysis used by B?A for water, soil a^d aediaants, and ai-r
generally dcceptable methods ar*d reptessr.t tha state of the art." We ass'-se
...'
the provisos on this statement from H3S can bs met fey S?A*
iwj
In view os the farther'information from SPA and of the most recent HSS opinion
regarding the methods used'by 2PAj we feel that the ststcmenc concerniag
habitabiliSy of she Love Caaal area, as expressed in the WHS avaluation dated
October 7> 193i} is accurate. In chat evaluation, we .stated'that "Che I-ove
C4nal ar«af outside Area 11, is aa habitable as the control sress with which
it vas compared.'-' Again, as srsced in that document, this jed^aartt r5^ardi^§
habitability ir.civdas the requireinST:-ts that Area 11 (the Canal sits itself and
the land occupied by the Cwo rings of homes Burroundifift it) be constantly
safaguardad agsir.st future leflkaae ffon the Canal and that cleanup is
for £sisrin§ contamirtAftion of local storra severs and their drsiTift£s traces.
-------
DEPARTMENT OF HEALTH & HUMAN SERVICES
Public Health Service
Office of the Assistant Secretary
for Health
Washington DC 20201
John W. Hernandez, Jr., Ph.D.
Deputy Administrator
Environmental Protection Agency
401 "M" Street, S.W.
Washington, D.C. 20460
Dear Dr. Hernandez:
I am sending you a modification of the September 18, 1981, Department of
Health and Human Services (HHS) evaluation of the health implications of the
Environmental Protection Agency (EPA) environmental monitoring at Love Canal.
The HHS scientists modified their earlier report after carefully considering
the National Bureau of Standards review of the environmental monitoring
program.
Since the modification represents a significant change in the tentative con-
clusions drawn by HHS in the earlier review, I am asking that it be attached
to the document I sent to you on November 18, 1981.
We will also continue to work with you and Dr. Dewling, your coordinator for
the final preparation and release of the EPA final report, to help interpret
our findings to the citizens and other interested parties.
As I stated in my last letter, I believe it is important for the Department of
Health and Human Services to review EPA's final report and press release
before these documents are made public.
Sincerely yours,
Edward N. Brandt, Jr., M.D.
Assistant Secretary for Health
Enclosures
-------
HHS EVALUATION OF RESULTS OF ENVIRONMENTAL CHEMICAL
TESTING PERFORMED BY EPA IN THE VICINITY OF LOVE CANAL
IMPLICATIONS FOR HUMAN HEALTH
MODIFICATION OF THE REPORT OF
SEPTEMBER 15, 1981
For CDC: Clark W. Heath, Jr., M.D.
Renate D. Kimbrough, M.D.
John A. Liddle, Ph.D.
For NIEHS: David P. Rail, M.D., Ph.D.
Terri Damstra, Ph.D.
Walter J. Rogan, M.D.
James D. McKinney, Ph.D.
June 14, 1982
-------
In September 1981, we prepared a report evaluating the human health
implications of the environmental chemical testing performed by the
Environmental Protection Agency (EPA) in the Love;Canal area during the summer
and fall of 1980. In that report we indicated that our evaluation rested "on
the assumption that the methods used by EPA in collecting, storing, and
testing specimens were satisfactory." The evaluation of the adequacy and
appropriateness of the methods used by EPA has been independently reviewed by
the National Bureau of Standards (NBS). The final NBS report regarding the
conclusions of that review, together with a revised draft of EPA's detailed
report on the total study, were made available to us by EPA on 12 May 1982.
In view of the criticisms and concerns expressed in the NBS report, we must
modify the tentative conclusions expressed in our earlier report.
The NBS review concludes that EPA did not adequately address the problems of
limits of detection and sensitivity of the analytical methods used. In the
instances where EPA detected contamination, it is reasonable to conclude that
chemicals were present. However, in the instances where EPA reports the
absence of contamination (about. 90 percent of the values reported), no strict
quantitative interpretation is possible, and judgments about the habitability
of the area cannot be based on these values. Thus, with the data now
available to us, no definite recommendations or conclusions as to the
habitability or the potential human health risks of the Love Canal area can be
made. The conclusions about the necessity for clean-up of areas shown by EPA
to be contaminated still stand.
-------
DEPARTMENT OF HEALTH & HUMAN SERVICES
Public Health Service
Office of the Assistant Secretary
for Health
Washington DC 20201
John W. Hernandez, Jr., Ph.D.
Deputy Administrator
Environmental Protection Agency
401 "M" Street, SW
Washington, D.C. 20460
Dear Dr. Hernandez:*
I am pleased to send you the report of the Department of Health and Human
Services, including all comments provided by outside consultants, regarding
the health implications of the environmental chemical testing conducted by
the Environmental Protection Agency (EPA) in the Love Canal neighborhood in
Niagara Falls, New York.
Our evaluation focused on the potential for adverse health effects resulting
from exposure to the levels of chemicals identified in the EPA environmental
monitoring project. The sampling plan and analytic procedures were not
reviewed.
Please let me know when and how you plan to convey the findings of the EPA
testing to the Love Canal area residents so that we can ensure that our
evaluation is released simultaneously.
I look forward to continuing our mutually cooperative efforts to respond to
these difficult environmental and public health issues.
Sincerely yours,
Edward N. Brandt, Jr., M.D.
Assistant Secretary for Health
Enclosure
-------
HHS EVALUATION OF RESULTS OF ENVIRONMENTAL CHEMICAL
TESTING PERFORMED BY EPA IN THE VICINITY OF LOVE CANAL
IMPLICATIONS FOR HUMAN HEALTH
For CDC: Clark W. Heath, Jr., M.D.
Renate Kimbrough, M.D.
John Liddle, Ph.D.
For NIEHS: David P, Rail, M.D., Ph.D..
Terri Damstra, Ph.D.
Walter Rogan, M.D.
October 7, 1981
-------
In the summer and fall of 1980, the Environmental Protection Agency (EPA)
performed extensive environmental chemical testing in the vicinity of Love
Canal and in comparison areas in Niagara Falls, New York. In the summer of
1981, the Department of Health and Human Services (HHS) agreed to provide EPA
with an independent evaluation of the results of those tests with respect to
their implications for human health (Appendix A). The HHS review, conducted
by the Centers for Disease Control (CDC) and the National Institute for
Environmental Health Sciences (NIEHS) of the National Institutes of Health,
made use of written opinions provided by 11 non-federal expert consultants, 10
of whom met at CDC in Atlanta on August 13, 1981, to discuss the EPA dat'a.
Prior to the meeting, the consultants were provided with the EPA data in
condensed form. The letter from CDC which transmitted these data (Appendix B)
asked that each consultant review the data with respect to four particular
questions:
1. Are the concentrations measured significantly different from levels
found in other areas of Niagara Falls?
2. Do the levels measured represent concentrations that could cause acute
or chronic adverse health effects in people living in the Love Canal
area beyond what might be expected under usual residential conditions
in the Niagara Falls area?
3. Were sufficient samples taken and analyzed to determine the extent of
contamination by particularly hazardous materials and to be able to
assess their potential threat to human health?
4. Based on available data, can you conclude that the area is not
habitable?
Presented below are responses by HHS to these four questions, taking into
account the written reports of the 11 consultants (Appendix C) and preceded by
a discussion of certain general considerations.
-------
Page 2 - Implications for Human Health
A. General Comment^
!• Format of the EPA Report. The results of the EPA testing have not
yet been assembled, in comprehensive, concise format suitable for standard
scientific publication. The present review was conducted using computer
printouts of test results. This was done so as not to delay the health
assessment of test results and so that consultants would have the
opportunity to weigh the data without prior interpretation or analysis.
Since the great bulk of testing (over 90 percent of the more than 400,000
measurements) found no detectable or only trace amounts of chemicals, HHS
opted to have EPA condense the data for review into a format focusing
primarily on those test results which had yielded finite values. However,
conclusions presented here regarding these data may need to be reevaluated
when EPA presents the data in full scientific format.
2. Methodology of Chemical Testing. The evaluation of data provided
here is made on the assumption that the methods used by EPA in collecting,
storing, and testing specimens were satisfactory. Independent review of
these aspects of the data is currently being performed for EPA by the
National Bureau of Standards (NBS). Should the NBS review suggest
methodologic deficiencies in the EPA data set, the health evaluations
presented here would require reassessment.
3. Consultant Opinions. While the written opinions of consultants
(Appendix C) vary in their interpretation of different aspects of the EPA
data and in their particular responses to the four questions, we would
judge that sufficient consensus is present to permit conclusions on the
major issues involved. In this context, it is important to stress that
-------
Page 3 - Implications for Human Health
all consultants agree that levels of chemicals detected in storm sewers and in
Area 11 (the Canal itself and the first two rings of houses surrounding the
Canal) exceed acceptable levels and represent a potential for increased health
risk if remedial actions are not pursued and if human access is not
controlled. All consultants also agree both regarding the difficulties
involved in assessing the large mass of data presented, especially in its
current format, and regarding the urgent need for EPA to develop as quickly as
possible a concise report of the data, adequate for peer-reviewed scientific
publication.
B. Response to Questions
Question 1. Levels of chemicals in the Love Canal declaration area,
excluding storm sewers and Area 11, are within the same range of low
values (parts per billion) as levels in control areas elsewhere in Niagara
Falls. For many chemicals, levels are below detectable limits in both
declaration and control areas. Distinctly increased levels are present at
the Canal itself and in some homes abutting it as well- as in material from
storm sewers at many points in the entire area.
Question 2_. In general, levels of chemicals in media (air, water,
soil) to which people are exposed in the Love Canal declaration area
(again, excluding storm sewers and Area 11) are well below established
regulatory or advisory exposure limits for those identified chemicals
where guidelines exist and, by inference, for closely related compounds.
In this context, these levels can be judged not to present risks to human
health different from those in the control areas.
-------
Page 4 - Implications for Human Health
This assessment must be interpreted with an important reservation that
applies to all such toxicologic judgments. Official safety limits have by
no means been established for all possible chemicals, and they are
developed for exposure to single chemicals alone, not for combinations.
Full data are therefore necessarily lacking on which to base truly
complete judgments of chemical toxicity in the Love Canal setting.
Question 3. Although the large number of environmental samples
obtained and analyzed by EPA appear to be more than sufficient for
numerical comparisons between declaration and control areas, the sampling
framework in which specimens were collected is not clear from material
presented. To permit a final judgment concerning adequacy of sampling in
the Love Canal area, whether for particularly hazardous materials or for
chemicals generally, it will be necessary for EPA to provide a full and
concise description of the study's sampling plan as well as estimates of
statistical power, given the numbers of samples analyzed in declaration
and control areas.
Question 4. Any judgment regarding the future habitability of the
Love Canal area rests on two important requirements. The first
reservation is that appropriate measures must be taken to clean up the
obvious contamination of local storm sewers and their drainage tracts.
Second, the security-of Area 11 must be reevaluated to guarantee permanent
containment of chemicals in the dump. To assure habitability into the
indefinite future, it is essential that optimal containment methods are
installed and maintained and that continuous safeguards are observed
-------
Page 5 - Implications for Human Health
to prevent further leakage from the site either through erosion of the
clay cover or through its displacement by movement of dump contents. Such
safeguards will require 1) surveillance of the site through regular
environmental testing of the site drainage system and 2) full maintenance
of both the drainage system and of the clay cover. Provided these
requirements are fully met, we conclude, based on the analytic
measurements of chemicals in the declaration and control areas presented
by EPA, that the Love Canal area, outside Area 11, is as habitable as the
control areas with which it was compared.
We reach these conclusions concerning the Love Canal area fully recognizing
differences of opinion offered by consultants who have assisted in this
evaluation. At the same time, we consider .it absolutely essential that EPA
prepare as soon as possible a full scientific report describing the
environmental testing performed at Love Canal. The complexity of the data set
and the far-reaching implications which this extensive environmental
investigation carries for future approaches in similar toxic waste problems
requires prompt and comprehensive communication.with the scientific community.
Attachments:
Appendix A, HHS/EPA Agreement
Appendix B, Letter from CDC to Consultants
Appendix C, Reports from Consultants
-------
HHS CONSULTANTS REVIEWING EPA LOVE CANAL REPORT
Dr. Steven Aust
Department of Biochemistry
University of Michigan
East Lansing, Michigan 48824
Or. Richard Browner
School of Chemistry
Georgia Institute of Technology
Atlanta, Georgia 30332
Dr. John Doull
Department of Pharmacology and Toxicology
University of Kansas Medical Center
Kansas City, Kansas 66103
Dr. Joseph Highland
Chairman, Toxicology Chemical Program
Environmental Defense Fund
1525 - 18th Street, N.W.
Washington, D.C. 20036
Dr. Robert Metcalf
University of Illinois
Departments of Entomology and
Veterinary Pharmacology
Urbana-Champaign, Illinois 61820
Dr. Samuel Mil ham
6428 Guerin, S.W.
Olympia, Washington 98502
Dr. Robert A. Neal
Chemical Industry Institute of Toxicology
P.O. Box 12137
Research Triangle Park, North Carolina 27709
Dr. Beverly Paigen
P.O. Box 145
Mount Desert, Maine 04660
Dr. Edo Pellizzari
The Research Triangle Institute
P.O. Box 12194
Research Triangle Park, North Carolina 27709
Dr. Robert Tardiff
National Academy of Sciences
2101 Constitution Avenue, N.W.
Washington, D.C. 20418
Dr. James Whittenberger
Harvard University
665 Huntington Avenue
Boston, Massachusetts 02115
-------
MICHIGAN STATE UNIVERSITY
DEPARTMENT Of BIOCHEMISTRY • BIOCHEMISTRY BUILDING KAST LANM.NG • MICHIGAN • 48824
August 19, 1981
Dr. Clark SV. Heath, 3r.
Director
Chronic Diseases Division
Center for Environmental Health
Center for Disease Control
Department of Health and Human Services
Atlanta, GA 30333
Dear Dr. Heath:
This letter constitutes my assessment of human health implications of environmental
data from the Love Canal area. I would like to present a general assessment of the data
and the meeting of August 13 before I address the specific questions in your letter of
August 6, 1981.
In general, I believe this study was fairly well conducted and that the results do
not convey reasons for alarm. However, I am particularly disturbed by two observations
that may, unfortunately, be related. The first involves the non-scientific manner in
which the data was presented, and the second concerns the fact that the EPA officials
were apparently unconcerned about the first. The callous nature of EPA officials towards
the environment they have the obligation- to protect is exemplified by their attitude
toward those asked to give recommendations to the agency. I can think of no better
way to start any EPA activity on a more sure road to its downfall.
First of all there are many examples of very poor science. The PCS data was presented
as ppb of a certain Aroclor mixture, which is scientifically unsound. If this was the
only feasible method then it should be explained. If peak matching was used to see if
the environmental pattern was the same as an Aroclor, what was the confidence level?
Is it possible that no PCS would be reported if the GC elution pattern did not fit any
of the standard Arocior GC patterns? This is an example of the lack of adequate explanations
also. Tables and graphs should always be able to stand alone such that they can be readily
understood by the reader. We never did get an explanation of some tables and maps.
I feel very uneasy about drawing any firm conclusions from the data presented or from
' the discussions held in Atlanta,
Another example of poor science is the lack of sufficient controls and definitions
of controls. The lack of sufficient explanations and definitions is apparent throughout
the report. The only aspect of the report adequately addressed is the quality assurance.
First of all the samples were not described. Secondly, no examples of raw data were
provided. Composites were uninterpretable. The maps and tables were not adequately
described, difficult to read, and frequently impossible to interpret. Then, at the meeting,
we were told to disregard some of the tables, presumably because they were in error.
But when asked to explain the error Dr. Deegan could not do so.
MSU if >"> tVHtnatier Ac tut*/Equal Of>fnriiutity Institution
-------
Dr. Clark W. Heath, Jr.
August 19, 1931
Page 2
I would have preferred to see examples of raw data, examples of data handling,
summary tables, and statistics. The lack of statistics was particularly disturbing. Any
statistics should have been accompanied by a rationale for using the particular statistical
method. This data is going to come under very close and complete scrutiny. And, various
people will be looking for various things. This was pointed out at the meeting by rhe
consultants. Some wanted to see raw data while at the other end of the spectrum some
only wanted summary tables. In order to satisfy each and every consultant it is necessary
to provide at at least examples of all aspects of the data.
I will now reply specifically to your questions in the letter of August 6.
1. Are the concentrations measured significantly different from levels found in areas
of Niagra Falls?
I am assuming that the question refers to areas 1-10 as compared to area 99, and
that "significantly different" does not mean statistically significantly different. Certain
chemicals certainly appear more frequently in areas 1-10 than they do in area 99. However,
the levels of most do not seem to be disturbingly high. It would appear that the migration
of chemicals did not occur to a great extent beyond area 11. I would recommend a thorough
evaluation of the data and the site such that migration to any area beyond area 11 be
identified and that remedial action be taken.
2. Do the levels measured represent concentrations that could cause acute or chronic
adverse health effects in people living in the Lovel Canal area beyond what might
be expected under usual residential conditions in the Niagra Falls, area?
If the remedial action which has and is being taken is as effective as it appears
to be, there would seem to be very little chance for acute or chronic adverse health
effects for people living in areas 1 through 10. However, the data should be carefully
analyzed and an on-site evaluation made to determine if any migration avenues extended
outside of area 11. I make this comment with reference to the identification of obvious
migration routes. The sand lenses must be thoroughly investigated to determine their
extent.
3. Were sufficient samples taken and analyzed to determine the extent of contamination
by particularly hazardous material and to be able to assess their potential threat
to human health?
The number of samples taken seems to be sufficient but it may be impossible to
guarantee that no small but excessively contaminated "hot spots" don't exist. This statement
is made, however, in light of very little explanation of sampling locations. It is my understanding
that an attempt was made to sample areas which would be most likely to be contaminated
but the fact cannot be guaranteed. Another serious fault with the study is the lack of
sufficient control samples. The number of times chemicals were found in control samples
is indeed less than the number of finite levels in the declaration area samples but the
number of control samples was much less.
The particularly hazardous substance of concern is most probably TCDD and the
number of analyses for this compound is limited. However, an area of contamination
was found and it must be dealt with. Fortunately, the area is not one normally contacted
-------
Dr. Clark W. Heath, Jr.
August 19, 1981
Page 3
by people living in the area. I am speaking specifically of the storm sewer system. But
its contamination must be noted and corrective action must be taken to prevent indirect
exposure to TCDD.
It must be noted that some concern should be expressed considering the number
of chemicals in this area. Although the levels do not seem alarmingly high, an inappropriate
mixture may exist. With the possible presence of both initiators and promoters the likelihood
of carcinogenesis may be increased.
4. Based on available data, can you conclude that the area is not habitable?
I cannot conclude that areas 1 through 10 are not habitable, although I must admit
that part of my inability to make this conclusion stems from having very little definitive
data or a complete analysis of the data. This hesitancy also stems from a lack of confidence
in analyses of environmental samples. It would have been desirable to be able to assess
the ability to duplicate data in all samples. This would be particularly important for
soil samples where recovery may be variably affected by binding of the chemicals to
the soil. However, my confidence in this conclusion is strengthened by the absence of
any highly contaminated samples. Also, the biota did not reveal the presence of any
bioaccumulated substances.
Finally, there were some very positive results of the study. The best example of
this is the fact that the theory that chemicals may have readily moved via the old swails
seems unproven. Also, there does not seem to be evidence for at least major migrations
by dump trucks. No excessively contaminated "hot spots" were found. Finally, the "natural"
barriers to migration seemed quite effective. I am speaking primarily of the barriers
provided by streets and the installation of sewers. It also seems apparent that the leachate
system that was installed is in fact effective.
Sincere
Steven D. Aust
Professor of Biochemistry
SDA/cmc
-------
GEORGIA INSTITUTE OF TECHNOLOGY
ATLANTA. GEORGIA 3O332
CHEMISTRY
August 19, 1931
Dr. Clark W. Heath, Jr., M.D.
Director, Chronic Diseases Division
Center for Disease Control
Atlanta, GA 30333
Dear Dr. Heath:
Following the meeting held at CDC on August 13, to review health
hazard data on the Love Canal, I am enclosing my comments and recommen-
dations on the material considered.
General;
The format of the analytical data presentation made effective com-
parison between the various Tables difficult. Consequently, it was not
really possible prior to the meeting to give a rigorous treatment to the
data and arrive at any firm conclusions. In addition, the lack of appro-
priate units to the primary computer print-out of data made evaluation
of these dimensionless numbers impossible.
During the course of the meeting, and also subsequent to the meeting,
the meaning and implications of the results contained in the reports have
become clearer, as a result of questions and discussions raised at the
time. Nevertheless, there remains a lingering concern that much of the
information contained in the reports is not fully clarified at this time.
While the position of EPA towards the data, namely a desire to present
raw values, with minimum interpretation by them is understandable, it is
probably not a helpful position for arriving at an informed opinion. Con-
sequently it is recommended that EPA provide a report based on the avail-
able material, but more in the form of a scientific paper. This could
concentrate on a Results and Discussion Section which would explain the
data contained in the Tables, provide statistical treatment where appro-
priate and discuss the implications of the data in the light of the statis-
tical treatment. As with my scientific paper, the discussion of the data
would be open to question and clarification by the review group. It is
felt that in the absence of such a report the magnitude of the task facing
the consultant group in reviewing this material will result in unclear
recommendations to HHS.
A meeting to review the total data, with as long a lead time as
possible, would be most helpful.
-------
Dr. Clark W. Heath 2 August 19. 1981
In the light of the foregoing reservations regarding comprehension
of the data, the following comments should only be considered as tentative
and subject to review.
1. Quality of Analytical Data.
The procedures used for sampling and chemical analysis, including
QA and QC testing appear to be of the highest quality. While certain
discrepancies certainly exist between various laboratories, these are
typical for trace organic analysis at these low levels. With the
proviso that the EPA procedures of running duplicates for only 10% of
samples is unacceptable to many analytical chemists, the overall
quality of the data can reasonably be considered to be excellent.
2. Comparison of Concentrations of Various Chemicals with Niagara
Falls Area 99.
In general, the data appear to indicate no significant pattern of
contamination for the toxic species examined in the Love Canal area
compared to the surrounding district, with the exception of the
Immediate vicinity of the Canal (Area 11). Area 11 appears to show
significant soil contamination compared to Area 99. Many of the
chemicals found at significant levels in air samples (e.g. benzene)
probably cannot be attributed to migration from the canal site, as
there is no evidence of dumping these chemicals there.
3. Toxic Hazards of Pollutants to Area Residents.
At the present time, with the remedial action that has been taken,
there appears to be little hazard to homes outside Area 11. Such
conclusions, however, could be vitiated by any reduction in the quality
of the pollution containment and monitoring procedures that are
currently in effect. In particular, the clay cap over the site will
need constant attention and maintenance to ensure that no adverse
leakage takes place, if settling occurs as water is drawn from the
treatment plane.
4. Habitability of Love Canal Area.
Area 11 should probably be considered permanently closed to human
habitation or general ingress, and the present houses removed. The
•potential for future exposure to toxic materials seems to involve too
high a risk to be justifiable.
Surrounding portions of the study area (Areas 1-10) appear to
offer no greater health risk than that experienced by the general
population of Niagara Falls, due to prevailing air pollution levels.
-------
Dr. Clark W. Heath £ August 19, 1981
I hope that these comments are of some value to you in considering
the response of HHS to the EPA study.
Yours sincerely,
/•,
Richard F- Browner
Associate Professor
of Chemistry
KFB/jap
"This report represents the opinion of the author. It carries no official
endorsement by the Georgia Institute of Technology".
-------
THE UNIVERSITY OF KANSAS MEDICAL CENTER
COLLEGE OF HEALTH SCIENCES AND HOSPITAL
RAINBOW BOULEVARD AT 39TH • KANSAS CITY. KANSAS 66103
SCHOOL OF MEDICINE AUQUSt 21 1 98 1 OPPARTMENT OF PHARMACOLOGY
SCHOOL OF NURSING ' (913)588.7140
SCHOOL OF ALLIED HEALTH
UNIVERSITY HOSPITAL
Clark W. Heath, Jr., M.D.
Director, Chronic Diseases
Division, CEH
CDC, PHS, HHS
Atlanta, Georgia 30333
Dear Dr. Heath:
In response to your request, I have reviewed the EPA. analytical and testing
results, the NYSDH environmental testing data, and the other materials
provided in preparation for our meeting on August 13, 1981, in Atlanta. In
responding to the four questions posed in your letter of August 6, I have
focused primarily on the EPA data, although there is consistency in many
aspects of the NYSDH and EPA data despite the differences in collection
period (prior to versus post-remedial measures) and sampling (air sampling
versus multi-media sampling).
The first question Is perhaps the most difficult to respond to since it
appears from the EPA data and the summary tabulations of the data (pg. k1,
51, 54, 55, 57, 60, 61, 72, 100, 109) that for most of the chemicals
analyzed the levels in the declaration area exceed the levels in the control
area or in study area 97- The key question here is whether the concentrations
measured in the declaration area are significantly different from levels
found in other areas. Although I have carried out some calculations using
the individual EPA data (2nd book) to attempt to validate and understand
the data presented in the summary tables^ I do not feel comfortable with
this process. During the meeting on August 13, it was suggested that the
data in the tables should be presented in a form that would permit the
casual reader to more fully understand and evaluate the data. I would
agree with this suggestion on the basis that the data as presented appears
to be somewhat misleading. Perhaps Dr. Deegan could work with CDC analysts
to accomplish this.
The second question asks whether the reported levels are likely to cause
adverse health effects. I have focused here primarily on the organics,
since the inorganics are highly variable and are probably more closely
related to geologic criteria than to material deposited in the canal. Both
the EPA data and the NYSDH data indicate that the levels of the organics
are low in comparison with TLV values or other criteria used for evaluating
risk, and my response to this question is that such cause and effect
relationships are unlikely.
-------
Clark W. Heath, Jr., M.D.
August 21, 1981
Page 2
The third question asks whether the sampling procedures were adequate. M'y
impression of the EPA data is that the number of control samples is rather
small, but since there is rather good consistency in the data from the
various media sources, I would conclude that the sampling was adequate, at
least in the area of interest (the declaration area). Some concern was
expressed at the Atlanta meeting concerning the integrity of the samples
during storage prior to analysis (leakage through septems, etc.), and this
is certainly a potential problem if it did occur. However, again on the
basis of sample consistency, it would appear to me that this does not
appear to have been a problem in the EPA data.
The final question asks about the safety of the area for habitation. Since
the canal area is obviously heavily contaminated and will remain so for
years, the use of this land in such a way as to prevent migration of the
buried agents or exposure of populations will present very long-term
planning difficulties. On the othe.r hand, the declaration area does not
appear to present any clear and evident hazard to residents as long as the
contamination of the canal area can be controlled. My concern with making
recommendations regarding the habitation of the declaration area derives
primarily from a concern about the ability of any legislative or regulatory
group to maintain proper control and surveillance of any residential area
over extremely long periods of time. Thus, I would not advise residents of
this area to move out of the area, but neither would I encourage new residents
to move into this or any area with the potential problems that could arise
as a consequence of living next to a land-fill containing large amounts of
hazardous materials.
The problems of Love Canal are not unique to Niagara Falls, and they
involve much more than the toxicologic or adverse health effects evaluations.
The benefits side of the risk-benefit equation is even more complex than
the risk side of the equation, but it cannot be ignored. In answering
questions such as those above, scientists and particularly toxicologists
need to distinguish between predictions or conclusions which are based on
an evaluation of the toxicity data base for an agent as opposed to opinions
or conclusions which are not based on any special expertise. My answer to
the last question probably falls in the latter category.
Sincerely yours,
Jofm Doull , M.D., Pti'.'D.'
Professor of Toxicology
JD/mkh
-------
ENVIRONMENTAL DEFENSE FUND
August 21, 1981
Dr. Clark w. Heath
Director
Chronic Diseases Division
Center for Environmental Health
Centers for Disease Control
Atlanta, Georgia 30333
Dear Dr. Heath:
Let me begin by again thanking you for the opportunity to
participate in the recent workshop held at CDC on August 13,
1981 to assess the human health implications of environmental
data from the Love Canal area. The rest of this letter will
serve as my report on that meeting.
Before discussing in detail"the conclusions I reached, I
think it is necessary to briefly describe my feelings as I
approached the evaluation of the data presented' to me. The
Love Canal problem has received enormous notoriety and if one
word could generally characterize both the scientific studies
performed and the review of these s-tudies, that word would have>
to be "inadequate." While health and environmental assessments
at Love Canal have been numerous, many questions remain about
the adequacy of the scientific methods used and the quality of
the data generated. Likewise, the review of data, such as
presented in the Lewis Thomas Report, can be heavily criticized
as cursory in nature and inadequate in scope. Therefore, I
approached the review of the data presented to me with the
belief that the efforts we were undertaking had to provide a
scientifically critical review of the data collection
methodology, the data analysis techniques and the conclusions
reached.
