NCD980602163_19961010_Warren County PCB Landfill_SERB C_Joel Hirschhorn - Responses to 3 December EPA letter-OCR

. ' ·. I •' .
Dec~bet 10, 1996 ... · ........................................................................ by FAX (five pages)
To: Technicll Committee
From: Joel HuschhoT!I
Subj~ct: 'R.espon~ to EPA Jetter of 3 0¢cembet, previously provided
In examining tlie EPA lc:tt~ to Bill Meyer I focused on the m.un points and I have
con~lude-a that I have no r~a:on to abandon or cban.ge my original &.cling$ and conclusion&. 1
have faxed :a request to the illihor of the EPA letter for a copy o(th~ origin.al contractor (Battelle
Menwri31 Institute) report, "11ich, :f proVlded, rn.ay reveal other -intere~1ini ir.fonnation It is
impona.nt to emphasize that only two documents have sarfaced regard.mg the origmal 1983
me1surcmeots b:;, EPA, the publlihcd paper from 1985 and one pag; of conclusions from what
'EPA no,•: says was a ,ont.ractor report prepared for it. Only the latter would have the testing
dewls allov1.,mg the most complete analysir. of what actually happened dwing the testing.
EPA has .said tlut uour statements that PCB cro1$s:ions from the Landfill were negligible at
th: ::irne ef the 5tudy were suppc,rted by the highest lcvd of t~hnical tevicw." The problem is
tha.t th~ paper itse~ as ! noted pre'\iously b.ad a logical inconsistency betwea, two of its
statffllents, one saying the 11n,;_uioni. were negligible and the other that the :mis.si.ons we!'t at low
leveis:. Cl~l,·, PCB emissions were mea~ed ud reported int.he p2:per. Ally r~ie\Wer of the
paper, how~··er. nuy h,ve missed th! importance of the two statell\eftts &11d, moreover, may ha-.,·e
mteipreted the statement about emissions from the WaI"TC!l County landfill being negligible as not
ne,cssanly related to health effects. ID fact, the paper presented no information about PCB levels
and heal'.1 effects. Tht paper wa.s trout pCB fflliisions from different land dhl)Osal situations.
aot poteutiaJ health impa:=rs.
EPA is nl)w defending its.el£by sayin& that the 1ct11aJ measurements orPCBs 11wtre likely
f,l~c-positive results.'' Tb.is .is a .tc:m.llklble statemeo.t. It wu not llUldc ir. the original paper, nor
is there any C"--i.drocc t·:> support this c-0ntmtion. Now, .EPA cla.ims t.bat "It would be cl:at to suy
Ulnlyti~al ,;bemi~ or l!Oyone with environmental monitoring e:q,crtise that these were likely false-
positive res:uhs. n Notice the q,.uimer '"likely." If ;,cie!tti.sts have a vny sound technical resson for
beiie-.mg that e'.l:FerimmtaJ data are false•positive res;uin (i.e., a &ding of 2 chemical when none
;s rellfy pr~sent), they either would not repon the data in a publication or they would clearly
uticufate why the data should bet '-'1~'e4 as false positives. Any objective analysis oft.be EPA
pubti~ed pape-r ¼-ill &how that there t!i no su.gsestion or proof that the reported ambient air PCB
lc:vcls were f!\se.pow:es. 'Fo.r EPA to now claim that ju data weYe likclJ false-positives is
ci.ismgennous at best. ~d for it to attempt to create explanations now but n~t in the publisbed
paf et is not noI'ID!.l profession'1 conduct.
A part oftlle c,ment =-onteution about fals.e-positives is th.at if the data \.\-Cre "real" ... and
.· ._ .......
