Sourot R*o«ptor Methodology tor
8oa« Chloriivatvd Hydrocarbons
SRI International, M«nl<> Park, CA
Prepared for
Bnvirona«rital Sciences Research Lab.
Research Triangle Park, NC
Jan 84
V I
1
Hi SfiygfMHg si Ssm&w
MBlkAMMiwiitoiiti
-------
i-LU-lttjIou
rlPA-600/3-84-023
January 1984
SOURCE RECEPTOR METHODOLOGY FOR
SOME CHLORINATED HYDROCARBONS
by
F. L. Ludvlg, E. M. Llitoo ..ad L. J. StUi
Atmospheric Science Center
SRI International
333 Ravenswood Avenue
Henlo Park, California 94025
Contract No. 68-020416
Project Officer
Jar is L. Cheney
E»1ss1on< Measurement and Characterization Division
r ental Sciences Research Laboratory
.sesw.ch Triangle Park, North Carol27711
ENVIROWtNTAL SCIENCES RESEARCH LABORATORY
OFFICE Of RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGP""
RESEARCH TRIANGLE PARK, MOST' ""
-------
mnxTss
P»A-60Q/>-a*-023
TCCMOCAL II VOffT DATA
«ww— Ufmt m
IBS
i liTU MotmnTiJ
•0CK3 UUKUI MRBODOOOCT It*
CXLOftXMTXO ITMOOMOM
RVOOTMTI
Jmmry 1984
I, WMOIMiM OOOAJHAAttOW COO*
I. HA*o«iiiiiia 6»sMd to estimate expected auirai grou'td-level
centrations of ^ rvitarl - rest downwind of the source area. A suitable
r-nnuT&i. 1.U4. i»g plant was sei.oO&*4 VV i.lcAa Materials Cco^ny near Wichita, Kausas) <**-*•
the aethod waa applied to estlaate eadssion rates for four different balogensted
hydrocarbons. The feasibility of the methodology was daaonatrated and suggestions
for i*prov»msata were made.
m »»oooa **»o oocuMawr »>**.Tsm
k OUCMfTOM
kLewnfitwoN* «mo«o iimi
«. cosati r^CM^
11 o<(TMi(urio«
KELEABI TO PUBLIC
It UCOhTy{Ult(fbi'irr
-------
NOTICE
TMi docuMtit has b#«n r*vi#w«d in accordance trlth
U.S. Environmental Protection Agency policy and
approvod for publication. Mention of trade uati
or cowercial producta does not constitute endorse-
ment or raco—endation foT usa.
11
-------
ABSTRACT
A source-receptoi methodology la described that can b« used Co
estimate Mission rata* of halogenated hydrocarbons fro* a manufacturing
plant whan accass Co cba plane Is not possible. An InarC trscar Is released
at a known rata from a vehicle traveling back and forth on a road out-
sida Che plant araa. Samples ara collactad downwind of Cba plant (ac a
distance of about 1 to 5 km) and analysed for Cha cracar and tha materials
of lntarast. Tha ralatlonahlp between Cha aalaslon ratas of Cha Cracar
and cha materials of lnCarasC, and cha aaasurad concantratlons has baao
darlvad. Tha aethod Is generally Insensitive"to meteorological conditions.
If appllad ac night or undar overcast conditions during cha da/. Ic la
suitable for estimating fugltlva Missions from aourcas within 10 or 13 ¦
of ground loval. Ic suae ba appllad with discration, if lncarfarlng
sourcas ara prasant In Cha aras. Onca asIssIon races have bean determined,
conventional Gaussian methods aay be used Co estimate expected maximum
ground-level concantratlons of cba materials of incarast downwind of tha
sourca araa. A sultabla manufacturing planr was selected (Vulcan Materials
Coapany near Wichita, Kansas) and tha aethod was applied to estimate
amission ratas for four different halogenated hydrocarbons, tha feasi-
bility of Che methodology * is demonstrated and suggestions for Improve-
ments were nade.
Ill
-------
coKTnrrs
Abstract . Ill
Flguraa vl
labia* rlt
1. Introduction 1
2. Thaory 5
3. Taat ilea 11
Crltarla uaad to aalact tha taat »lta 11
CudU«ti taat altaa 12
Taat alta—tba Vulcan matarlala
coaptajr, Wichita, boua . . . 14
4. Taat Procaduraa 21
Caaaral 21
Mataorological obaar-vatlona 22
Traear ralaaaa procaduraa 22
Saapla collact loo procaduraa 24
Saapla analyala procaduraa 25
5. Taat luulta 27
Cacaral 27
Taat of 8-9 Auguat 1981 27
Taat of 10-11 August 1981 30
Taat of 11-12 August 1981 36
Taat of 12-13 Auguat 1981 43
6. Intarpratatlon of Kasulta 49
Caoaral approach 49
Eatlaatad aaiaalon ratal 50
7. Suaairy and Eaco*endatlooa 63
Catlaitid mIuIobi 63
Suggastsd luprovaaaata In tha
nathodology 63
Rafarancaa 67
v
-------
FIGUJtES
Hu>b»r Eats
1 Topographic up of tha Vulcan plant sits ........ 15
2 AarLul photograph of tha iru surrounding tha
Vulcan plant 16
3 Tha parchloroathylcna and chloroaathana unlta
at tha Vulcan plant . t 17
4 Tha Vulcan plant viaved fro* tha north 17
5 Aarlal photograph of tha Vulcan plant 19
6 Sampling locations and tracar ralaaaa lint for
tha night of 8-9 August 1981 29
7 Sampling locations and Cracar ralaasa list for
tha night of 10-11 August 1981 35
8 Sax pi leg locations and tracar ralaasa 1 Id a for
tha night of 11-12 August 1981 40
9 Sampling locations and tracar ralaasa Una for
tha night of 12-13 August 19B1 ..... 45
10 Concantratlons on 11 August 1981, 0000-0100 CSX .... 31
11 Concantratlons on 11 August 19C1, 0100-0200 c$T . . 5j
12 Concaatratlona on 11 August 1981, 2200-2300 CST .... 55
13 Concantratlons on 11 August 1941, 2300-2*00 CST .... 58
14 Concantratlons on 13 August 1981, 0400-0500 CST .... 60
vi
-------
TABLES
Wuabar fm