1525 18ch Screec, N.W. Washington, DC 20036 '£? 202-833-1484
OFFICES IN: NEW YORK, NY (NATIONAL HEADQUARTERS); WASHINGTON. DC; BERKELEY, CA; DENVER, CO
-------
Dr. Clark w. Heath
August 21, 1981
Page 2
It was clear from the outset of Dr. John Deegan, Jr.'s
presentation that he did not expect us to carefully review the
data but rather to rely heavily'on his personal conclusions in
order to establish our conclusion as a panel. In fact, one of
Dr. Deegan's assistants, when speaking to me, stated that it
was EPA's belief that we would "trust" what they said and
answer the four questions posed based on EPS's data review.
Since I do not believe it is our role as scientists to merely
accept^what we are told, I approached the assignment given us
by asking three basic questions. These questions were:
1) Are the data and analysis presented
clear, accurate and if so, what do they show?
2} How accurately does the picture created
by the data reflect the current situation at
Love Canal and how likely is that picture to
remain constant in the future?
3) Can the four questions posed be
answered based on the data presented and if so,
how would I answer them?
My comments are presented below in answer to these three
questions.
To begin to answer the first question, are the data clearly
and accurately presented, one need go no further than the
printed material we received. As Dr. Whittenberger noted early
in our discussion, the data were often in an unreadable and
unintelligible fora and certainly not in standard, proper
scientific format. Moreover, once the material was deciphered/
it was evident that we had been presented selected segments of
the raw data base, rather than a compilation or summary of all
the data that had been gathered. The rationale offered for
this presentation by Dr. Deegan was that EPA hoped to avoid
biasing the conclusions of the reviewers. While I appreciate
his concern, the material presented was clearly inadequate to
conduct a proper scientific analysis.
In response to criticisms of the format, Dr. Deegan
suggested that any reviewer who wished could refer to the
entire raw data base in order to conduct his or her own
analysis. Such a suggestion could not be taken seriously,
given the amount of time it would have taken for any individual
to have made sense out of the vast amounts of raw data that had
been collected. If any further analysis of these data is
deemed necessary, I would suggest that the data be provided to
us in a completely revised format, a format similar to that-
-------
Dr. Clark w. Heath
August 21, 1981
Page 3
used_in the publication of a peer reviewed article. To avoid
biasing the conclusions of the reviewers, the data could be
presented clearly and accurately, statistical analyses could be
provided and the reviewers left to draw their own conclusions.
The problems that arise in analyzing segments of raw data
often presented in an illegible manner are demonstrated by the
following two examples. Dr. Deegan, in making his
presentation, provided the review panel with his personal view
of the meaning of the data. One conclusion that Dr. Deegan
drew was that the extent and level of chemical contamination in
the declaration area was no higher than in the control area.
Referring to page 41 of the document entitled "Environmental
Monitoring at Love Canal, 1980" which had been distributed to
us, I question the veracity of that conclusion. Reviewing the
data, I found that the presence of organic chemicals in the
declaration area far exceeded their presence in the control
area. In all cases, the levels of chemical contamination in
the declaration area exceeded those in the control area and
thus I found it difficult to reconcile the data with Dr.
Deegan1s conclusion.
Upon questioning, Dr. Deegan indicated that the table
presented on page 41 was inoperative and really should not be
referred to for data analysis:'" Rather, he suggested that the
large compendium of data by areas within the declaration area
and for the control area be referred to instead. In a few
minutes, during the lunch hour, I reviewed the data in these
tables. For each area I counted the number of organic
chemicals found in soil samples and compared these findings to
the control area. The results are presented below.
Area Number of Organic Contaminants
Found in Soil
Control 2
Area 1 11
Area 2 16
Area 3 14
Area 4 20
Ar e a 5 6
Area 6 3
Area 7 16
Area 8 2
Area 9 34
Area 10 10
Area 11 30
-------
Dr. Clark W. Heath
August 21, 1981
Page 4
In all cases, it is clear that the number of organic chemical
contaminants found in soil in each area of the declaration area
exceeds the number of such chemicals found in the control
area. When presented with this conclusion, Dr. Deegan
suggested that a statistical analysis of this data showed the
apparent differences to be insignificant. Unfortunately,
neither this statistical analysis nor any other analysis were
provided in the materials we received. Consequently, it is
impossible to judge whether the differences are insignificant
or in fact indicate a significant contamination of the
declaration area compared to the control area.
In further discussion of this issue, Dr. Deegan indicated
that the statistical analysis that had been performed evaluated
each chemical individually for its presence in the declaration
area and the control area. It is my scientific judgment that .a
number of these chemicals should have been considered together
because they are likely to have emanated from a single source.
For example, alpha/ beta" and gamma BHC and lindane, if found,
should all have been considered together. The failure to find
all four in one soil sample is easily explained by the fact
that these chemicals may have diffused widely from the point
source from which they emanated and the soil sample taken was
extremely limited and likely therefore to have missed a number
of chemical constituents. If a further review of this data is
to take place/ I would suggest that proper data analysis be
presented so that the reviewers can evaluate the method of data
analysis and the conclusions reached.
A similar confusion arises in an evalution of the data on
organic air pollutants found in the control area and in the
declaration area. Again, after Dr. deegan presented his
conclusion that the extent and level of chemical contamination
found in the declaration area and control area were no
different/ I quickly compared the levels of o-dichlorobenzene
measured in living areas in both the control area and the
declaration area. The results of that comparison are presented
below.
Area Mean Level Found
(ug/m3)
Control none reported
Area 1 23.4
Area 2 9.3
Area 3 13.1
Area 4 4.9
Area 5 4.6
Area 6 6.5
-------
Dr. Clark w. Heath
August 21, 1981
Page 5
Area 7 6.4
Area 8 29.4
Area 9 6.5
Area 10 4.1
Area 11 5.1
As in the case of the soil analysis previously presented, it is
clear that in every area of the declaration area, living area
air was more highly contaminated with o-dichlorobenzene than in
the control area. In fact, by referring to the large
compendium of data provided to us, one would conclude that
o-dichlorobenzene was not found at all in the control area.
Dr. Deegan quickly corrected this misunderstanding by
referring to page 74 of the document entitled "Environmental
Monitoring at Love Canal, 1980." He noted that in sampling for
o-dichlorobenzene this chemical had been detected at- the
detection limit in 10% of the cases, as compared with 43.4% of
the cases in the declaration area. I failed to make this
comparison, largely because the table provided is illegible.
When questioned whether the incidence of o-dichlorobenzene
contamination in the declaration area was statistically
significant compared to the control area, Dr. Deegan reported
it was. Consequently, it was hard to understand how he reached
a conclusion that the control-area and the declaration area
were equally contaminated. Dr. Deegan quickly pointed out that
for p-dichlorobenzene the reverse pattern of contamination was
observed. He suggested that it therefore might be appropriate
to consider evacuating the control area. Such sarcastic
statements have no place in a proper scientific disucssion.
I am left with the conclusion that there is a great deal of
ambiguity about the meaning of the data collected by EPA. The
data analysis referred to above certainly suggests that
chemical contamination of the declaration area exceeds
contamination in the control area. While this was not Dr.
Deegan's conclusion, it will remain mine until I can be shown
that these and similar analyses are in some way inappropriate
or incorrect.
An additional point to consider is the quality of data
presented to us and the conclusions that can be drawn from it.
I remain quite concerned that the quality control experienced
in these studies is less than adequate. This is not to suggest
that the methodologies chosen by EPA are less than
state-of-the-art. At this point in time I have not personally
reviewed the appropriateness of each test methodology employed
by EPA, and must rely upon conclusions of experts like Dr.
Browner that, in terms of analytical techniques, the methods
selected by EPA are appropriate. However, Dr. Browner in his
-------
Dr. Clark W. Heath
August 21, 1981
Page 6
discussion of this issue alluded to the fact that he could make
no judgment about the sampling procedures used by EPA nor the
adequacy of the treatment of the samples prior to analysis.
The materials presented to us clearly indicate that there was a
significant variability in the analytic capabilities of the
laboratories employed to conduct the environmental analysis at
Love Canal. Moreover, several significant questions still
remain regarding the effect of unusually long storage periods
prior to sample analysis. For example, on page 142 of the
document entitled "Environmental Monitoring at Love Canal,
1980" there is a discussion of the effect of holding times on
the analysis for volatile organic chemicals. The report states
that "only minimal loss of purgeable organics" occurred when
the holding time for analysis was 18 weeks after collection
rather than 4 to 5 weeks. It is unclear what the term "minimal
loss" really means, and I would wish to see data regarding the
number of chemicals analyzed and what percentage of loss that
actually occurred. Furthermore, and perhaps more important, is
the conclusion stated in the report that "losses, if any, in
the first few weeks are indeterminable at this time."
Therefore, the magnitude of loss of volatile organic chemicals
that occurred during the first 4 to 5 weeks of storage is
unknown and it is highly likely that under certain holding
conditions a significant portion of volatile organics would
have been lost.
One final point needs to be made in discussing the analysis
of the data presented to us and that concerns the number and
identity of the control areas chosen. In presenting his review
of the data, Dr. Deegan indicated that different numbers of
samples were taken for analysis from different environmental
media. For example, if my notes are correct, of 19 sites
selected for surface water sampling, 14 samples were taken from
within the declaration area and 5 from the control area.
Similarly, of 79 different sites sampled for groundwater
analysis, 68 were from within the declaration area and 11 were
from the control area. The ratio of samples obtained from the
test area (areas 1 through 11) to the control area (number 99)
suggests that the sampling pattern heavily favored the
collection of samples from within the Love Canal area. While I
am not a statistician, I believe that a skewing of sample
selection so heavily toward the experimental neighborhood may
make the study relatively insensitive compared to a study where
a better balance between samples collected from the
experimental area and the control area is achieved. In further
discussions of this data I would like an analysis by a
statistician as to the significance of the effect of the
sampling pattern on the sensitivity of the study.
-------
Dr. Clark W. Heath
August 21, 1981
Page 7
In addition, I believe we need a much better explanation of
how the control areas were selected, for what criteria they
were matched compared to the Love Canal declaration area, and
whether in fact any of these sites themselves were
contaminated. Dr. Paigen raised the point that one of the
sites, namely Control Site C (see page 4 of the document
"Environmental Monitoring at Love Canal, 1980"), the site in
the vicinity of 7320 Packard Avenue is in close proximity to,
or actually overlaps, a known hazardous waste dumpsite:, If
this is true, then a thorough analysis needs to be provided to
assure us that the control site is clearly distinct from the
area known to be a chemical dumpsite and that no contamination
from that site would be detected in sampling at the control
area. Moreover, nowhere in the materials presented are we
informed as to which values for control readings come from
which of the control sites. I would find it very helpful to
know where the control data comes from, especially after
learning more about the nature of each of these control sites
and their appropriateness to serve as controls for the Love
Canal declaration area.
Whether one concurs, with the conclusions reached by Dr.
Deegan or finds that the data suggest something entirely
different, the question of rehabitation necessitates a judgment
on how predictive the current--picture 'is for the future. This
issue was addressed by several of the participants in the
working group, including Dr. Doull, Dr. Kim, Dr. Pellizzari and
myself. Dr. Pellizzari commented that we are only looking at a
"microcosm of the total reality." Dr. Kim cautioned that it
might be advisable to continue monitoring for some time into
the future, before making a judgment on the habitability of the
site. He questioned the advisability of predicting future
conditions based on our current knowledge.
I strongly concur with the sentiments of those participants
noted above because I too have great reservations about the
applicability of the current data to predicting the future
situaton at Love Canal. As one example of the possible
changing nature of the conditions at the Canal, I offer the
observation I have made on visits to several other dumpsites
where clay caps have been used as part of the remedial cleanup
p'lan.
At the Kin BUG dump in New Jersey and Gratiot landfill in
Gratiot County", Michigan, clay caps were put in place, topsoil
applied, and the area seeded. Within 12 to 18 months, both
caps showed significant deterioration resulting from subsidance
and erosion. There is no guarantee even in the short term, no
-------
Dr. Clark w. Heath
August 21, 1981
Page 8
less in perpetuity, that the cap at Love Canal will be properly
maintained. There are no funds designated for long term
maintenance. Consequently, the cap at Love Canal may
deteriorate with time. Dr. Kim has already reported that when
cracks developed in the current cap prior to sodding, one
could detect the escape of volatile organic chemicals. I
believe, therefore, that any discussion of the possible future
uses of the Love Canal area must address concerns such as the
maintenance of the cap and explore the likelihood for changes
in the nature of the conditions that exist today.
Finally, we must ask whether the four questions we
addressed can be answered based on the available data, and if
so, how we would answer them. The first question asks "Are the
concentrations measured significantly different from levels
found in other areas of Niagara Falls?" This question is
difficult to answer, based on the available data. The analysis
presented above suggests, that the number of chemical
contaminants and, in many cases, the level of contamination in
the Love Canal declaration area exceed those in the control
area. However, since we were not provided with adequate
descriptive information about the control areas, it is
impossible to conclude whether they are representative of the
Niagara Falls area in general. In order to properly answer
this question one would have to have a more complete analysis
of the data collected and would need more extensive information
on the control areas and their relationship to other areas in
the Niagara Falls vicinity.
Question 12 asks "Do the levels measured represent
concentrations that would cause acute or chronic adverse health
effects in people living in the Love Canal area above what
might be expected under usual residential conditions in the
Niagara Falls area?" I suggest that this is an impossible
question for anyone to answer because of our very limited
knowledge of the toxicological effects of lifetime low level
exposures to a wide variety of chemical toxins. Moreover, in
many cases, the toxicological risk posed by exposure to many of
the chemicals reportedly disposed of in Love Canal is unknown.
Therefore, it is impossible to make an accurate judgment about
the likelihood of health impairment if exposure to these
chemicals occurs. In addition, in order to answer this
question one would have to assume that the data generated in
the control areas are representative of the conditions in the
general Niagara Falls area. At this point we certainly don't
know that to be a fact.
As discussed at the workshop meeting, there is an another
element of information that needs to be considered in
attempting to judge the toxicological risk of habitation of the
-------
Dr. Clark W. Heath
August 21, 1981
Page 9
Love Canal area. That element is the data generated on the
adverse health effects that have already resulted from living
at Love Canal. It is the conclusion of both the State of New
York and Dr. Beverly Paigen that an increased incidence of
adverse birth outcomes resulted from residing in the Love Canal
community. In addition, Dr. Paigen1s work also suggests that
other forms of health impairment (i.e., the incidence of
asthma) were elevated in the Love Canal community. The
available health effects data has been criticized for various
scientific inadequacies. The environmental data we have been
asked to review is, in my judgment, flawed in a variety of
respects as well. Consequently, the best judgment that can be
made about the risk to human health from residing in Love Canal
would be a judgment based on both the evidence we have from
environmental monitoring and from health effects testing.
Even if we were able to accurately predict the adverse
health consequence of exposure to all the chemicals in the
various environmental media, we would be hard pressed to
determine whether the reported level of chemical contamination
is accurate or is predictive of future conditions. With regard
to the current situation, it is unclear from the data
presentation what margin of error exists for the reported
levels of chemical contamination . Further analyses would have
to be provided in order for anyone to 'feel confident in judging
the true magnitude of current contamination reported within the
declaration area.
Concerning future patterns of contamination, it is my
belief that, based on the data before us, we are unable to
predict, with any accuracy what the future will be. Disruption
of the clay cap as noted above might lead to far greater
contamination of the -declaration area. Suggestions voiced at
the workshop that the area designated by EPA as Area 11 might
become a park or a public playground indicate that the
integrity of the clay cap in that area might be at far greater
risk than merely from erosion, subsidance and lack of
maintenance. Any activity that would have people using the
canal site as a playground or park would most surely jeopardize
the integrity of the clay cap.
Moreover, even Dr. Deegan noted that for certain
environmental media/ contamination within the declaration area
far exceeded that in the control area. Most notably Dr. Deegan
spoke about the contamination of sewer lines and sediments.
While Dr. Deegan alluded to the fact that EPA had just
announced plans for the cleanup of the sewer lines and
sediments, it should be remembered that this announcement of
cleanup is not new. Under the pevious Administration, the
Regional Administrator for Region II announced a program that
was also going to clean up the sewer lines. However, that
-------
Dr. Clark W. Heath
August 21, 1981
Page 10
program was not undertaken. Consequently, one should not
assume that the promise of cleanup programs means that cleanup
will occur. Finally, on this same point it should be noted
that while Dr. Deegan stated that the barrier drain system was
operating as designed and "the material was moving back faster
than it had ever moved out," there is no documentation for such
a statement in the materials presented to us. This is a
significant point because previous reports about the
functioning of the barrier drain system were less optimistic
about its working success.
The third question posed was "Were sufficient samples taken
and analyzed to determine the extent of contamination by
particularly hazardous materials and to be able to assess the
potential threat to human health?" At the workshop, you
attempted to have the consultants answer this question with
respect to dioxin contamination. In my judgment, insufficient
discussion took place about this question . Differences of
opinion were voiced as to the potential threat posed by dioxin
exposure, but the direct intent of the question (i.e., the
sampling or review of the sampling regime) was not discussed.
I agree with those who expressed concern over the high
levels of dioxin in soil samples taken in the Love Canal
declaration area. As one participant noted, it certainly is
difficult to explain to the' public that parts per billion
levels in soil are an insignificant risk when the Commissioner
of Health of the State of New York has warned people against
eating fish contaminated with parts per trillion of the same
chemical. Obviously, the exposure that may result from eating
fish contaminated with dioxin as opposed to living in an area
with dioxin-contaminated soil is not the same. However, it
should be remembered that young children, for example, often
play in soil and very young children may in fact ingest soil.
The last question posed to the -consultants was "Based on
available data, can you conclude that the area is not
habitable?" This question might as well have been asked in the
following way: "Based on available data, can you conclude that
the area is habitable?" Had the latter question been posed, I
think.the answer clearly would have been "No." The
environmental monitoring data presented to us is, in my
judgment, an insufficient data base on which to conclude that
the Love Canal declaration area is habitable and that
individuals should be encouraged to move back in. Because the
question is worded as it is, any belief that the data base we
have to rely on is inadequate would force one to answer the
-------
Dr. Clark W. Heath
August 21, 1981
Page 11
question in the negative. The conclusion that most likely
would be drawn from that response is that the respondant
believes that the declaration area should be rehabitated. I
certainly do not believe that the evidence we were asked to
consider allows one to conclude that the Love Canal declaration
area is a safe place to live. I strongly believe that CDC and
the Department of Health and Human Services would be making a
grave error if they concluded from a review of this data that
the Love Canal declaration area was safe to live in, and that
individuals should be encouraged to relocate there. At best,
our judgment must be that the data base presented to us, taken
alone, is insufficient to reach a conclusion on habitability
that would be scientifically supportable. If forced to offer a
more definitive opinion, then one would have to say that no one
should be encouraged to relocate in the Love Canal area based
on what we know today.
I recognize that a number of families in the declaration
area have chosen to stay rather than to move out. That is
clearly their choice to make. I believe we have a
responsibility as scientists to -- as best as we can -- explain
the significance of the data that have' been gathered, and to be
honest and straightforward about the numerous uncertainties
that exist. If these individuals continue to choose to stay,
that is their right. Providing such advice is clearly
different from actively supporting a conclusion that the
declaration area should be reinhabited.
I hope, these comments will be useful to CDC in making
a recommendation to SPA regarding the meaning of the
environmental assessment data. At the workshop you stated that
the final report that will be prepared by the Department of
Health and Human Services for EPA will have our letters to you
attached as appendicies. I believe it is essential that our
comments be included because I want to be sure that the views
of all consultants are clearly and accurately presented.
As I indicated at the meeting, I would be happy to
participate in another meeting if that would be useful to CDC.
I would, however, request that if such a meeting is to take
place, we be provided with more time to review any data we will
be expected to discuss. Again, thank you for the opportunity
to participate in the workshop and to serve as an expert
consultant to CDC on this matter.
Sincerelv, _ ..
t ' r\.• L &• /TcAW'"-"*-* l<7\ /.
flghland, p*^ / ^—
Chairma>
Toxic Chemicals Program
JH:ec
-------
University of Illinois at Urbana-Champaign
*/
INSTITUTE FOR 408 3oulh Coodwin Avenus
Urbcra. Illinois 61301
ENVIRONMENTAL STUDIES Au ^ 1981 ,2,7) 333-4,73
Clark W. Heath, jr., M.D.
Director, Chronic Diseases Division
Center for Diseases Control
Atlanta C-A, 30333
Dear Dr. Heath:
Enclosed please find my report "Threat to Human Health Among Residents of
Love Canal Area,. Niagara Falls, New York". This is the report requested
from my service as a consultant • to Center for Diseases Control following
the meeting on the above subject held in Atlanta, Georgia on August 13.
I hope that the submission of this report satisfactorily completes
my consultant assignment.
Sincerely,
Robert L. Metcalf
Professor Biology, Entomology,
Environmental Studies
-------
REPORT TO
CENTER FOR DISEASE CONTROL
ON
THREAT TO HUMAN HEALTH AMONG
RESIDENTS OF LOVE CANAL AREA
NIAGARA FALLS, NEW YORK
Robert L. Metcalf
Institute for Environmental Studies
University of Illinois, Urbana, Illinois
INTRODUCTION
Analyses of the problems in environmental toxicology around Love
Canal in Niagara Falls, New York are complicated by the vast amount of
analytical data made available from various aspects of environmental
monitoring. Many of the pollutants monitored including heavy metals,
polynuclear aromatics, halogenated aliphatics, acrolein, acrylonitrile,
benzene, toluene, xylenes, and polychprinated biphenyls; represent the
broad range of chemical pollutants typically found in the urban industrial
environment. The amounts detected in Niagara Falls, are more or less
typical of urban industrial environments in New Jersey, Louisiana, Texas,
Arkansas, and West Virginia (List of Compounds found in Ambient Air,
CDC). Without in any way minimizing the undesirability of their presence
in the air and water consumed by humans, it is concluded that these
"typical" pollutants are not adequate indicators of the possible hazards
of Love Canal dumping on the habitability of adjacent residential areas.
-------
-2-
Therefore, to properly analyze the environmental toxicology in the
residential areas in the vicinity of the Love Canal chemical waste dump,
I have concentrated on data indicating the presence of organic chemicals
representing typical and persistent toxic waste products from Hooker
Chemical Company operations during the years 1942 to 1953. According
to data furnished by the company to the New York State Health Department,
at least 21,700 tons of such wastes were dumped into Love Canal during
this period. Hooker Chemical Company pioneered during and immediately
after World War II in the production of a large variety of chlorinated
aromatic compounds including 2,4,5-trichlorophenol and chlorobenzenes;
benzotrichloride and other chlorinated toluenes; photo-chlorination of
benzene to produce lindane and other isomers of hexachlorocyclohexane;
and the highly reactive hexachlorocyclopentadiene for Diels-Alder diene
syntheses of the insecticides aldrin, chlordane, heptachlor, endosulfan,
and mirex.
The toxic chemicals selected as suitable indicators for this evalua-
tion are shown in Table 1 together with available data on TLV values for
human industrial exposures to contaminated air and ADI values for contamina-
ted drinking water. In the subsequent Tables 2-7, the available analyses
for the maximum detected concentrations of these indicatorNchemical
pollutautsin various environmental situations, i.e., human living areas,
drinking water, shallow wells, soils, sanitary sewers and sumps are
tabulated for the Love Canal dump site (11), for the residential areas
immediately adjacent in the West (2, 3), North (4), and East (6, 8),
and the miscellaneous control area (99).
-------
-3-
Soil; The daca in Table 2 for pollution of storm sewer sediments
clearly indicate the very extensive pollution of Love Canal (Area 11)
with a wide variety of toxic chemicals. These have been transported
by leaching in large amounts through the storm sewer system with the worst
pollution being found in Area 4 to the North. However, the large ppm
quantities of the several pollutants found throughout the adjacent areas
suggests a rather massive movement of the sub-surface soils throughout
the entire storm sewer system. The serious nature of the contamination
of the adjacent aras from this source is clearly shown by the tetrachloro-
dibenzo-p_-dioxin concentration of 672 ppb recorded in Area 4 together
with 20,000 ppb of y-hexachlorocyclohexane (lindane insecticide). We
understand that EPA will undertake to clean up the storm sewer contamination
and that capping of the Love Canal area and the provision of a leachate
treatment facility should substantially decrease the movement of the toxic
pollutants through the storm sewer system.
The data for soil contamination (Table 3) shows evidence of migration
of dichlorobenzene to Area 2. Further analytical data for other hazardous
pollutants in the "declaration zone" would appear desirable.
Water; The sump water (Table 4) and shallow well data (Table 5) if
representative and reliable appear to indicate substantial ground water
pollution only in the Love Canal (Area 11) . This is essentially what would
be expected because of the relatively deep burial (ca 30 feet) of the
21,700 tons of pollutants. The deep well data is essentially negative for
the key chemical pollutants in all sites. The water sampling data for
storm and sanitary sewers (Table 6) is, however, indicative of movement of
toxic chemicals in water from the Love Canal burial site to the adjacent
-------
-4-
declaration areas. The presence of the highly toxic tetrachlorodibenzo-
£-dioxin in Area 2 and hexachlorocyclopentadiene in Area 8 is particularly
disturbing. Rather large amounts of dichloro-and trichlorobenzenes and
7"-hexachlorocyclohexane in storm sewer water of Area 2 and lesser amounts
in Areas 4 and 8 also demonstrate the spread of the pollutants from Love
Canal into the "declaration area" through water. Thus additional and
regular monitoring of storm and sanitary sewer water for the presence of
the key 'pollutants is highly desirable.
Air; The air sampling data from Living Areas (Table 7) is difficult
to interpret because of the small number of key pollutants monitored. The
data are almost entirely restricted to dichlorobenzenesandchlorotaluenes
that are among the most volatile of the key pollutants. The presence of
these in Living Areas, however, is greater (at maximum values) in the
"declaration areas" 2, 3, 4, 6, 8;. than recorded in Area 99. These
values especially for dichlorobenzenes are 4.4X greater in Area 6,
6.5X greater in Area 4, 11. 5X greater in Area 2, 12. 5X greter in Area 3,
and 26X greater in Area 8. These substantially increased values must
be significant and air monitoring in Living Areas for other volatile key
pollutants especially hexachlorocyclopentadiene and hexachlorobutadiene
should be made to confirm air movement.
ANSWERS TO QUESTIONS POSED:
Q-l - Are the concentrations measured in the declaration area significantly
different from levels found in other areas of Niagara Falls?
A-l - As shown in Tables 2-7, concentrations of the specific pollutants
associated with the Love Canal waste disposal are relatively very
high in the Love Canal Area (11) in storm sewer sediments (Table 2) ,
soil (Table 3), water-sump (Table 4), and water-shallow well (Table 5)
-------
-5-
The evidence for leaching and migration away from the Love Canal
dump area is strongest with storm sewer sediments where relatively
large amounts of important pollutants such as Y-hexachlorocyclohexane,
hexachlorobenzene, tetrachlorobenzene, dichlorobenzenes, and
trichlorobenzenes were found in Areas 3, 4, and 8 and to a lesser
extent in Area 2. Particularly ominous, of course, is the very high
dibenzo
level of the extremely toxic tetrachloro-p-dioxin (TCDD) found in
j**
Area 4, Migration of these substances will undoubtedly be decreased
by the remedial construction undertaken but cannot be discounted as
a long term source of contamination of the "declaration area" by
highly toxic and relatively non-volatible pollutants. Thus, there
is reliable evidence of concentrations of key pollutants in areas
adjacent to Love Canal significantly higher than in the other areas
of Niagara Falls.
Q-2 Do the levels measured represent concentrations that could cause
acute or chronic adverse health effects in people living in the
Love Canal beyond what might be expected under usual residential
conditions in the Niagara Falls area?
A-2 It is impossible to answer this question in the-negative. The presence
of at least 21,700 tons of hazardous pollutants in the Love Canal and
the evidence summarized in A-l that detectable amounts of these are
to be found in the adjacent Areas 2, 3, 4, 6, 8 suggests that people
living in these areas will be exposed to trace chemical insults from
a variety of highly toxic chemical pollutants for generations.
Several of the key chemical pollutants including tetrachlorodibenzo-
p-dioxin, hexachlorocyclopentadiene, and hexachlorocyclobutadiene
-------
-6-
are extraordinarily toxic. Others such as y-hexachlorocyclohexane
(lindane) and hexachlorobenzene have been strictly regulated or
proposed for regulation by the World Health Organization, EPA, OSHA,
etc. All-of these chemicals are suspect carcinogens. The epidemio-
logical studies of birth defects, abortions, and chromosomal aberra-
tions made in the Love Canal and adjacent areas, while subject to
disputations, are significantly disturbing to demand a strongly
conservative attitude about promoting the exposure of thousands of
persons, especially young children to these trace chemical pollutants
over a lifetime.
Comparisons of environmental health standards for air quality i.e.
8-hour TLV-TWA values, and drinking water i.e. ADI values; with
maximum air pollution levels and water contaminations recorded for
the Love Canal area, are somewhat reassuring. The maximum detected
levels for most of the key pollutants for which the data are available
are very substantially less than accepted standards. However,
dibenzo
tetrachloro-£-dioxin, the most toxic of all the pollutants is a
notable exception and analytical data are so sketchy for the highly
toxic hexachlorocyclopentadiene and hexachlorocyclobutadiene that
together with inadequate study and standards for these pollutants the
results are virtually meaningless. The use of 8-hour TLV's for air
pollutants for evaluation of the 24-hour lifetime exposures that are
found in the declaration areas is also very misleading. For lifetime
exposures the spectrum of sensitivity of the exposed population
including the very young and the aged is much wider than that of an
occupationally exposed population. The threat of mutagenic, carcinogenic,
-------
-7-
and teratogenic chemicals is very much greater. All of these condi-
tions suggest the need for a safety factor of at least 1000 for the
very hazardous chemicals concerned, that should be superimposed
upon any considerations of TLV values.