Dec 10 '96 16:15 F·. OS
they ccrtai!:ly ·were real enough tor the EPA p!!rSOns to report and yublisb them •· then the more
vo!atile Atoclor 124 2 rat.her :ha.n the rr~as.,,ired l 260 would h¥ve been mearured.. But there are
varic,ur. pcs.sib)e exptaoation:. of v..hy the less volatile and not the mote volatile Aroclor ~ould be
found in J.fflbient rir ltlOnitoring. These possible reasons include: vwbl~ types of PCB emi.ssions
nom th.~ landfill at different time~ making a small number of meas:urements in amhiet2t air nor
reflective of the cumulative types of emissions over time; a greater loss cf-the more volatile
Arc,;lor prier to the time of the measurement eq,eciAJh· because ofthe mny months that the
wc..'1.cs were c,q>osed to the ttmosphere 'When they were tirst dumped on the roads and continuing
through their diaposal in the open pit; diff'c:rcnt dlV:iro.nrncutaJ interact.ions dwing the air trau.sport
of the vapors. leP.dmg to different variation& ofthe PCBs reach.int a given point ofme.uurement;
the greater water solubility of 1242 vers1u t 260, by a factor of about 100, that .in tbe very wet
l'tWm.tls mtide t.he landfill could exp!am why ~Dsiderablc 1242 could be trapped in the water and
not relea~ to the air (i.e .. th~ difference in watet solubility is much greater than the difference in
v.tpor pres..~re); moie of the 1242 tnay have been biodegraded over the entire time the wastes
exmed prior to th.e gas v11porization.
The letter from EPA no:.e, that the more nomal higher ll!\-·el of 1242 over 1260 wu found
i1:1 th~ m..1m and upper !each.ate vent ~ipes. But of course it docs not ootc: that tb.e relationship
r.-erv,1een the 1,11-·o A.roclors was not the noi1'1Jal t;pe m three other categories reported in the
orig.in.I publishe-0 paper (lower leachate port and two small vents). Most interestingly, for
example, one of the mw1 vents (E) had Aroclor 1260 leveb twice th&t of 1242~ and 2.t levels tliat
-&'.ht! detection i!.!ue is irrelevant. Clelrly these arc erois'1on.s measured in the small veot ,s,,het: a
bubbk h11d fonncd in the top p~stic liner. Cf more 1260 thlll 1242 clearly c.me ftom this source
~f1andfill gas, then wh>· docs EPA make ~.ich an issue of fin.dmg more 1260 than 1242 at any
otf.sitc location? EPA also maxes the mistake of •~~ing that a non•detect finding i~ !qml to
z.ero concC!ltration:, \\<ben in. tact that is a scientmctlly incorrect and improper uwmption to
mah At ill the other atcl)ient testing locatioM where no positive l~els of PCBs were reponed,
the vthies r.nay in uct he )Omething less thlll the detection limit.
The letter from £PA also attempts to make a point that the ambient 1ir levels found "did
t::i,")~ app= correlate (sic) '\\ith proximity to the vents." In actual fact, there 1' problem in EPA 's
mteTpret:uion because immediately after the statement cited, the paper tlso said "The two
m.axim,1.11:l levels were found 98 m downwind of the main vent.'' In other words, of the four
,:,osr.i\·c finding, of PCBs m ambient ,ir reported as .. 11, 12, .SO, and 71 ng/mJ", ~ccordmg to EPA
the teAJings of SO and 71 (now cwmed u not re.al, but inalyzed nevertheless by EPA) were net
at th~ ho.isc. which was, according to the paper, some 1000 m away. Aocordmg to the paper, the
high~st levd mea"1ted was at the fencelin.e and logically so ·w2i the second highest level found.
Th:t.t leaves two very ~br but lower PCB level$ foWld it two other locations., which accordini
to Ttble IT of the papet were at the nciuby house and beside the roam vent. The only
meisurement, tt,ereforc, that is som~what in'egular is the lower level fo\llld beside the main vcut.
because it seems much too low (ie., ii there are a sufficiE11tly luge number of meamrements one
expects to find the hight~ ooncentntious at the ~urce, with steadily lower concentrations at
increasing -iisunc.es a"'t) fi-om the source). But the anomaly is the very low reading next to the
2
Fax :91<_:;-257-11J00 Dec 10 '96 16:16
JI12in v~t., especially btcau&t much higher 14'Vels were me.a~ed inlide the main -..,·ent and EPA
~aid that there was A pcsitive ga, ftow with PCBs bemg emitted from the main vent . However,
su~b low reaclings where higher levels are expect~ cu be explained by, for example. vari';hle
emission rates, variable wmd co:1.ditions., and por-su.'bly different time, of meuurements at different
locations
mcieed, EPA has a big problem trying to e,cpwn why the high levels found inside the main
v.?ttt did iot translate into much higher levels "beside main vent" ~hieb presumably was.
ae<:~rd;ng to the 6.gw-c showiog th.e sampling locations, only l m away. The problem of course is
that variable v.ind conditions will lead to vastly different levels meaSUTcd at any given location,
unless many more meuurements arc taken over time. The original pap~r noted that wind speed
\'aried from 0.04 to 6.6 mis, and that the arobicm air: monitoril"...g was oot done on the same days
that. testmg took place in the vents. The atter fact St.t~ests that highly variable emissions from
the l.andhll were oc.curring, expl.ainmg why a Vet)' low reading might obtained near the main vent .