1 Suggaatad Valuta Cor tha Conatanta
In Equation (2) 6
2 Valuaa of k for Typical Exparlnanul Coadltiona 9
3 Troducara of Chlorl&atad Hydrocarbona .* 13
4 srfc kluii Oata 23
5 Concentration of Rafaratica Standarda 26
6 Obaarvad Waatbar Coadltiona In Vlchlta Aru,
8-9 Auguat 1981 28
7 Maaaurad Goncaotratloca, 8-9 Auguat 1981 31
8 Obaarvad Waathar Coodlcloaa la Wichita Art*,
10-11 Auguat 1981 34
9 Kaaaurad Concantratlona, 10-11 Auguat 1981 37
10 Obaarvad Waathar Condition* in Wichita Araa,
11-12 Auguat 1981 39
11 Maaaurad Concantratlona, 11-12 .Sjgupt 1981
12 0ba«rva4 Waathar Condi dona In Wichita Araa,
12-13 Auguat 1981 . 44
13 Kaaaurad Concantratlona, 12-13 Auguat 1981 46
14 Calla loo lata Eatlaataa for 11 Auguat 1981,
0000-0100 CST 52
15 talaaloo Uti Eatlaataa for 11 Auguat 1981,
0100-0200 CST 54
16 Eailaaton Rata Eatlmataa for 11 Auguat 1981,
2200-2300 CST 36
*11
-------
frjabsr fin
17 blsslon Lat• EsClmatss for 12 August 1981,
2300-2600 CST 39
18 Ealsslon Kid Estimates for-13 August 1981,
0400-0500 CST 61
19 Suhm(7 of Emission SUt« Estla*t*s for
Vulcaa Plant M
vlll
-------
SECTION 1
imODUCTION
Tha O.S. InvlroDaaotal Protect loo Agency (EPA) la concerned that
•o«e hydrocarbon compounds, especially halogenated hydrocarbon compounds '
that say be haxardoua at low concentrations, sight hava an Immediate
local Impact dovnviad of the plant* chat uauftctur« Cham. Bowavar,
thara haa been a lack of data that would allow the prediction of either
the amounts of thaaa materlala that ara released to tha acmoaphere or tha
concentrations that might ax1st downwind of tha plant* under various
meteorological coadltlona.
Savaral factora hava contributed to thla lack of data. Plret, thaaa
compounds do act come from clearly ldantlflad sources such as stacks or
vaata. Instead, thay ara llkaly to come fro* more widely dlatributed
sources auch aa leeks from valves, plumbing, Incompletely aaalad con-
tainers,' and filling and emptying operations, or during tranaport from
ooa placa to anocber. Tha magnitude of source* of thla cypa la extremely
difficult to measure or even estimate. Tha second, and almsst equally
Important, reason why ao f«w data are available concerning tha magnitude
of aourcea of halogenated hydrocarbon* la that tha procaaaaa and equipment
u*»d to produce the aatar1*1a ara frequently proprietary. Slallarly,
tha aanufacturara ara quite accretive, for competitive reaaona, about
thalr production and aalaa ratea. In such an environment. It la not
surprising chat tha manufacturer are very reluctant to allow outeldera
on the grounda of tha aanufacCurlr.j plants to maka measurements chat
would be necessary Co characterise source strengtha directly.
The problwm has been Juat such rs dot' station of aout^.
strength*. Onea a source strength Is known, maximum ground-level coneen-
tratlona can be estimated aa a function of downwind dliinca and meteoro-
logical conditions. Tha EPA recognized a need for a oou.ce-rmceptor
methodology chat could be used to estimate source strength and to predict
maximum ground-level coneeatrationa, and, for the reasons cited above,
required that tha method not rely upon measurements made within Che
manufacturing complex. Originally, measurements in the study described
in Chls report were Co be made within some manufacturing plant and used
Co verify Che performance of e different methodology that required only
measurments from outside the manufacturing area. The original intact of
the study described In thla report waa to meet the following objectlvea:
(1) Select a suitable manufarturlng site that producaa two
>e three following chlorinated hydrocarbons: methylene
^ ^oride (CH2CI2), trlchloroothylene (CjHClj), and
-r chloroethylene (CjCl^).
1
-------
(2) Perform the ntcnucjr tests In the plant to establish
the total fugitive emission ratea of thaaa compounds.
(3) Develop sampling strategies for fence-line and down-
wind measurements of tha ambient concentration of
thaaa compounds.
(4) Develop a diffusion/transport Methodology to predict
tha aourca strength and tha downwind concentration
fro* tha fanca-llna measurements.
(5) Verify tha model using source acrangtha estimated
from tha In-plant measurements.
(6) Develop a strategy for the uat of the Measurement
methodology.
Four of tha objectives were met. Objectives 2 and 5 ware not
achieved, because bo suitable plant was found Chat would allow on-site
measurements. Nevertheless, the project yielded a ¦•thodology Chat caa
be uaed to estimate Che fugldve emissions from within a plant when acceas
to tha plant la not possible. The source-reccptor methodology doea ut
uae Measurements made literally at the fence line, unless the fence Is
sufficiently far from the sources that Mixing will have produced reasonably
uniform distributions through the loweat 10 m or so. Thla report describes
that M^rhodology and its application Co a specific manufacturing plant.
