NCD980602163_19961216_Warren County PCB Landfill_SERB C_Joel Hirschhorn correspondence, Dec 1996-OCR

December 16, 1996 ............................................................. by FAX (5 pages)
To: Technical Committee
From: Joel Hirschhorn
Through: Doris (for distribution prior to WG meeting as muc.b as possible, and to all Working
Group members at meeting of Dec. 17)
Subject: Review of orieinal EPA tontnctor report on PCB air emb§iow-
.ln Tesponse to my request~ I received the Final Report on Ambient Monitoxing For PCBs At The
Warren County (North Carolina) Landfill by Battelle from Dr. Lewis and have prepared the
following comments on it so that the Working Group can better understand the situation. ln
general I have found even more problems and reasons to be concerned about the entire histor1 of
this siruation.
~h latest review, analysis and conclusions strongly support the position of the Iccf~11
Committtt.in its re~Qi.rlition that it is imperative that full disclosure of tht dao2t.f..LQf.tb.e
landfi,U be openly comntunicated to the public and the state iovernrnent through a11other
press cogference, letters to the Governor and Secretao· Bowes. appropriate actlons being
taken by the state (:milJJ$ carbon filter installation and new mopitorin,:). and.a well
planned r.omn1unity meeti.nc that the Governor is invited to participate in. I w;ml..1!2
.emp_hasizc that the most important recommendation presented by E..r..t\~~r in
l,283, calling for future periodic monitorine for PCB air release:,. was NOT implcmcnted..b.):
1be statur EPA. In mv professional opinion this was a ullous and scientific;1JlDm£fil!P..d
decisio11 by the state and EPA that sas:ri/iced protection of public health to deri2Y1
tQnctrns, like wantine to avoid ,:ettin,: data that wou,Jd conclusively demQJll.~7Jl1.U2
lVarren County residents and the broader public that the Wa.rnn County FC1LLLrulfill
was dangerous, unsafe. and a 11ource of uncontrolled environmental rn~'!.J_he
recommendation for future periodic monitorin1 and the test results from 1983 mmvd h1.YC
Joci,cally resulted in the ~tate taking the proper precautionary step of instalJing at.r.Mn
filters oo all landfill vent!. Thi" matter de,41un~ the dronic.~t pos~ihlc action h_" th~
\Vorking Group to ufe~acd the interests of Warren County rcsident~~hat efhctiv~
1kruificatjon of the laodfiU is actu3Uy accomplished as soon as possible, The Goyern.QI
should be •·equestcd to determine why the recommendation of EPA's contractor to cond~~
future monitoriuf for PCB emissions wa, not implemented,
The report was dated August 16, 1983,.indicating that complete data were available to EPA
relatively soou after the onsite testing had been done iu January and Fehrulry 1983 . Of the 40+
page report, only page 2, Section 2, Conclusions had been found in the state files provid~d me,
and which I previously commented on.
1
Page 3 of the report was SeC1,ion 3 Recommendations and included the following : "In the future,
rua recommended that periodic monitoring be performed to determine tht. trend in PCB
emission rates from the gas vents and leuhate access ports on the landfill site • .'' (emphasis
added) The single paragraph also included a recommendation to use somewhat different sampling
technology, called high volume PCB samplers, because they offered ''significantly lower detection
limits." It appears that these recommendations were never implemented by either EPA or the
state. The recommendation for future testing is particularly significant because there were
relative.ly few positive detections of PCBs in ambient air, indicating that the contractor recognized
that in later times PCB emissions and .releases offsite were plausible:. In other words, the
contractor was correct in recognizing the appropriateness of experimentally, through further field
testing, detennin.ing the "trend" of PCB emissions. This implicitly recognized the positive
detections of PCBs in vents and ports, as sources of environmental releases, as well as initial
detections of PCBs in ambient air. It also recognized the reasons why the study had been
conducted.
The contractor report contained no information on the design or construction of the landfill,
which the published paper had presented. But the contractor report did note that: "Local
residents and the Warren County Health Departm.ent have expressed cone-em about the possibility
of airborne PCB emissions from the lB..Ddfill being transported to neighboring areas, thus
threatening the public welfare." It also noted that "The study was performed at the requeSt of
the No.rth Carolina Division of Health Services ... " This makes it even more questionable why the
full comractor report was not found in state files. Five specific study objectives were presented~
including determining if PCBs were present at the house approximately one-haJfmile away. Of 42
specific measurements in ambient air, however, only two were for the nearby house.
There is absolutely no statement iu tb.e entire report that in any way supports the contention
receotJy communicated by EPA that the positive findings of PCB air emissions were false
positives.