The answer t£ question 2_ ±s_ that the concentrations o_f_ key pollutants
detected in the Love Canal area could cause adverse health effects
beyond those in usual residential conditions in the Niagara Falls area.
Q-3 Were sufficient samples taken and analyzed to determine the extent
of contamination by particularly hazardous material as to be able
•to assess their potential threat to human health?
A-3 Although a large number of samples were taken and analyzed and the
values obtained seem to be reliable within the present state of
the art, sampling and analysis were not rigorous enough to describe
adequately the human health threat from the highly toxic key pollutants
of Love Canal. The largest gaps in the sampling scheme are for the
highly toxic hexachlorocyclopentadiene and hexachlorobutadiene, and
dibenzo
tetrachloro-p_-dioxin. More effort should be expended on routine
^
and regular sampling for the other key pollutants such as lindane,
hexachlorobenzene, pentachlorophenol, pentachloronitrobenzene
and the more volatile di-,tri-,tetra-chlorobenzenes and phenols, and
chlorinated toluenes.
Q-4 Based on available data, can you conclude that the area is not
habitable?
A-4 In the writer's opinion the immediate Love Canal (Area 11) should
be completely barred to human access. Suggestions that it become
a "park area" are irresponsible. It is much more difficult to
assess the habitability of the "declaration areas" especially Areas
2, 3, 4, 6, 8 adjacent to Love Canal. Concerns aroused by the
-------
-8-
pollutant monitoring and epidemiological studies cannot be erased
by the data presently available. Additional monitoring data is
important to determine the effects of the remedial capping and
barrier construction instituted in the Love Canal and to fill the
voids in the analysis of key pollutants as indicated above. The
writer does not believe that it would be responsible to advise
inhabitants that the area is totally safe. Certainly he would not
want to expose himself, his family or his friends to these hazards
of chemical pollution through lifetime residence there. There is
in addition a measurable hazard in lifetime exposre to the polluted
atmosphere of the city of Niagara Falls itself.
-------
COMPARISON OF MAXIMUM DETECTED CONCENTRATIONS OF KEY
LOVE CANAL POLLUTANTS WITH SAFETY STANDARDS
maximum values
Love Canal Area
tetrachlorodibcnzo-jg-dioxin*
hexachlorocyclopentadiene
hexachlorobutadiene
pentachlorophenol
2,4, 5- tr ichlorophenol
y-hexachlorocyclohexane
hexachlorobenzene
pentachlorobenzene
pentachloronitrobenzene
tetrachlorobenzene
trichlorobenzenes
jj-dichlorobenzene
o-dichlorobenzene
o-chloro toluene
chlorobenzene
AIR
g/m3
Love Niagara
Canal Falls
(114) 0.41
(500)
(74) 9.6
(84) 43.7
(100) 100
(7650) 226
(172)
WATER
ppb
Love
Canal
(5.1)
734
17 (50)
,
120 (5000)
190 (52)
190
130
2700
(10)
TLV Suggested no-adverse
jjg/m3 U effect level H£0 2J
ppb
0.0007
100
240
1500 21
5.25
500 7
7
7
670,000
40,000
450,000 93.8
300,000
250,000
350,000
— American Conference of Government Industrial Hygienists (1980).
71
— National Academy of Sciences (1977) Drinking water and health.
*
Suspect carcinogen.
Figures in ( ) Kim, C. S., R. Narang, A. Richards, K. Aldous, P. O'Keefe, R. Smith, D. Bilker, B. Bash,
J. Stack, D. W. Owen "Love Canal" Chemical Contamination and Migration Oct. 15-17 (1980)
-------
POLLUTANT ANALYSIS LOVE CANAL AND ADJACENT AREAS OF NIAGARA FALLS, NEW YORK
STORM SEWER SEDIMENTS
pollutant
11
(Love Canal)
maximum concentration - ppb
map area
2346
(W) (W) (N) (E)
8
(E)
99
(control)
tetrachloro-jv-dioxin
hexachlorocyclopentadiene
hexachlorobutadiene
pentachlorophenol
tetrachlorophenols
trichlorophenols
dichlorophenols
Or-hexachlorocyclohexane
hexachlorobenzene
pen tachloroni trobenzene
i
tetrachlorobenzenes
trichlorobenzenes
dichlorobenzenes
chlorobenzenes
chlorotoluenas
9570
260
109100
495900
440000
63100
468600
457700
680700
5.4
- -
79
.1.300
305
21
3590
490
672
20000
15700
1700
4100
3900
0.50
0.40
250
.
-------
POLLUTANT ANALYSIS LOVE CANAL AND ADJACENT AREAS OF NIAGARA FALLS, NEW YORK
SOIL
pollutant
11
(Love Canal)
maximum concentration - ppb
map area
234
(W) (W) (N)
6
(E)
8 99
(E) (control)
tetrachloro-p_-dioxin
hexachlorocyclopentadiene
hexachlorobutadiene
pentachlorophenol
tetrachlorophenols
trichlorophenols
dichlorophenols
'Y-hexachlorocyclohexane
hexachlorobenzene
pentachloronitrobenzene
tetrachlorobenzenes
trichlorobenzenes
dichlorobenzenea
chlorobenzenes
chloro toluenes
407
10485
247
2642
377
316
9
-------
POLLUTANT ANALYSIS LOVE CANAL AND ADJACENT AREAS OF NIAGARA FALLS, NEW YORK
WATER - .SUMP
pollutant
11
(Love Canal)
2
(W)
maximum concentration - ppb
map area
3
(W)
(N)
6
(E)
8
(E)
99
(control)
tetrachloro-p_-dioxin
hexachlorocyclopentadiene
hexachlorobutadiene
penfcachlorophenol
tetrachlorophenols
trichlorophenols
dichlorophenols
/Y-hexachlorocyclohexane
hexachlorobenzene
pentcichloronltrobenzene
te trachlorobenzenes
trichlorobenzenes
dichlorobenzenes
chlorobenzenes
chloro toluenes
0.6
710
30
8500
2000
230
1300
A 500
1340
ZSUU
7600
14
-------
POLLUTANT ANALYSIS LOVE CANAL AND ADJACENT AREAS OF NIAGARA FALLS, NEW YORK
WATER - SHALLOW WELLS
pollutant
11
(Love Canal)
maximum concentration - ppb
nap area
2346
(W) (W) (N) (E)
8
(E)
99
(control)
tetrachloro-jg-dioxin
hexachlorocyclopentadiene
hexachlorobutadiene
pentachlorophenol
tetrachlorophenols
trichlorophenols
dichlorophenols
A"-hexachlorocyclohexane
hexachlorobenzene
pentachloronitrobenzene
tetrachlorobenzenes
trichlorobenzenes
dichlorobenzenes
chlorobenzenes
chloro toluenes
734
507
17
135
190
320
- 193
6650
.
*
,
0.01
-------
POLLUTANT ANALYSIS LOVE CANAL AND ADJACENT AREAS OF NIAGARA FALLS, NEW YORK
WATER - STORM AND SANITARY SEWERS
pollutant
11 '
(Love Canal)
2
(W)
maximum concentration - ppb
map area
346
(W) (N) (E)
8
(E)
99
(control)
tetrachloro-p_-dioxin
haxachlorocyclopentadiene
hexachlorobutadiene
pentachlorophenol
tetrachlorophenols
trichlorophenols
i
dichlorophenols
O^-hexachlorocyclohexane
hexachlorobenzenc
pentachloronitrobenzene
tetrachlorobenzenes
trlchlorobenzenes
diehlorobenzenes
chlorobenzenes
chloro toluenes
120
11
239
5.4
.
79
1300
305
i-l
i
3. A
87
41.6
1.3
94
86
5.3
379
228
13
-------
POLLUTANT ANALYSIS LOVE CANAL AND ADJACENT AREAS OF NIAGARA FALLS, NEW YORK
AIR - LIVING AREAS
pollutant
11
(Love Canal)
2
(W)
.maximum concentration - ppb
map area
3
(W)
(N)
6
(E)
8
(E)
99
(control)
tetrachlpro-p_-dioxin
hexachlorocyclopentadiene
hexachlorobutadiene
pentachlorophenol
tetrachlorophenols
trichlorophenols
dlchlorophenols
>>-hexachlorocyclohexane
hexachlorobenzene
pentachloronltrobenzene
tetrachlorobenzenes
Urichlorobenzenes
dichlorobenzenes
chlorobenzenes
chlorotoluenea
5.58
3.15
6.47
•
•
40.72
8.72
43.27
6.22
0.098
0.41
23.72
7.23
15.30
6.74
89.88
3.26
S.54
3.52
5.76
-------
IOHNSPELLMAN &^%*/ AUM OBB-
Covernor
STATE OF WASH!NCTO\
DEPARTMENT OF SOCIAL AND HEALTH SERVICES
Olympid. Wishntstun 985<>-i
August 17, L981
Clark W. Heath, Jr., M.D.
•Director, Chronic Diseases Division
Center for Environmental Health
Centers for Disease Control
Atlanta, Georgia 30333
Dear Dr. Heath:
Before giving formal responses to the questions you pose about the
Love Canal dump site, I would like to have the EPA raw sampling
data summarized and interpreted as in the format of a scientific
article.
I have come to some preliminary conclusions after perusing the raw
data, reading the available published studies, and listening to the
presentations and discussion at the Atlanta meeting.
Since the EPA sampling was done, in August-October 1980, one year
after the site had been capped, these samples represent historical
remnants of Love Canal pollutants.
It is currently impossible to reconstruct the pollution situation at
earlier times. Thanks to the nature of the soil and terrain, most of
the pollutants dumped in the Canal did not migrate very far. The only
exception to this is the contamination of storm sewers, catchment
basins, waterways and streams draining Ring I homes. The capping of
the Canal and the installation of the barrier drain system and treat-
ment plant seem to be working to prevent further pollution.
With the exception of the anecdotal reports of gross contamination of a
dozen Ring I and II houses, and sampling results adjacent to sand lenses,
there is surprisingly little objective evidence that serious chemical
pollution of the environment occurred. There is currently no good
evidence that this pollution has caused human health impairment. The
pollutant levels measured are far below those causing acute effects
and probably could not be linked to chronic health effects if, indeed,
such effects were demonstrated.
The EPA sampling effort was extensive and seems to be state of the art.
If the integrity of the clay cap and covering turf can be maintained
and the barrier drain system properly operated, I see no reason why the
-------
Letter to
Clark W. Heath, Jr., M.D.
August 17, 1981 -2
area cannot be rehabilitated. With the exception of the dozen or
so obviously contaminated houses, I'd go as far as having no
objection to the other houses in the area being reoccupied.
When the summarized EPA sampling data are available, I'd like to
see them.
Thanks for giving me the opportunity to offer advice.
Sincerely,
Samuel Milham, Jr., M.D.
Head, Epidemiology Section LB-15
Health Services Division
-------
CHEMICAL INDUSTRY INSTITUTE Of TOXICOLOGY
PRESIDENT. ROBERT A. NEAL, PH.D.
VICE PRESIDENT. DIRECTOR OF RESEARCH. JAMES E. CIB5.O.N. Vii.D.
SECRETARY AND ADMINISTRATIVE MAiNACER, DONALD A. HART. f.n.I).
P. 0. 15OX 12IJ7
RESEARCH 1RIAMJLK PAUK,
NORTH CAROLINA 2771)')
) 541-207!)
August 18, 1981
Clark W. Heath, Jr., M.D.
Director, Chronic Disease Division
Center for Environmental Health
Center for Disease Control
Atlanta, GA 30333
Dear Dr. Heath:
Given the heterogeneity of the group who reviewed the EPA monitoring data,
it is clear that there will not be a consensus concerning what these data mean
relative to a health risk to the residents of the Love Canal area. I think
it is quite clear that the analytical data was generated in as scientifi-
cally a rigorous manner as is possible. All of the analytical chemists
present at the meeting were of that- opinion, including Stephen Kim of the
New York State Department of Health. The validity of these analytical data
is also suggested by their general concordance with the data generated earlier
by the New York Department of Health. Some of the panel members had difficulty
evaluating the EPA analytical data in the form in which they were presented. I
had less difficulty in that regard, primarily because I had spent considerable
time comparing the data from the various sampling regions of the declaration
area with the control data. It is my opinion that, with the exception of the
canal area itself (including some ring one houses), the analytical data do not
indicate that residents of the declaration area are exposed to levels of chemicals
by way of air, water or soil which are, in a practical sense, different than
exposure, by those same routes, of people living in the control areas. It is
also highly unlikely since they are,, in general, much lower than allowable
concentrations in the industrial work environment, that persons in the control
areas are experiencing any significant health risk from exposure to the levels
of various chemicals detected in the EPA monitoring. It can be argued that
persons in the declaration area are exposed to toxic chemicals, as a result of
their presence in the Love Canal, which were not monitored for by the New York
Department of Health-or EPA. However, I do not believe that argument has any
merit in a practical sense. The EPA monitoring examined for compounds of known
and unknown toxicity which they knew had been disposed of in the Love Canal.
Unquestionably, there are compounds in the Love Canal which were not examined
for in the EPA monitoring and for which there is no toxicology data base.
However, unless these compounds posses considerable volatility, the persons in
the declaration area are not likely to be exposed to them. I say this because,
with the exception of about 10% of the houses in ring one and the effluent
-------
Clark W. Heath, Jr., M.D.
Page Two
August 18, 1981
of some storm sewers adjacent to the canal, there is no evidence of migration
of chemicals from the immediate area of the canal. In addition, the human
health monitoring or the animal studies which have been done on site have not
yet produced any evidence of adverse health effects or the potential for adverse
health effects in residents of the declaration area. I repeat that it can, and
has been, argued that persons living in the declaration area may be exposed to
chemicals which have not yet been identified in the monitoring program. It is
further argued that these chemicals may have an adverse health effect. My
answer to this is that there comes a time when we must stop delaying decisions
relative to the health risks to persons in the declaration area based on unknown
and improbable eventualities. I believe the time for decision making in this
regard has come. Based on the analytical data, the results of human health
monitoring and on-site animal studies it is my opinion that no practical health
risk exists for persons living in the declaration area (excluding the immediate
area of the canal and ring one houses). The finding of low birth weight children
in the "wet" vs. "dry" areas of the declaration area is offered as an argument
that adverse health effects are occurring. I do not find these data at all
convincing. First of all, the numbers are small. Secondly, there is no evidence
that there was movement of chemicals from the canal into these areas. Finally,
the fact that the incidence of low birth weight babies of families living in ring
one and ring two houses was not significantly different from control incidences
suggest the effect seen in the "wet" area is either a spurious result or unrelated
to living in close proximity to Love Canal.
With those general comments as a background, let me direct some
remarks to the specific questions you posed in your letter of August 6.
Question 1
As I indicated above, the raw data from the EPA analytical monitoring
does not suggest that the compounds which were monitored for were in higher
concentrations in the declaration as compared to the control areas of
Niagara Falls (excluding some ring one houses and the immediate area of the
canal).
Question 2
Again, as noted above, the analytical data currently available do not
indicate that persons living in the declaration area are exposed qualitatively
or quantitatively to chemicals which are, in a practical sense, different
from the exposure of people living in control areas of Niagara Falls.
Question 3
In my opinion, sufficient sampling for known substances of high toxicity
have been performed. The results of this sampling do not suggest the presence
-------
Clark W. Heath,
Page Three
August 18, 1981
Jr., M.D.
of sufficient amounts of these substances in areas outside of some ring one
houses and the immediate canal area, which pose a health risk to people living
in these areas.
Question 4
In my opinion, people living outside of the ring one houses are not exposed
to levels of the chemicals present in the Love Canal which pose any significant
health risk. Further, the lack of significant migration of chemicals from the
canal (excluding some ring one houses and storm sewers adjacent to the canal)
in the 30 plus years of the presence of significant amounts of chemicals in
the canal does not suggest the possibility of future exposure of persons living
in the declaration area from chemicals present in the immediate area of the
canal. The fact that additional safeguards against migration of chemicals from
the canal by various routes, including storm sewers, have been put in place,
adds additional assurance that significant migration is highly unlikely in
the future.
Sincerely,
Robert A.
President
Neal
RAN:earn
-------
Review of EPA Monitoring Program at Love Canal
August 16, 1981
by Dr. Beverly Paigen
Summary
The consultants to CDC were asked to evaluate a subset of the data obtained by
EPA's environmental chemical testing program and to answer the following
questions:
1. Are the concentrations of chemicals significantly higher than other
areas of Niagara Falls?
2. Could these concentrations cause adverse health effects?
3. Were enough samples taken of very toxic chemicals such as dioxin to
be able to assess its health threat?
4. Is the area habitable?
The data had several problems in the presentation of material and in the
adequacy of measurements. Putting these aside, I believe it was possible to
judge that organic chemicals were higher in Love Canal study area (the entire
Love Canal neighborhood excluding Rings 1 and 2) than in other areas of
Niagara Falls. Organic chemicals were higher in storm sewers, in stream
sediments, in sumps, in soil, and in home air. Only in drinking water were
organic chemicals similar in Love Canal and control area and this is based on
a small amount of data.
The data are not adequate to permit an answer to questions 2 and 4 regarding
health and habitability for reasons discussed below. In answer to question 3,
there were insufficient samples taken to adequately evaluate whether dioxin
levels constitute a health threat in the study area even though the small
number of samples had surprisingly high levels of dioxin.
Finally, additional analyses of data would be helpful in determining the
extent of contamination at Love Canal and these are listed.
I. Adequacy of data
A. Problems for the consultants in reviewing data
The data were difficult to evaluate because of the massive amount of
data, the short time to review, the lack of statistical treatment, the
inability to read some tables and maps, and the lack of sufficient
summary tables and paragraphs. Although I can appreciate the
difficulties in assembling and reproducing such large amounts of data, I
am not comfortable as a scientist in answering questions concerning
public health based on data I cannot read (pp 72-74, 76-78), data with
obvious errors (maps on pp 62-70 show stream sediment samples taken
where there is no stream), unexplained symbols on maps and tables, and
little aggregation of data or statistical analysis (i.e., the entire
second book of sampling areas has no summaries at all).
-------
Page 2 - Review of EPA Monitoring Program at Love Canal
Several points were particularly disturbing. There was no
description of control sites (i.e., Packard Avenue may be near another
dump site). The EPA program was designed to test for preferential
migration (p 8), yet there was no data analysis concerning that
hypothesis. The separation into areas 1-11 had some rationale, but
there is no discussion of this nor analysis by area.
The picture was further confused when Mr- Deegan of EPA, at the
August 13 meeting of consultants, stated that, in his opinion, the data
demonstrated no evidence of chemical contamination beyond Ring 1 except
for storm sewers and stream sediments. When a consultant asked Mr.
Deegan about the soil data on p 41, Deegan replied that table was "no
longer operative." The consultant then spent the lunch hour tabulating
data by hand from the larger data bases and indicated to Deegan that
organic chemicals were frequently found in Love Canal and rarely in
control area soils. Mr- Deegan claimed that the differences were "not
statistically significant."
As a scientist asked to make a judgment based on data, I simply
cannot in good conscience ignore tables of data simply because one
person states they are no longer operative nor can I accept one person's
statement that the differences are not statistically significant when I
am not shown the statistical methods or significance tests and when the
data available in hard copy to review lead to the opposite conclusion.
B. Problems with using the data base to answer questions concerning
health and habitability
The original plan was to couple the environmental testing programs
with a health study. Had that plan been carried out, the two studies
taken together could have been used to answer questions of health and
habitability. As it stands, the current data base is not appropriate to
answer such questions for the following reasons:
1. Love Canal has over 200 chemicals present and only a subset were
measured.
2. Little or nothing is known about the toxicology of a large
percentage of the chemicals present because they are waste
products or process intermediates (for a fuller discussion see
Rail et al. "Report of Subcommittee to Evaluate Health Effects
Near Hazardous Waste Sites." HHS 1980).
3. Little is known about the synergism or antagonism that might
exist in such a complex mixture of chemicals.
4. Even for those chemicals that were measured and for which
toxicological data exist, most studies are based on workers or
experimental animals. Almost nothing is known concerning the
concentrations of chemicals that would be safe for the human
fetus, yet adverse reproductive outcomes were the health effect
best documented at Love Canal.
-------
Page 3 - Review of EPA Monitoring Program at Love Canal
The primary ways in which chis particular data base could be used to
answer questions of health and habicability are two-fold. First, if the
levels of one particular chemical were high enough, one could state it
was unsafe. Second, if one particular area showed no evidence of
chemical contamination, then one could tentatively conclude the area was
safe if the sampling was adequate and if the subset of chemicals measured
are an adequate reflection of the whole set.
C. Problems with Quality Control
EPA had a good quality control plan for this study with spiked
samples, duplicate samples, and inter- and intra-laboratory
comparisons. The results of the quality control program are
described on pp 111 and following. Several points were made by EPA.
1. The "meaning of trace is obscured by the variance in minimal
detection limits among laboratories" (p 119)- Trace from one lab
could be 30 times higher than trace from another lab (i.e., see
tetrachlorobenzene on p 130).
2. Some laboratories performed much better than others ranging from
33 percent acceptable results to 100 percent acceptable results
(p 133). These two facts led to a cautionary statement that
"users of Love Canal data should be aware of differences in
detection limits .... and performance" among laboratories (p 120).
3. The magnitude of the testing program "literally overwhelmed the
national capacity for low level chemical analysis" (p 121). Thus
many samples were held too long and this fact becomes important
for volatiles and semivolatiles. For water many samples arrived
with poorly fitted or leaking Teflon cap liners (p 122). Water
samples were held too long. Soil samples were held too long and
the effect of holding was stated not to be a problem between 4-18
weeks but "losses in the first few weeks are indeterminable at
this tine" (p 142). For air holding time was not given in the
summary data. Mr. Deegan referred the consultants to the full
data base in Atlanta, but we could not find holding time there
either^ Holding time for volatiles in air is quite critical.
4. Some chemical measurements were not to be trusted, i.e.,
phthalates and methylene chloride in soil and benzene and toluene
in air.
5. Reproducibility and accuracy of measurements for metals in soil
and sediment were fairly good but were not good for volatiles and
semivolatiles (p 148).
-------
Page 4 - Review of EPA Monitoring Program at Love Canal
6. The reproducibility data for air was calculated in a way that is
not familiar to me and which obscured the real differences (p
160). For example: if sample 1 has duplicates a and b and b is
50 percent higher than a; and sample 2 has duplicates a and b and
b is 50 percent lower than a; I would calculate that the average
difference between duplicate samples was 50%. EPA calculated
that the average difference was 0 percent f-SQ+SO = Q). I
2
would like to see a more conventional analysis of reproducibility -
II. Tentative conclusions based on the data
The data base is not adequate to answer questions of health and
habitability but there are several questions that can be answered or could
havebeen answered by an environmental monitoring program. The conclusions are
only tentative given the problems with presentation of material and quality of
data base discussed under I above.
A. Is the Love Canal area contaminated?
The storm sewer, stream sediments, soil, sumps and air in homes in
Love Canal are all clearly contaminated compared to the rest of Niagara
Falls. Only the drinking water appears to be the same as the rest of
Niagara Falls and that is based on very limited data.
EPA did not give any statistical analysis of the data so I examined
the data for organic chemicals and simply asked whether each organic
chemical was more frequently found in Love 'Canal samples than in control
samples. I then used a simple statistical test, the sign test which has
relatively low power, to ask the probability that, for example, out of 13
separate tests, Love Canal was higher than control 13 out of 13 times.
This is given by (1/2)13 or 0.0001, a significant difference. This is
not a sophisticated analysis; it would be possible to use other tests
that involve a combination of probabilities (such as Fischer or Wallis).
In addition, a strong argument could be made for combining the chemicals
into groups such -as all trichlorophenols or all BHC or all organics.
For storm sewer sediments, there were 21 Love Canal area samples and
1 control and 14 organic chemicals gave positive readings. In 14/14
cases, Love Canal is higher than control (p 54). For storni sewer water,
there were 9 Love Canal samples and 1 control and 13 organic chemicals
were detected (p 55). Love Canal was higher than control in 13 out of 13
(12/13 controls were 0). In the sanitary sewers, only 1 Love Canal water
sample was taken (no control samples). Twelve organic chemicals were
found at levels comparable to or higher than storm sewer water (p 57).
The stream sediments in Love Canal were contaminated but the stream
water did not show evidence of contamination. There were 4 Love Canal
samples and 5 control samples. For the sediments, 12 out of 12 chemicals
were higher in Love Canal; in the stream water only 2 organic chemicals
were detected (toluene and heptochlor) and Love Canal was lower than
control.
-------
Page 5 - Review of EPA Monitoring Program at Love Canal
In Che sumps, there were 105 Love Canal samples and 6 control area
samples (p 51). Love Canal was higher in 9/9 cases (6/9 controls were 0).
In the soils, there were 113 Love Canal samples and 8 control samples
(p 41). Love Canal soils were higher in. 12/12 organic chemicals. All 12
control samples were 0.
In drinking water, there were 31 Love Canal samples and 5 control
samples. A total of 7 organic chemicals were detected; Love Canal samples
were higher in 3 and control samples were higher in 4. Thus there was no
apparent difference in Love Canal and control, a result that is not
surprising since the same municipal water is supplied to both areas.
The data given for shallow aquifer wells and deep aquifer wells was
not sufficient to permit an evaluation.
There were over 100 Love Canal samples of outdoor air (p 72) and over
80 Love Canal samples of basement air Cp 73). However, there were no
controls and so one cannot determine whether Love Canal is different than
Niagara Falls.
For living area air, there were over 400 Love Canal samples and 30
controls (p 74). After eliminating benzene and toluene based on
information in the quality control section; 11 out of 12 organic chemicals
were higher in Love Canal homes than in control homes. The exception is p
dichlorobenzene, which is completely missing from the table on p 75
causing me to wonder, if there is an error. All of the air data were
difficult to evaluate because they were almost unreadable.
In conclusion the data show that Love Canal is clearly higher in
organic chemical contamination than the control area in storm sewer,
stream sediments, sumps, soils, and living area air. No determination
could be niaca for sanitary sewers, basement air, or outside air since no
control data exist. Drinking waster was the same for both Love Canal and
control and indeed comes from the same municipal source. Surface water;
surprisingly, did not show any evidence of contamination although the
sediments did. Perhaps this is because organic chamicals would tend to
stay in sediments or be released into air, but would not be dissolved in
water.
B. How far has the contamination occurred?
Since the data indicate that Love Canal is contaminated, it becomes
relevant to determine how far contamination has occurred.. Unfortunately,
no analysis by area was given. The areas of particular concern are 5 and
1 which are the extreme edges of the neighborhood and where the greatest
percentage of current residents are located. Both areas 1 and 5 have a
few houses on swales or streams and these appear to be giving quantifiable
levels of chemicals. I would like to see an analysis with and without
those houses.
-------
Page 6 - Review of EPA Monicoring Program at Love Canal
C. By what routes did contamination occur?
No analyses wera prese-nted that would permit such a determination, but
it may be possible. I would like to see an analysis by 1) area, 2)
wet/dry, 3) swale/non-swale, and 4) air in home with and without sumps.
Several homes in the area have flooding in the yard from storm sewers.
Are the soils, air, or sumps higher in such homes? Is there any
correlation between positive readings in soil and air or sumps in the home?
D. Has remedial work resulted in a reduction in contamination?
Mr. Deegan stated that back flushing of chemicals from the soil was
occurring to a greater extent than expected. This is an important finding
and I would like to see the data.
I would also like to see a comparison of homes done by-both New York
State and Love Canal to compare soil, air, and sumps.
E. Is there correspondence between the geochemical evidence of
contamination and the adverse health effects?
Tne adverse reproductive, outcomes found by New York State should be
compared to the geochemical data.
III. Questions posed by CDC
A. Are the concentrations of chemicals significantly higher than in other
areas of Niagara Falls?
As discussed above, the answer is yes.
B. Could these concentrations cause adverse health effects?
Due to the limitations of the data base discussed above, I do not
believe this question can be answered.
C. Were enough samples taken of very toxic chemicals such as dioxin to be
able to assess its health threat?
Only 4 soil samples in the study area were tested for dioxin and all 4
were negative. This is an insufficient number of samples. Eleven storm
sewer sediment samples were taken and all had dioxin levels.
D. Is the area habitable?
As discussed above, I do not believe this particular data base permits
an answer to that question.
-------
Page 7 - Review of EPA Monicorir.g Program at Love Canal
However, as a parent of young children, I would not want to live in
any Love Canal house that 1) was along a straam with contaminated
sediments, 2) had a history of basement or yard flooding from backed up
storm sewers, or 3) had a sump pump. Any of these situations could lead
to periodic exposures much higher than those detected by the EPA
monitoring programs.
IV- Further Analyses
The following information would be useful in a further evaluation of
the data base.
1. Holding time for air.
2. Evaluation by area.
3. All positive readings for organic chemicals on a map by media and
with various media combined.