Suggestions are also Included for Improving the Method and Its application.
The theory underlying the Methodology Is discussed In the Section 2
of this report. Briefly, the methodology relies upon the release of an
Inert, nontoxic tracer gas at a known rate outside the manufacturing
site. Concentrations of the tracer are measured downwind of the site
to establlah the magnitude of the atmoepherlc transport and dilution
processes. This value Is used in turn with simultaneous downwind measure-
aents of Che concentrations of the chlorinated hydrocarbons, allowing an
estlsute to be made of the magnitude of their sources within the plant.
Once these source strengths have been established, conventional iij^irslon
modeling techniques can be used to eotLiate downwind concentrations for
various meteorological situations.
This report also describes the application of the Methodology to a
specific site and llluatrates some of the Important features of the
rethod. The description begins with the selection of a candidate test
site, followed by a discussion of the various parts of the test procedure:
(1) The requirements for meteorological observations
(2) The procedures used for tracer releases
(3) The procedures led for collecting and analysing tracer
samples.
(4) The pr',cr-Jar*8 used for collecting and analysing
.~hior Lnatsd hydrocarbon compounds.
2
-------
Tuti ware conducted at tl\« Vulcan Material* Coapaay plaat naar
Wichita, K«ni«i, on the nlghtc of 8-9, 10-11, 11-12 and 12-13 AuguaC 19S1.
This report d*Krtk*i tit* condltloai that prevailed during theae teate
and interprets the rasulta Chat wtx« obtained to provide eatlaatae of
aourca anlaaloaa rataa for aoaa chlorinated hydrocarbooa. The report
concludes with a review of the Mthodolojy and a lueury of the r-«ulta
obtained in tha teat application.
3
-------
SECTION 2
THEORY
The theory underlying the source-receptor methodology used in this
study Is presented below. The derivation provides a relationship
between the dilution of a tracer released froa a ground-level line source
and the dilution of materials released froa a nearby ground-level area
or point source. First, consider the Causslan forsula for a ground-level
concentration from an infinite ground-level line source:
2Q,
C, -
* ilntu 0( y 2*
(1)
where
C - ground-level concentration (g ®"^)
Ql ¦ line-source ealsslon rste (g a~*s"*)
4 ~ angle between wind and line source
u - windspeed <¦ »"*)
oE " standard deviation of Causslan concentration
distribution In the vertical at a distance froa
the source parallel to the wind direction (a).
The vertical standard deviation og can be approxlaated bj a function
of th« fallowing fora:
«8(x) - axb (2)
where x is the distance froa the source, parallel to the wind direction.
Table 1 gives values of a and b suggested by Butse and Zlaaeraan (1) for
use In the range froa about 0.5 to 5 ka downwind.
If we substitute froa Eq. (2), Eq. (1) becomes:
2Ql
C1 " /— b
slnf nj 2*uax
-------
TA&LE 1. SUCCESTED VALUES FOR THE
CONSTANTS IN EQUATION (2)
Ataospharlc Stability
Con»tanta
a
b
ExCraaaly unacablc
0.25 x 10°
2.09
Moderately ucatabla
0.049
1.11
Slightly unatabia
0.10
0.93
Neutral (day)
0.26
0.69
Neutral (night)
0.25
0.63
Stabla
0.20
0.60
SOURCE: Buaae and Zlaner&an (I)
-------
The foraula for an trti source (of Infinite extent In the croit-
wtnd direction) can be obtained by Integrating Eq. (3) along x:
c»-
w*(xi
-b 1-b
" d
)
for b i 1
(4)
>ln+ y/Tfl Ua(l-b)
where Xy and X. are che distances to the upwind and downwind edges of Che
area aource ana (g s~*) 1* Che area-source Mission rate. The
"line sources" that are integrated to get Eq. (4) can be oriented arbi-
trarily. For convenience, we choose Co orlenC thea at the sane angle Co
the wind ~ as the tracer line source. Dividing Eq. (4) by Eq. (3) gives
- 'D • -b
Cx " Q^l-b)
Solving for gives:
CA^1
(1-b)
,1-b
,1-bi
(6)
All quantities on the right side of Eq. '6) can be Measured experi-
mentally except b, which depends on aCaospherlc stability. For experl-
aents conducted aC night, Che a .aosphere will be either stable or neucral
according Co the coosMnly used Methods for classifying atmospheric sta-
bility (2). If we limit th<* -'ownwind distances at which aeasureaenes
k, ^ all between 0.5 and 5 ka, Chen che
appropriate values of b range froa about 0.6 for stable conditions to
about 0.7 for neutral conditions (1).
Although a value based on aeteorologlcal factors can be chosen for
b, there is soae uncertainty, so it is important to evaluate the sensi-
tivity of Che relationship Co Che choice of value for b. Ue begin by
defining a factor containing all Che b Cera* as follows:
1-b
f_l-b _l-b\ b
Xu " Xd ) * "
(7)
Substituting Eq. (7) In Eq. (6) gives:
7
-------
(8)
Table 2 gives valuea of B for b • 0.6 and b ¦ 0.7, by using different
combinations of x, Xu, and Xj. It la apparent from tha tabla that tba
calculations ara not vary sensitive to atmospheric stability Within
tha llalto of experimental accuracy, b could ba aat equal to 0.65, and
Eq. (6) would than become:
Some assiaptlona inherent In Eq. (9) should ba understood:
(1) Eq. (9) appllaa for nighttime conditions or neutral
daytime atmospheric conditions.
(2) Concentration ara measured 0.5 to S b downwind of
tha sources.
(3) Concentrations ara measured near tha canter of tha plume
where tha assumption of an infinite crosswlnd extent of
tha sources Is most nearly valid.
(4) Tha crosswlnd extant of tha sources la Large compared to
the dimensions of a point-source pluae at the downwind
distance where tha concentrations are measured, e.g.,
- At 5 km, about 600 ¦ for neutral atablllty and
about 300 ¦ for stable.