There js a discrepancy between the details of the positive .findings as preseuted in the published
paper versus the contractor report. The published paper indicated four positive detections in
ambient air and clearly gave the locations of tluec of the:m as: beside main vent, fence Jine
downv.-ind, and nearby house. The contractor report gave: beside main vent, two for fence line
downwind, aud onsite upwind of main vent. The footnote in the published paper rodicating that
one of six measurements at the nearby house was above detection limit is inconsistent with data 111
the report that indicates that only two measurements were taken at the nearby house. What is
most perplexing at this time is, therefore, that the published paper showed a positive finding of
PCBs at the nearby house, while the contractor report showed all .findings as less tban the
detection limits. But the procedure used by the contractor was to consider less than dctectiou
limit fiudings as equal to the detection limit. Following the contractor~s procedure the average
concentration for Aroclor 1242 and 1260 for the nearby house would be 6 and l 0 ng/cm.
respectively. In my previous comments I used what was reported in the published paper, namely
a positive finding of 10 ng/cm for 1260, which is still supported by the actual fi.ndings in the
2
contractor report. Let me emphasize that the contractor's methodology of assuming a finding of
Jess than the detectiou limit is equal to the detection limit is far more appropriate than assuming
that such a nondetect .finding is equal to zero, which seems to be EPA's position. Nevertheless,
th.ere remains a discrepancy between the published paper's clear notation that a positive finding of
Aroclor J 260 was found at the nearby house versus the contractor report that indicates that all
four readings on the two days of testing for the two Arodors were nondetects. 'What deserves
attention by .EPA is why it reported, in the published paper, the positive finding of PCB at the
nearby house.
The variation in detection limits was made understandable in the contractor report, because the
data given showed that sampling times and, hence, sample volumes varied somewhat, with
unanticipated lower volumes resulting in higher detection limits in some cases. In all cases. where
positive detections in ambient air were found, the detection li.mjts were the lowest, making the
findings more reliable.
Although the published paper presented key PCB data in terms of ranges and averages, the data
presentation in the contractor report was different and, in my opinion, more useful and
informative. One problem in the published paper i.s that some data were reported as "ND" which
is standard notation for non-detected, while other data were reported as "<" meaning less than the
folto\.\ing numerical number which normally would be the actual detection limit (DL). 1n fact,
there was no reason to .report NDs at all in the published paper, and in the contractor report data
were reported either as positive detects or as <DL for PCB concentrations. The point is the NDs
represented a sp~cific test in which no reliable data were obtained. But the published paper said
that ND represented "nondetectable." This definition misrepresented the actual meaning of some
data. For example, for the Jnwer leachate access port, the published paper indicated the low end
of the two ranges were ND, but this actually referred to a test where the data were not usable.
The illference was that a ND low end represented a zero level rather than no reliable data. The
published paper reported averages of0.04 and 0.05 uglcm for 1242 and 1260, respectively, which
were calculated by assuming that ND = 0 and that <DL = DL. But using the actual data in the
report, and eliminating the ND values results in the averages being .06 and .08 ug/cm,
respectively. Clearly> the corrected averages portray a significantly higher level of PCBs from thl?
lower leachate access port.
Another systematic source or widerrcporting PCB levels was the unusual procedure of not
rounding off numbers following the nonnal practice that if a digit is 5 or greater, than the
preceding digit is increased by one. For ex.ample, .056 should become .06. But in the published
paper, some averages and other figures arc too low, because rounding was not done properly.
For example, the average PCB levels for small vent E were reported as 0.24 and 0.4, hut should
be 0.24 and 0.5.
There is an important lack of details about the weather conditions existing at the time of sampling.
For example, for Jan. 30 the notation was ''no sampling due to weather c~ditions," \.\ithout,
however. explaining what the weather conditions were. The relative humidity for the following
3
two days on which testing was done indicated very high levels, suggesting that perhaps rain or
snow was a factor during the testing period. In general, the collection of meteorological data was
incomplete, and the report presented DO detailed analysis of how weather condhio.ns might have
affected the testing results, both positive and negative detections. lo my opinion; the weather
conditions were probably not the most conducive for transporting PCBs offsite which could exist
at other times. This supported the study's recommendation for additional future testin.g. For both
periods during which monitoring was done for the nearby house the wwd was not blowing from
the landfill in the direction of the house. This fact and the very limited number of measurements
taken at the house indicate, in my opinion, poor study design.
The contractor report made an important observation about the finding of Aroclor 1260 but not
1242 in ambient air: "It is puzzling that Aroclor 1242, the major component of the vent emissions,
was not also detected in these samples." lbis is the issue raised recently by EPA However, the
contractor does not attempt to explain why the noted obsexvation occurred or how it can be
explained, and cenainly never taises the possibility of f.alse positives.
In eight cases reliabJe data was not obtained because sample pumps malfunctioned, witl, the only
reason givell as "due to the low temperature and high humidity conditions that prevailed duriug
the sampling period. 11 This supported monitoring at a later time under different climatic
conditions.
Tue published paper did not note that the two rughest ambient air detections were for the same
locatiou at the same time, but for two heights of 4 aud 15 feet, with 71 and 50 ng/scm,
respectively. The level of PCBs at the higher height is significmt.
The contractor report made it clear that gas flow rates from vents and leachate access ports were
measured by EPA .personnel on ''March 2, 1982." Presu.m.1bly this ·was an enor (not caught in
peer review), and was really in 1983, but it was definitely after the contractor performed the
major portion of the ""ork. And while the published paper some data on the gas flow from the
main vent, it did not present the detailed data in the most understandable and sign.ificaat way. The
contractor report's presentation of the data showed that gas flow rates increased siguHicautly
from 9:45 AM to 3 :00 PM, with a 32% increase over that period. This would logically suggest
that g3s flow increased as temperature increased (no climatic data was given for this day in the
report), supporting my contentiou that air releases of PCBs increase with increasing temperature
and also supports the report's recommeudation for subsequent monitoring at the landfill. The
mod.ding work by the contractor cited by EPA only used the average of the data obtained for
<tifferent times, "nhich does not, therefore, produce results indicative of maximum PCB release
rates.