4. List of organic chemicals not likely to be found in consumer
products.
5.. Map of control sites and other dumps in Niagara Falls.
6. Statistical evaluation of data.
7. Readable tables of data for air.
8. Analysis by wet/dry.
9. Correlation of chemical data with adverse reproductive outcomes.
10. Discussion of p dichlorobenzene—the only chemical found more often
in cortcrol than in Love Canal.
11. Readable figures for daily variation in air readings.
12. Comparison of New York and EPA data for some locations.
V. Final questions
The most striking thing about the EPA monitoring program was the high
percentage of negative findings and the low levels of chemicals found
compared to New York data.
The question that comes to me is whether exposure at Love Canal is
periodic in nature and whether the health impact may occur by transitory
exposure to higher levels than those found by EPA. According to Vianna gt
al, the peak in adverse reproductive outcomes occurred in the mid-sixties
-------
Page S - Review of EPA Monitoring Program at Love Canal
when as many as 50 percent: of all pregnancies on 99th Street ended in mis-
carriage. In the mid-sixties, basements were being dug for the homes on
97th Street. Perhaps the exposure of large amounts of contaminated soil
to the air on 99th Street resulted in a transitory high exposure for
residents on. 99th Street and beyond. Consistent with this hypothesis is
the peak of adverse reproductive outcomes claimed by the Love Canal
residents during the remedial construction in 1979 and 1980. Perhaps Love
Canal may be habitable as long as-no digging occurs. Another peak
exposure may have occurred following the unusually heavy precipitation in
1977. This exposure caused the residents to organize and to seek help and
it was this type of exposure that the remedial work was designed to
prevent.
-------
RESEARCH TRIANGLE INSTITUTE IX ^
POST OFFICE BOX 12194 <~—-—T^_
RESEARCH TRIANGLE PARK. NORTH CAROLINA 27709
CHEMISTRY AND LIFE SCIENCES GROUP
August 27, 1981
Dr. Clark W. Heath, Jr.
Director, Chronic Diseases Division
Center for Health
Centers for Disease Control
1600 Clifton Road, N.E.
Atlanta, Georgia 30333
Dear Dr. Heath:
Enclosed is a copy of the review on the EPA Love Canal Study.
I enjoyed the opportunity to provide some assistance to you in tlais
program and hope that it will be helpful to the overall goals. If
I can be of any further help please give me a call at (919)541-6579.
Sincerely,
E. D. Pellizzari, Ph.D.
Director for Analytical
Sciences Division
EDP/mlu
Enclosures (as stated)
(919) S4t-6000 FROM RALEIGH. DURHAM AND C H A P E U H I t_ 1_
-------
REVIEW OF EPA LOVE CANAL STUDY
E. D. Pellizzari
Research Triangle Institute
Research Triangle Park, North Carolina
Introduction.
The emphasis of this review was directed toward assessing
the study design and analytical methodology employed for acquiring
information concerning toxic chemicals in the various media. The primary
questions to be addressed in the review were expanded to provide more in
depth probes to the viability of the study design and results. These
aspects are debated below.
Purpose of the Review
The major thrust of this review is to address the following primary
questions:
(1) Were sufficient samples taken and analyzed to determine the
extent of contamination by particularly hazardous materials
and to be able to assess their potential threat to human
health?
(2) Are the concentrations measured significantly different from
levels found in other areas of Niagara Falls?
(3) Do the levels measured represent concentrations that could
cause acute or chronic adverse health effects in people living
in the Love Canal area beyond what might be expected under
usual residential conditions in the Niagara Falls area?
(4) Based on available data is the area not habitable?
-------
Specific Issues Reviewed
The primary intarrogatives not only have far reaching ramifications
but also contain subtle, detailed issues that need examination prior to
drawing any valid conclusions. Some specific issues are enumerated and
addressed here as part of this review.
(1) Was the study design adequately structured to meet the main
objectives implicit in the primary questions?
Based upon readily accessible information the implied
intention of the study was to determine whether a significantly
greater risk of exposure to chemicals from the Love Canal site
exists now or in the future for inhabitants of the declaration
area as compared to inhabitants of other areas within Niagara
Falls, NY. In essence the design should be a probability
study. To demonstrate significant differences between a "test
case" and a "control case" sufficient, representative, experi-
mental data from both cases need to be acquired. The problem
becomes predicting what is "sufficient" when initially setting
up the study design. This is particularly difficult if the
samples analyzed yield many non-detectable values. Probability
functions can be used to decide upon "representative" data.
Also, what constitutes significant, i.e., is the purpose to
measure differences between the two areas that are greater
than factor of two or an order of magnitude.
The monitoring of ambient air for volatile chemicals
occurred at 62 sites in the declaration area and 4 sites in
the "control" area over a two month period. Both outdoor and
indoor (living room and basement) ambient air was sampled as
-------
12 hr integrated samples (two/day). Unoccupied premises were
sampled which had been closed to enhance the possibility of
elevated levels of chemicals. Although sufficient numbers of
samples were collected during the period of September and
October, 1980 at each site, more sites designated in the
"control" area would have been desirable, especially when
testing for significant differences between the declaration
and control area. The degree of representativeness of the
four sites in Niagara as constituting a "control" might be
questioned.
Monitoring of water was primarily through "grab" sampling
instead of integration with time. Drinking water was obtained
from pre- and post- municipal treatment systems serving the
communities designated as the declaration and control areas.
This is an indirect approach. Acquiring drinking water samples
from each home selected for air monitoring may have revealed
whether any small ruptures in water lines existed in contaminated
areas. Nevertheless, the likelihood of such an occurence
probably is small. Ground water was taken from about 170
wells, including shallow and deep aquifers from the declaration
and control areas. In this case their were more sampling
sites from the control area (15) than in the air sampling
design. Water samples from storm sewars (28 sites), sumps in
basements (54 sites) and creeks/rivers (19 sites: 14 declaration,
5 control) were also collected. The rationale for selection
of sampling sites within each area was not clearly stated in
the report provided, nor whether the objectives were achieved.
-------
Sediments were collected from 19 streams, 3 sumps and 24 storm
sewers. One sediment sample from the Leachats Treatment Facility was
also taken. Soils were taken from 171 locations (9 in control area) as
6 ft. vertical cores.
The major exposure route is postulated to be ambient air, since
vaporization of surface chemicals would lead to their inhalation. The
importance of soil and sediment cannot be over emphasized since they can
provide historical (time integrated) documentation of fate and transport
of chemicals. The number of sediment samples examined probably are
representative of the streams and sewers; however, an insufficient
number of sumps were selected. Sediments from sumps in both occupied
and unoccupied homes in the declaration area should have been included
in the study, particularly since previous evidence is suggestive that
major pollution of basement air came from the basements sumps.
Soil also can be a major route in "pica" exposure in young children.
Moreover, it can be a sink for chemical emission (fallout) and subsequent-
ly involved in dynamics- of transport. Vertical cores were not analyzed
in sections. As such, considerable dilution could occur from such
compositing to the extent that any chemicals potentially present as a
small discrete zone (say at the ground surface) were diluted below the
detection limits of the techniques employed.
(2) Was the most exhaustive list of toxic chemicals potentially
occurring at Love Canal sought in the study design?
-------
The selection of methods for sampling and analysis of
air, water, soil, sediment and biota obviously should be based
upon the best available technology. The techniques employed,
however, essentially limited the study design to a finite list
of target compounds. One should be cognizant of the fact that
comprehensive monitoring for all chemicals in the five matrices
was not conducted nor is it possible with the current analytical
methods in the analyst's arsenal of tools. This may be inap-
propriately gleaned as a serious limitation of this study.
The technical approach employed emphasized quantification of
target compounds. Nevertheless, an equally important aspect
should have been the determination of any "unknown" chemicals
that may be present. The 'qualitative identification of as
many chemicals as possible with state-of-the-art technology
would increase ones confidence that the chance of potential
exposure to toxic chemicals is minimal if the findings are not
significant. This feature of the study was short changed
since state-of-the-art technology (see below) was employed for
only some of the environmental media.
(3) Were the sampling methods selected for the various environmental
media - air, water, soil, biota, sediments - viable for the
purposes intended to support the study design?
In as much as its possible to assess this facet of the
study it appears that the sampling techniques employed were
adequate. One minor exception on soils was noted above,
however. More detailed discussion on the documentation of
accuracy of sampling will be discussed under Quality Control/
Assurance (QC/QA) below.
-------
(4) Were the analytical methods that were employed state-of-the-
art?
The methods chosen, for sampling and analysis of ambient
air for toxic organic vapors were in this case state-of-the-art.
Neither method (Tenax-GC/MS or Polyurethane foam - GC/ECD and
HPLC/EC) is a "standard method". Also, the application of
these complimentary methods still do not yield a comprehensive
picture of the ambient air composition. The full potential of
either method was not realized in this study. The Tenax-GC/MS
method lends itself to assessing sampling accuracy by the use
of deuterated surrogate chemicals. This feature was not
exploited (To be discussed further under QC/QA). The polyure-
thane foam technique (PUF-GC/ECD) employed antiquated pack
column .gas chromatography. The use of capillary column techno-
logy would have improved the limits of detection for the
target compounds (halogenated aromatics including PCBs and
polychlorinated naphthalenes) as well as ensured more specifi-
city to isomer identification and quantification. Thus,
collection with polyurethane foam is at the frontier of the
scientist's capability while packed column gas chromatographic
analysis is at least a decade outdated.
Similar criticisms can be made of the use of the Federal
Register Methods for analysis of water samples. The strength
of these methods rests with the degree of validation (both
intra and interlaboratory) that each has been subjected to
over the past several years. Nevertheless, the use of selective
detectors with gas chromatography is limited to quantification.
-------
Broad spectrum analysis (identification of non-targeted chemicals)
should have been given more emphasis to decrease the chances
of missing important chemicals in this study. Capillary
column GC/MS should have been employed rather than packed
column technology, again to increase the accuracy of identifi-
cation and the breadth of analysis capability.
Similar deficiencies can be cited for the methods employed for
analysis of soils, sediments and biota.
(5) Were experienced and reliable research/service laboratories
employed for the sampling and analysis of air, water, soil,
sediment and biota?
Analysis of air samples were performed by two laboratories
with no prior experience with the Tenax-GC/MS method. This
was evident in the Quality Assurance sample results which
appeared to be less precise during the first few weeks of
analysis by one laboratory. Laboratories performing analysis
using the PUF-GC/ECD method were experienced with the technique.
All other laboratories had several years prior experience
in performing priority pollutant analysis in water samples and
to a lesser extent in soil, sediment and biota.
(6) Were their sufficient Quality Control and Assurance practices
imposed on the sampling, sample analysis, and data analysis to
ensure the integrity of the analytical data reported?
The QC/QA practices employed were extensive and one of
the relatively stronger features of this study. Unfortunately,
"Standard Reference Materials" are not available from the
National Bureau of Standards. Thus, reliance was on performance
-------
evaluation solutions, controls, blanks, etc., to assess the
sample preparation and analysis steps.
Quality assurance applied to the sampling and storage
step for each environmental matrix was limited to the use of
spiked sampling devices (containers) for determining recovery
(i.e., accuracy). N£ internal standards were used to assess
potential malfunctions of the sampling step and subsequently
storage, work-up and analysis. This is a serious limitation
of all methods. A small percentage (^ 10%) of the samples
should have included internal standards during sampling. For
example, deuterated benzene (dfi), dp-chlorobenzene, etc.,
could have been added to the ambient air sampling device
(protocol: Tenax-GC/MS) to monitor the accuracy of sampling
and analysis without interfering with endogeneous chemical
analysis. The use of isotopic compounds would more closely
mimic the endogeneous compounds than the use of spiked, unsam-
pled devices. The isotopic concept is well documented and
widely accepted by analytical chemists in many different areas
of research.
Much of the QC/QA practices were devoted to establishing
precision. As such criteria were used to establish "out-
of-control" guidelines. This aspect of the study is well done
and substantially documented. Somewhat disappointing were the
number of data points classified as out-of-control and thus
not usable. Storage times exceeded recommended maximum holding
times for water, soil, and sediment. This may have accounted
for the high rejection level. More important, a substantial
-------
quantity of data for water samples could not be accounted for
(by this reviewer) on the basis of samples collected. This
problem needs to be clearly rectified.
(7) Is the overall quality of the data adequate to support the
study objectives?
For the most part the precision of this data appears to
be plus or minus a factor of two (2 a) for the sampling and
analysis techniques employed for each environmental medium
except soil which is somewhat larger. To determine whether a
significant difference exists between the declaration and
control areas this level of precision can be one of the limiting
factors. The accuracy for each S/A method is more difficult
to discern, in fact no definitive statement should be made
since internal standards were not employed as surrogates.
(8) Does the data appear to be consistant across all environmental
media studied?
An estimated 90% of the 400,000 measurements performed
yielded results below the detectable limits of each method.
This is the most important observation to be cited of the
data, i.e., its consistency across all environmental media.
Such overwhelming negative evidence suggests that the conclu-
sions to be drawn will be substantially founded. It is highly
improbable that all S/A methods employed were inferior and
incapable of detecting the target chemicals if present.
(9) Does the data provide any insight to significant differences
between the declaration and control areas?
-------
Even though the statistical frequency tests between the
two areas are not significant, the sample size employed in the
control area may have been too small since the number of
values reported below the detection limits was large (possibly
skewed data).
-------
Conclusions
The following responses are offerred to the primary questions
raised concerning this study:
(1) More sites representative of the "control" area in Niagara
Falls should have been selected in this study for making
comparisons to the declaration area.
(2) Data available for review did not indicate that significant
differences in concentrations of the target compounds between
the declaration area and other areas of Niagara Falls; however,
this conclusions is subject to the qualifications cited in
(1).
(3) If a clustering health effect phenomenon was presently latent
in the declaration area it is not clear how the statistical
•
design of this study would have detected this possibility.
The results of this study design suggest at best the minor
differences in concentrations between the declaration and
control area would have to cause acute or chronic adverse
health effects at an elevated rate in the Love Canal area
relative to Niagara Falls.
(4) Based on the available data from this study design the area
appears to be habitable. More importantly, the data should be
reanalyzed to determine whether any one household might be
considered at a greater risk (risk assessment analysis) rather
than comparing statistical means, medians, geometric means
between the two areas and extrapolating potential risk to the
entire area. Perhaps each location (household) in the declara-
tion area should be individually evaluated.
-------
Finally, this reviewer strongly recommends that the Final Report on
the Love Canal be written in acceptable scientific format with one
additional peer review of the report prior to its release. In its
present form it is unacceptable for release.
-------
NATIONAL RESEARCH COUNCIL
ASSEMBLY OF LIFE SCIENCES
2101 Constitution Avenue Washington, O. C. 20413
EXECUTIVE DIRECTOR
BOARD ON TOXICOLOGY AND
VIROiVMENTAL HEALTH HAZARDS
August 26, 1981
Clark W. Heath, Jr., M.D.
Director, Chronic Diseases
Division
Center for Environmental Health
Centers for Disease Control
Atlanta, GA 30333
Dear Dr.. Heath:
I am responding to your letter of August 6th requesting
review of monitoring data from the Love Canal. I have reviewed
the entire package of data and offer my conclusions in the
format stated in your letter:
1. Ar.e the concentra t io.ns measured significantly different
from levels found in other areas of Niagara Falls?
The data indicate that the concentrations of some organic
and inorganic substances are higher in the soil and sewer
sediment in the Love Canal area than in other areas of Niagara
Falls. However, for purposes of assessing risks to human
health, it is unclear from these data whether human exposure is
significantly different from one population to the other.
2. Do the levels measured represent concentrations that
could cause acute or chronic adverse health effects in
people living in the Love Canal area?
The levels measured in the various media would not be
expected to produce acute toxicity unless, for example, large
quantities of contaminated soil would be consumed (an unlikely
prospect). On the other hand, the risk of chronic injury is
real for exposure to the carcinogens and mutagens because the
risks are assumed to have no measureable population threshold.
The risks from these substances are associated not only with
concentration level but also with duration of exposure. Since
there is no longitudinal information about exposure, risk
estimation is necessarily uncertain.
ft* National Research Council is the principal operating agency of the National Academy of Sconces and the National Academy of Engineering
to serve government and other organizations
-------
C. W. Heath,
August 26, 1981
Page 2
Jr. , M.D.
Were
3.
the
materiaIs
sufficient samples taken and analysed to determine
extent of contamination by particularly hazardous
and to be able to assess their potential threat
to human health?
There is little doubt that the sampling was exhaustive
geopgraphically and that all media were analyzed extensively-
These data provide an adequate picture of contamination in the
area at the time of sampling. However, to assess risks to
human health, two assumptions are required: First, the
concentrations in media such as soil represent a measure of
exposure of humans (for sewer sediment this is difficult to
conceptualize); and second, the exposure presently assumed will
continue for a large portion of the lifespan of the residents
(longitudinal data would be advisable to address this issue).
While perhaps impractical, data from actual human exposures
would be the most reliable to make assessments of risks to
human health. If those direct exposure data were available, an
exposure algorithm could be constructed to estimate total body
burden. This information could then be used to estimate risk.
4. Based on available data, can the area be considered
"not habitable?"
From these data, I would conclude that some form of
clean-up or containment is required so as to minimize possible
exposure from, for example, contaminated soils. However from
these data alone it is not possible for me to conclude that
health hazards of such unusual proportions exist to-raake this
area "uninhabitable." There are societal and legal
considerations thac must be taken into account in judging
hab stability.
I appreciate the opportunity to respond to your request.
Sincerely yours,
Robert G. Tardiff,
Executive Director
Ph.D.
-------
HARVARD UNIVERSITY
SCHOOL OF PUBLIC HEALTH
JAMES L. WHITTSNBERGSil, M.D.
Professor of Physiology
James Stevens Simmons Professor of Public Health
Clark W. Heath, Jr., M.D.
Director, Chronic Diseases Division
Center for Environmental Health
Centers for Disease Control
Atlanta, GA 30333
Dear Clark:
665 HUXTINGTOS AVENCE
BOSTON, MASSACHUSETTS 02115
August 19, 1981
Before attempting to answer the questions posed in your letter of August 6, 1981,
I wish to comment in general about the meeting and about the materials provided
to the consultants. Overall it was one of the more frustrating experiences I've
had as a consultant. On trying to read the draft report prior to the meetingr I
found so many graphs and tables that were illegible or uninterpretable that I
gave up trying to form any conclusions from the submitted material.
It would have been very helpful to have had in advance some of the comparison
tables that some of us tried to construct for ourselves by searching through the
printout during the meeting. For example, I thought it would be interesting to
compare one air pollutant measurement in the living areas of the homes sampled
in each of the geographic areas studied. Eventually I had a simple table showing
that the chemical was present in all geographic areas except for the control
area. It was annoying to be told repeatedly, in answer to questions, that "the
data are all there on the table for you to examine." I was informed later that
EPA had made additional comparison tabulations which had not been accepted by
CDC. If this is true, I wonder about the reasons. I can respond better if I
have a scientific presentation in essentially the form presented for publication
in a peer-reviewed journal - the objectives, a statement of hypotheses to be
tested, the study design, the methods in detail, including sampling strategy,
the results, and the author's interpretations. I don't accept the author's
interpretations unless he has persuaded me that he has used the best possible
methods and analysis. I had the feeling that most such information came out
piece-meal at the meeting, and we would have had a more fruitful day if these
issues had been presented clearly at the beginning, or if we had had better
information in advance.
On Page 17 of the EPA report is reference to "two comparison volumes." These
were never identified so far as I know, although they contain data for Area 97,
which is an interesting comparison area in some of the tables. The air monitoring
data for Area 98 might also be of interest.
Whatever additional comments I have about the study will come in my attempts to
answer the questions.
Question 1: Are the concentrations measured significantly different from levels
found in other areas of Niagara Falls?
-------
Dr. Eeath - 2 - Aug. 19, 1981
The more I reflect on this question, the more I wonder what it means. If taken
literally, the answer is no, because many of the "concentrations measured" were
found in "other areas of Niagara Falls." Presumably the intent was to compare
either Area 11 (the Canal) or the Declaration Area (Areas 1 through 10), with
the "control area" (Area 99). Since Niagara Falls has so many hazardous waste
dumps, active or inactive, and since the control areas were defined in terms of
proximity to known dump sites, the concept of "control area" is questionable in
Niagara Falls, since dissemination or translocation of chemicals from dump sites
is certain to have occurred by various construction activities as '.veil as by
natural processes. There are high concentrations of inorganics in sampling sites
in many parts of Niagara Falls, including metals like arsenic, lead, mercury, and
cadmium. My conclusion is that the Study Area (which does not include Area 11),
is not significantly different from other areas of Niagara Falls, provided that
one takes into account that storm sewers and outfall sediments can have high
concentrations of contaminants from Love Canal.
2. Do the levels measured represent concentrations that could cause acute or
chronic adverse health effects in people living in the Love Canal area beyond
what might be expected under usual residential conditions in the Niagara Falls
area?
Again, I assume the question refers to the Study Area, excluding Area 11. Con-
sidering contamination of air in living area and contamination of drinking water,
I think residents of the Study Area are not at greater risk than might be expected
under usual residential conditions in Niagara Falls.
3. Were sufficient samples taken and analyzed to determine the extent of contami-
nation by particularly hazardous materials and to be able to assess their
potential threat to human health?
As noted at the meeting on August 13, 1981, the sampling and analytical methods
used by EPA and its subcontractors were "state of the art" and probably as well
carried out as is possible in a study of this magnitude and complexity. EPA has
been very sensitive to quality control issues since their problems with CHESS
(a large program of epidemiologic studies of air pollution several years ago)
and their-problems with pesticide testing; I believe they took quality control
measures seriously in this study and did as well as anyone could do considering
the number of subcontractors and the state of development of specific methods.
I think sufficient samples were taken except possibly for the control area.
Only four houses had air sampling because it was so difficult to persuade
occupants to vacate their houses for 3 months. If the data were consistent for
a given house sampled repeatedly, for a week, for example, it might have been
better to have more houses sampled for much briefer periods'. However, I don't
see this as a strong criticism of the EPA study.
4. Based on available data, can you conclude that the area is not habitable?
I assume again that we are talking about residential areas outside the rings of
houses that were largely evacuated in 1978, and assume further that remedial
measures will be continued to prevent further dissemination of chemicals still
retained in the Canal, or previously deposited in storm sewers or outfalls.
-------
Dr. Heath - 3 - Aug. 19, 1981
Under these assumptions, I believe residence in the Study Area would be no more
hazardous than in other residential areas of Miagara Falls.
In summary, I think EPA has done a very good study under difficult circumstances.
In my opinion, the study is an adequate basis for judging the habitability of
the areas neighboring Love Canal. Unfortunately, the findings are not directly
relevant to assessment of human exposures pre-1978, since the study was performed
after remediation. I believe that neighborhoods 200 ft. or more from the old
Canal are just as habitable as other residential areas in Niagara Falls, provided
remedial measures are continued, to contain the residual hazardous wastes.
Sincerely,
g
James L. Whittenberger, M.D.
JLW :mc
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
REGION II
26 FEDERAL PLAZA
NEW YORK NEW YORK 10273
Dr. Clark W. Heath
Director, Chronic Diseases Division
Center for Disease Control
Department of Health & Human Services
1600 Clifton Boad, N.E.
Atlanta, Georgia 30333
Dear Dr. Clark:
I have received Dr. Brandt's June IS, 1982 letter concerning the Department of
Health & Human Services' (HHS) review of the Environmental Protection Agency's
(EPA) Love Canal report. That letter indicates you have relied heavily on the
National Bureau of Standard's (NBS) review in reaching your conclusions. The
Love Canal report has been modified to answer the key questions raised toy NBS
and to clarify the description of the quality assurance and quality control
program and results.
Before HHS makes its final decision concerning their evaluation of the Love
Canal report, I would appreciate your reviewing the final text. You should
consider not only the issues raised by NBS, but the physical and chemical
properties of the chemicals and the knowledge accunulated to date concerning
"how these substances migrate. Based on collected Love Canal data and the
"property" factors, EPA has concluded that, except for the storm sewers,
there are no chemicals in the Declaration Area attributable to migration
from Love Canal, and that the frequency of occurrence and concentration of
chemicals detected in the Declaration area is no different than the rest of
Niagara Falls.
As you are aware, it is NBS's contention that detection limits and precision
and accuracy should be specified before samples are analyzed. It is EPA's
experience that such specificity is not possible given the present state-of-
the-art in GC-MS organic chemical analytical methods.
NBS also is of the opinion that all reported measurement results should
be accompanied by detection limits and precision and accuracy estimations
derived from field samples. EPA agrees that it is preferrable, where
possible, to provide method detection limits (MDLs) and precision and
accuracy estimates for every chemical measured. As part of the Love Canal
study, measured MDLs and precision and accuracy estimates were derived for
38 organic chemicals and 16 inorganic chemicals out of the approximately 150
expounds that were measured (see Appendices C, D and E of the Love Canal
Report).
In the professional judgment of EPA scientists, it is reasonable to assume
that the MDLs and precision and accuracy of the approximately 95 other
organic chemicals are similar to the 38 for which we have values. (See
Table I.) EPA does not claim that there is field or laboratory data that
-------
-2-
supports this judgment and, clearly, it would be preferrable to have
additional information of this type, but we do not believe that it is
essential.
Again as a practical matter, such information is generally not obtained,
either in other studies or tests performed by EPA, analyses performed by
other Federal agencies, including the Food and Drug Administration, and
private industry. EPA believes that although the quality control and
quality assurance performed during the Love Canal study was neither per-
fect nor the stringent ideal specified by NBS, it was acceptable and
generally more detailed than most field studies where trace levels of
organic chemicals are analyzed.
As mentioned above, EPA did perform the kind of MDL and precision and
accuracy estimates suggested by NBS on 38 organic chemicals. Also, it
is important to note that since the Love Canal study EPA's EMSL Labora-
tory, Cinn., has developed single laboratory MDLs for the "other" 95
organic compounds. Thus, we do have a data base which can assist us
in decision making. The physical and chemical properties of the 38
chemicals includes those which are most mobile (e.g. benzene, chloro-
banzene, and toluene) and which are relatively soluble in water and
moderately to highly volatile. Other chemicals adsorb to soil and are
less mobile (e.g. & BHC, PCB JL242 and dioxin) and have low solubility
and are highly adsorptive. Of course, the metals also have different
physical and chemical characteristics that can act as effective tracers
of many of the inorganic chemicals. Since the MDLs and precision and
accuracy are known for these inorganics, definitive conclusions can be
drawn about how far they have migrated from Love Canal. It is our
opinion that the available organic and inorganic data demonstrate
that chemicals have not migrated beyond the fenced-in area, except
through the storm sewers.
Further evidence of this characteristic migration pattern of organic
chemicals is supported by a review of the literature. Enclosed are a
few of the articles which document this pattern.
In summary, the data from Love Canal, the literature and EPA's experi-
ence at other hazardous waste landfills indicate that chemicals did not
migrate past the first ring of homes. Given this conclusion, it is
essentially irrelevant whether there is less confidence in the meaning
of "none detected" for the 95 other organic chemicals in the Declaration
Area. Also, it is important to point out that the procedure for deter-
mining the method detection limit, which is set forth in Environmental
Sciences and Technology, Volume 15, Number 12, December, 1981, pages
1426-35, provides a high degree of assurance that positive identifi-
cations are real identifications. Under this procedure the probability
is less than 1% that a value reported as "none detected" (which includes
both trace and below detectable values) is, in fact, a value greater than
twice the method detection limit. This represents the worst case situation.
-------
-3-
As you knew, EPA. is planning on releasing its report on or before
July 14, 1982; therefore, I vrould appreciate your final response
concerning our position.
Sincerely yours,
Richard T/ Dewling, Hi.D.
Deputy Regional Administrator
-------
\
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
ENVIRONMENTAL MONiTOK'NG AND SUPPORT LABORATORY
CINCINNATI, OHIO" 45263
ADMINISTRATIVELY CONFIDENTIAL
DATE: July 9, 1982
SUBJECT: Method Detection Limits (MDL) for Water, Soil,
and Sediment Samples
FROM: William L. Budde, Chief /,} f*>,? ^/Ot.L
Advanced Instrumentation Section ^-^^^ -/O^^<~--
Physical and Chemical Methods Branch
Robert L. Booth, Acting Director *.<£. *<£>&
Environmental Monitoring and Support /
Laboratory - Cincinnati
TO: Richard Dewling
Deputy Regional Administrator, Region 2
U. S. Environmental Protection Agency
26 Federal Plaza
New York, New York 10007
During June 1980, when the Love Canal monitoring program was designed, it
was recognized that cost considerations precluded MDL determinations by all
contract laboratories for the full list of targeted organic compounds.
Therefore, a subset of 20 compounds—nine for Method 624 and eleven for
Method 625, was selected for MDL determinations by all laboratories (Table
C-l, p. 229 of Volume I, Environmental Monitoring at Love Canal).
The subset compounds were selected to represent specific structural classes
and groups of analytes with similar physical and chemical properties. The
subset compounds are shown numbered from 1 to 20 in the attachment. Under
each compound is listed the compounds with similar properties. Among
similar properties, the following were considered most signficant: chemical
structure, volatility, water solubility, acidity, chromatographic behavior,
and mass spectrometric fragmentation,at 70 eV. All groups were based on
judgments of the interrelationships of these properties.