Also, the p:iper indicates that the ambient air monitoring occurred on different days, explaining
why PCB ievels would chal'Jge significantly for different locations.
The original EPA paper said that ambient air data were given in its Table lI and that "only
four of th~ .39 ambient .tlr s.a.:i.ples MUl)-zed contained detectable quantities of PCB" and that
"three ~mples were pofflive for Aroclor 1260 only (at 11, 1~ SO, and 70 ng/m')." The problem
is that in T,blf' II there 4re three footn0tes for e•eh of the three data showing po&:itive l.c'vels of
PCBs detected. as follows:
besi.de nwn vent: "1 One of liix measurements above detection limit."
fenc: line, downwind:"" Two of' 13 mea~ements above d~tection limit (0.05 and 0.07
ug/mJ)_''
neub)· house: •• One of six meuureme11t~ tbove detection limit."
There is a d~r~ancy betwCCtJ these footnotes and the body of the paper's tc::xt. The total
t!umber of measurement seems like 25, not 39. Another problem is th.t the tabulated data
i:ldicate that the detection limit was lO t1glm'. but the second footnote indicates that the det~tion
Hrnr~ was from SO to 70 ngirn3, and both the original paper and the recent EPA lett~ said th.at the
dc:to,..,-iion limits were frotn IO to 20 ng/m'· while the p ■ge of c;onchisions from EPA 's contractor
repo.rt said the detection limits were 6 to 10 ng/m.1. Only cJose examination of the original
~boratory d;.ta sheets th.at might bt in tht Battelle ccm.uactor report coiild resolve these
discrepancie,.
EPA', letter claims wt "the vent ptpes were not the source of the PCB s detected in air,
even 1f the measured air values were real.'' First. EPA is acknowledging that the data nay
acr.ally be accurate positives and, secon.d, EPA is speculating th3t some other source o.f PCBs
existed around the landfill. In fact, in its origjoa.1 publi..~ed paper, EPA said that "air levels were
.tor ne1t b1ckerou.nd 11 at the Wmen County landfill, which &denti!-:;ally is a very different
arg-.mient than its present ooe base4 on claiming that the findings were falS&positives. In other
3
Dec 10 '96 16 =17 r'. o?
word&, origin.ally EPA disregarded the levels of .PCBs it meuured m ambient air by arguing they
w:re "at or uear backgoWld" and now by arguing that they were not really positive findmgi. But
in fact, ev-:u the £PA ar~t that the mea~ed (real) PCB levels were only background docs
not sand up under close scrutiny.
EPA cites the general d&ta for the U.S. about hishl}· variable ambient ait levels of PCBs,
\\id: the maximum be:!l8 10 ng/ml, whieh is actually less than all four positive findings reported in
tht paper (11, 12, SO ud 71 ng/m') More imponantly; the rangc@ivm in EPA'slctter (md
grveo origin.ally in EPA's contractor report, but not the pub1ished paper) does not distinguish the
typ~s of iocations at which PCB s htve be~ measured In ua, tht bigber level$ of ambient PCB
lc,.·eis correspi>nd to more industrial and urban areas, and areas near any type of incineration. of
w~e. not the rural a.re.i .at which the Warren CoUllty landfill u located. Lcgicall}·, the low e:tid of
the: range n:ported, na:nely S n.g/m', might be appropriate for ruch • rural location, but the correct
scientJic method in rucb an mvetttigation is actually to take measurements at a suba;iantial di&tance
from. tha likely point source (i e .. the la.ttdfill) but within the local geographic area . Then, this
experimental backs::-ound levef appropriate for the mJdy should be used. But assuming that the
low end of the national rl!lgc may be appropriate, the four positive findings arc significantly
higher, m.1.dumining EPA's argumeot.
The !etter from EPA cl.aims that the 't'fhe Warren County site was brand new and should
h:.ve b~ emitting ~t a mal.imimi rate." This is sheer nonsense and perlups more than any othe-r
EPA statement demonstrates that EPA penon1 are trying very hard to cover up the truth. My?