The study design was deficient because the sample volumes for the small vents and the leachate
access ports was the same as for the main \'ent, which resuhed in higher detection limits fo1 the
fonner vers.us the Jatter1 because the quantities of PCBs in the former were substautially smaUer.
Such a situation could have been predicted, and at the very least this problem would have been
4
corrected in subsequent monitoring by increasing the sampling rate and volume (as was done for
the ambient monitoring).
1 found it unusual that several significant statements made in the Conclusions section "'ere not
contained in any form \\ithin the body of the report, particularly Section S Results and Discussion.
This is highly unusual, because normally there is a more extended discussion in the bod}' of the
report that supports sumrmry conclusions. For example, the comments iu the conclusions about
.PCB emission rates being reduced in the future are not supported by any information or analysis
given in the body of the report. Not are the comments completely consistent with the report's
major recommendation for future periodic monitoring. Other statements in tbe conclusions are
inconsistent with the details given in the report, namely that ambient PCB levels were only found
"at or below minimum detection lmrits" while, in fact, the report's data showed positive detections
as high as seven times the detection limit. Based on my professional experience, th.ese kinds of
unsupported and inconsistent staterocnts in conclusions result from the contractor's client (EPA)
making changes during a final review of the draft report in order to protect or serve its o~n
interests.
As to peer review that Dr. Lewis has claimed is so significant, I want to point out that the peer
review process has many limitations, including the fact that a primary report that forms the basis
for a published paper is not provided peer reviewers, so that soroe shortcomings in the paper
cannot be properly assessed.
s
•'
DATE:
TO:
'5'919 541 352i ~ERL/AMRD d IB
U.S. ENVIRONMENTAL PROTECTION AGENCY
NATIONAL EXPOSURE RESEARCH LABORATORY
AIR l\1EASUREl\1ENTS RESEARCH DIVISION (:rvIAIL DROP 44)
RESEARCH TRIANGLE PARK, NORTH C . .<\ROLINA 27711
TELEPHONE [phone] FACSIMILE 919-541-3527
11 December 1996
Bill Meyers
E-MAil., [e-mail]
FAX l\1ESSAGE
NUMBER OF PAGES: 2
Fax No.: 715-3605
Telephone:
FROM: Dr. Robert G. Lewis
USEPA (MD-44)
Research Triangle Park, NC 27711-2055
SUBJECT: Response to Hirschhorn's Latest Comments
MESSAGE:
14] 00J /0 0.'.!
I will not dignify Dr. Hirschhorn by responding to his latest attack on the quality of the EPA
monitoring study at the ·warren County landfill and my scientific interpretations. It is obvious that nothing
will persuade rum to abandon the faulty reasoning that he has committed himself to in order to convince
the citizens of Warren County that they are at risk.
I will, however, provide you with my biographical sketch in support of my credentials and suggest
that you ask him to do th~ same. I can send you my full C.V. (28 pp), if you wish. I believe that I am
v.~dely recognized throughout the world as an expert in environmental monitoring, especially for organic
compounds in air. I wrote the defining chapter on sampling for organic chemicals in air for the American
Chemical Society's Professional Reference Book, Principles of Environmental Sampling (L. Keith, ed.),
and am currently writing a book on the subject for the ACS. I have also authored or coauthored eight
ASTM standards and two ISO standards on monitoring methods, including one on the methodology
(updated) used in Warren County (ASTM D 4861). Furthermore, Battelle Memorial Institute, which did
the field v:ork at the landfill for us, is one of the leading environmental research institutions in the U.S. I
would suggest that you get the opinions of two of the world's leading PCB air monitoring experts as to
the validity of our conclusions and Dr. Hirschhorn' s arguments. They are both very familiar with my work.
Dr. Terry F. Bidleman Tel. [phone]
ARQP Fax. [phone]
Environment Canada E-mail: [e-mail]
4905 Dufferin Street
Downsville, Ont., Canada M3H 5T4
and
Prof. Ronald A. Hites (current editor of ES&1)
School of Public Health and Environmental Affairs
Indiana University
10th. & Fee Lane
Bloomington, IN 47405
Tel. [phone]
Fax. [phone]
E-mail: [e-mail]
'a919 541 3527 \ERL/ AMRD /~lB
Biographical Sketch of Robert G. Lewis
ROBERT G. LEWIS is a Senior Scientist in EPA's National E>..'])osure Research Exposure
Laboratory, Research Triangle Park, North Carolina. He is responsible for overseeing research
focused on new sampling and analytical methods for pesticides, PCBs, dioxins and related
semivolatile organic chemicals and for their application to total human exposure assessment.
Dr. Lewis received his Bachelor of Science in Chemistry from the University of North Carolina,
Chapel Hill, North Carolina in 1960 and his Ph.D. in Physical-Organic Chemistry in 1964 from the
University of Wisconsin, Madison, where he was a National Science Foundation Fellow and a
>Jational Institutes of Health Fellow.