The MDLs for the 18 pesticides measured by Method 608 were also determined
by a contractor laboratory (Table C-3, p. 231 of Volume I). In addition,
the Environmental Monitoring and Support Laboratory - Cincinnati
(EMSL-Cincinnati) has measured MDLs for most of the analytes cited in
Methods 624 and '625. These values have been published along with the
procedure for their determination (Environmental Science and Technology,
1981, 15, 1426-1435). In the enclosure, the EMSL-Cincinnati MDLs are given
for your convenience. These values generally support the representative
nature of the subset compound for each group.
-------
For soil and sediment samples (discussed in Appendix D, pages 247-250 of
Volume 1) MDLs can be considered equivalent to those observed for water.
Since standard reference materials are not available, it was not possible to
determine recoveries of organics from the sample matrices. However,
extraction conditions were selected that provided far more efficient removal
of organics from these samples than natural conditions could affect in the
Love Canal area. Accordingly, the probability of very low recoveries for
the soil and sediment samples having a significant effect on the MDL values
is very small.
Finally, the concern for the need to assign values to the "non-detected"
analytes (this includes trace amounts and below detection limit
designations) has been addressed in the attachment under "worst case
range." These ranges for the twenty representative compounds were derived
by taking two times the MDL of the lowest and highest MDL values cited in
Table C-l of Volume 1. In so doing, the statistical probability is that far
less than 1 percent of the "non-detected" analytes would, in fact, exceed
worst case range. Therefore, it can be stated with high confidence that no
analytes cited as present in trace amounts or below detection limits in the
Love Canal Monitoring Report could, in fact, have been present at the one
mg/L or Kg (ppm) or higher level.
Enclosure (1):
As Stated
cc: Courtney Riordan w/enclosure
-------
Concentration, ng/L or ppb
Method 524
1. Benzsne
2. Chlorobenzene
benzyl chloride
2-chlorotoluene
3-chlorotoluene
4-chlorotoluene
3. Chloroform
methylene chloride
chloromethane
vinyl chloride
trichlorofluoromethane
bromochloromethane
bromodichloromethane
4. bromoform
bromomethane
di bromoch1oromethane
1,2-dibromoethane
5. 1,1,2,2-tetrachloroethane
(sym-tetrach1oroethane)
1,1-dichloroethane
chloroethane
6,
7,
1,
1,
1,
1,
2-dichloroethane
1,1-trichloroethane
2-dichloropropane
1,2-trichloroethane
carbon tetrachloride
trichloroethylene
1,1-dichloroethene
cis-1,2-dichloroethene
trans-1,2-dichloroethene
2,3-dichloropropene-l
trans-1,3-dichloropropene-l
MDL Worst Case
(EMSL-Cincinnati) Range (L-H)
4.4 5 - 52
6.0 4 - 34
6 (estimated)
6 (estimated)
6 (estimated)
6 (estimated)
1.6 3-58
2.8
2.8 (estimated)
2.8 (estimated
2.8 (estimated)
2.8 (estimated)
2.2
4.7 4 - 84
4.7 (estimated)
3.1
4.7 (estimated)
6.9 3 - 62
4.7
6.9 (estimated)
4.1
3.8
6.0
5.0
2.8 5-74
1.9 3 - 52
2.8
4.7
1.6
2.2
5.0
-------
Concentration, yg/L or ppb
MDL Worst Case
(EMSL-Cincinnati) Range (L-H)
8. tetrachloroethylene
9. toluene
o-xylene
m-xylene
p-xylene
ethyl benzene
Method 625
10. 1,4-dichlorobenzene
1,3-di ch1orobenzene
1,2-dichlorobenzene
bis(2-chloroethyl)ether
bis(2-chloroisopropyl)ether
N-nitrosodi-n-propylamine
hexachlorobutadiene
N-n i trosod imethy1 ami ne
bis(2-chloroethoxy)methane
11. 1,2,4-trichlorobenzene
2,4-dichlorotoluene
1,2,3-trichlorobenzene
1,3,5-trichlorobenzene
isophorone
4-chlorobenzotrifluoride
(trifluoro-p-chlorotoluene)
12. 1,2,3,4-tetrachlorobenzene
hexachlorobenzene
hexachlorocyclopentadiene
pentach1oron i trobenzene
2,4,6-trichloroaniline
1,2,4,5-tetrach1orobenzene
tetrachlorotoluenes
4.1
6.0
6 (estimated)
6 (estimated)
0.9
7.2
5.0
4.4
1.9
5.7
6.3
8
0.9
5 (estimated)
5.3
1.9
9 (estimated)
9 (estimated)
9 (estimated)
1
1
1
2.2
1,9 (estimated)
0.5
1.9
0.5
0.5
0.5
0.5 (estimated)
0.5 (estimated)
(estimated)
(estimated)
(estimated)
5 - 56
12 - 19
10 - 68
4 - 64
1 - 34
-------
Concentration, yg/L or ppb
13. 2,4,5-trichlorophenol
2-chlorophenol
3-chlorophenol
4-chlorophenol
phenol
2,4-ch'methyl phenol
2,4-dichlorophenol
2,3,5-tri ch1oropheno1
4-chloro-3-methyl phenol
14. pentachloropheno1
3,3-dichlorobenzidine
15. 2,6-dinitrotoluene
2,4-dinitrotoluene
nitrobenzene
16. 4-nitrophenol
2-nitrophenol
2-methyl~4,6-dinitrophenol
2,4-dinitrophenol
17. 2-chloronaphthalene
4-chlorophenylphenylether
N-nitrosodiphenylamine
4-bromophenylphenylether
1,2-diphenylhydrazine
18. beta-benzenehexachloride
hexachloroethane
delta-BHC'
heptachlor
aldrin
mi rex
heptachlor epoxide
DDE
dieldrin
endosulfan sulfate
ODD
chlordane
DDT
MDL Worst Case
(EMSL-Cincinnati) Range (L-H)
2.7 5-48
3.3
2.7 (estimated)
2.7 (estimated)
1.5
2.7
2.7
2.7 (estimated)
3.0
3.6 7 - 60
16.5
1.9 4-50
5.7
1.9
2.4 5-42
3.6
24
42
1.9 4-34
4.2
1.9
1.9
22
4.2 8 - 19
1.6
3.1
1.9
1.9
4.2 (estimated)
2.2
5.6
2.5
5.6
2.8
4.2 (estimated)
4.7
-------
Concentration, yg/L or ppb
MDL Worst Case
(EMSL-Cincinnati) Range (L-H)
19. fluoranthene
phenanthrene
anthracene
pyrene
chrysene
naphthalene
acenaphthylene
acenphthene
fluorene
benzo(a)anthracene
benzo(kjfluoranthene
benzo(b)fluoranthene
benzo(a)pyrene
indeno(1,2,3-dc)pyrene
dibenzo(a,h)anthracene
benzo(g,h,ijperylene
20. di-n-butylphthalate
diethylphthalate
butylbenzylphthalate
di(2-ethylhexyl)phthalate
dimethylphthalate
di-n-octylphthai ate
2.2
5.4
1.9
1.9
2.5
1.6
3.5
1.9
1.9
7.8
2.5
4.8
2.5
3.7
2.5
4.1
2.5
1.9
2.5
2.5
1.6
2.5
4 - 40
5 - 158
Notes
A. AeryIonitrile and acrolein were Method 624 analytes for qualitative
analysis only. Quantitative analyses and Method Detection Limits for
these analytes are defined in Method 603.
B. Benzidine was a Method 625 analyte for qualitative analysis only.
Quantitative analyses and Method Detection Limits are defined in Method
605.
C. Alpha BHC, gamma BHC, endosulfan I, endosulfan II, and endrin were primary
analytes for Method 608 and were confirmed in Method 608 extracts with
Method 625 GC/MS procedures.
-------
In order to evaluate the potential effects of choice of minimum detection
limit (MDL), EPA independently determined the MOL according to published procedures
(Environmental Science and Technology, 1981, 1426). The variability reported by
the laboratory with the highest variability for the lowest level of the calibration
check sample (for example the highest variability shown for the 1,2,3,4 tetra-
chlorobenzene 750 ng sample in the case of polyurethane foam) was used as a
"worst case" estimate of variability for this determination. The results in
are shown in the second column of the table.
Column 3 of this table is the upper one-sided tolerance limit (<*=0.99) for
99.9% of the distribution of the population about the MDL given in column 2. This
is interpreted to mean that if we were to repeat this experiment 100 times under
similar conditions, we would expect at least 99.9% of the population would be
less than the tabled value, 99 times. In other words, it is extremely unlikely.
that a value greater than given in column 3 would be Tabled below minimum
detectable or trace in the data set.
The information for these estimates of MDL comes from Tables E-3, E-8, E-9,
and E-10 of the report Environmental Monitoring at Love Canal Volume I.
C]>C
-------
Table: Estimated Highest Detection Limit and Val'j= .r.:r 'J^par 0.999 Tolerance
Limit for Air Measurements at Love Canal
Est. Highest n Upper Tolsrancs
Compound Detection Lirit (UT/:.^) Llni":^ (i.;g/ri^}
Benzene 11 32
Carbon tetrachloride 5 14
Chlorobenzene 8 24
o-chlorotoluene 6 17
1,2-dibromoethane 6 18
o-dichlorobenzene 6 17
1,1,2,2,-tetrachloroethylene 9 26
Toluene 16 50
1,2,3,4-tetrachlorobenzene 0.6 2.0
Pentachlorobenzene 0.8 2.8
Hexachlorobenzene 0.1 0.4
Lindane 0.1 0.4
2,4,5-trichlorophenol 0.1 0.4
Element
Antimony 0.015 0.039
Arsenic 0.010 0.024
Beryllium 0.001 0.001
Cadmium 0.001 0.002
Chromium 0.013 0.033
Copper 0.006 0.015
Lead 0.017 0.041
Nickel 0.002 0.004
Zinc 0.202 0.508
-------
II
N
B
S
-------
9 t98?
UNITED STATES DEPARTMENT OF COMMERCE
National Bureau of Standards
Washington, D.C. 20234
OFFICE OF THE DIRECTOR
Dr. Richard T. Dewling
Deputy Regional Administrator
Environmental Protection Agency
Room 900
New York, New York 10278
Dear Dr. Dewling:
As requested on July 8, 1982, in a meeting of representatives of the
Environmental Protection Agency (EPA), the Centers for Disease Control
(CDC) of Health and Human Services, the Department of Justice (DoJ),
and the National Bureau of Standards (NBS), I am providing written
comments on the NBS review of the Analysis for Organic Chemicals in
the EPA Love Canal Monitoring Study. These comments are based on an
extensive NBS review of documentation on this subject provided to NBS
by EPA from August 18, 1981 through February 17, 1982, and a limited
review of the approach to setting method detection limits described in
Appendix C, "Limits of Detection/Quantitation," Volume 1 of the final
report.
Additional NBS comments are intended to supplement but not replace the
NBS review and are appropriate for two reasons:
o EPA provided to NBS on June 28, 1982, additional information
on limits of detection for the water analysis program in a
revised Appendix C of their-final report.
o CDC is having difficulty interpreting the significance of the
NBS review. The NBS review was aimed in part at helping
EPA to adequately document numerical data over the full
range of concentrations measured (parts per billion
range and above). CDC has indicated that their concerns
are with concentrations in the parts per million range and
above.
As I stated to you and CDC, the methods of analysis used by EPA for
water, soil and sediments and air are generally acceptable methods and
represent the state of the art. As we stated in our review, although
there are difficulties in implementation, the number, nature and
frequency of analysis of quality assurance samples specified by EPA
should have been adequate to maintain quality control.
Most of the recent discussion between EPA, CDC and NBS has centered on
limits of detection for the soil and sediments, air and water analysis
programs. As EPA clearly recognized in their report,- they were not
able to establish limits of detection for soils and sediments. With
respect to the limits of detection for the air program, EPA
established and presented limits of quantitation for targeted
compounds and limits of detection were determined by the participating
laboratories. The documentation of the performance of the air
monitoring laboratories is generally acceptable. The exception to
this statement is the problems that arose from TENAX contamination
that are discussed in the EPA report and the NBS review.
-------
In order to respond to NBS concerns, raised in its review, on minimum
detection limits for the water analysis program, EPA has presented an
approach based on method detection-limits. It is the assessment of
NBS that this is a reasonable approach. In our view, the method
detection limits obtained from this approach can be used for data
analysis requiring minimum detection limits. We have three provisos
to make with respect to our opinion.
(1) EPA should present the raw data and the chain of logic used
to arrive at the method detection limits appropriate to
each laboratory doing analysis.
(2) EPA should compute the actual minimum levels of
detection for the measurements as made in the
laboratories. That is, they should replace the
"zeros" and "traces" in their data with "not more thans".
(3) EPA should demonstrate that the 38 compounds for which
method detection limits have been stated are representative
of the targeted compounds for the Love Canal water samples.
CDC has asked if NBS can certify the data contained in the EPA report.
NBS cannot. Our comments pertain to the procedures and methods used
by EPA. The data, as is true with any technical report, are the
responsibility of the authors—in this case EPA. Representatives of
EPA indicated in the meeting that, using an approach that NBS assesses
as reasonable, EPA will provide data to CDC on minimum detection
limits for the water analysis program.
You and CDC have asked if NBS endorses the statement that the NBS
review did not identify any problems that would affect the conclusions
of the monitoring program. NBS cannot comment on the significance of
the problems that NBS has identified to the conclusions of the EPA
report because such an evaluation requires knowledge of health effects
and mechanisms of chemical migration and degradation in addition to
knowledge of chemical analysis. NBS is knowledgeable of chemical
analysis only.
If I understand the thrust of recent questions, both EPA and CDC are
looking to NBS for a general characterization of the EPA study rather
than the comments on particular issues to which we in our review have
confined ourselves in accordance to the original assignment from EPA.
NBS is not able to provide a general characterization of the EPA study
because there are not quantitative statements of the precision,
accuracy and minimum detection limits for data requirements for the
study. In the absence of such a quantitative statement, general
characterizations will have to come from the experts drawing
conclusions from the study—namely, EPA and CDC.
Sincerely,
Raymond G. jammer
Deputy Director
-------
UNITED STATES DEPARTMENT OF COMMERCE
National Bureau of Standards
Washington, D.C. 20234
OFFICE OF THE DIRECTOR
6 1982
Dr. John Hernandez
Deputy Administrator
U.S. Environmental Protection
Administration
Washington, D.C. 20460
Dear Dr. Hernandez:
Enclosed is an original and three copies of the National Bureau of
Standards' Review of Material Provided by EPA on the Analysis for
Organic Chemicals^ in the EPA Love Canal Monitoring Study. As we
discussed on April 9, 1982, the NBS Review is brief and addresses
the seven tasks, as modified, in the charge to NBS from EPA. A
more detailed discussion of the NBS Review of the EPA study is
contained in Appendix A.
The NBS Review is of the draft report which was received by NBS on
December 17, 1981, as modified by supplementary material received
through February 17, T982. We hope this Review will be helpful to
you in the preparation of your final report for the study. We are
available to discuss or clarify any aspect of our Review that you
or your staff may find useful.
Sincerely,
Raymond G. ammer
Deputy Director
Enclosures
-------
NBSIR 82-251 1
REVIEW OF MATERIAL PROVIDED BY
EPA ON THE ANALYSIS FOR ORGANIC
CHEMICALS IN THE EPA LOVE CANAL
MONITORING STUDY
May 1982
U.S. DEPARTMENT OF COMMERCE, Malcolm Baldrige, Secretary
NATIONAL BUREAU OF STANDARDS, Ernest Ambler, Director
-------
Review of Material Provided by EPA on the Analysis
for
Organic Chemicals in the EPA Love Canal Monitoring Study
NBS Review Panel
H. S. Hertz L. T. McClendon
H. H. Ku J. K. Taylor
W. S. Liggett E. White V
W. H. Kirchhoff, Chairman
May 10, 1982
-------
Table of Contents
Page
Foreword 1
Introduction 3
Review of the Analysis for Organic Chemicals in
the EPA Love Canal Monitoring Study 4
Appendix A: Responses to the Template Questions 11
I. Goals and Objectives for the Acquisition of Organic
Analytical Data from the Love Canal Study 11
II. Monitoring Protocols 16
III. Quality Assurance Protocols 21
IV. Performance of the Monitoring Program 26
V. Performance of the Quality Assurance Program 32
VI. Data Reduction and Analysis 38
VII. The EPA Audit of the 6C-MS Computer Records 41
Appendix 8: Documents Received by NBS 44
-------
FOREWORD
On July 20, 1981, EPA requested NBS to review the analysis for organic chemicals
performed by EPA in its Love Canal Monitoring Study. In response to that
request, this review has been prepared by a panel of NBS scientists with
expertise in organic analytical chemistry, quality assurance and statistics.
"Hie review is based upon written material, provided by EPA to NBS, related to
the acquisition of organic analytical data for samples collected from the Love
Canal vicinity by EPA contractors.
The charge from EPA, accepted by NBS on August 17, 1981, was to carry out the
following seven tasks:
1. Review the sample collection and analytical protocols for organic
chemicals in air, water, soils, sediments and biota to determine their
appropriateness for identifying and measuring the substances of interest
to EPA at Love Canal.
2. Review the adequacy of the quality assurance/quality control protocols
for all of the media in Task 1 to assure the validity of substance
identification and analytical measurements.
3. Review the results of the quality assurance program presented in the EPA
Love Canal report to assess the performance of the analytical program.
4. Review as necessary any of the reports of the on-site laboratory audits
carried out by EPA.
5. Review the results of the EPA audit of GC-MS tapes to evaluate the
quality of the analytical program.
6. Review any of the monitoring data as necessary to help in the overall
review of the program.
7. Prepare a report on the overall adequacy of the EPA analytical and
quality assurance protocols to meet the organic chemical monitoring
goals (accuracy and precision) as set forth in the EPA Love Canal
Monitoring Plan.
On August 18, 1981, NBS received an initial set of documents for review, and on
August 25, NBS was given an oral briefing by officials of EPA and
representatives of the prime contractor, the GCA Corporation, on the conduct of
the Love Canal Monitoring Study. At the time NBS initiated its review, no final
report on the Love Canal Monitoring Study was available. As a result of the
unavailability of such a report, Tasks 3 and 7 were modified to:
3. Review the results of the quality assurance program as revealed in
documents provided to NBS describing the performance of the analytical
program.
-------
and
7. Based upon written material provided by EPA, prepare a report
summarizing the reviews conducted by NBS in Tasks 1 to 6, paying
particular attention to conclusions drawn by EPA concerning the
precision and accuracy of organic analytical data from the Love Canal
Monitoring Study.
In order to give structure and specificity to the review, the NBS panel devised
a set of questions to serve as guidance for the evaluation of the organic
analytical portion of the EPA Love Canal Monitoring Study. This set of
questions, designated in correspondence between EPA and NBS as the template for
the review, was transmitted to EPA on October 13.
On December 9, 1981, NBS submitted its review of material received from EPA.
That material consisted of a variety of documents including contractor plans,
tables of data derived from samples taken from the Love Canal vicinity, tables
of data derived from internal and external performance check samples, quality
assurance audit reports, miscellaneous letters and reports, and a preliminary
draft of portions of a final report. This latter document was designated as
OEM-LC-2 by NBS. The December 9 review identified a number of deficiencies in
the Love Canal Monitoring Study on the basis of the written material provided by
EPA. In the cover- letter to EPA accompanying the December 9 review, NBS stated
that the responses to the template questions would be submitted to EPA in
January 1982.
On December 15, EPA requested NBS to review additional material, and on
December 17, EPA provided NBS with a copy, designated #9, of a draft report.
NBS was informed that those sections of the draft report which dealt .with the
quality assurance aspects of the study were still in the process of being
revised.. NBS was requested to work with the revised versions. Since the
additional material influenced the NBS responses to the template questions, NBS
decided not to transmit the responses until the additional material had been
reviewed. By February 17, 1982, the last of the additional material provided by
EPA, which included revisions of the draft report and a copy of the final report
of the prime contractor, was received. At the request of EPA, NBS briefed EPA
on the contents of this review on April 15, 1982.
The present Review covers all material received, with the exception of the
aforementioned OEM-LC-2, which was replaced by the draft report. All documents
received by NBS are catalogued at the end of this Review.
Finally, the authors of this Review wish to acknowledge the advice and
assistance of Prof. Michael Gross, Mr. Hugh Huffman, and Dr. James Sphbn.
-------
INTRODUCTION
This is a review by the National Bureau of Standards of the analysis for organic
chemicals conducted by the U. S. Environmental Protection Agency in the Love
Canal Area of Niagara Falls, New York. In conducting this review, NBS has been
guided by the seven tasks specified by EPA in the charge given to NBS.
To give additional structure to the review and to ensure that all appropriate
aspects of the Organic Analytical Program of the Love Canal Monitoring Study
were considered, NBS devised a series of questions concerning various aspects of
the program to serve as a detailed guide for the review. This set of questions
has been designated in correspondence between NBS and EPA as the "template" for
the review and the responses to the template questions are included as Appendix
A of this Review.
In this Review, NBS has commented upon deficiencies or limitations that NBS has
identified in the Love Canal Monitoring Study. Not all of these deficiencies or
limitations influence the conclusions drawn by EPA in its draft report and some
have already been described and interpreted by EPA. The purpose of including
such comments is to call to the attention of readers particular limitations
which might not be immediately apparent or which should be considered by those
who wish to use the EPA data to draw independent conclusions. When deficiencies
or limitations are identified which might affect the conclusions drawn by EPA in
its draft report, they are so noted.
Finally, since the NBS review was to be concerned solely with the analysis for
organic chemicals, other aspects of EPA's Love Canal Monitoring Study have been
excluded, by agreement, from this review. Specifically, these are:
1. The choice of sample locations and media sampled. NBS did review,
however, how the sample collection protocols might have affected the
chemical integrity of the samples being collected.
2. The reduction of acquired data leading to contamination level summaries
and distribution patterns. NBS did review, however, those statistical
procedures associated with the determination of precision, accuracy,
and limits of detection in the analysis of environmental samples for
organic chemicals.
3. The choice of compounds which would be determined in the samples
collected from the Love Canal vicinity. NBS did consider, however, the
effect of selecting a list of targeted substances on the overall
analytical program.
4. The analytical programs for inorganic chemicals and radioactivity.
5. The conclusions drawn by EPA from the Love Canal Monitoring Study. NBS
did consider, however, whether or not the quality and reliability of
the organic analytical data might affect the conclusions.
-------
REVIEW OF THE ANALYSIS FOR ORGANIC CHEMICALS
IN THE EPA LOVE CANAL MONITORING STUDY
EPA STUDY GOALS
The EPA Love Canal Monitoring Study was undertaken to accomplish three
explicit goals:
1. Determine the current extent and degree of chemical contamination
in the area defined by the emergency declaration order.
2. Assess the near-term and long-term implications of groundwater
contamination in the general vicinity of Love Canal.
3. Provide an assessment (from an environmental contamination
perspective) of the habitability of residences included in the
emergency declaration order.
EPA STUDY DESIGN FOR THE ANALYSIS FOR ORGANIC CHEMICALS
The monitoring effort involved the analysis of environmental samples for
more than 100 targeted substances. The substances targeted for analysis
were selected from those known to have been buried in Love Canal, those
previously observed in the environment near Love Canal, and the EPA list of
priority pollutants. In addition to the targeted substances, the EPA'draft
report cites as one of the major safeguards in the monitoring study the
identification of the 20 most abundant non-targeted substances in each
environmental sample.
Four media—air, water, soil and sediments, and biota--were to be
independently studied using gas chromatography-mass spectrometry (GC-MS) as
the primary analytical technique. A detailed quality assurance program was
designed for which the number of quality assurance samples to be analyzed
equaled the number of field samples. From the analytical data, EPA was to
generate a validated data base for the environmental samples.
SUMMARY OF THE NBS REVIEW
In response to a request from EPA, NBS has reviewed the analysis for
organic chemicals in the EPA Love Canal Monitoring Study. The review is
based upon written documentation provided to NBS during the period
August 18, 1981 to February 17, 1982. (See Appendix B for the list of
documents received.) The review was confined to those subjects discussed
in the Foreword and the Introduction to this report.
-------
EPA's stated goals and objectives, as they applied to the analysis for
organic chemicals, were appropriate in that media, substances, potential
patterns of contamination, and temporal variation were considered. They
were not, however, quantitatively explicit. EPA did not define at the
outset of the study the limits of detection or the precision and accuracy
required to permit comparison of levels of contamination in the Declaration
Area with levels which constitute an environmental hazard or with levels
currently found in U.S. cities.
In the implementation of the monitoring program, the requirement for the
identification of the 20 most abundant non-targeted substances was
generally not met. Consequently, the possibility of contamination by non-
targeted substances has not been conclusively excluded. This deficiency
does not invalidate, however, the investigation of targeted substances.
The quantitation limits realized by the water monitoring laboratories for
individual analytes varied by more than a factor of ten. Few laboratories
reported quantitation limits as good as those reported by the EPA
laboratory responsible for the water monitoring effort. Comparisons of
quantitation limits with data provided by EPA from other studies indicated
that the contract laboratories were not uniformly performing at the state
of the art in the water monitoring program.
In any study involving measurements, conclusions are bounded by measurement
error and limits of detection. EPA has not fully used its quality
assurance data to derive estimates of precision, accuracy-, and limits of
detection. EPA has not incorporated estimates of precision, accuracy, and
limits of detection into its validated data base. Unless measured values,
including "none detected," are accompanied by estimates of uncertainty,
they are incomplete and of limited usefulness for further interpretation
and for drawing conclusions.
SPECIFIC FINDINGS
In the remainder of this Review, specific comments are grouped according to
the seven tasks in the charge to NBS from EPA. Because of the variations
in the approach taken in the analysis of samples from each of the four
different media, these comments often pertain to only- one of the four
media. Additional detailed comments can be found in the answers to the
template questions appearing in Appendix A.
1. "Review the sample collection and analytical protocols for organic
chemicals in air, water, soils, sediments and biota to determine their
appropriateness for identifying and measuring the substances of
interest to EPA at Love Canal,"
Tne approach of identifying a list of targeted substances and
requiring a search for the 20 most abundant non-targeted substances
-------
was sound. Targeted substances were chosen from appropriate lists and
from the results of earlier Love Canal investigations. The criteria
established for the selection of targeted substances were reasonable.
However, EPA has not explicitly related included or excluded
substances to specific criteria. The substances sought were
representative of broad classes of organic chemicals.
Chemical identification and quantitation by gas chromatography-mass
spectrometry (GC-MS) was appropriate to the general goals and
objectives of the study and represented the best overall technique for
monitoring organic chemicals in environmental samples. NBS is not
aware of any fundamentally different analytical techniques for the
identification and quantitation of organic compounds which would have
been more suitable for the study as planned. For some individual
substances, alternative analytical methods might have been more
suitable (for example, the use of a collection medium other than TENAX
for benzene and toluene--a problem'recognized in EPA's draft report).
The sample collection and analytical protocols were generally complete
for the air and water monitoring efforts and have been, to varying
degrees, widely used in environmental monitoring for several years.
The protocols for soil and sediments and for biota were modifications
of the water methods and were to be evaluated during the course of the
study. (The biota monitoring effort was experimental in nature and
the data resulting from this effort were not used by EPA to draw
conclusions from its study.) Some ambiguity existed in the details of
the protocols which could have affected the quality of the analytical
data.
The monitoring program design did not anticipate the possibility that
a significant fraction of the samples might not contain any detectable
levels of analytes. Consequently, the significance of the limits of
detection and quantitation and how these limits might affect the
conclusions drawn from the study were not adequately considered.
2. "Review the adequacy of the quality assurance/quality control
protocols for all of the media in Task 1 to assure the validity of
substance identification and analytical measurements."
The number, nature, and frequency of analysis of quality assurance
samples specified by EPA should have been adequate to maintain quality
control.
The quality assurance plans for the four media (air, water, soil and
sediments, and biota) were similar in scope but varied in detail. The
quality assurance plans for the participating laboratories were, in
general, complete and followed the outline given in the overall
Quality Assurance Plan. These plans did, however, vary in detail from
one laboratory to another. The variations in detail could lead to
-------
difficulties in comparing data obtained from different laboratories
and different medi a.
The Quality Assurance Plan encouraged but did not require the use of
control charts. Uniform reporting of corrective actions taken
when quality control was lost was not required. It will be difficult
at best to evaluate further any data reported by a laboratory without
an accompanying quality control chart (or equivalent data).
The Quality Assurance Plan did not address how the quality assurance
data would be used to qualify the environmental data, that is, to
assign uncertainty limits to each measured value, including "none
detected."
Estimates of the limits of detection are critical to the conclusions
which can be drawn when the majority of samples contain no measurable
concentration of analytes. The Quality Assurance Plan did not
anticipate such a possibility.
3. "Review the results of the quality assurance program as revealed in
documents provided to NBS describing the performance of the analytical
program."
The quantisation limits realized by the water monitoring laboratories
for individual analytes varied by more than a factor of ten. Few
laboratories reported quantitation limits as good as those reported by
the EPA laboratory responsible for the water monitoring effort.