- *r. 2 V- iL 300 u ¦ ,-al and about 150 ¦
for stable.
(5) The angle between the wind and the line source Is relatively
large, about 45° or gr«atsr.
(6) Tha separation between tha line source and the area source
should be kept as small as possible, preferably within
about 251 of the distance between the line source and
the samplers.
In general, the above conditions that apply to tha line source wart
Ht, except for the requirement that the wind direction be at a large
angle to the line source. This was not met for the night teat of 8-9
August 1981. Tills test, and the test of 7-4 August 1981, ware only
marginal with regard to the crosewlnd extent of the are* from which
fugitive Missions were expected, from the standpoint of meeting the
8
-------
TABLZ 2. VALTES OF B FOR TYPICAL EXPERIXDfTAl CdHDITIOHS
X
{¦)
X
u
(•)
Xd
(¦)
<¦
B
-1)
tUlaClv*
Dlffarcac*
(I)
b - 0.6
b - 0.7
4000
4600
4200
5'22 x 10~3
5*26 x 10~3
0.7
4000
3800
3600
4-77 x 10~3
4-74 x 10~3
0.8
2000
2400
2200
5-44 x 10~3
5-51 x 10~3
1.3
2000
1800
1600
4-53 x 10~3
4-46 x 10~3
1.6
1500
1900
1700
5-57 x 10~3
5-68 x 10~3
2.0
1500
1300
1100
4-37 x 10'3
4-27 x 10~3
2.3
9
-------
asaumptlons underlying Eq. (9), Che laat three teats were Che beat. The
data analyses presented later focua on these three caaea.
point or very small area at the surface, with a source acren^ih q The
concentration at the center of the plume (C > la given by:
It la also poaalble to laauao that the plant emissions come frem a
It
c, • T5^r • <"»
y t
where o„ la Che atandard deviation of the Cauaalan concentration dlatrl-
butlon In the crosswlnd direction; oy la approximately proportional to
xJ-9 (2). The conatant of proport loml ley depend* on atablllty. It
rangea from 0.062 for moderately acable atmospheric condition* to 0.13
for neutral conditions, when both Xp and Oy are expreased In the tut unit*
(a); Xp la the downwind distance from the point source.
If the point source and the line aource are close together compared
to the downwind distance, than (X_/X)b • 1 for the values of b discussed
•arllsr. For example, If 0.85 - vX_/X) - 1.15, then (Xp/X)b will be
between about 0.9 and 1.1. Using tnls fact, we can determine the ratio
between Eq*. (1) and (10), and solve for Qp to give:
„ „ 0.9
P Q1 P
^p " K sln$
(U)
where
10.32 for a
0.25 for a
0.16 for i•
neutral atmosphere
^lightly stable atmosphere
jaderately *tab3» -sphere
Eq. (11) in not quite as desirable as Eq. (9) for estimating source
strength because it is subject to appreciable dependence on atmospheric
stability. However, it does not require an estimate of the size of the
area sourca in order to datenine total ealaslon rates, tn the analyses
that follow both Eqs. (9) and (11) have been used.
10
-------
SECTION 3
TOE TEST SITE
Tha EPA wanted this study Co be performed at a sita that manufactured
two of the three following chlorinated hydrocarbon*: methylene chloride
(CHjClj), trlchloroethylene (C^HClj), and perchloroethylaae (CjCl^).
There or* only a limited nualver of auch manufacturing alt** la tha Onlted
State*; a Use of th«*« alt** wa* coapilftd and aach alte waa evaluated
according to tha crlterl^ discussed below.
CRITERIA USED TO SELECT THE TEST SITE
Although It wa* not anphaslzed In tha preceding aact Ion, It la
esaential to ba abla to aatabllah that aaaaured concantratloaa of tha
halocarbooa of lntarcat coma from tha plant la queatlon. Therefor*, tha
tru aurrounding tha manufacturing alt* must ba fraa of othar aourcaa of
tha same material*. Such aourcaa would Interfere with Interpretation of
tha data and make It difficult to estimate source atrangtha within tha
plant ltaalf. Therefore, on* of tha cr It aria for selecting tha taat
aita waa that thara ba t>o othar naarby aourca* of chlorinated hydrocarboaa.
Tha uaa of a tracar to aatlmata dilution rataa requlrea that tha
tracer ba released aa cloaa to the aourcaa of interest as pcsslbls. It
also requires that samples of aablent air ba collected or measured both
upwind and downwind of tha manufacturing aita, prafarably vlthln a faw
kilometers of tha aita. Inasmuch at tha wind can blow ftoa any direction.
It 1* daalrabla to hava tha araa around cha plant ba accaaalbla for a
few klloaeters In avary dlractlon. A. flat, open tarraln with a-tny lightly
travalad roada la most daalrablo.
''n. c>.uc*r methodology is no;,t appropriately applied at night.
Therefore, a plant that oparataa on an around-the-clock achadula la daalr-
abla ao that the cast* can ba conducted at night whan wind* ara light
and tha air la atabla. Thia criterion wa* not v*ry raatrlctlv* In tha
aita ••lection proceaa, because all tha plant* that war* coa*ldarad do
oparate on auch a achadula. Nevertheleee, It muet ba considered.
The diffusloo/tranaport methodology a*suae* that mo*t eal**lons
occur naar ground level and that mixing through the lower layer* la about
tha aaaa ao that of the tracar. The aathodology woult not be applicable
to a plant where appreciable quantities of the materials ara aalttad
froa tall stacks. This criterion wa* not restrictive for the plants
conaldarad in the aelectloa proca**, bacauae virtually all tha emissions
at thasa plants take place within 5 to 10 a of tha surface and with little
11
-------
buoyancy. When questioned, Che manufacturers stated that all atack
(acta art processed through incinerators before venting, to there should
b« no emissions froa the stacks'.