F oUov.ing six years v.,'ith industry, Dr. Lewis joined the EPA in 1971 and held the positions of Section
Chief from 1971 to 1978 and Branch Chief from 1978 to 1993 . He has done substantial research on
monitoring and exposure assessment methods for organic chemicals in air and other environmental
media :tnd in human and biological tissues and fluid.;. He established and headed for 12 years EPA' s
first ambient air methods development program. He was a member of EPA' s first task force on PCBs
in 1972 and in 1983 was appointed by the Governor of North Carolina to the Intergovernmental
Working Group on PCB Detoxification.
Dr. Lewis has authored or co-authored more than 200 journal articles, books, book chapters, and
other scientific articles and reports on various matters of environmental concern. He bas been
awarded the EPA Bronze medal for Commendable Service m'i.ce and has received four Scientific and
Technological Achievement Awards for outstanding journal articles.
Dr. Lewis is a member oftbe American Chemical Society (since 1958), Air and Waste Management
Association, American Association for Aerosol Research, American Society of Testing and Materials
(Fellow), International Standards Organization, Phi Beta Kappa, Sigma Xi, Alpha Chi Sigma
(Chemistry), Phi Lambda Upsilon (Chemistry) and Delta Phi Alpha (German). He has been listed in
American Men and Women of Science since 1966. He has served on numerous committees and work
groups concerned with environmental monitoring, exposure assessment, and the chemical profession.
He is currently chairman of ASTM D22.05 .02 on Organic Chemicals in Indoor Air, U. S.
Representative to ISO on Indoor Air, Research Area Manager for Environmental Measurement
Methods for the EPA Residential Pesticides Exposure Research Program and an advisor. to the EPA
National Human Exposure Assessment Study (NHEXAS), NCI/EP A/NlEHS Agricultural Health
Study, the Lower Rio Grande Valley Environmental Study (NAFTA), the Great Lakes Air Deposition
Project and the Pesticide Spray Drift Task Force.
i
l
~
. .,, ,.·
. ,•:•·.
D~err..ber 10, 1996 ............................................................................ by FAX (five pqes)
To: Technical Committee
From: Joel Hu-schhorn
Subject : 'llelpon~ to EPA Jetter of 3 Oeoember, previously provided
Io examining the £PA lett~ to Bm Meyer I focused on the mllll points and t have
ooncl1.1de--i! that I have no rta:on to abandon or change my original &.dings and ooncluaion&. 1
have fu:ed • request to the r.uthor of the EPA Jetter for a c.op>· o!th.e original contractor (Battelle
MtmDrial Institute) report, which, :f provided, may fe\/eaJ other mterestin_g ir.forrnation. It is
impon•nt to emphasize that only two doc:wnents h.v, A1.rfaced rcgardm1 the original 1983
measurements b;, 'EPA, th~ published paper ftom 198.S and one pas; of conclusions n-0111 what
ePA now says was I co11t1actor repon i,repued for it. Only the latter would have the testing
de!4ils allo....,'m.g the moEt complete analysis of-what actually happened during the testing.
'EPA h..u said that "our statements that PCB eroit,Sions from the l.a.nd.fill were negligible at
the dme ef the 5tady were suppc,rted by the highi:st le'\-d of t~bni~ review." The problem is
tba.t th~ peper itse~ u r noted pTe'\iou.sly had a logi~ inconsistency betWffl:1 two of its
m:atm:ents, o.ne s.aying the smi ... uion, were ~egligible .and the other that the m:xiS&c-~5 we!'e at low
levei~ Cle.arty, PCB emissions were mur.:ured snd reported in the paper. A21y re\iewer of die
pape( however, m.iy h,ve missed th! importance of the two statements U>d, moreover, may have
!nterpreted the statement about emis!ions from the WaJTe!l County landfill being negligjble as not
ne,esunly related to health effects. In fact, the paper presented no information about PCB levels
and heal-Ji. effects. Tht -paper wa.s t1'out PC':B emission~ from different land disposal situations.
oor poteutial health impa:=rs.
EPA is new .. def~ding its.el(b)· saying that the 1ct11aJ measurements o{PCBs "were likely
r,,~c-positivc rcsvh$." Tb.is .is a rc.miJkAble sutemeot. It wu not m.d.c m the original piper, nor
is there any C".idcncc to support thii c-0ntmtioo. Now, I!PA cll.ims that 11Jt would be cl:at to any
&rlalyti~al chcmi~ or anyone with ~omnental monitoring e,cpertist that these were likely false-
positive res:uhs." Notice the q,.uimer "likely." If £Cie!1tisu have a vi%)· sound tecb.nical reason for
belic-.ing that e~:rimen.taJ d!ta are f.\lse•positive result, (i.e., a &ding of 2 chemical when 11011.e
is re~Hy prt1:sent), they either would not repon the data in • publication or they would clearly
uti~ate why the dau should b~ \.'le\.\·t4i as false positives. Any objective analysis of tbe EPA
pub.tis.bed papen-·will &how tba.t there tfi no ,uuestion or proof that the reported ambient air PCB
1-:vcls were ftlse.pow.-es. 'For £PA to now claim that k, data were likcl;· false-positives is
cismgennou.s at best. ~d for ii to attempt to create explanations llOW but not in. the publisbed
paf tr is not normal profession&l conduct.