Comparisons of quantitation limits with data from other studies (where
available) indicated the laboratories were not uniformly performing at
the state of the art.
The limits of quantitation to be achieved by the air monitoring
laboratories were set at concentration levels comparable to levels
observed for some U.S. cities.
As demonstrated by the audit of the GC-MS computer records, the water
and the soil and sediments monitoring laboratories did not generally
identify non-targeted substances. This aspect of the performance of
the air monitoring laboratories was not evaluated by EPA.
In isolated instances, contamination of the samples during sample
collection or analysis may have obscured possible trends and patterns
in the environmental data.
The transcription of the data and the verification of the accuracy of
the transcription and transmittal of the data into the final data base
were adequately demonstrated.
The documentation provided by EPA indicated that the quality control
limits set in the Quality Assurance Plan widened during the study.
7
-------
The degree of change was greater for the water program than for the
air program. The reasons for the observed changes have not been
explained in the documents provided by EPA.
The estimation of the precision and accuracy of the numerical values
of concentrations of targeted substances has not been adequately
documented. Only a portion of the available quality assurance and
environmental data has been used in arriving at these estimates.
The estimates are generally unconfirmed and estimates for some of the
analytes have not been given.
The estimation of the limits of detection and quantitation are also
not adequately documented. Because the conclusions of the study rest
on comparisons of the frequency of compound detection-, the values
obtained for the limits of detection of individual substances can
influence significantly such conclusions.
The rejection of data has neither been clearly described nor
adequately justified.
4. "Review as necessary any of the reports of the on-site laboratory
audits carried out by EPA."
The reports of on-site laboratory audits reviewed by NBS described
audits conducted by EPA and by the prime contractor during the early
stages of the Love Canal Monitoring Study. These reports indicate
that all laboratories-were having some difficulties and some
laboratories were having great difficulty at the start of the study.
In particular, some of the laboratories were experiencing difficulties
in adjusting to the use of capillary columns.
The prime contractor maintained telephone contact with participating
laboratories during the course of the study, but there is no evidence
of further monthly on-site visits as prescribed by the Quality
Assurance Plan. Such on-site audits would have confirmed whether or
not initial difficulties had been overcome, whether or not laboratory
contamination of samples was a problem, whether or not good laboratory
practices were being followed, and whether or not analog data (e.g.,
chromatograms) from the GC-MS runs indicated laboratories were
operating at the expected level of performance.
•5. "Review the results of the EPA audit of GC-MS tapes to evaluate the
quality of the analytical program."
The EPA audit of GC-MS computer records was not an evaluation of the
quality of the entire analytical program, but only an evaluation of
one aspect of that program, namely the interpretation of the computer
8
-------
records. The audit evaluated a sample of the computer records
generated in the water and in the soil and sediments monitoring
programs. No audit of the air or biota monitoring programs was
performed.
The conclusions of the audit have not been completely justified, e.g.,
the claim that missed compounds were predominantly in heavily
contaminated samples is not supported.
The implications of the audit—that significant difficulties existed
at the limits of detection—to the conclusions of the study are not
discussed in documents provided by EPA.
Eighty samples (water, soil and sediments) were audited for the
identification of non-targeted substances. In 58 of these samples,
both the audit laboratory and the analytical laboratory identified no
non-targeted substances. In the remaining 22 samples, the audit
laboratory identified 84 non-targeted substances while the analytical
laboratories found only 1. This result indicated that, for the
samples audited, most laboratories did not identify non-targeted
substances.
6. "Review any of the monitoring data as necessary to help in the overall
review of the program."
Precision, accuracy, limits of detection, and limits of quantisation
have not been incorporated into the validated data base. TTrat is, the
measured values, including "none detected," in the validated data base
are not accompanied by estimates of their uncertainty.
The information given is not sufficient to assist others in the
interpretation of the Love Canal data. Information on 1aboratory-to-
laboratory variability is incomplete. In particular, recovery factors
have not been given nor confirmed for most analytes in water and in
soil and sediments samples.
In the presentation of information, the distinction between plans and
attained performance is not always clear. Examples selected to
support the conclusions of the study have not always been confirmed as
being truly typical of the data in general.
Conclusions of no discernible patterns or differences or of no
significant contamination are bounded by the limits of detection and
quantisation. In the presentation of its conclusions, EPA has not
addressed explicitly how those conclusions are influenced by the
limitations in its analytical data. Unless measured values, including
"none detected," are accompanied by estimates of uncertainty, they are
incomplete and of limited usefulness for further interpretation and
for drawing conclusions.
-------
"Based upon written maternal provided by EPA, prepare a report
summarizing the reviews conducted by NBS in Tasks 1 to 6, paying
particular attention to conclusions drawn by EPA concerning the
precision and accuracy of organic analytical data from the Love Canal
Monitoring Study."
The foregoing discussion, along with additional, detailed comments
appearing in Appendix A, constitute the report summarizing the reviews
conducted by NBS in Tasks 1 to 6.
10
-------
APPENDIX A
RESPONSES TO THE TEMPLATE QUESTIONS
Template questions were designed by NBS to serve as a guide for its review, to
give structure to the review, and to ensure that all appropriate aspects of the
analysis for organic chemicals in the Love Canal Monitoring Study were
considered. NBS has not always answered each subquestion but has used each
group of questions as guidance for its review. All questions are included for
completeness.
I. Goals and Objectives for the Acquisition of Organic Analytical Data from the
Love Canal Monitoring Study.
Scope: To evaluate the consistency of the goals and objectives for the
acquisition of organic analytical data with existing state-of-the-art
methodology.
A. What were the goals and objectives of the Organic Analysis Project? Has
EPA addressed in the written documentation whether or not these goals
were met?
The goals and objectives of the Love Canal Monitoring Study were presented
on Page 1 of the draft report reviewed by NBS. As stated therein, these
were:
1. Determine the current extent and degree of chemical contamination in
the area defined by the emergency declaration order.
2. Assess the near-term and long-term implications of groundwater
contamination in the general vicinity of Love Canal.
3. Provide an assessment (from an environmental contamination
perspective) of the habitability of residences included in the
emergency declaration order.
These goals were defined in somewhat more detail on Pages 53 and 54 of the
draft report as follows:
11
-------
1. To characterize in each environmental mediun the incremental extent
and degree of chemical contamination in the Declaration Area
directly attributable to Love Canal.
2. To determine potential temporal variability in contamination and
infer the causal mechanisms (for example, changes in climate)
influencing the observed contamination patterns.
3. To determine if swales, sewer lines, and other geological features
(for example, sandy soil deposits in the form of sand lenses) had a
significant effect on the migration of toxic substances from the
canal.
4. To determine the presence and direction of ground-water flow in the
area, and evaluate the effectiveness of the remedial construction
performed at Love Canal.
5. To investigate the use of locally available biological systems as
potential indicators of toxic substances present in the environment.
6. To obtain integrated multimedia measurements of environmental
contamination.
7. To provide from an environmental perspective an assessment of the
relative habitability of Declaration Area residences and the short-
term and long-term implications of observed environmental
contamination.
In the section on implementation beginning on Page 54, immediately following
the statement of the objectives of the study, it is stated:
"The EPA studies were initiated by first identifying the data
requirements of the overall objectives and then designing data
collection mechanisms appropriate for such activities."
These identified data requirements have neither been described further nor
have they been presented in the report beyond the non-specific statements on
Page 61:
"The common objective of these plans was to collect and analyze a
statistically adequate number of samples to characterize accurately
Declaration Area contamination caused by Love Canal, and to minimize the
effects and uncertainties associated with the constrained sampling
period."
and Page 62:
"...a primary goal of qualitative accuracy for organic analyses (that
is, correct identification of detected substances) was established."
12
-------
These statements do not address in quantitative terms the data requirements
of the goals and objectives of the study. For example, the data
requirements could have been made more quantitative by defining the
concentrations of targeted compounds which had to be detected and/or
quantified in order to compare levels of contamination in the Declaration
Area with levels known to constitute an environmental hazard and/or with
levels currently found in U.S. cities.
On Page 70 of the draft report, statements concerning the expected
performance of analytical methodology are made:
"Precisions better than 50 percent RSO [relative standard deviation]
were expected in water and air; precisions better than 100 percent were
expected in other media."
"Furthermore, it is also known that in water, the minimum method
quantitation limits expected for the methods used are in the range of 1
to 10 micrograms per 1-iter (parts per billion) Quantitation limits
below these values were neither required nor expected of the analytical
subcontractors, except as noted previously for 2,3,7,8-TCDD and certain
pesticides."
With the exception of the 2,3,7,8-TCDD, these were expectations rather than
data requirements. Without more quantitative specificity, it is difficult
to see how EPA could design a monitoring program to meet its requirements.
In the presentation of its conclusions, EPA does not address explicitly how
limitations in the analytical data affected the conclusions. Whether or
not this is a direct consequence of the absence of quantitative precision
and accuracy goals cannot be determined. Nevertheless, 'those who wish to
use the Love Canal data or wish to interpret the conclusions of the Love
Canal Monitoring Study must read and understand the entire report in detail,
including the Appendices, to gain even a qualitative sense of the
limitations of the data gathered in the study.
The "primary goal of qualitative accuracy" is partially addressed in the GC-
MS audit performed by EPA and is discussed in more detail in Section VII of
this Appendix.
B. Would the approach of selecting a list of targeted compounds in any way
hinder EPA from observing, identifying, or quantifying significant
quantities of other compounds? Should EPA review all of the GC-MS tapes
to determine if significant chemicals might have been missed.?
The approach of selecting a list of targeted compounds should not have
hindered EPA from observing, identifying, or quantifying other compounds
which are amenable to GC-MS. Furthermore, this approach is appropriate to
the study and the list is reasonable. By restricting the analytical
13
-------
methodology mainly to GC-MS, those compounds which are either too polar or
too nonvolatile to analyze without derivatization would not be detected. To
derive the list of targeted compounds, EPA selected from a list of chemicals
which were known to have been dumped in the Love Canal, chemicals found in
previous studies of the Love Canal, and certain chemicals generally
recognized as toxic, associated with industrial production, or observed in
other studies as environmental contaminants—the so-called "priority
pollutants". EPA should state in its report which compounds were selected
from which lists. The list of targeted compounds was not all-inclusive.
For example, a previous study of the Love Canal area (J. Barkley, et al.,
Biomed. Mass Spectrom. 7(4), 139-146 (1980)) cites estimated levels of
1,1,1-trichloroethane and trichloroethylene in air inside homes in the Love
Canal area, but these compounds were not in the list of targeted air
compounds in Table 1-2, Page 1-4 of the draft report.
All laboratories were instructed to identify the 20 most abundant substances
beyond" those appearing on the target list. The audit of the GC-MS computer
records (discussed in Section VII below) indicated that the requirement for
identifying non-targeted compounds was generally not met. Further review of
the GC-MS computer records for non-targeted compounds missed in the initital
analyses would be useful only if it were to become necessary to evaluate the
significance of these compounds.
In the description of the development of the list of targeted compounds, it
was noted:
"...the intentional inclusion of specific substances on the target list
that were known to serve as effective and efficient tracers of
subsurface migration of leachate was designed to permit a comprehensive
determination and assessment of migration patterns from the canal
source."
Nowhere in the report are these compounds identified or justified as
efficient tracers, nor were any results presented which give the reader of
the report an indication of whether or not these efficient tracers migrated
through, the subsurface area in the Love Canal region.
C. Were the accuracy and precision goals for the organic data clearly
established at the outset and have the criteria for these goals been
clearly explained?
Section 3.3.2, Pages 68 to 70, of the draft report contains a description of
the precision and accuracy goals for the study. This section deals
primarily with the accuracy of compound identification and provides some
references to the precision which might be expected for quantitative
analyses such as those conducted by EPA in its Love Canal study. Precisions
of 2 to 13 percent relative standard deviation (RSD) for water analyses, 50
percent for air analyses and 100 percent for other environmental media were
14
-------
quoted in the draft report as generally being acceptable. These estimates
of precision did not serve as limits to be used in a quality control
program, but as guidelines for what could be expected from a study such as
this. In its analyses of the data acquired from the Love Canal samples, EPA
accepted, in some cases, data which had uncertainties as high as one or two
orders of magnitude (cf. Table II1-5, Page II1-28 of the draft report).
Criteria (reviewed in Sections II and III below) were established for the
acceptance or rejection of data and were based on laboratory analyses of
quality assurance samples. However accuracy and precision goals for the
organic data were not established at the outset. The draft report includes
statements of precision and accuracy which varied from one procedure to
another and from one laboratory to another.
D. If inadequacies in the precision and accuracy goals are identified in
answering the above questions, will these inadequacies affect our
judgment of the quality of the organic analytical data?
The establishment of precision and accuracy goals, the training of
laboratory personnel to meet those goals, and the use of quality control
charts based on those predetermined goals would have improved data quality.
Without precision and accuracy goals which have been formulated on the basis
of the overall goals of the study, it is difficult to determine if these
overall goals were actually met. Moreover, the lack of clarity in relating
the precision and accuracy estimates of the acquired data to the overall
conclusions of the report will make decisions based on those conclusions
difficult to justify. This lack of clarity is illustrated by the following
excerpt from the discussion of air contamination on Page 183 of the draft
report:
"Even though maximum concentration levels are often of considerable
interest to individuals, because in some way they may be thought to
represent 'worst case' estimates of environmental contamination,
problems of interpretation exist. The reason for this is because both
the occurrence and reliability of the obtained maximum values may be
plagued by measurement problems. To illustrate this point, it is often
the case that maximum concentrations are reported by only one analytical
laboratory and on one particular date, whereas other analytical
laboratories do not report concentration levels anywhere near such
maxima (and in some cases do not even report concentration levels above
the minimum detection value)."
Statements such as these allude, in a general way, to the lack of precision
and accuracy in the data but are not supported by quantitative information
in the draft report. They give little guidance to the user of the data on
how the lack of precision and accuracy affects conclusions that might be
drawn.
15
-------
II. Monitoring Protocols
Scope: To evaluate the consistency of the monitoring protocols with the
goals and objectives of the Organic Analysis Project and with existing
state-of-the-art methodology.
Have the sources of the protocols used by EPA been identified? Are
these protocols well documented? Had they been evaluated and did EPA
consider these evaluations in their selection? Are there additional
references EPA should have considered in the selection of these
protocols?
The analytical methodology used in the study is described in the document
"Quality Assurance Plan: LC-1-619-206, Appendix B" by the prime contractor
for the EPA study, the GCA Corporation.
Those methods which were used in the water analyses, Methods 608, 624, and
625, had been, prior to the initiation of the study, published in the
Federal Register. These methods were in the process of being validated by
inter laboratory tests, and such validation may now be complete. To the best
of our knowledge, such validation was not available at the outset of the
Love Canal Monitoring Study and hence could not be used as a guide for
quality control. When Methods 608, 624, and 625 were published in the
Federal Register in December 1979, EPA asked for and received comments on
the methods. These comments have been considered by EPA, were found in some
cases to be germane, and if incorporated would have led.to improvement in
the clarity of presentation or in the actual performance of the methods.
These modifications were not incorporated into the methods used in the Love
Canal Monitoring Study. Nevertheless, Methods 608, 624, and 625 have been
widely used and in our opinion represent a reasonable choice for the study.
The methods used for the air analyses are well referenced in the draft
report and are described in detail in the aforementioned Appendix B of the
Love Canal Quality Assurance Plan. To our knowledge these methods have not
been formally validated.
The sample preparation steps in the methods used for soil, sediments and
biota samples were experimental in nature and had to be developed as part of
the Love Canal Monitoring Study. They were, consequently, unevaluated
beyond the evaluation appearing in the draft report. A number of the
protocols which, taken together, comprised the methods for the soil,
sediments and biota were still optional at the time Appendix B of the
Quality Assurance Plan was released. Although the methodology was fixed by
the end of the study, it is unclear whether all the data acquired for the
soil, sediments and biota were obtained using the methodology described in
the draft report.
16
-------
There are no fundamentally different analytical techniques for the detection
and quantisation of organic compounds which would have been more suitable
for the study as planned.
B. Does EPA discuss the precision and accuracy to be expected from the
chosen protocols? Does EPA document the minimum detectable limits for
their targeted compounds? Are the precision, accuracy and minimum
detectable limits claimed by EPA reasonable and consistent with our
knowledge of or experience with the methodology? Are these claims
supported by reference to available literature?
Since none of the methods used in the study had been subjected to a formal
inter!aboratory validation prior to the study, expectations concerning the
precision, accuracy and minimum detection limits had to be drawn from the
reported experience of analysts who had used the various methods.
For air methods, EPA established limits of quantitation for targeted
compounds; limits of detection were to be determined by each laboratory.
For water methods, the limits of quantitation for targeted compounds were to
be determined by each laboratory; expected limits of detection were reported
in the Federal Register description of the methods.
For many of the protocols, expectations of precision, accuracy, and
detection limits were unavailable and were to be determined during the
course of the study from the quality assurance data. The protocols did not,
in all cases, explain how such determinations were to be made. Moreover,
such guidance as was given varied from one environmental medium to another.
For example, precision was to be determined for air monitoring methods from
data on duplicate measurements and sample splits and the accuracy from data
on spiked TENAX tubes and polyurethane foam plugs, while for total organic
carbon determinations in water, the methods called for the analysis of
blanks and replicates but had no requirement for the determination of
accuracy.
Not all claims on expected precision, accuracy, and detection limits were
supported by appropriate references. The requirement that they be
determined by the individual laboratories during the course of the study is
correct and always necessary.
C. Can we identify targeted or non-targeted compounds which the monitoring
protocols would be likely to miss? Can we recommend ways in which EPA
could justify the completeness of its identification of compounds which
it has not presented in the written documentation supplied to NBS?
17
-------
In general, the protocols were capable of detecting all targeted compounds.
Because of the composition of TENAX, and the allowed lapse time between
collection and analysis of samples, the use of TENAX for the analysis of
benzene and toluene in air is inappropriate at the levels encountered. This
problem has been recognized by EPA and is commented upon in its draft
report.
The protocols are likely to miss certain non-targeted compounds! In
particular, compounds which are too polar or not sufficiently volatile will
not be detectable using GC-MS.
D. Are the protocols as written complete? Do they omit steps or
precautions that would be expected'to affect the identification or
quantification of certain organic compounds? Were all of the following
steps clearly described in the protocols, were they complete?
1. Sample Collection Techniques
2. Sample Storage and Transportation
3. Chemical and Physical Manipulation of the Samples Prior to Analysis
4. Calibration
5. Analysis
6. Records
The protocols were sufficiently compTete in the sense that they could be
followed by competent analytical chemists with experience in organic
analysis. The protocols for the sample collection, for storage and
transportation, and for record keeping were satisfactory. Some ambiguities
in the remaining steps could have resulted in actual differences in how the
protocols were executed. Such differences could introduce difficulties when
comparing data from the various laboratories.
Instances of incompleteness or ambiguity in the protocols are listed below.
(1) The analytical procedure for volatile, organics on TENAX allows for the
use of an internal standard, dependent on the availability of the
apparatus needed to perform spiking (Quality Assurance Plan, Appendix B,
Section 1.2.2C). This apparatus is not defined in the procedure and the
addition of an internal standard is optional. Comparability of data
between laboratories requires comparable methods of internal standard
addition.
(2) Mass spectrometric confirmation of gas chromatography-electron capture
(GC-EC) and high performance liquid chromatography (HPLC) measurements
for pesticides collected on polyurethane foam plugs are to be conducted
on composited or selected samples. The choice of these samples is not
18
-------
specified. Furthermore, Subsection F, Section 1.3.1, Appendix B of the
Quality Assurance Plan of this analytical procedure directs the analyst
to "confirm results as required by combined GC-MS" but does not specify
what constitutes "as required."
(3) Tine analyst is given the option of choosing from four methods, involving
different extraction schemes and packed or capillary columns for the
analysis of semi-volatile organic compounds in water.
(4) The analytical procedure for the determination of total organic carbon
in water samples req-uires a chemical preservation step to be performed
in the laboratory within two hours of collection. The QA procedure
submitted by ORB Associates specifies that samples to be analyzed for
total organic carbon and total organic halogen content are to be
transferred to the'GCA sample bank within four hours of collection.
(5) The analytical method for the determination of pesticides in sediments
calls for the use of a procedure entitled "Sample Preparation and
Analysis of Bottom Sediments" with the exception that part 5 of the
method has been modified and part 4 of the method is to be disregarded. '
There is no statement of what is to be done in place of part 4. (Part 4
is the section dealing with sample preparation and extraction
procedures.) One could assume that Method 608 is the replacement, but
this is not specified. In this same bottom sediments procedure, the gas
chromatographic step calls for proceeding as described in Section 11.A.
There is a Section 11 of Method 608 which deals with gas chromatography
but there is no subsection 11.A.
E. Do the protocols adequately address the prevention of sample
contamination from impure reagents, the sampling process, contaminated
sample containers, etc.?
With few exceptions, the protocols dealt with the problems of analytical
contamination in a reasonable manner. These exceptions included the
problems of benzene and toluene contamination of TENAX mentioned in II.C
above and problems with vagueness or inconsistency in instructions on how to
correct for common sources of analytical contamination. For example, the
quality assurance procedure for the analysis of semi-volatile compounds in
soil and sediment required that the method blank not show any signal
corresponding to compounds on the list of targeted compounds, but if common
laboratory contaminants such as phthalate esters (which were on the targeted
list) are found, the reported values for the corresponding compounds in the
environmental samples must be reduced by the amount found in the blank.
This instruction is vague in that levels of contamination for which method
blank corrections are valid are not given.
19
-------
F. In the documentation provided by EPA, is the intended use and purpose of
the data from the control samples described? Is this intended use
reasonable? Could it have been expected to achieve its intended
purpose?
The control samples, that is, samples from the Control Area, were designed
for comparison of levels of contamination in the Declaration Area with
levels outside that area. (As such, they were not intended as analytical
controls and should not be interpreted as such. They are, within the
context of a review of the analytical program, to be treated as
environmental samples, indistinguishable from any other environmental
samples.) Their use and purpose are described and the intended use is
reasonable. Determination of whether or not they were sufficient in number
and kind to have achieved their intended purpose is beyond the charge from
EPA to NBS.
G. Are there any means, not addressed above but available to EPA at the
beginning of the study, by which- EPA could have improved upon the
selection, description or justification for use of the selected
monitoring protocols?
Improvements in the methodology have been addressed in the sections above.
20
-------
III. Quality Assurance Protocols
Scope: To evaluate the consistency of the quality assurance protocols with
the accuracy and precision objectives of the study and to evaluate the
adequacy of these protocols to maintain quality control over the reported
A a+ 3
data.
A. Was the quality assurance plan sufficient to maintain quality control?
If not, what were the major flaws in the plan? Were there any necessary
elements missing?
A major feature of the Love Canal Monitoring Study was the quality assurance
program for the study. The prime contractor, GCA, prepared an extensive
quality assurance plan, and all but one of the participating subcontractors
prepared quality assurance plans in"response to the GCA plan. These plans
contained a large variety of quality assurance measures including document
and report control, laboratory control standards, replicate analyses,
internal standards, calibration checks, and surrogate standard additions.
Quality assurance must be recognized as consisting of two factors: quality
control and quality assessment. The quality control aspects of the plans
were generally adequate, especially those portions which dealt with the
collection and handling of samples and maintenance of records, including
chain of custody. The Quality Assurance Plan was deficient in the manner in
which it addressed quality assessment. The interrelation of individual
items and the manner in which quality assurance data was to be analyzed and
used were not clearly described.
Protocols varied between media. Examples of this variability from Section 7
of the Quality Assurance Plan will be described in the following paragraphs.
(1) For air samples collected on TENAX cartridges, the use of an internal
standard was optional and this standard when used was to be added to the
TENAX cartridge. Internal standards should have been mandatory. For
air samples collected on polyurethane foam plugs, an internal standard
was mandatory, but was added to the sample extract. Moreover, the
procedure stated "if a suitable compound can be determined a single
internal standard will be added to each sample extract." The standard
should have been specified and should have been added as early as
possible in the analytical procedure. For the determination of dioxin
(TCDO) in air samples, chlorine-37 labeled TCDD was added to each sample
as an internal standard. This was appropriate.
(2) For those air samples collected on polyurethane foam plugs, a three
point calibration curve of all compounds was required each day; a
single-concentration, mixed standard was required after every fifth
21
-------
sample; and one spiked foam plug was to be analyzed.per analysis batch
per day. For those air samples collected on TENAX, only a single
calibration check was required per eight hours of analysis. No
discussion is presented describing why the two protocols were so
different.
(3) For the determination of semi-volatile organic compounds in water,
deuterated internal standards were required (but the draft report
contains no mention of their use). For the determination of volatile
organic compounds in water, internal standards (non-deuterated) were
optional. For the determination of pesticides in water, the analyst is
referred to Section 7 of Method 608 which recommends the use of
fortified samples. The use of internal standards should have been
required.
(4) For the determination of volatile and semi-volatile organic compounds in
water, surrogate compound additions were required. Recoveries of
surrogate compounds were to fall within specified ranges. (The ranges
described in the draft report were substantially different from the
ranges in the Quality Assurance Plan.) For pesticides, the use of
fortified samples was recommended (by reference to Method 608).
(5) For the determination of pesticides, volatile organic compounds, and
semi-volatile organic-compounds in water, laboratory control standards
were required. Control limits were established for the recoveries of
the laboratory, control standards. For pesticides, the control limits
were specified as "±2 standard deviations." For volatile and semi-
volatile organic compounds, the control limits were specified as ranges
of "percent of true value" for individual compounds.
(6) For the analysis of pesticides in soils and house dust, a recovery check
on a spiked sample using a list of specified pesticides was required.
The same recovery check was not part of the sediment procedure and •
should have been.
(7) For the analysis of total organic carbon and total organic halogen
content in water, specifications were presented for method blanks and
replicate analyses, but none for a laboratory control standard, an
internal standard or a surrogate standard; hence, there was no accuracy
check.
In addition, control limits associated with many of the quality assurance
procedures were to be determined during the course of the study. These
limits, if they were determined, have not been reported. Performance as
determined from calibration check samples and performance evaluation samples
was to be evaluated as the study progressed. This evaluation process was
not described in the Quality Assurance Plan.
The above remarks indicate the quality assurance plans were quite extensive
and detailed, but were lacking in consistency from one medium to the next
and lacking in detail on how the quality assessment was to be performed.
22
-------
B. Were criteria established for the rejection or acceptance of data? Were
these criteria reasonable, and if not, how could they have been
improved?
Two levels of review for rejection or acceptance of data were included in
the Love Canal Monitoring Study. The Quality Assurance Plan called for
contract laboratories to monitor their own performance and to invalidate
data according to certain criteria. Laboratory performance was to be
reviewed by the prime contractor and by EPA. Criteria for acceptance or
rejection by the prime contractor or by EPA of data already validated by the
contract laboratories were not part of the Quality Assurance Plan, but were
described in the draft report. Furthermore, these criteria were not
consistent with those in the Quality Assurance Plan, as will be described in
the following paragraph. In contrast to the statement on page 64 of the
draft report, the Quality Assurance Plan did not require the contract
laboratories to maintain and submit quality control charts.
The criteria used by EPA for acceptance or rejection of data differed from
the criteria specified for the contract laboratories in the Quality
Assurance Plan. For example, Table 7.2.1 of the Quality Assurance Plan
specifies control limits for the recovery of surrogate compounds in water of
81 to 100 percent for fluorobenzene, 88 to 118 percent for p-bromo-
fluorobenzene, 18 to 58 percent for 2-fluorophenol, 33 to 8"8~ percent for 1-
fluoronaphtha.lene, and 34 to 98 percent for 4,4' -dibromooctafluoro-
biphenyl. Table III-5, Page 111-28 of the draft report indicates there were
no upper control limits for the recovery of these five surrogates and that
EPA invalidated all analyses for which more than one' of the surrogates was
lower than the lower control limits of 68 percent, 60 percent, 1 percent,
2.8 percent, and 8.3 percent respectively for the five compounds mentioned
above.
The EPA criteria for validation of data were less stringent, in practice,
than the criteria in the Quality Assurance Plan. Whether this indicates the
plans were unrealistic or the implementation of the Quality Assurance Plan
was inadequate cannot be determined..
C. Were the quality assurance plans of the individual contractors uniform
and consistent with each other and with the overall Quality Assurance
Plan?
The quality assurance plans of the individual contractors followed an
essentially consistent outline which was appropriate to the study. The
individual plans, however, differed in detail.
Not all quality assurance plans were available or complete when NBS began
its review. In particular, to our knowledge, no plan was filed by Research
23
-------
Triangle Institute (RTI) which was one of the three QA contractors and which
had responsibility for the spiking of the TENAX cartridges used in the QA
program for the analyses of air samples. The QA plans from TRW, one of the
laboratories performing analyses for organic compounds in water, and Wright
State University, the only laboratory performing dioxin analyses, were
received late by NBS and it is not clear whether these plans had received
approval from EPA prior to the initiation of the Love Canal Monitoring
Study. The plans received from CompuChem/Mead and Accurex Corp. contained
notations that additional information had been requested. This information
was received by NBS well after completion of the chemical analyses. It is
not clear whether these additions were included in their QA plans or were
descriptions of practices which had been followed. The plans from PJB and
Advanced Environmental Systems were lacking in a number of elements required
by the overall QA plan of the prime contractor, but no indication was given
whether revised plans or additional information had been requested. The QA
plan for Southwest Research Institute provided no information on the
analytical procedure or the quality control procedure for the preparation of
the polyurethane foam plugs which were used in the air analyses. The plan
from Battelle Columbus Laboratories lacked a summary of analytical methods,
equipment maintenance^ and trouble-shooting procedures. Finally, none of
the QA plans addressed'the identification of the 20 most abundant, non-
targeted compounds.