Tha final criterion used for (electing a teat alte van that the
Manufacturer be cooperative and allow acceea to the plant. Thla parti-
cular criterion proved impossible to satisfy. Although Boat of the
manufacturers vera v11ling to discus* the planned teats and to consider
allowing access to the plant, none was actually willing to grant such
access. The reasons for tvalr unwillingness, which were cited in
Section 1, aro certainly valid froa their perspectives. It should also
be noted that the operator of the plant that was finally selected for
this study was cooperative enough to furnish some information on plant
producta and operatlona.
CANDIDATE TEST SITES
Only seven plants in tha United States produce the chlorinated
hydrocarbons of interest. These are listed in Table 3. The table alao
identifies which of the compounds are produced at each of the sites.
The operatora of theae plants were contacted, naps of tha areas
surrounding the plants were obtained, aid cllmatological data for the
regions were studied during the selection process. Initially, SRI
believed that on* or more of the operating coapanies night be willing
to allow acceas to their plants, or that the EPA could expedltously
negotiate access on our behalf. Under the aasuaption that access could
be obtained, SRI focused on ."hree particular plants, which were visited.
These plants were PPG Industries, Inc., Laka Charles, Louisiana; Stauffer
Chaalcal Company, Louisville, Kentucky, and Vulcan Materials Company,
Vlchlta, Kansas. The latter plant was finally chosen because It best
a«t the selection criteria—other than the criterion regarding cooperation
by the operator. This site will be discussed in detail In the next
sect ion.
The Louisville and Lake Charles sites were not selected becauae
they would have been aore difficult for field operations than the Wichita
site. The marshy terrain surrounding the Lake Gharlus site limits the
number of available roads In sons directions froa "Mus plant. The Ohio
River is v.-i'.le plant, so that itzCL-jslvc ' .ouiu
have been required to get froa the plant to eoa* sampling areas. Although
It would have been possible to conduct tests at the Louisville and Lake
Charles sites, surroundings limited the number of wind directions under
which tests could have been conducted and would have Introduced soae
logistical problems that were not encountered at the Wichita site.
Inasmuch as the purpose of the program was to develop and demonstrate the
source-receptor aethodology. It would have "been counterproductive to
Introduce unnecessary complication*.
There was another aore serious reason for rejecting the Louisville
site. It was not entirely clear that there were no other sources of
12
-------
TABLE 3. PRODUCERS OF CHLORINATED KYDROCARBOKS
Company and
Location
Products
c«2ci2
c2hci3
c2cia
Dow Chwlcal Coapaay, USA
Fraeport, TX
~
/
Plaquemine, LA
/
/
Ethyl Corporation
Baton Rouge, LA
J
~
PPC InduJtriea, Inc.
Lake Charles, LA
/
/
Stauffer Chemical Company
Louiaville, KY
/
/
Vulcan KaCerlala Company
Geianar, LA
V
/
Wichita. KA
~
/
13
-------
chlorinated hydrocarbon* in nearby Industrial areas. This la also why
th« Dow Chemical plant at Frerport, Texas, and the PFC plant at Laka
Charlaa, Louisiana, were not given serious consideration. The application
of the methodology in an area of Large-scale petrochemical production,
such as Freeport, Texas, could pose some jerlous difficulties. SRI does
not believe that applying the methodology ln such areas would be impossible,
but it would likely require careful selection of the condltlona under
which the tents were conducted nd appreciably sore measurements upwind
of the plant to establish ambient background conditio!.*. It would alto
require a careful analysis of land use and manufacturing operations in the
region, and a more thorough interpretation of the collected data. It
seeaed premature to introduce these complications before the method had
been demonstrated under more favorable condltlona.
The reasons foe rejecting the other plant* listed ln Table 3 were
similar to those dlecussed above. Briefly, the Dow Chenleal plant at
Plaqueaine, Louisiana, is located next Co the Mississippi River, which
would have posed serious logistical problems. The Vulcan plant at
Celsmar, -Louisiana, and the Ethyl p * —. c at Batoa Rouge, Louisiana, were
not given serious consideration because SRI had been refused entry
earlier in the project. By t'.»e Cine it became evident that in-plant tests
would not be possible, SRI had already selected the Vulcan plane at
Wichita as the best site.
TEST SITE—TOE VULCAN MATERIALS COMPANY. WICHITA, KANSAS
The Vulcan Materials Company plant was chosen as the site for
testing the source-receptor methodology. The plant is located about
35 m southwest of the center of Vichlta, Kansas. Figure 1 is a topo-
graphic map showing the area near the plant site. The area around the
plant is quite flat; within 2 ka of the plant, the elevation changes
over a range of less than 15 m (about 1,280 to 1,320 feet). The site It
serviced by a railroad, which also serves a grain elevator about 1 km
to the northeast. A wastewater treatment plant and a fluorocarbon manu-
facturing plant (both opereted by Vulcan) are adjacent to the Vulcan site
and a power plant is located about 2 b to the northeast. A swtll
chemical plant is located about 0.3 ka south of site, hut it
does not use any ox the &aterials of interest.
Other than these activities, the area surrounding he Vulcan plant
is generally devoted to farming for at least 5 ka in all directions.
As the nap shows, there is a grid of roads at 1.6-km intervals around
the plant, but thesi are generally not heavily traveled, especially
during the night and early morning hours when the tests were conducted.
Figure 2 is an aerial photograph (taken February 1977 by the Kansas Cee
end Electric Co.) of Che immediate vicinity of the plant. Figures 3 end 4
are ground-level photographs of the plant that were takae during this
project. Figure 3 shows the "perchloroethylene" -nit (produces perchloro-
ethylene and carbon tetrachloride) on the left and the "chloroaethaae"
unit (produces methylene chloride, chloroform, end carbon tetrachloride)
on the -^ht. Figure 4 shows the plant end the surrounding (ermlaad.