A part of the cwtent :.onteutior; about fal~e-positrves is that if the data >were "real" •· and
1
. ·--· .. •.·
Fax:[phone] Dec 10 '96 16:15 P.05
they ~enaicly were real enough tor the EPA parsons to report and rublisb them ~· thm the more
vo!atile Atoclor 1242 rather thtn the rr~uured 1260 would h,ve bcm measured.. But there are
various: pcssi'ble explanation. of ~y the less vo~tile and not the more volatile Aroclor could be
found .in ~bient air monitoring. These possible reasons include: varisblt ~es of PCB emwion,
!om tht land.fill at different times., making a small number ofmeamremenu in ambie!lt air not
refeetive of the cumulative types of emissions over time; a greater Joss of the more volatile
A.rc,;lof prior to the time of the measurement especiaDy becauu of'the maay months th•t the
wa.i.tcs were cx:posed to the atmosphere whim they were first dumped on the roads and continuing
through their di,posaJ in the open pit; diffc:rcnt environmental interawons dwiDg the air tran.sport
of the vapor!. leJ.dmg to different variation£ of the PCBs reachint a pea point ofme.ullfement;
the g.re3ter water solubility of.1242 veuua 1260, b)· a factor of about JOO, that m the very wet
m.atemls miide the landfill could e,cplain why ~11sider1ble 1242 could be trapped m dlc water and
no~ relea~ to the lir (i. e . the difference ill water solubility it much greater than the difference m
v.apor pres..~re); m.ote of the 1242 may have been biodegraded over the ~tire time the wastes
exiited prior to the gas v,aporizarion.
The letter from EPA note, that the more norm.al highet level of 1242 over 1260 was found
in th~ m.1m and upper leachate vent pipes. But of c.ourse it docs not oot~ that tb.e relationship
hetween the 1,,...0 Aroclo,s was not the nomial t}pe m three other categories reponed in the
origin;l published paper (lower leachate port 1.11d two small vents). Moil interestingly. for
exa1Dple, one of the mw1 vents (E) had Aroclor 1260 levels twice th.at of 1242, and at levels that
the detection i~sue is inelevant. Clearly these are emiscfons measured in tbe smaU veut wheie a
bubble had formed in the top plastic liner Cf more 1260 than 1242 clearly came &om this source
~flandfill gas, then why does EPA make r.ich c issue cf fin.dmg more 1260 than 1242 at my
off.site location.? EPA al.so male es the mhuke of •~ing that a nnn•detect Anding i~ eq,al tt>
zero concC!ltration ,. ·when in fa.ct that is a scieniliically incorrect and improper assumption to
n:.ai.e At ill the other ac:\bient testing locations wbere no positive ltvtls of PCBs were reponed,
tbe values m•y m f.act be ~mething less tlwi the detection limit.
The letter from EPA also atttmpts to mau a point that the ambient 1ir levels found "did
o,"lt appez con-e:late (sic) \\ith pro~· to the vents." In actual fact, there i£ problem in EPA 's
ir!terpretation because immediateily after the statement cited, the _paper tlsc said "The two
ma,um,.1.11:1 levels were found 98 m doY..'?lwind of the maio vcnt.'1 In other words, of the four
posr.i\·c finding, of PC.Bs m ambient ,ir reported as •1 I. 12, SO, and 71 ng/mJ.t, accordmg to EPA
the reaJinr;s of SO and 71 (now cwmed u not rul, but inalyzed nevertheless by EPA) were net
at t~c ho~sc. which was, according to the paper, ,ome 1000 m away. According to the paper, the
high,st level mea.uted was at the fenceline a11d logically iO waa the second highe&t level found .
Thit leav~s two very simibr but lower :PCB level$ found at two other local.ions., -wb.ich according
to Ttble IT of tbe papet were at the nearby howe and beside the main vent. The only
meuuren1ent, therefore, that i1 somewhat in'egular is the lower level found beside the main vcut.
because it seem~ mucb too low (ie., ii there are a sufficiet:1tlyl1J'ge number of muNremtnts one
expects to find the highest oonceotntious at the ~urce, with steadily lower concentrations at
increasing •iisur1c-es a"'t) fi-om the source). But the anomaly is the vety low reading nex't to the
2
Fax:[phone] Dec 10 '96 16:16 F·. 05
mam vent, especially because much luper levels were.me.asured in.Ii.de the main ,·ent and EPA
said that thert wn a pcsiti\'e gu ftow with PCB& bemg emitted from the main -vent However,
sui;l! low readjugs where higher levels are expected cu be explained by, for example~ varit;ble
cwission rates, variable wmd conditions., and poHt"bly different timea ofineuurements at different
locstions
mdoed, EPA w a big problem trymf to e,c:pwo why the bish levels tound inside the main
v~ did iot transla1e into much highe1 levels "beside main vent" "'bicb presumably was.