The quality assurance plans should always be signed and dated by the
submitting laboratory and the approving authority before work begins. This
was not done consistently in the Love Canal Monitoring Study.
D. Was the number, nature and frequency of quality control samples or
performance audit samples appropriate to achieve monitoring goals, to
take corrective action, and/or to maintain quality control?
The number, nature, and frequency of quality control samples as required by
the Quality Assurance Plan should have been adequate to maintain quality
control. Evidence exists, however, that quality control was not maintained
in some cases. The Quality Assurance Plan did not anticipate the large
number of samples which would be found to contain no levels of contamination
above the limits of detection. When a large number of samples are expected
to contain concentrations below the detection limit, sample splits and
duplicate samples are of limited usefulness. These difficulties will be
discussed further in Section V dealing with implementation.
24
-------
E. Did the plan adequately and clearly prescribe corrective action?
Instructions from EPA and the prime contractor to the contract laboratories
for corrective action (CA) were clearly and adequately described in the
overall Quality Assurance Plan with two exceptions: (1) The overall plan
did not require contract laboratories to maintain and submit quality control
charts. (2) There existed ambiguity in the description of the mechanism for
reporting corrective action to the prime contractor. Figure 11-1, Page 11-3
of the Quality Assurance Plan, is an example of a form which was to be
completed and submitted describing corrective action taken whenever loss of
quality control was identified. On Page 11-4, however, it was stated
"subcontractors are not required to follow this system exactly or to use the
pictured CA form." For such a large and complex study conducted in such a
short period of time, a uniform system should have been required. On the
basis of the written, corrective-action procedures appearing in the
individual subcontractor QA plans, it is apparent that corrective-action
procedures were not consistently adequate. For example, IIT stated
corrective action (unspecified) would be implemented as required but did not
state that it would report the need for corrective action to GCA. PEDCO
also gave only a vague statement of the nature of corrective-action
procedures and the documentation of corrective action.
25
-------
IV. Performance of the Monitoring Program
Scope: To evaluate, from the information provided by EPA, whether or not
the implementation of the monitoring program met the goals and objectives of
the Organic Analysis Project and were consistent with proper implementation
of the monitoring protocols and with commonly accepted good laboratory
practices.
A. What evidence does EPA present to support claims of proper
implementation of the monitoring protocols? Is the evidence presented
that the contract laboratories were capable of performing their assigned
tasks convincing?
Based on the material reviewed, the sample collection, storage and
transportation, document control/chain of custody procedures, and data
management were generally implemented according to plan. Field and quality
assurance samples were distributed according to plan. An acceptable
fraction of the samples was analyzed. Some difficulties with the analysis
of the samples were encountered and these difficulties are discussed in the
remainder of this Section and in Section V below.
B. Is there any reason to believe that significant concentrations of
targeted or non-targeted compounds were missed in the analyses of the
samples? What additional information should EPA provide to justify
claims that compounds were correctly identified and that significant
concentrations of targeted or non-targeted compounds were not missed?
The GC-MS audit of computer records indicated that most non-targeted
compounds were probably missed. Pages VI-12 to 13 of the draft report
state:
"In most of the 22 samples [of th_e 80 samples audited] containing non-
target compounds, the audit laboratory reported finding 1 or 2 compounds
while the analytical laboratory reported none. There were 5 samples
audited where 6 to 20 compounds were reported by the audit laboratory
but none were reported by the analytical laboratory.,.."
The audit also indicated inconsistency between laboratories in the
determination of the presence of targeted compounds when the concentration
of those compounds was near the detection limit.
Pages II1-14 to II1-16 contain a discussion of the limits of detection and
quantitation for organic compounds in water. Table III-l, Page 111-15,
contains a summary of the limits of quantitation, and the accompanying
discussion states:
26
-------
"There was considerable variance among the analytical
laboratories in method quantitation limits for a given analyte,
and the data suggest that some laboratories were not operating
consistently at the state-of-the-art possible with the methods.
This is neither unusual nor unexpected."
Comparison of the numbers in Table III-l with:
(a) Table II-l, Page II-5 which contains the range of median values of
compounds found in drinking and surface waters in various industrial
locations,
(b) Table 11-4, Page 11-8 which contains values of volatile organic
compounds found in drinking water in a number of cities, and
(c) Table II-9, Page 11-34 which contains values of selected volatile
organic compounds found in national surveys of water supplies,
indicates that the limits of quantitation for volatile organics, determined
by three of the five laboratories performing water analyses in the Love
Canal Monitoring Study, were an order of magnitude higher than values
reported by laboratories participating in the studies quoted in Appendix II
of the EPA draft report. Also, for the volatile organics, the limits of
quantitation for the five contract laboratories were 1, 2, 4, 8 and 9 times
higher than those of EPA's own laboratory in Cincinnati. For the
semivolatile organics, little data exist against which to compare the
results of this study, but the limits of quantitation for the four contract
laboratories were 1, 8, 9, and 14 times higher than those of the EPA
laboratory. In its draft report, EPA recognizes the variability in limits
of quantitation reported by the contract laboratories. This variability
will make the Love Canal monitoring data difficult to interpret.
The limits of quantitation set for the air monitoring laboratories appear to
be adequate for quantifying levels of compounds which have been reported for
some U.S. cities.
C. \Is there any reason to believe that significant contamination of the
samples from impure reagents, the sampling process, or contaminated
tools, or containers may have occurred? Does EPA present evidence that
this has not occurred? Is it convincing and if not, what additional
documentation is needed?
EPA has presented evidence that artifactual contamination, that is,
contamination of the samples from impure reagents, the sampling process, or
from contaminated tools or containers, was a problem for certain samples and
for identified analytes.
(1) For those air samples collected on TENAX, significant blank problems
associated with benzene, toluene, and 1,1,2,2-tetrachloroethylene were
encountered. In Table V-4, Page V-13 of the draft report, EPA indicates
27
-------
the frequency and concentrations of these compounds found on blank TENAX
cartridges. The accompanying text states:
"To be relatively certain that an obtained single value
was not due to blank contamination, the field concentration
should be greater than three standard deviations above the mean
values reported in Table V-4 for these three compounds."
From the data in Table V-4, "three standard deviations above the mean
values" were 7, 29, and 22 micrograms/cubic meter for benzene, toluene
and 1,1,2,2-tetrachloroethylene respectively. According to the draft
report analytical values obtained for these compounds below these levels
would be suspect. (Averages of several analyses would have somewhat
lower limits depending on the number of samples included in the
analyses.) In Section 4.2.6, Pages 167 to 195 of the draft report, EPA
describes the results of the air monitoring program. The only
significant differences between the Declaration, Canal and Control Areas
are the concentrations -of o-chlorotoluene, o-dichlorobenzene, and
chlorobenzene. Yet, for tFe remainder of tHe discussion, only benzene,
toluene and 1,1,2,2-tetrachloroethylene are discussed. We question the
appropriateness of using these three compounds, for which blank problems
call into question the reliability of their determination, as "typical"
examples. The median concentration plots in Figures 54, 55 and 56 on
Pages 188 to 190 were chosen1to illustrate the lack of pattern in the
distribution of these compounds, but the maximum values for these
medians are below the above-mentioned limits on TENAX blanks.
(2) On Page II1-29, the contamination of bedrock B wells by use of ordinary
hydrant water during the drilling process is described. This led to the
exclusion of "28 ground-water, Method 624 samples" from the validated
data base. Chloroform was identified as the contaminant and its source
was attributed by EPA to improper purging of the wells, but no
quantitative data were presented to support this claim. A total of 340
bedrock well samples were included in the validated sample set. The
effect of excluding the 28 samples on the conclusions drawn by EPA is
not discussed in the draft report.
(3) On Pages II1-29 to 31, the contamination of water samples by methylene
chloride, bis(2-ethylhexyl) phthalate, and dibutyl phthalate is
described. This contamination resulted from the presence of these
chemicals in the atmosphere of the analytical laboratories and from
plastic containment materials. As a result, all reports of these three
analytes were removed from the validated data set. The same problem was
encountered with the sediment and soil analyses.
With the exception of item (1) above, the incidence of artifactual
contamination should present no problem so long as conclusions drawn
concerning Love Canal are based on patterns observed for groups of chemicals
and not for individual substances. The above examples of contamination were
discovered because so many of the samples which were collected had no
concentrations of analytes above the detection limit. For the same reason,
28
-------
it is unlikely there are other cases of widespread, but as yet undetected,
art ifactual contamination by targeted compounds.
0. Is there any evidence to indicate that participating laboratories were
incapable of adequately following the protocols or did not adequately
follow the protocols? Is there any evidence of carelessness in
execution of any of the steps from sample collection to analysis?
As part of the quality assurance program, site visits were made to the
participating laboratories and reports of these visits were reviewed by NBS.
The reports indicated that most laboratories were having some difficulties
at the outset, and a few were having great difficulties. In particular,
some of the laboratories were experiencing difficulties in adjusting to the
use of capilVary columns. No follow-up visits were made, or if they were,
reports of such follow-up visits were not provided. It is not known if
appropriate corrective action was taken in all cases.
The data in Table 6 on Page 110 of the draft report are inconsistent with
the number of targeted compounds which were actually sought in water,
sediment and soil samples. Table 6 contains a sunmary of the frequency of
detection of substances in the validated Love Canal samples. The table also
contains the "number of substances (analytes) measured" and the number of
samples which were analyzed. Dividing the number of analytes by the number
of samples gives the average number of analytes per sample. These average
numbers are, for the air analyses, consistent with the number of targeted
analytes listed in Table 1-2 on Page 1-4. For the water and sediment
analyses, these average numbers range from 20 to 30 for almost all of the
samples analyzed, whereas Table 1-1 on Pages 1-1 and 1-2 indicates that 135
analytes were to be sought. Similarly, the average number of analytes per
sample of drinking water obtained from Table 12, Page 196, is 25 to 28.
This inconsistency should be corrected.
Performance qualification of all the laboratories prior to initiation of the
study, as was done for air monitoring subcontractors, would have diminished
the number of problems encountered but could have delayed the initiation of
the study. This has been acknowledged by EPA.
E. Does EPA, in its report, adequately document the performance of the
contract laboratories? If not, what additional documentation should EPA
provide?
Appendix III of the draft report does not adequately document the
performance of the water monitoring laboratories. Four types of performance
checks were used. Laboratories analyzed laboratory control standards (LCS)
and quality control (QC) samples daily, performance evaluation (PE) samples
monthly and surrogates for every analysis. The laboratory control standards
will be discussed in Section V.B. The QC and PE samples were concentrates
of selected compounds in organic solvents which were added to reagent water
29
-------
and analyzed. The identities of the compounds in the QC samples may or may
not have been known to the performing laboratories; the concentrations were
unknown. The identities and concentrations of compounds in the PE samples
were unknown to the performing laboratories. The surrogates were known
amounts of known compounds added to each sample. No summary of laboratory
performance on QC samples is given in the draft report and should be. The
performance on QC and PE samples was not used to determine precision and
accuracy for the methods but should have been used, at least, to confirm the
estimates of precision and accuracy.
Performance on PE samples is summarized in Table 111-4 on Page 111-22 of the
draft report as percentages of analyses submitted by the contract
laboratories which were judged "acceptable." The definition of "acceptable"
is not given and should be. The percentage of all analyses of PE samples
found to be unacceptable cannot be determined because the numbers of
analyses corresponding to the entries in Table III-4 are not given. Thus,
it is not possible to compare the performance on PE samples with the overall
performance, i.e., the percentage of Love Canal data which was validated.
The performance on PE samples should be consistent with the data validation
process and this consistency should be documented and discussed. As Table
III-4 now stands, it appears that the number of PE sample analyses judged as
unacceptable (ranging from 0 to 75%} is considerably larger than the number
of samples invalidated on the basis of surrogate recoveries (17 samples).
This apparent inconsistency should be clarified.
The documentation of the performance of the air monitoring laboratories
appearing in Appendix V of the draft report is generally acceptable though
Table V-5 on Page V-20 would be more readily interpretable if it contained
the ranges of the concentrations of the compounds to which the standard
deviations pertained. Inconsistencies exist in the main body of the draft
report in the description of the performance of the air monitoring
laboratories. On Page 174, is the statement:
"The analytical accuracy...for the TENAX samples was within ±10 percent
(as indicated by analytical recovery of substances purposely introduced
into samples at concentration levels known only to EPA)."
On Page 184 is the statement:
"As can be seen from the results displayed in Figures 52 and 53,
variability in concentration levels were observed over time. As noted
above, however, much of the variability observed in the sample results
could be accounted for by the systematic differences in analytical
laboratories and by measurement error." -
The variability in Figures 52 and 53 ranges from 0 to 40 vg/m3, well beyond
the ±10% "accuracy" claimed on Page 174. Moreover, there is no quantitative
justification given of the statement "...could be accounted for by the
systematic differences in analytical laboratories and by measurement error."
These statements, as they stand, appear to be in conflict in their
description of the performance of the air monitoring laboratories.
30
-------
The documentation of laboratory performance would be improved if the
performance of individual laboratories on all of the different classes of
quality assurance samples were summarized, tabulated, and compared with
criteria established for performance in the Quality Assurance Plan.
31
-------
V. Performance of the Quality Assurance Program
Scope: To evaluate from the information provided by EPA whether or not the
implementation of the quality assurance program was adequate to maintain
quality control of the Organic Analysis Project and whether or not accuracy
and precision claims, if given, have been convincingly justified.
A. Were the quality assurance samples sufficient in kind, number and
reliability to test adequately the performance of the participating
laboratories? What sources of measurement error, if any, might EPA have
missed or have had the most difficulty in evaluating?
In general, we believe the quality assurance samples were sufficient in
kind, number and reliability to test adequately the performance of the
participating laboratories. They should also have been sufficient to
maintain quality control provided appropriate corrective action was taken.
The large number of samples containing no analytes above the limits of
quantisation limited the usefulness of replicates for evaluating the
comparability of analyses performed in different laboratories (e.g.,
estimating inter laboratory precision).
The measurement errors most difficult to evaluate are those associated with
the soil, sediment and biota samples. The quality control samples used for
the monitoring of these samples were water samples which do not duplicate
the matrix effects presented by soils, sediments and biota. Thus, not all
of the steps in the analytical methodology were tested by the quality
assurance program. This is acknowledged by EPA in its draft report.
Has the rejection or acceptance of data from a participating laboratory
been clearly explained-and justified? Can we suggest ways in which the
rejection of data might distort the representativeness of the data?
The rejection of data has not been clearly described and adequately
justified.
The fractions of samples rejected for various reasons in arriving at the
'final set of validated samples are summarized in Table 1 on Page 11 of the
draft report.
(1) In Table 1 EPA indicates 410 samples out of 2457 water samples which
were analyzed were rejected. The extensive discussion on the data
validation process for water samples on Pages 111-25 to 29 states 17 of
the Method 624 and 625 samples were rejected. Discussions for the
32
-------
other analytes are vague but indicate similarly small numbers of
rejected samples. Therefore, it appears about 300 samples were
rejected without a corresponding, satisfactory explanation in the draft
report.
(2) The percentage figures in the fourth and fifth columns of Table 1 do
not agree with the totals listed in the left three columns. Moreover,
they do not agree with the numbers of rejected samples appearing in the
appendices. For example, from Table 1, a comparison of samples
analyzed with samples validated for soil, sediments and biota give 27,
4, and 8 samples rejected, presumably on the basis of the validation
procedure. Appendix IV, Page IV-23 reports these numbers as 9, 7, and
7 respectively.
(3) The footnote to Table 1 should apply to the colunn headed "Percent
Other" and not "% Rejected by QA/QC."
Beyond these observations, the following difficulties with the
justifications of the rejection or acceptance of data were encountered:
(1) Page 111-24 contains a discussion of the errors which could result
from holding samples to be analyzed by Method 624 beyond the
maximum allowable holding time of 14 days. Experiments to determine
the effects of holding time were performed using samples containing 100
micrograms/liter concentrations of selected analytes. The report
should address and justify the significance of these experiments to
samples containing 10 micrograms/liter, that is, concentrations near
the limits of quantisation. •
(2) The discussion on data validation procedures (Pages II1-25 to 33) and
on estimates of data accuracy (Pages III-36 to 38) for the water
monitoring effort should contain a description of the percent
recoveries for the analytes studied. The reporting of data without
correction for the lack of complete recovery of particular analytes has
been justified on the basis that the analytical methods do not provide
a,procedure to correct for these losses. Recovery information is
critical for proper assessment of data resulting from the Love Canal
Study. If the mean recovery for a particular analyte were 50%, then,
on the average, all data for that particular analyte would tend to be
low by a factor of two. This could be important when comparing the
distributions of two different analytes. The only indication of such
recoveries is given in Table II1-5 on Page II1-28. These data were
derived by EMSC and are probably not representative of the performance
of the contract laboratories.
(3) The description of the rejection of the 28 ground-water, Method 624
samples from bedrock B wells should be clarified. How did the
Environmental Research Laboratory in Ada, Oklahoma determine these
wells were not purged adequately prior to sampling?
33
-------
(4) As discussed in Section III.8 above, criteria were established by EPA
for validation of data by the water monitoring laboratories based on
recoveries of surrogate compounds. EPA used the recoveries of the
same compounds but substantially wider acceptance limits for its
validation of sanple data which should have been rejected (as required
by the Quality Assurance Plan) by the contract laboratories. These
wider limits were based on a statistical analysis of the recoveries
measured by EMSC in 5% of the water, soil and sediments samples. The
reasons why wider acceptance limits were used are not discussed in the
draft report. A comparison between the number of samples, for which
the data would have been validated had the original control limits been
used, and the final number of validated samples would likely give
greater insight into the quality of the water data than the tabular
data presented in the draft report. As mentioned in Section III.E
above, laboratory performance on PE and QC samples should be consistent
with the sample validation procedure.
Similar, though not as serious, concerns exist for the air monitoring
data.
(5) Table III-6 on Page 111-35 of the draft report presents relative
standard deviations for replicate measurements of Method 624 and Metho'd
625 analytes in laboratory control standards (LCS). Table 7.2.1, Pages
7.2.2 to 7.2.4 of the Quality Assurance Plan lists control limits for
the analyses of laboratory control standards. If the analysis of the
laboratory control standard yielded values outside the control limits,
the laboratory was directed to check analytical procedures, analyze a
second reference sample and, if still out of control, invalidate
results. A comparison of Table II1-6 with the control limits for the
LCS's indicates the performance for certain of the laboratories was out
of control for a significant fraction of the time. For example, the
control limit for benzene (item 6 of Table 7.2.1 on Page 7.2.2) was 79
to 107 percent recovery,' a range of 15% around the average of 93 per-
cent recovery. The relative standard deviations of the recovery of
benzene for two of the six laboratories included in Table II1-6 were
well in excess of this ±15% control limit. Similar statements can be .
made about other compounds listed in Table III-6. EPA should, in its
report, explain this apparent discrepancy in the quality control
program.
(6) The quality assurance program for the air samples addressed the problem
of artifactual contamination of TENAX by benzene and toluene if samples
were analyzed more than 30 days after final cleaning of the TENAX. The
significance of such contamination could be better assessed if
experimental results were presented showing variability in background
levels of these compounds as a function of time after cleaning.
Finally, EPA has not used the full range of quality assurance data available
to it to justify its acceptance of data. The incomplete use of quality
assurance data is discussed in other sections of this Appendix.
34
-------
What conclusions has EPA drawn from the sample splits and the
intercomparison of data from single sampling sites as to the precision
of the data? Are such conclusions justified? If not, what conclusions
could be drawn and/or what justification should be presented to support
such conclusions?
EPA has drawn few conclusions concerning the precision of the data from the
sample splits and intercomparison of data from single sampling sites. This
is primarily a consequence of the small fraction of samples containing
concentrations of analytes above their detection limits.
Triplicate samples were collected by the EPA contractors in the water
monitoring program. Five percent of the samples taken consisted of such
triplicates. Seventy-five percent of the samples analyzed with Method 625
contained no analytes above the quantisation limit. This would imply that
only 25 of the triplicate samples had quantifiable concentrations of
analytes present (25% of 5% of the total number of validated water samples
listed in Table 1 on Page 11 of the draft report). In the draft report, EPA
states that similar observations were made with the Method 624 and the
Method 608 analytes. EPA concluded that this number was too small to give a
reliable estimate of method precision. The actual number of triplicate
samples containing quantifiable concentrations of analytes should be
reported by EPA along with estimates of precision derived'therefrom and
these estimates compared with the estimates in Table II1-6 on Page 111-35
and with additional estimates made from quality control and performance
evaluation samples.
Duplicate samples collected in the air monitoring program were used to
estimate analytical, precision. The precision was based on the standard
deviation of the observed differences in reported values. This procedure
has the desirable feature of taking into account all aspects of the
monitoring process. The three compounds listed in Table V-5 on Page V-20 of
the draft report for which a large number of duplicates were obtained were
the three compounds identified in this study to be artifacts on the TENAX
columns: benzene, toluene and tetrachloroethylene. The standard deviations
appearing in Table V-5 for these three compounds reflect both the
variability in analysis and in the actual concentrations of the three
compounds as artifacts on TENAX. Therefore precision estimates deduced from
Table V-5 for these compounds are not transferable to other compounds.
These difficulties in interpreting the significance of data from duplicate
samples have been pointed out by EPA on Page V-12 of the draft report.
No data are presented on those replicates which might give insight into the
measurement process near the limit of detection, i.e., for cases where the
detected compounds in at least one of the replicates are reported at trace
concentrations. Comparison of replicates for which at least one of the
compounds was reported as below the quantisation limit should be tested for
consistency with the limits of quantisation reported by EPA.
35
-------
D. What are some possible limitations of the quality assurance program?
The major difficulty which will be encountered in further analyses of the
data acquired at Love Canal will be coping with 1aboratory-to-1aboratory
variability in performance. No additional limitations in the program have
been found beyond those discussed in the answers to the other questions of
this section. In summary, EPA has not used all of the quality assurance
data it has collected in assessing the performance of the analytical
laboratories, i.e., tables of precision, accuracy, and detection limits are
based on limited sets of quality assurance data and have not been
confirmed.
Were the abilities of the participating laboratories to detect trace
levels of targeted and non-targeted compounds adequately tested by the
quality assurance program? What evidence does EPA present to justify
claims of laboratory performance? Is additional evidence necessary to
support such claims and, if so, what?
The quality assurance program did not adequately address the question of
minimum detection limits. Each of the participating laboratories was asked
to estimate its limit of quantisation (for the water program) or detection
(for the air program). These estimates have been tabulated in the draft
report. The ability of the participating laboratories to detect trace
concentrations (that is concentrations between the limit of detection and
the limit of quantisation) was not formally evaluated as part of the Love
Canal Monitoring Study. As stated in Section IV.B above, the participating
laboratories did not appear to be achieving the levels of detection with
Methods 624 and 625 which had been achieved in other studies. Greater
attention should be paid to the estimates of the limits of detection. These
estimates should be confirmed with other quality assurance data wherever
possible. The methods used to estimate and confirm the limits of detection
should be clearly described. This point is critical because all conclusions
of the draft report are bounded by the limits of detection achieved in the
study.
The ability to detect and identify non-targeted compounds was evaluated by
the subsequent SC-MS audit (see Section VII below).
Beyond the acceptance or rejection of data, what conclusions'has EPA
drawn from the analysis of quality control samples, field blanks, and
samples from control areas as to the precision and accuracy of the
reported data, the overall performance of the participating
laboratories, or the minimum detectable limits of targeted or non-
targeted compounds? Are these conclusions justified? If not, what
36
-------
conclusions could be drawn and/or what evidence should be presented to
justify such conclusions?
EPA has used a portion of the quality assurance data to estimate precision,
accuracy, and limits of detection for the acquired Love Canal data.
Estimates of these quantities were supplied by EPA or by the contract
laboratories with little documentation as to how the estimates were made.
The conclusions of the study may be sensitive to patterns and comparisons
near the limits of detection, and therefore the determination of these
limits must be described with greater clarity and the limits confirmed.
Precision estimates were based on d-ifferent types of data for air, water,
soil and sediments, and biota samples. For the air samples, the precision
estimates were based on comparisons of duplicate samples (see Section V.C
above). For soil and sediments and for water samples, the estimates of
precision were based on the analysis of laboratory control standards (LCS).
Not all targeted analytes were included in the laboratory control standards.
Table III-6 on Page 111-35 lists the relative standard deviations (RSD) for
the recovery of compounds in the LCS's. Some entries are missing,
suggesting lack of adherence to quality assurance protocols. Table III-6
would be improved if it included the number of LCS's for each compound for
which relative standard deviations were calculated. As noted above, EPA
should also use the data from quality control and performance evaluation
sanples to confirm the claims of precision and accuracy appearing in the
draft report. These same comments apply to the precision estimates for the
analyses of soil and sediments.
The accuracy of air methods was estimated from extensive studies of
calibration check samples. The accuracy of water methods was estimated from
the data validation procedures (discussed elsewhere in this review) and the
laboratory reports on the analyses of LCS samples. EPA did not estimate the
accuracy of soil, sediments, and biological data because the relative matrix
effects of spiked versus natural samples were not known.
Given the imprecision of the analyses as tabulated in the draft report, it
is inappropriate to comment further on the presence or absence of bias.
Because EPA has not fully utilized its quality assurance data, it cannot
substantiate the claim on Page 67 of the draft report:
"As a result, the Love Canal data are among the most carefully
validated environmental measurements, and (given the constraints
previously mentioned) are representative of the current state-of-the-
art in environmental measurement methodology in terms of precision,
accuracy and specificity."
37
-------
VI. Data Reduction and Analysis
Scope: To determine if the data reduction and analysis are consistent with
and supported by the quality of the reported data.
A. Has evidence been presented to demonstrate the accuracy of
transcription, collection and tabulation of data? If not, what
additional supporting documentation is needed?
The report of the prime contractor, GCA, presents compelling evidence that
sufficient care was taken in the transcription of the data and in the
verification of the accuracy of the transcription and transmittal of the
data'into the final data base. We have no reason to believe any
significant errors were introduced into the validated data base through
errors in the transcription and transmittal of data.
B. Has EPA assigned accuracy and precision limits to the reported data?
Have these precision and accuracy claims been adequately justified by
the quality assurance program and the available literature on the
monitoring methods used? If accuracy and precision limits have not
been assigned, does the documentation provided by EPA indicate that
they could be assigned? If so, what documentation should EPA provide
to justify precision and accuracy claims?
EPA has not incorporated precision and accuracy estimates into the
validated data base.
EPA has used data generated by the quality assurance program to estimate
precision and, in some cases, accuracy for the determination of particular
analytes in particular media. It has been noted in several sections of
this Review that EPA has not used, as fully as it could, the data generated
by its quality assurance program to confirm its reported estimates of
precision, accuracy, and limits of quantitation. EPA has not compared the
precision and accuracy estimates derived from the quality assurance program
of the Love Canal Monitoring Study to estimates derived in other studies
and published in the scientific literature.
C. Has EPA adequately interpreted the significance of "none detected" or
"below detection limit"? Has EPA listed and supported detection limits
for its targeted compounds? Has it supported "below detection limit"
claims with documentation from its quality assurance program? Are
there ways in which EPA could present the organic analytical data to
38
-------
make "below detection limit" more useful or should it continue to
report such data in such a manner?
EPA has provided, in its draft report, estimates of the limits of
quantitation and, in some cases, limits of detection for particular
analytes in particular media. Methods used to arrive at these limits are
not described and should be. Wherever possible, additional data from the
quality assurance program should be used to support the reported estimates
Additional comments on the limits of detection and quantitation can be
found in Sections IV and V of this Appendix.
EPA has not discussed in its draft report how these limits of detection
could have affected its conclusions, a central issue since most of EPA's
conclusions are based on comparisons of "detects" with "non-detects." If
the limits of detection are well below the level at which EPA would
recommend some alternative actions to those appearing in its
recommendations, then EPA should so state. If, on the other hand, the
conclusions would be altered if more sensitive or more accurate data were
available, EPA must consider this in its report. In any case, the
conclusions drawn in the EPA report must include a statement of the
concentration levels at which the conclusions are valid.
D. What conclusions has EPA drawn from intercomparison of sample splits
and sample campaigns (the collection of a number of samples from a
given location over the duration of the study)? Are the conclusions
justified?
The use of sample replicates has been addressed in Section V.C above.
Because of the small fraction of samples containing levels of analytes
above the detection limit, the conclusions drawn were limited and were
used to estimate levels of precision for some of the methods and some of
the sampled media. EPA has not compared variability at a given site over a
period of time with analytical variability. Such comparisons may not prove
to be illuminating, but would be well worth investigating.
E. Has EPA adequately justified the rejection of data and, if not, should
some data which have been rejected be reincluded in the validated data
base?