14
-------
JOL
'<¦!_/
urn.
OJ
0
' ' I
•CALf
Fl|ur« 1. Topographic up of th« Vulcan plant sltt.
IS
-------
Fljurt 2. AarUl photograph of tha irM (urrouadU| tba Tulcu
plant. (Ffeotofrapfc provided by Kansas Caa and
KlscCrlc Company.)
16
-------
Figure 3. The perchloroethylene and chloronethane
units at the Vulcan plant.
The Vulcan plant viewed froa the north
17
-------
The Vulcan plane Is a medium-sited plant occupying a retangular
space approximately SCO m In the east-west direction and 600 a In the
north-south direction. However the area Involved In solvent operations
Is ouch smaller. Figure S, a section of the aerial photograph shown In
Figure 2, Indicates areas where operations night be conducted that would
be related to the emissions of the solvents of concern. The outside
boundaries of these sources are approximately 225 ro In the east-west
direcclon and 150 m In the north-south direction, but emissions of any
given material are not likely to be present at all locations within these
boundaries. Later, for the purpose of estimating the total emissions,
'.t is assumed that the total emitting area is about 25,000 m*; this it
aSout 60S of the solvent area described above. As noted earlier. It Is
be.' leved that the emissions are all likely to occur within 5 to 10 m of
the surface and at ambient temperatures, 90 there is not likely to be
any significant buoyant rise.
IB
-------
SECTION 4
TEST PROCEDURES
GENERAL
This section describes Che test procedure* that were applied at the
Vulcan manufacturing site near Wichita, Kansas. These procedure", would
generally apply for all sites. The need to collect samples upwind and
downwind of the plant, as well as the methods for collecting and analyzing
these samples would not change. The theory underlying the source-receptor
methodology dictates that Che best results vlll be obtained during the
nighttime hours, because the method Is not particularly sensitive to
prevailing meteorological conditions when the atmosphere Is stable, as
it often Is at night. The theory also dictates chat the tracer releases
be made In a line Chat Is as close to the manufacturing plant as possible
and as nearly perpendicular to the wind direction as possible.
Several considerations suggest that the sample collection be dona as
near the source as possible. Concentrations should be higher near the
source; hence, the results of the chemical analysis should be more reliable.
Collecting the samples near the source should also minimize the lttaral
spreading effects that negate the assumption of an Infinite line source,
which was used in deriving some of the equations for the methodology.
However, one very important factor dictates that Che samples not be col-
lected too close to the source. The theory presumes that both Che Cracer
and che materials of interest are released at approximately the same
low altitude. Vertical mixing will,tend to minimize the effects of
differences in Che release heights of the various materials at distances
of 1 km or more from Che source, where Che verdcal spread—as measured
by the vertical standard Aviation, oz, of a Causslan concentration dls-
v'-~ should b r'.. . 'at 15 m, even under moderately stable conditions.
This would be enough to provide a reasonably uniform concentration dis-
tribution (when averaged over about an hour) through the lowest 10 to
20 m of the atmosphere for materials that have been released within thac
same layer. As noted earlier, the assumption of an infinite line source
makes it desirable to mlnimlre the lateral spreading. During nighttime
conditions, the lateral spread—as measured by the horizontal standard
deviation, Cy, of a Gaussian distribution from a point source—will
generally be 200 m or less to downwind distances of about 3 km. Therefore,
the downwind samples should usually be collected at distances between
1 and 3 km downwind.
21
-------
METEOROLOGICAL OBSERVATIONS
One of Che ruiom for developing a methodology baaed on Che release
of a tracer gas was to minimize the need for meteorological observations.
Conventional, eaaily made meteorological observations are not directly
related to all tha factors that affect the diffusion and transport of
emissions released from a plant. The relationship between conventional
measurements and atmospheric stability is particularly subject to uncer-
taintlea. Nevertheless, there are operational requirements for some
meteorological observations when using the tracer technique. Futheriaore,
meteorological observations provide an Important source of backup Infor-
mation as well as information that can be used to Interpret concentration
data.
Host important to both the data Interpretation and operational plan-
ning of the study are wlndspeed and wind direction observations. These
were measured at an altitude of about 10 m in the immediate vicinity of
the test site. The measurements were made with a propeller-vane anmometer
at SRI'a mobile laboratory site, indicated in Figure 2. The temperature
and humidity were also measured at the same site, at a height of about 5 m.
In addition, SRI obtained copies of the official hourly weather
observations made by the U.S. National Weather Service at Mid-Continent
Airport in Wichita, Kansas, about 8 km north of the test area. These
observations included temperature, humidity, wlndspeed and wind direction,
and visual estimates of cloud amount and type. Airport wind directions
represent more or less Instantaneous measurements, subjectively averaged
over about a minute or two, while those at the site were averaged over
the hour, beginning at the time shown in the table. Cloud amount and
type are Important to the estimation of atmospheric stability class.
TRACER RELEASE PROCEDURES
The tracer chosen for this study .was sulfur hexafloride (SFj),
coaoonly used because it is totally inert and nontoxic, and because it
occurs only at very low concentrations ir the atmosphere. The tracer
wea released from the back of a rv^vinr -• ';iele through a flow-limiting
nsedle valv* ji a flow vtlve ensured a relatively
constant emission rate. The flow rate was measured at nominal half-hour
intervals using a Singer Modal 115, dry test meter. Flow rates were
very constant (±3Z) over an 8-hour period. The total enlsalons during
a teat were determined from the weight of the SF$ cylinder before and
after the toat. The average emission rate was determined from the
following relationship:
22
-------
where
Q. " Che average line-source emission rate of SF^
1 (g m-Vl)
W - the total weight of SF^ released, as determined
by weighing (g)
L - the length of the road segment over which the
vehicle drove while releasing SF^ (m)
t - the time Interval over which the SFg release
took place (s)
It is possible to use an average emission rate, determined as
described above, for these tests because the samples that are collected
are also averaged over extended periods of tlae <1 h). If the averages
obtained this way are to be reasonable approximations of the true
average, the release vehicle must traverse the road segment at least
six or seven tines per hour. (In the experiments reported, it was
traversed approximately AO times per hour.)