'"~rding to the figure sb.owiog the sampling locations. only I m away. The problem of course is
that variable wind condition, will Jead to vastlv different levell mealUl'cd at any given location,
unless many more meuuretnent, arc taken ov~r time. Tbe original pape-r noted that wind speed
\·aricd from 0.04 to 6.6 r.r,,/s, and that the atnbient air monitoring was not done on the same days
that testm~ took place in the vents. The utter fa.ct suuests tha.t highly variable emissions from
the lmdhll wete occurrins, expl.ammg why a vet) lbw reading might obtained near the main vent .
Aho. \he paper indi~ates that the mient air monitoring ocCWTed on different days, explaining
why J»CB lev~ls would chaDge signmcant)y for di.tfcrent locations.
The origmai EPA paper wd that ambient air data were given in its Table 11 and th.at "only
four ofth~ 39 ambient ,air sa:q,le, aul)-z~ contained detectable quantities of PCB" and that
"three ~mples were positive for Aroclor 1260 only (at 11, 12, SO, md 70 a.g/m')." The problem
is th.at in T ab1.-ll there ,re three foomotes for etch of the three data showing pogjtive levels of
PC'Bs detected. as follows:
beside main vent: "1 One of i.ix measurements above detection limit."
fence line, downv.iind: "h Two of' 13 mea~ements Above detection limit (0.05 and 0.07
\lg,'mJ)_.,
nurby house : •• One of six m.easurtments above detection limit."
There is a di.screpa.n.cy betwccn these footnotes and the body of the paper's text. The total
cumber of.measurement seems like 25, not 39. Another problem is th.t the tabu.lltcd data
i:ldicale that the detoction limit was lO ngtm3, but the second £ootnote indicates that the det~tion
lhnr~ was from ~O to 7:) ng/ni', and both the original paper and the reoent EPA letter &aid that the
dc:te,.,-tion limits wece .fro01 IO to 20 ng/m3· while the p ■ge of c;onclusioas fiom EPA 's coDtractor
repo.rt said the detection limit• were 6 to 10 ng/mJ. Only cJose examination of the original
11.boratory d;t1 shHts that might be in the Battelle ccmuactor report co-uld reso'l\'C these
discrepancie&.
EPA', letter cl1ims th.It "the vent ptpes were not the source of the PCDs detected in air,
even 1fthe measured air values were re,1.1' First. EPA is acknowledging that tb.e data may
&er.ally be accurate positives ind, second, EPA is speculating that some ether source nf PCBs
eXlSled around the landfill. In fact, jn its origmal publi.Ebed paper, EPA said that "air levels were
at or nur backe.rou.nd" at the Wure:n CoW1ty landfill, YAlich "-ientm~aDy is a very different
ar,-.mient tbm its present ooe based on r.laiming that the fi.odings wer:e falsie-positives . In other
3
•.
Fax:91'3-257-1000 Des 10 '96 16:17
word&, originally EPA disreguded the levds of PCB, n meuured iD ambient air by arguing they
were "at or ucar backgound11 and now by arjtling that they were not really positive findings. But
in fact, e\'-;u the £PA ar~t chat the a,euu,red (real) PCB lC\-tl5 were only bacqroUDd docs
not sund up under clo,e 1C?Utmy.
EPA cites the ceneral data for the U.S. about hishl)• Vlriable ambimt air levels of PCBs,
\\ith the maximum be:!J8 10 n,Jm', which is actually less than all four positive findings reponed in
tht s,aper ( 11, 12, ,o ud 71 ng,'m') More importantly; the range given in EPA 's Jetter (md
10\'ell origmally in EPA's contrietor report. but not the published paper) doelii not distinguish ihc
typ~s ofiocations at "1lich PCBs have be~ meamred In ua. the higher level$ of ambient PCB
le\'el.!1 con-espond to more industrial and urban areas, and areas near any eype of incineration. of
w.ste. not the rural areJ at which the Wan-en Co\Dlty land.fill iJ located. Lcgicall)·, the low =d of
the range reponed, na:nely 5 n.g!m', might be appropriate f.or mch I rural looatiou, but the ~orrect
scimt~c method m ~cb an i:avutigation i.( actually to take measurements at a subsuntial di£tance
from the lik.tly point source (i e .. the landfill) but within the loeal geographic vea. Then, this
t!xperimer.tal backs:ound level 1pptopriate for the fflldy lhould be used. But assuming that the
low e!ld of the national rar,gc my be a.ppropriate, the four positive findmgs are significantly
higher, undumining .EPA', arpmect.
The letter &om EPA cwms that the ''The Warren County site wu brand new and should
h:\e been emitting ;t a maJ.imum rate." This is sheer nonsense and perhaps more than ~Yother
EPA .statt'ment d:mon.stntes that EPA pcnon1 ue trying very hard to cover up t.be truth. My?