The rejection of data has not always been clearly explained (see Section
V.B). For data whose rejection has been clearly explained, the rather wide
tolerances allowed for validation make it unlikely that acceptable data
were rejected by EPA.
F. Has EPA adequately justified the inclusion of data in the validated
data base and, if not, what kinds of analyses should EPA perform on the
39
-------
sample data and the quality assurance data to provide such
justification?
EPA has not addressed, in its draft report, the probability of false "non-
detects" in its validated data base. The GC-MS audit has given some
insight into the fraction of compounds missed or misidentified at
concentrations near the detection limit. Such considerations become
important only when the conclusions of the study are likely to be affected.
The statements and recommendations made in Section VI.C addressing the
significance of limits of detection apply here as well.
G. Has EPA incorporated precision and accuracy statements into its
validated data base? If not, does it have sufficient information to do
so? If it has, has it been done correctly?
»
Precision and accuracy statements have not been incorporated into the
validated data base. Some precision and accuracy statements have been
given in the appendices to the EPA draft report and those wishing to use
the Love Canal data for further study must refer to these estimates when
interpreting the data.
As indicated earlier in this Review, EPA should use to the fullest extent
the available data to-confirm these estimates of precision and accuracy.
H. How have recovery factors been included in the validated data file? Is
it clearly stated how this was done? Was the procedure justified?
Data in the validated data base have not been 'corrected' by division by
recovery factors, which is consistent with current accepted practice.
Recovery factors have been presented for some of the targeted compounds for
the air monitoring program but not the water monitoring program. Recovery
factors should, be presented for all targeted compounds whenever possible.
Recovery factors for the air monitoring program were calculated from
performance on calibration check samples. In a similar manner, recovery
factors for the water monitoring program analytes should be obtainable from
laboratory control standards, and from performance evaluation and quality
control samples. Recovery factors are invaluable when comparing
environmental data from several laboratories.
I. Should any of the reduced data tables be published as they now stand?
In discussions with EPA, following submission of the study template to EPA,
it was determined and mutually agreed upon that this question was beyond
the scope of the review requested by EPA.
40
-------
VII. The EPA Audit of the GC-MS Computer Records
Scope: To determine if the EPA audit of the GC-MS computer records
provides usable information on the quality of the organic analysis data
and to suggest ways in which information obtained from this audit might be
used.
A. What sources of error would have and/or has such an audit uncovered?
Does the audit suggest that significant numbers of such errors might
be present in the validated data file? Could such errors affect
substantially any conclusions which might be drawn by EPA or anyone
using the data?
The EPA audit reviewed pnly the interpretation of GC-MS computer records
(tapes) and was not a review of all the steps of the analyses conducted
by the contract laboratories. EPA reviewed the computer records for 5% of
the water and sediment samples. The two indices developed for the audit
give a measure of the agreement between two laboratories. They do not
distinguish between misidentifications and missed identifications. For
example, if only one compound was present and was detected by both
laboratories, but was identified incorrectly by one laboratory, the same
indices would be obtained as in the situation where two compounds were
present but each laboratory correctly identified only one compound and
each identified a different compound. EPA gives no additional indication
of the nature of the differences uncovered by the audit. For example,
were differences primarily in the identification of two targeted compounds
of similar chemical structure?
The use of only three comparisons between EMSC and AERL to establish the
standard for agreement against which all other agreements will be compared
does not seem reasonable.
These concerns aside, the GC-MS audit provided valuable insight into
discrepancies in compound identification and into the effectiveness of the
identification of non-targeted compounds.
The discussion on Pages VI-11 and 12 concerning the concentration levels
at which disagreement occurred is convincing. No evidence is presented,
however, to support the claim "the great majority of discrepancies
...[were]... in samples that were identified as heavily contaminated."
Assertions concerning whether missed identifications were primarily in
samples already identified as contaminated as opposed to samples
considered free of contamination, if false, could influence the
conclusions of the report. The seriousness of this concern would also
depend on the concentration levels at which EPA is basing its conclusions
and recommendations as discussed in Section VI.C above.
The EPA draft report cites as one of the major safeguards in the
monitoring study, the identification in each environmental sample of the
41
-------
20 most abundant non-targeted substances. Eighty samples (water, soil and
sediments) were audited for the identification of non-targeted substances.
In 58 of these samples, both the audit laboratory and the analytical
laboratory identified no non-targeted substances. In the remaining 22
samples, the audit laboratory identified 84 non-targeted substances while
the analytical laboratories found only 1. This result indicated that, for
the samples audited, most laboratories did not identify non-targeted
substances.
Finally, the results of the SC-MS audit of water, soil, and sediment data
are not applicable to other media. An independent audit of air data is
necessary if similar information is sought.
B. Does the audit indicate there are major differences in the ways
the laboratories analyzed their 6C-MS tapes? How would such
differences manifest themselves? Was the audit adequate to detect
such differences? Should it have been?
The audit showed that laboratories differed in the identification of
compounds near the detection limit. In its analysis of 'the audit, EPA
attributed these differences to differences in computer algorithms and
analysts' judgements.
The differences would manifest themselves as misidentifications or
missed identifications by the analytical and/or audit laboratories.
The audit appears to have been adequate to detect these differences.
Should all of the GC-MS tapes corresponding to the validated data
file be reanalyzed using a single program? Can we specify which
program? Should GC-MS tapes corresponding to rejected data also
be reanalyzed? Could faults in the analysis of the GC-MS tapes
have led to incorrect rejection of data from or inclusion of data
into the validated data file?
Variability introduced by use of different programs is likely to be less
important than the variability introduced by the different analysts who
reviewed the data.
Further audit of the GC-MS records, including review by a team of analysts
using identical criteria, would, appear to be necessary, so long
as conclusions drawn by EPA in its Love Canal Monitoring Study are
sensitive to the presence of undetected contaminants at concentrations
near the detection limit.
42
-------
Faults in the analysis of the GC-MS tapes are not likely to have led to
incorrect rejection of data from or inclusion of data into the validated
data file because the criteria for rejection were basically unrelated to
possible problems with compound identifications.
43
-------
APPENDIX B
DOCUMENTS RECEIVED BY NBS
Item
Bound and Titled
Document Title
1.
2.
3.
4.
5.
Quality Assurance Plan Love Canal Study
Quality Assurance Plan Love Canal Study
Appendix A (Sampling Procedures)
Quality Assurance Plan Love Canal Study
Appendix B (Analytical Procedures)
Quality Assurance Plan Love Canal Study
Appendix Q (Subcontractors' QA Plans)
Love Canal Monitoring Program
6CA QA/QC Summary Report
Unbound and Titled
6.
7.
8.
9.
10.
11.
12.
Environmental Monitoring at Love Canal, 1980
(Designated by NBS as OEM-LC-2)
Data and Control Charts by Level
[Air] — 'Four Documents'
Audit of Gas Chromatography/Mass Spectrometry
Data (GC/MS) Provided by Love Canal Project
Analytical Laboratories
GCA Report Volume I, Proposed Work Plan Love
Canal Monitoring Project
Additional EPA Love Canal Materials for NBS Review
(Correspondence)
Supplemental Documentation of EPA Love Canal Quality
Assurance Program (Correspondence)
Revised Version of Audit of
Gas Chromatography/Mass Spectrometry Data
Provided by Love Canal Project Analytical
Laboratories (Item 8 above)
Date Received
at NBS
18 Aug. 81
18 Aug. 81
18 Aug. 81
18 Aug. 81
28 Jan. 82
18 Aug. 81
21 Oct. 81
21 Oct. 81
21 Oct. 81
21 Oct. 81
19 Nov. 81
19 Nov. 81
44
-------
13.
14.
15.
16.
17.
18.
19.
20.
21.
Unbound
U.S. EPA Environmental Monitoring at Love
Canal (Draft Report, Designated Copy #9)
Appendix I
List of Substances Monitored at Love Canal
Appendix II
Comparative Data and Existing Standards for
Substances Monitored at Love Canal
Revised Sections 3.3, 3.3.1, and 3.3.2 for
EPA Draft Report
Quality Assurance for Water Samples
Revised Appendix III for Draft Report
Quality Assurance for Air Samples
Revised Appendix V for Draft Report
Report on the Audit of Gas Chromatography/Mass
Spectrometry Data Provided by Love Canal
Project Analytical Laboratories
New Appendix VI for Draft Report
Quality Assurance Plan, Love Canal Study,
Appendix Q, Revision 1
Quality Assurance for Soil, Sediment,
and Biota Samples
Revised Appendix IV for Draft Report
and Untitled
17 Dec.
17 Dec.
17 Dec.
28 Jan.
28 Jan.
28 Jan.
28 Jan.
28 Jan.
17 Feb.
81
81
81
82
82
82
82
82
82
22. Laboratory Audits - Trip Reports
Love Canal Study
(Designated by NBS as OEM-LC-1) 18 Aug. 81
23. Love Canal Study Area: Sampling Sites
(Designated by NBS as OEM-LC-3) 18 Aug. 81
24. Aggregate of Data from Various Sampling Areas
(Designated by NBS as OEM-LC-4) 18 Aug. 81
25. Aggregate of Data from Various Sampling Areas
(Designated by NBS as OEM-LC-5) 18 Aug. 81
26. Aggregate of Data from Control, Canal, and
Declaration Areas
(Designated by NBS as OEM-LC-6) 18 Aug. 81
45
-------
27. Aggregate of Data from Control, Canal, and
Declaration Areas
(Designated by NBS as OEM-LC-7) 18 Aug. 81
28. Graphical Displays of Air Monitoring Data as
a Function of Sampling Date 6 Oct. 81
29. Quality Assurance Control Charts & Tables
for Water Samples 6 Oct. 81
30. Quality Assurance Data (Tables) for Love Canal
Soil, Sediment, and Biota 21 Oct. 81
31. Soil Holding Times (Internal EPA Correspondence) 21 Oct. 81
32. Holding Time Study for Purgeable Compounds
in Water (Internal EPA Correspondence) 21 Oct. 81
33. Report on Audit of PJB Laboratories (Internal
GCA Correspondence) 21 Oct. 81
34. Report on Audit of Battelle and PEDCO
Laboratories (Internal GCA Correspondence) 21 Oct. 81
35. Collection of Public Comments Received by
EPA Regarding Methods 624 and 625 and EPA
Response to those Comments 28 Jan. 82
46
-------
NSS-1UA iHgy. 2-«oi
u.s. OSPT. OF COMM.
BIBLIOGRAPHIC DATA
SHEET (See instructions;
1. PUBLICATION OR
REPORT NO.
NBSIR 82-2511
2. Performing Organ. Report
3. Publication Date
May 1982
4. TITLE AND SUBTITLE
5. AUTHOR(S)
6. PERFORMING ORGANIZATION (If joint or other than NSS. see instructions)
NATIONAL BUREAU OF STANDARDS
DEPARTMENT OF COMMERCE
WASHINGTON, D.C. 20234
7. Con tract/ Gran t No.
S. Type of Report & Period Covered
9. SPONSORING ORGANIZATION NAME AND COMPLETE ADDRESS (Street. City. Store, ZIP)
U.S. Environmental Protection Agency
401 M Street, SW
Washington, D.C. 20460
10. SUPPLEMENTARY NOTES
| 1 Document describes a computer program; SF-185, FIPS Software Summary, Is attached.
11. ABSTRACT (A 200-word or less factual summary of most significant information. If document includes a significant
bib/ioyrapny or literature survey, mention it here)
ABSTRACT
This report is a review conducted by the National Bureau of Standards
at the request of the U.S. Environmental Protection Agency (EPA) of
the analysis for organic chemicals conducted by EPA in the Love
Canal Area of Niagara Falls, New York.
12. KEY WORDS (Six to twelve entries; alphabetical order; capitalize only proper names; and separate key words by semicolons)
Love Canal, Monitoring, Organic Analysis, Environment, Pollution
13. AVAILABILITY
| ] Unlimited
rXl For Official Distribution. Do Not Release to NTIS
| j Order From Superintendent of Documents, U.S. Government Printing Office, Washington, D.C.
20402.
| ] Order From National Technical Information Service (NTIS), Springfield, VA. 22161
14. NO. OF
PRINTED PAGES
15. Price
USCOMM-OC 5043-P80
-------
Ill
E
P
A
-------
EPA Observations on the Review of the
National Bureau of Standards of the Love
Canal Monitoring Program
Abstract
The NBS Review highlighted many technical aspects that must be
considered in detail when employing state-of-the-art analytical
methods for the first time in operational programs, particularly
programs being carried out under severe time constraints. NBS
considered the methods used by EPA to have been the best that were
available for the study. The Review underscored the importance of
determining the detection limits and the precision and accuracy of
these methods.
The Review identified many issues that were not adequately
addressed in the initial drafts of documents provided to NBS by EPA.
However, the Review did not identify any technical limitations,
deficiencies in laboratory performance, inadequacies in quality
assurance activities, or other shortcomings in methodological
approaches or procedures that would affect the conclusions of the
monitoring program. The deficiencies in the early drafts have been
corrected, and the revisions of these documents incorporate or
respond to all suggestions and questions raised by NBS.
General
During the fall of 1981, EPA requested the National Bureau of Standards
(NBS) to review and critique a number of initial drafts of a variety of
technical documents prepared during 1980 and 1981 in support of the Love Canal
Monitoring Program. The NBS comments are set forth in the May 1982 report
"Review of Material Presented by EPA on the Analysis for Organic Chemicals in
the EPA Love Canal Monitoring Study" which also identifies the EPA draft
documents that were considered. The Review identified many issues that were
not adequately addressed in these initial drafts and greatly assisted EPA in
preparing the final documentation to support the study. The NBS comments were
particularly useful in EPA's effort to document more fully the technical
basis for the conclusions presented in EPA's final report of the study
"Environmental Monitoring at Love Canal," May 1982, and in particular, in
helping to ensure the completeness and clarity of the associated quality
assurance documentation. In this regard, the initial drafts of Appendices C,
D, and E of Volume 1 of the final report which describe the quality assurance
program have been substantially rewritten to incorporate many suggestions of
NBS.
-------
The NBS has extensive experience in the development of high precision
scientific measurements, in general, with a long-standing interest in
advancing the scientific frontiers of analytical methods. Furthermore, NBS
had not been involved in monitoring activities at Love Canal and therefore was
in a position to bring an independent perspective to a review of EPA-'s
monitoring efforts. This combination of research experience, interest, and
fresh perspective resulted in many useful observations in the NBS Review and
complements internal EPA reviews which drew on EPA's more extensive experience
in conducting environmental monitoring programs.
As recognized in the NBS Review, the analytical methods used by EPA were
at the state-of-the-art. In some instances, the Love Canal effort was
advancing the state-of-the-art. For example, the methods for air and water
analyses had not previously been subjected to multi-laboratory validation
studies. With regard to the analysis of soil and sediment, the initial weeks
of the study were devoted to determining, through abbreviated multi-laboratory
tests, which method of several options would be most appropriate. For
analysis of water as well as soil and sediment samples, fused silica capillary
columns were used on a significant scale for the first time.
In view of EPA's regulatory mission and the need to respond promptly to
complex technical issues at Love Canal, there was no opportunity to conduct a
methods evaluation program and to determine in advance the full capabilities
of the methods to be used. Rather, it was necessary to determine and document
those capabilities during the study. Despite the uncertainties inherent in
using methodologies that had not been thoroughly tested, NBS and EPA agree
that the methods that were employed were the most appropriate given the
objectives of the study.
Use of relatively new methods on a large scale strained the manpower
skills of contractor laboratories throughout the country. Given the limited
experience with these technologies, EPA technical specialists devoted
substantial efforts to ensure that the participating laboratories immediately
achieved an acceptable level of performance. In general, the contractor
laboratories responded very well to this unprecedented technical challenge as
indicated by the relatively low rate of data rejection, the few instances of
laboratory performance that required major corrective action, and the general
consistency of data generated by many laboratories.
The Love Canal program was the most extensive environmental monitoring
program ever undertaken at a single site. It included the analysis of more
than 6,000 environmental samples and of an equal number of quality assurance/
quality control samples, involving a total of about 150,000 individual
measurements. The effort was of necessity compressed into an 80-day sampling
period and an analytical phase that lasted about six months. It was clear
from the outset that there was to be no opportunity to collect additional
samples nor to conduct repeat analyses. Thus, the primary emphasis during the
initial months was to ensure that samples were collected properly and analyses
performed satisfactorily. Procedures were established to ensure that these
efforts were technically sound. The procedures were set forth in the con-
tractual requirements involving the prime contractor and the subcontractors.
-------
From the outset attention was directed to the complete documentation of
the program, including further elaboration of procedures set forth in
contractual requirements when necessary. However, time pressures frequently
required actions before all documentation had been completed, and in these
cases frequent consultations between EPA and contractor technical specialists
were essential in clarifying the details of technical approaches. As
reflected several times in the NBS Review, the written procedures supplemented
by technical consultations achieved their objective. For example, instances
of unacceptable laboratory performance were uncovered and prompt corrective
actions taken. A number of contamination problems were identified and
adjustments made. In a few cases, excessive holding times for samples prior
to analysis were discovered and the implications promptly evaluated. Given
this focus on the real-time operational aspects of the program, the early
drafts provided by EPA to NBS needed substantial improvement to complete the
written documentation supporting the effort.
Response to Specific Concerns of NBS
As suggested above, most of the specific concerns raised by the NBS
Review can be traced to shortcomings in the initial drafts prepared in 1980
and 1981 that were provided to NBS by EPA. The documents prepared since that
time have been substantially modified, and they incorporate many NBS
suggestions. The Review also expressed general concerns over the limitations
of the analytical methods to detect extremely low levels of organic chemicals
and the possible impact of these limitations on the study's conclusions. As
explained in Volume 1, EPA considers the technical capabilities of the methods
adequate to fully support the conclusions presented. Finally, NBS raised
several issues concerning performance of the contractor laboratories and the
problems inherent in coordinating a mutei-laboratory program. EPA considers
that the overall performance of the laboratories was quite good and that while
there were complex coordination problems, many benefits were derived from
obtaining complementary data from multiple laboratories as well as from
several environmental media*
The principal issues raised in the NBS Review are set forth in the main
body of the Review. Elaboration of some issues and identification of several
other issues are contained in Appendix A of the Review. Summarized below are
all the issues raised in the main body of the Review (pages 1 to 10 of the
Review) with a brief EPA commentary on each issue. In several cases the
commentaries also address additional issues raised in Appendix A. Whenever
appropriate the commentary identifies the section of Volume 1 of the EPA final
report which contains additional information on the topic.
1. "EPA's stated goals and objectives, as they applied to the analysis for
organic chemicals . . . were not . . . quantitatively explicit. EPA did
not define at the outset • • • the limits of detection or the precision
and accuracy required ..." (Page 5 of NBS Review)
-------
Since the analytical methods had not been previously subjected to
multi-laboratory validation studies, data were not available to make
a priori determinations of precision and accuracy or limits of
detection. These determinations were made using the multi-
laboratory data generated during the Program, and the procedures
and determinations are documented in Appendices C, D, and E of
Volume 1. Existing Federal Register methods provided general
guidance for water analysis, but they were based on packed columns,
not the fused silica capillary columns (FSCC) used in this study.
2. ". . . the requirement for the identification of the 20 most abundant
non-targeted substances was generally not met." (Page 5)
A principal reason that the laboratories did not identify many
non-target compounds is that they were generally not present in the
samples at detectable levels. Specifically, the audit showed that
in about 73 percent of the water/soil/sediment samples non-target
compounds were not present. In the great majority of the cases when
non-target chemicals were found to be present but not reported, they
were present at only trace levels, and the sample was heavily
contaminated with target compounds. The reporting omissions did not
affect project findings.
3. "The quantisation limits realized by water monitoring laboratories . . .
varied by more than a factor of ten. Few laboratories reported
quantisation limits as good as those reported by the EPA laboratory . . .
Comparisons of quantisation limits with data provided by EPA from other
studies indicated that the contract laboratories were not uniformly
performing at the state of the art- . . ." (Page 5)
For GC/MS methods the detection limits were all in the range of 1 to
50 micrograms per liter, and nearly all were less than 30 micrograms
per liter. For certain pesticides, the limit of detection was in
the low nanograms per liter range. These values are representative
of the current technology limitations in analytical chemistry.
Indeed, the range is very acceptable since it is comparable to the
range obtained in other studies. (See, for example, Environmental
Science and Technology. Volume 15, Number 12, December, 1981, pages
1426-35.)Given the different missions of EPA and contractor
laboratories, it is not surprising that the EPA laboratory with its
highly specialized staff achieved a higher level of performance. We
are unaware of other published studies that had used fused silica
capillary columns which indicated different ranges of quantisation.
4. "EPA has not fully used its quality assurance data to derive estimates of
precision, accuracy, and limits of detection. EPA has not incorporated
estimates . . . into its validated data base." (Page 5)
-------
As set forth in Appendices C, D, and E of Volume 1, EPA has fully
used its quality assurance data to derive estimates of precision,
accuracy, and uncertainty. However, so few replicates had positive
results, estimates of precision and accuracy were based on results
obtained from analyses of performance evaluation samples and/or
laboratory control standards.
5. "Unless ... 'none detected' (values) are accompanied by estimates of
uncertainty, they are incomplete and of limited usefulness." (Page 5)
The procedure for determining the method detection limit, which is
set forth in Environmental Science and Technology. Volume 15, Number
12, December, 1981, pages 1426-35, provides a high degree of
assurance that positive identifications are real identifications.
Under this procedure the probability is less than 1% that a value
reported as "none detected" (which includes both trace and below
detectable values) is greater than twice the method detection limit..
This represents the worst case situation.
6. "EPA has not explicitly related included or excluded (target) substances
to specific (selection) criteria." .(Pa9e 6)
Criteria for selecting target compounds are stated in Volume 1 and
were (a) chemicals known to have been deposited in Love Canal, (b)
the priority pollutants previously defined by EPA as exhibiting
toxic properties in water, and (c) other chemicals of toxicological
concern believed to be present in the area as a result of studies by
New York State and analyses of leachate material. These target
compounds included chemicals with a wide variety of physical and
chemical properties that influence their rates of migration through
different environmental media.
7. "For some individual substances, alternative analytical methods might
have been more suitable (for example . . . other than TENAX for benzene
and toluene)". (Page 6)
Given the magnitude of the effort, it was not feasible to tailor
individual methods to very many individual chemicals. The
contamination problems associated with TENAX had not been documented
prior to the Love Canal effort. Nevertheless, even anticipating
such contamination, TENAX was the best method then available.
Future research efforts should take into account this limitation in
the use of TENAX. The limitations of TENAX are further discussed in
Appendix E of Volume 1.
8. "Some ambiguity existed in the analytical protocols." (Page 6)
We are unaware of ambiguities. To insure that the written protocols
did not raise ambiguities, EPA technical specialists met with
representatives of the laboratories at the outset of the effort to
review the protocols in detail.
-------
9- "Program design did not anticipate the possibility that a significant
fraction of the samples might not contain any detectable levels of
analytes." (Page 6)
EPA clearly anticipated that a larger number of samples would
contain measurable levels of contaminants. However, the same
procedures would have been employed had there been greater awareness
that most of the positive findings would be near the levels of
detection. In interpreting the data when very low levels are
encountered, greater reliance must be placed on the consistency of
the data and on possible patterns of contamination. In this effort
the data obtained from the different media were consistent and were
in accordance with the findings of the hydrogeological
investigations.
10. "The quality assurance plans . . . vary in detail from one laboratory to
another. The variations in detail could lead to difficulties in
comparing data obtained from different laboratories and different media."
(Page 6)
All laboratories were required to satisfy a minimally acceptable
quality assurance program specified by EPA. All laboratories met
this reqairement. As would be.expected, a number of the
laboratories imposed more stringent requirements of their own.
Since all laboratories had satisfied the minimum requirements,
comparisons of results were possible.
11. "The Quality Assurance Plan . . . did not require the use of control
charts." (Page 7)
The contractual arrangements required the use of quality control
charts, and they were used.
12. "The Quality Assurance Plan did not address how the quality assurance
data would be used to qualify the environmental data." (Page 7)
The criteria and procedures for qualifying data are set forth in
detail in Appendices C, D, and E of Volume 1.
13. "Estimates of the limits of detection are critical to the conclusions
which can be drawn when the majority of samples contain no measurable
concentration of analytes. The Quality Assurance Plan did not anticipate
such a possibility." (Page 7)
See 9 above.
14. "Laboratories were not uniformly performing at the state of the art."
(Page 7)
See 3 above.
-------
15. "Laboratories did not generally identify non-targeted substances."
(Page 7)
See 2 above.
16. "In isolated instances, contamination of the samples during sample
collection or analysis may have obscured possible trends and patterns."
(Page 7)
Field and laboratory blanks were analyzed together with the
environmental samples to determine whether accidental contamination
was introduced during the sample collection or the laboratory
analytical procedures. In a few instances these blanks indicated
such accidental contamination, and corrections were made in the data
on those environmental samples which were subjected to the same
accidental contamination. While it is conceivable that additional
accidental contaminants may have been present in a few samples, it
is highly unlikely that undetected accidental contamination occurred
to the extent that trends and patterns of contamination were
obscured. The trends and patterns observed were consistent among
all media and were consistent with the physical and chemical
properties of the chemicals which affect their migration rates.
17. "The quality control limits set in the Quality Assurance Plan widened
during the study." (Page 7)
The quality control limits did not change. The use of fused silica
capillary columns required different limits than the limits for
packed" columns specified in.Method 625. Appropriate control limits
were established at the outset and uniformly applied throughout the
program.
18. "The estimation of precision and accuracy ... has not been adequately
documented. Only a portion of the . . . data has been used in arriving
at these estimates. The estimates are generally unconfirmed and
estimates for some of the analytes have not been given." (Page 8)
See 4 above.
19. "The estimation of the limits of detection and quantitatioh are . . . not
adequately documented." (Page 8)
The data are available in project files. While not considered
essential for inclusion in this report, they are publicly available
and will be reported in the scientific literature at a later date.
The limits of detection and quantisation derived from actual sample
analyses are discussed in Appendices C, D and E.
-------
20. "The rejection of data has neither been clearly described nor adequately
justified." (Page 8)
See 12 above.
21. "There is no evidence of ... monthly on-site visits as prescribed in
the Quality Assurance Plan." (Page 8)
All participating laboratories were initially audited through on-
site visits. Subsequent visits were generally unnecessary because
the prime contractor, through daily review of quality control data,
took timely action via telephone when questions arose. When
additional visits were needed, they were conducted. In addition to
the reviews by the prime contractor, the data were further examined
and qualified by EPA through a deliberately designed redundant
system of data review, including analysis of surrogate compound
recoveries for each sample.
22. "No audit of the air or biota monitoring programs was performed."
(Page 9)
The results of the biota monitoring program were not a significant
factor in reaching the study conclusions. Therefore, a diversion of
limited resources to audit the GC/MS tapes generated during the
program was not warranted. The Love Canal compounds likely to be
detected in air were well identified from analyses of the head space
in liquid samples taken from the French Drain and were on the target
list. An audit of the air program did not seem warranted in view of
the unlikelihood that additional Love Canal chemicals would be
detected and the consistency of the data that had been reported.
23* "The claim that missed compounds were predominantly in heavily
contaminated samples is not supported." (Page 9)
Documentation is contained in project files and is publicly
available. Individual sample analyses from the data audit are too
cumbersome for inclusion in the report. Appendix F of Volume 1
presents a summary of results.
24. "The implications of the audit to the conclusions of the study are not
discussed . . . For the samples audited, most laboratories did not
identify non-targeted substances." (Page 9)
See 2 above.
25. "The measured values ... in the validated data base are not accompanied
by estimates of their uncertainty." (Page 9)
See 4 above.
8
-------
26. "Recovery factors have not been given nor confirmed for most analytes in
water and in soil and sediment samples." (Page 9)
Recovery factors for laboratory control standards are set forth in
Appendices C and D of Volume 1. Performance evaluation standards
for organic chemicals in soil and sediments were not available, and
therefore recovery factors for analytes in soil and sediment could
not be estimated. Recovery factors are not essential in determining
patterns and trends.
27. "Examples selected to support the conclusions of the study have not
always been confirmed as being truly typical of the data in general."
(Page 9)
Since it was not feasible to include all of the voluminous data in
the descriptions of the findings of the study, examples were used to
assist in supporting and clarifying the findings. The criteria
used in selecting examples are set forth in Volume 1. For example,
compounds of varying solubility and migratory propensity were
included in the graphical presentations to help ensure that any
significant migration patterns would not be overlooked. Also, in
some discussions examples were chosen to represent each of the
analytical categories (pesticides, volatiles, semi-volatiles,
inorganics). In one case, the highly toxic compound dioxin was
given special attention.
28. "EPA has not addressed explicity how . . . conclusions are influenced by
the limitations in its analytical data." (Page 9)
See 9 above.
29. "Nowhere . . . are • - • results presented which give . . . indication of
whether . . . (chemicals which might serve as) tracers migrated through
the subsurface area." (Page 14)
Compounds which are more soluble in water such as phenols, nitrates,
fluorides and many volatile organics did not indicate patterns of
subsurface migration as would have been expected if such migration
had, in fact, occurred.
EMSL-LV: 07-01-82
------- |