There 1b the tacit (and reasonable) assumption that the meteorologi-
cal conditions are uncorrelated with the location of the release vehicle.
For all of the test data considered In this report, a segment of
road approximately 0.82 km in length was used. This road is Immediately
west of the plane, approximately 225 m from the center of the hydrocarbon
manufacturing area (see Figure 2). It was possible to make quick U-turns
at the Intersections at each end of the road segment, because traffic
was light during night hours on the road. Because the U-turns could be
made
-------
SAMPLE COLLECTION PROCEDURES
Bag Samplers
The downwind simples were taken at ground level by using the Auto-
cide samplers that SRI routinely uses for this type of work. These
samplers manufactured by Environmental Measurements Inc. (EMI), are
multiple-bag samplers containing 12 pumps and an electronic timing circuit
by which up to 12 1-hour bag samples can be obtained sequentially. The
EMI sampler Is approximately 60 en In diameter and 110 cm high, and
weighs 11 kg. The sampler Is operable over a temperature range of
-10° to 50° C.
Each sampler 1b battery operated and uses a timing circuit that
actuates the pumpa intermittently to conserve battery life. The start
tine for sequential sampling (for the nlghc operation) can be set in
advance on each timer. As a result. It was possible to service the
samplers In advance, before they were placed at the sampling sites in
the early evening.
The EMI samplers were initially designed to actuate the pumps on a
duty cycle of 2 s on and 30 s off to produce a S-llter sample each hour.
The timing circuits on a set of these samplers owned by SRI have been
modified to obtain a duty cycle of 80 ma on and 1 a off. In addition,
the bag sice has been reduced to 2 L. These modifications provide a
nore representative hourly sample than can be obtained with the factory-
supplied pump duty cycle, while still providing more than adequate sample
volume.
From our experience with these samplers, SRI has found that Tedlar
is the best material for these bags. It has no background of chlorinated
hydrocarbons and it 'has very low permeability, so samples can be stored
without contamination or loss. The bags are fabricated at SRI and a
large number of thea were In stock for use in the study. The bags were
routinely checked for leakage during the test program. SRI has developed
a routine bag cleaning process that was used between each sampling. The
cleaned bags were tepted .ni a random basis to ensure that: ~-1 ' . been
cleaned properly «iiid rh,., thc/e ''as .vd inadvertent contats ;• it- . Each
of the bags was identified with a label listing the sampler number and Ch«
time tliar the sample started.
Sample Locations
The samplers wera placed at ground level, at power poles whenever
possible. They were chained to Che poles for security. For the first
three tests, the samplers were placed at approximately 300-*eter intervals.
For the last test, the spacing vas decreased to approximately 130 m,
because a preliminary review of the data Indicated that the plume might be
passing becween samplers.
The arc covered by the samplers was dictated by the wind direction
and the prediction of wind direction from the U.S. National Weather Service.
The sampler locations sre stawn on the maps presented in Section 5.
24
-------
The samplers were placed at the designated locations by 2200 CST,
and they were collected the following morning, beginning at 0700. After
the samplers were returned to the noblle laboratory site, the bags were
removed, checked for proper Inflation, and hung in special racks pending
analysis. Bags from the previous night's run were cleaned and used to
reload the samplers.
SAMPLE ANALYSIS PROCEDURES
Chlorinated Hydrocarbons
The SAaples were analyzed iiaing two Perkln-Elmer Model 3920 gas
chrooatographs (CCs), each with dual columns, Injectors, and detectors.
Each CC was equipped with a 1-meter Injection loop, which was cooled
with liquid oxygen to trap the halocarbons from S0-ca^ gas sampler. It
was then heated with boiling water to transfer the concentrated sample
Into the CC. The columns were 10 ft x 1/8 in (304 x 0.32 cm) 80/100 mesh
SP2100. They were operated lsothermally at 50° C with a total analysis
time of 20 sin.
Tracer Cas
SF(, analysis was performed using a System Science and Software
Model 215 KVP Envlronmerer, which is an electron-capture CC that operates
at ambient temperatures. This CC uses an internal vacuum pump to flush
and fill a 2-cm-* gas sampling loop, which is valved to introduce the
samples. The output of the electron capture detector is sensed by an
electrometer circuit, which Is coupled to a peak read-and-hold circuit
for digital peak maximum display, and a buffer amplifier that drives the
analog output. In the test, the analog output was connected to a Hewlett-
Packard Model 3390A computing Integrator, which automatically labeled
the sample, the peak elution time, and the peak area. The only modifica-
tion used from the standard production model was the addition of a Brooks
flow controller to allow finer control of the carrier gas flow rate.
Calibration
Calibration standard gases (Scott-Marin) were diluted with ultra-
high purity nitrojsn to make a series of calibration gases for the
chlorinated hydrocarbons. In addition, there was a series of sue
predlluted standards for SF$. The concentrations of these standards
are given in Table 5.
Data Reduction
A data acquisition system was connected *o the output of the CC.
This system was designed to actively record the elution time and area
of each peak. The peak area was then used to calculate the concentration
of the species, during tbe test program, an intermittent failure
occurred in the duta acquisition system, causing it to record Incorrect
areas for some of the peaks. This failure was not discovered until
25
-------
detailed data analysis was performed. For Chose cases where an
incorrect ara* was recorded, SRI attempted to estimate the area from
peak height. This w*s only partially successful, so any data based on
peak height are considered unrellxble.