The~e ue variovs sound, sdcntific reasons why lmd..611 ~D rates were not n~ssarily the
highest at the timt of the .EPA measurements, in~ludmg: the tempenture at the time wu low (-1
to • l 4"C) relative to summer periods ~ium high emis:lion rates would occur; the complex mi.'<Nre
of materials &ctua.lly placed m the laJ:dnll ud the dispo~l process nece"arily trap, PCB vapors
that take time to find triiMport pathw2ys; the large amount of water present m the materials
buried 1.1'1 !lie landfill from heavy rainfalls would slo'Wly sink and accumulate to the bottom of the
l.aDdfi!l .nd the internal dr)ing proce$S (for some of the buried waste) would more easily allow
PCB vaptJrs to escape over time.
lastly, f.PA is arguing that its computer modebng (the details of'wbich were not given in
the piblished paper but m.-y be in the contractor report) sbowcd that eilcuhtcd (not measurtd)
le..-els of PCBs at some distuce tW&}' from the land.fill would be "far below the dctcctio11
c:1pabiliry of the sampl!r employed " But such dispersion modeling is based on mmy technical
asswn;,tions about the ~urce of e:niscio:1~ and many clli!1Atic conditions that CINlOt be usessed
at this p~int. Implicitly: in It& original published paptr, EPA was atpmiE that the theoreifoal
modeling .3howed dau below The PCB levels actual/>• 1ttetU1AHd 2t cli.stao.ces away from the
laud.fill. But EPA did t!,1Jt actually say in Jts published paper that the modeled data were below
bacl-:grou.nd level.I . M?rt importantly, th! eore bari~ for EPA'i modeling were only three
meuuremmts i,(the gas flov. from the main vent and wme un'°'tated data for what was meaS1Lted
in t.'ie .ff.Wtl VCDt. leading to what EPA descnbed u an l\'cr&ge gaseous PCB emission rate. Toe
p10blem. of c-our~. is that there were only a small number of meuu.reme:nts dwin.a a wmter
,..-_
F CB u.lORKING i3ROUP Fax:[phone] Dec 10 '95 16:18 P.08
period. reducing the reliability of the spmfic •• used by EPA iD its modeling and undcmJiaicg
u.y objective confidence in the results of the modeling. In truly objective and fair work, field
mea.i,--urements o{ air levels of a to>eic substance are used to evaluate whether the modeling is
accurate and reliable. But in th.a case. there was not even an e,cpliQt analysis by EPA ofiu o~n
positive meam.~ts and it, model..~g results.
In cond.tuion, the UDPQt&a;at dcficicpey of tbe EPA work ,r11 Cb•t they oner
d.e.tid~d...lbat then wu ,umcicot uuon, to perform much more complete tcstin1 for PCB
cmi,uiocs from the \\iarrcn Cau.qty landfill, EPA, of courte, htd I eonflict or intcrc,t,
bcq1,uc it bad approved the ~men County PCB landfill and the obj«tirc of the 1983
,o,dy ~ccnu to hue been to prove that aover.nmcot eanctioncd PCB 1,ndfills were "11fc"
relative to wt(.O..otr...Olltd toxic wut.t JU" with PCB, that rcguir,td dt1DJ1.D, Since tbc
~r,:c.a_County PCI J,adflY wssJbr cleanup for tbc aciri-P1J PCB road dpmpinr •itu,atio»
1.b..1.1-I.P A.a1unarcd and fund«!. it clc.arb· wu oat in £PA•, intereu to rtoort u,nificapt
and pot.tntialti dannroue 1,vth of PCB air cmi,,sion• from ita controlled landfill, Also,
.t.rA ,., pubtiahea PAPtt, submitted foe pu.blication ;o Duemhtr 1984. about two years after
the Jaodf'ill wu conr;tz:utted, described the \Varreo Count):'. landCtll has baviu the
pcrlorakd .oh>, 5.XU~m that we now know '1:11 oat insulled. l;PA 's 4cu:ripti0Jl of Im
laudtlU .u :;UtO:D[~tbe:-art was consi1teot with what EPA ,note.d to portray, namm a
'-IUJ..trollt:d landfill that b,-~o.m»ari.ton would show how bad the uncQntrolled Indiana PCB
s,tcs~c. But tPA•s dcuriptton !!'U inaccurate and, pabuu, iotcgtionally ,o it aa lbc
m.tc new dairru, tP.~Jia~Jf appro..,ed a t:blDH iD tbc dc,iga of the Warren Counu; 11od!!U
cJ11, wouJd have bad to Q'"" in J98l and that allowed the statt to omit wine the
pctfwtc4 Jue.bate can«tion 11ioe system at thr bottom or ilte landfill.
5
I I I t
l.2:03:96 i 3:01
I
ft
~
DATE:
TO:
'6'919 5-H 352i NERL: Aillli> : ~B
U.S. ENVIRONMENTAL PROTECTION AGENCY
NATIONAL EXPOSURE RESEARCH LABORATORY
AIR MEASUREMENTS RESEARCH DMSION (MAil. DROP 44)
RESEARCH TRIANGLE PARK, NORIB CAROLINA 2771 l
TELEPHONE [phone] FACSNILE 919-541-3527
E-MAIL lewis . [e-mail]
FAX~SSAGE
3 December 1996 NUMBER OF PAGES: 3
(including cover)
Bill Meyer Fax No.: 715-3605
Telephone:
FROM: Dr. Robert G. Lewis
USEP A (MD-44)
Research Triangle Park, NC 27711
SUBJECT: Hirschom Report on Warren County PCB Landfill
MESSAGE:
14!001 .-003
I was out until today and have just reviewed the above-referenced report that you faxed to
me yesterday.