The:-e are various. sound, sdcntific reasons why landfill cmis..41ion rates were not necessarily the
highest at the time of the EPA measurements, inc;ludmg: the temperature at the time was low (-1
to • l .i"C) refative to swnmer periods v.he:n high emisgioo rates would occur; the complex mi:<ru.re
of materials actually placed in the la.ndfill tnd the dispo~l proceas nece,.arily traps PCB vapors
that take time to end tr:tMport pathwiys; the latge amount ofwaterpreieJ!t m the materials
buried u, tlie landfill from heavy ram.falls would slowly sink and accumulate to the bottom of the
l.a.odfiH :md the internal drying process (for gome of the buried W!lste) would more easily allow
PCB vapors to escape ov~ time.
Lastly, EPA is arguing that its computer modeling (the details of'which were liOt given in
the published paper but m~y be in the contractor report) showed that calculated {not me&sured)
leve!.s of PCBs at some distADoe away from the landfill would be 11far below the detection
c1pabiliry of the sampl!r employed •· But such disp=rsion modeling is based on mmy technical
as~tions about the l>Ource of e:mssions and tnJJ1y cbtic conditions that cannot be assessed
at this p.'.)int. lmpli~itly. in rt& otignw publimed paper1 EPA was arguin~ that the theoretical
model.mg £owed dau below the PCB levels actuall>· 1tteasMHd 2t distau.ces •"-'iY from the
latidfill. But EPA did n.•)t actually say in its published paper that the modeled data wer~ below
background level&. Ml)r~ import.mtly, th! core bu:is for EPA') modeling were only three
measurements ,,fthe gas flo"' from the main vent Jnd wme tm1,t2ted data for what was measured
in the .nwt> veot. leading to what EPA des:cnbed u an average gaseous PCB emission rate. Toe
ptoblem. of cour~. is that there were only a small number ofmeuuremeats dwing a winter
F CB o.lCIF:n NG GROUP Fa.x :919-257-1000 Dec 10 '96 16 =1 8 F·. 08
period, reducing the reliability of the specific figure used by EPA .m its modeling and undermining
my objective confidence in the results of the modeling. In truly objective and fair work, field
me~--urements of air levels of a toxic substance ate used to evaluate whether the modeling is
accurate and reliable. But in this case, there was not even an explicit analysis hy EPA ofiu oY.n
positive measu.~ts and its modet~g results.
l!u.9nchHion. the i,,mpgrta;at deficiency of the EPA work waa U,at they oner
d.e.tid~d.Jhat then wu sufficient uu.o.n. to perform much more complcU testio& {or fCB
emtuiocs ftam the »·m:cn County Iandfiil, EPA, of course, hed a conflict of interest.
bcq:;uc it had approved the »:men County PCB landfill 4nd the objective of the 1983
,u,dy ~ccm1 to hayeJ>.ecn to prove that 1ov~11ment tan.t.tiontd PCB hmdfills were "safe"
;-ciatiye to utt.to.n.tr...oUc-d toxic wuttJittl with PCBa that ceguircct clc;MUJ, Since the
\trL('!.e.tLCPunty PCB LJpdfiU mu thr, cleanup for tbc acie:i,,nal PCB ta•d rtumpwr dtuatio.o
tMJ.E.PA 111proyc:d and funded, it clc.atb' was not in EPA 's intereu to r~ort s,i:nificant
;nd potentialf)' dan,crau! ltvels of PCB air cmi,sion1 from ita controUcd Iandfi.H. Alao •
.tr#+.'! publiahed paptr, 3uhmittcd for publication ip December 1984, about two ytars after.
the landfill wu ~on:;trutie:d. described the \Varren Connty landfill bas haviu the
pcrloraud Pilll :UU1m that we now kQow ms got instilled, tPA 's 4cu;tiption °( lhc
landfiU ;,,us,te-o(~the:ut was coruisknt with what EPA wsnte.d to portray. namdv a
~.tt:!2lle:d landfill that hl'.-'.o.mllJlrison would show how bad the uncontrolled Indiana PCB
.s.itu~c. But EPA •s .dcuription wu inaccurate and. perhaps, intentionally so if. 11 the
ru.tc: nnw dainu, tP.4.Jiutf anpro-.,ed a tbauu jn the de,ign of the W,cnn County JcoclfilI uie, would have had to ~c.ur in J98l and that allowed the statt to omit usinc the
pmonted Jeubatc colfoct;ion 11ipe system at thr bottom of tbe landfill.
5