The 1983 EPA paper ( on which I was lead author) that Dr. Hirschom attacked in his
report was published in Environmental Science & Technology, a peer-reviewed journal widely
recognized as the leading journal in the world on environmental monitoring. Prior to publication,
it underwent rigorous technical review within the EPA and by two or more leading experts
outsid~ the Agency. Therefore, our statements that PCB emissions from the landfill were
negligible at the ti.me of the study were supponed by the highest level of technical review.
It is clear from the data presented in Table II of our paper that no significant PCB
concentrations could be measured in the air surrounding the landfill, even at one meter distance
from the main vent pipe. As the paper states, "only four of39 ambient air samples analyzed
contained detectable quantities of PCB." These were positive only for Aroclor 1260 (the least
volatile of the two Aroclors monitored) and "did not appear to correlate with proximity to the
vents." It would be clear to any analytical chemist or anyone with environmental monitoring
expertise that these were likely false-positive results. The levels detected were 0.01 to 0.07
µg/m3, at or very near the method detection limit (0 .01 -0.02 µg/m3). Had they been "real," the
more volatile Aroclor 1242 should have been found at much higher concentrations, as it was in
main and upper leachate vent pipes (see Table I). The lack of correlation of the measured ambient
air levels with proximity to the vent pipes, taken with the fact that Aroclor 1242 was the only
analyte identified in the air even though Aroclor 1260 was at much higher concentrations in the
vent gas, conclusively demonstrates that the vent pipes were not the source of the PCBs detected
in air, ~ven if the measured air values were real. It should also be noted that PCBs were
l
l2 t (IJ :96 · 13:01 'a'919 5-H 352i NERL/ AMRD : ~ [4)002 .-003
ubiquitous in the ambient air in the United States, typically at 0.005 to 0.01 µg/m3, at the time of
the Warr en county air monitoring effort.
Our air monitoring results were also consistent with both the air emissions models applied
to the vent measurements. As we stated in the paper, the models predicted that no measurable air
concentrations of PCBs would result at 14 meters from the main vent and beyond.
In his analysis of the data presented in our paper, Dr. Hirsbom refers to measurements of
"Arocblors" (sic) obtained "at" the main vent (120.2 µg/m3 for Aroclor 1242). The vent pipe
measurements shown in Table 1 were made within the vent pipes, not in the air at or near the
vents. Hence, they cannot be compared with the ambient air measurements made at the
uncontrolled sites in Indiana and reported in our paper. There were other major differences
between the Warren County and Indiana sites. The Warren County site was brand new and
should have been emitting at a maximum rate. The Indiana disposal sites were old (before 1972,
as stated in our paper). The PCB-contaminated soil in Warren County was underground, in the
thermic temperature regime, and should not have been affected by ambient air temperature (i.e.,
emission rates should have been the same whether it was summer or winter at the time of
monitoring). In Indiana, PCB-contaminated surface soil was abundant; thus, emissions were
greate: in summer when the sun heated the soil. As we said in our paper, the PCBs measured in
the vent pipes in Warren County were be transported by methane from decaying organic matter
(primarily grass excavated along v.ith the roadside soil) and should have declined greatly once
that matter had decayed. Therefore, Dr. Hirshorn's statement on p. 4 of his repon. that the "high
levels found at the Warren County Landfill were quite comparable, and perhaps even greater than
the levels found above the other three uncontrolled landfills" is absolutely without merit.
As lead author of the ES&T paper, the interpretations of our findings were principally
mine. However, the report provided to EPA by Battelle Memorial Institute, who did the air
monitoring for the Agency under contract, reached the essentially the same conclusions (see
attachment).
Anachment
'
"0'91\J 5-ll 35l7 !\ t.K L : .'UlKU : .lli:i
SECTION 2
CONCLUSIONS
1f!J l.J l_J .j: l_l l_l ;j
The conclusions drawn for results of this study are su1T111ari2ed below.
(1) The principal source of emissions from the landfill originate from
the main vent pipe. The average PCB concentrations of Aroclor 1242
and Aroclor 1260 measured in the main vent emissions during the
study were 123,000 ng/scm (~100 ppb) and 2,000 ng/scm (_...2 ppb),
respectively. These concentrations are substantially lower than
the current occupational standards for workplace atmospheres which
range from 0.4 to 0.8 ppm.
(2) Ambient air PCB levels on and surrounding the landfill site (even
as close as one meter from the main vent) were found to be at or
below minimum detection limits (6 to 10 ng/m3) for the sampling
method. PCB levels generally present in the atmosphere throughout
the U.S. are in the range of 5 to 10 ng/m3.
(3) Mathematical modeling predicts that ambient air PCB concentrations
on and in the vicinity of landfill resylting from the main vent
pipe emissions may be approximately 106 to 109 times lower than the
detectioM limits for sampling method used in this study. PCB
contributions to the ambient air from the landfill based on the
model predictions are insignificant when compared to general
ambient air PCB levels in the U.S.
(4) It is anticipated that the low PCB emission rate from the landfill
will be reduced still further as decay of organic matter producing
the methane and other gases emanating from the vents subsides. The
reduction of hydraulic pressure by removal of water from the site
should also reduce ·emission rates substantially.
2