The CC analyses provided measurements of CCI4 concentrations, as
well as those for the other thrte compounds. Although not required by
the contract, results based on the CCl^ measurements are presented in
the following sections.
TABLE 5. CONCENTRATION OF
REFERENCE STANDARDS
Cat
ConeentratIon
c2hci3
1.2 ppb
C2C14
1.2 ppb
W!3
1.25 ppb
cci4
1.3 ppb '
SF6
25.6 1 5 ppt
55.7 i 5 ppt
105 t 5 ppt
511 ± 5 ppt
.1050 1 50 ppt
10600 1 500 ppt
26
-------
SUCTION 5
TEST RESULTS
GENERAL
This section describes the conduct of the four successful tests and
the concentrations observed during those tests. An Interpretation of
these observations will be presented In Section 6, along with dilution
ratios and inferred emission rates. In this section, each of the four
tests is presented separately, beginning with an overview of the meteoro-
logical conditions that prevailed during the test and followed by a
description of the tracer release and the sampling array. Any information
about samplers that were moved during the evening to better measure the
patterns that were expected from the observed wind direction is also
included. Finally, the observed concentrations for each test are tabulated.
TEST OF 8-9 AUGUST 1981
Meteorological Conditions
Table 6 summarizes the prevailing meteorological conditions during
this test. It can be seen that the skies were clear until about 0500
CST on the morning of 9 August 1981. The temperature fell, in response
ro radiative cooling under clear skies. Winds tended to be light,
generally less than 3 m s"* at the site and at the airport. The winds
were blowing from directions between about 200° and 250° (approximately
from south-southwest to west-southwest). Light winds and cloudless
skies generally accompany stable atmospheric conditions, so -it is
reasonable to assume that the AtiROsphr *c > stable during ihlr: even J r".
Test Operations
Figure 6 shows the locations where samples were collected during
this test. With winds generally blowing from the southwest, the
samplers ware located northeast of the plant. The road segment over
which the SF$ releases took place is also shown in Figure 6. With a
southwest wind, it'was not possible to find a road segment that was
normal to the wind. The data collected on thla evening were not as
usaful as those collected on some of the other evenings, becauae of the
location of the line-source tracer release relative to wind dlrectlona.
27
-------
TABLE 6. OBSERVED WEATHER CONDITIONS IN WICHITA AREA,
8-9 AUCUST 1981
Hour
(CST)
Sky
Cover
(tenth*)
Wind
Relative
Humidity
(X)
AC Site
Mid-ConCinent
Airport
Temperature
(°F)
Direction
(°>
Speed
(as"1)
Dlrc-tlon
(°)
Speed
-------
SAMPLER LOCATIONS
V '
¦vrt
y moved
•14 between
000 and .
>** , • >0000 and
X ^^Voioo CST
f
I MOBILE
laboratory' /
i a£K/g£3\ sm
:\o
rlkU*
Figure 6. Sampling location* and tracer release line for
the tiljhc of 8-9 Auguat 1981.
ia
-------
Observed Concentrations
Table 7 summarizes the concentrations of the tracer and the
chlorinated hydrocarbons that were measured. The sampling locations
refer to those marked In Figure 6. It should be noted that some of the
samplers were moved during the course of the test in order to provide
an arc chat was aore nearly centered on the downwind direction. Complete
analyses were not performed on samples for whlcii the SF5 results indicated
little effect frrm the plant area.
The concentrations varied considerably from one location to another
and from one time period to another. This high degree of variability Is
frequently observed; it appears to reflect a corresponding variability
In the emissions from the plant. The Importance of this variability to
the design and execution of future experiments will be discussed later.
Although none of the data collected during this evening were suitable
for estimating the emission rates, they do indicate the concentration
levels chat are observed In the vicinity of the plant.
TEST OF 10-11 AUCUST 1981
Meteorological Conditions
Table 8 summarizes the meteorological conditions during this test.
As was the case during Che 8-9 August test, the winds were light, generally
less than about 2 m s~^ at the site. (They were slightly stronger at
the airport.) Wind directions and sky conditions differed considerably
from those of 8-9 August. Ulnds were Initially from the east-southeast,
but around midnight they switched to north-northeast. Overcast skies
accompanied this wind shift. Initially, the overcast was altocumulus
at about 3,000 m. Later, a layer of clrrostratus was observed at about
7,500 m. The cirrostratuB overcast was dense enough to be considered
opaque by the observer. Overcast skies are generally accompanied by
neutral stability (2).
Test Operations
The location of the samplers and the route over which the line-
source tracer release was laadt .»re shown in Figure 7. Initially, when
the wind direction was from the east, a line of samplers was placed west
of the site, as shown in the figure. However, as the wind developed a
more northerly component, it was deemed necessary to add samplers south
of the plant. Thus, samplers were Installed and operated after about
0115 CST at the locations marked 19, 20, 1, and 2. Unfortunately, these
samplers were faulty, and the samples collected could not be used. It
would probab'y have been vise to realign the tracer release route at
about the same time that Che additional samplers were placed. This
would have provided <1 line-source release that was more nearly perpendic-
ular to the wind direction during the later hours.
30
-------
TAB LI 7. MJUIIIU COKOnUTIOM, AUCUST LMl*
L
OUtkMk
Rmt
(CST)
2
1
1}
20
19
18
i)
12
6
7
CCi
2000
r*
1.72
.91
1.20*
2.0J*
.91
12.0
1.0**
.9J*
2100
t
2.21
).06
».•)
4.94*
20.1
1.71
l.M*
1..74
2200
i.2l
S.61
9.21
.01*
r
2.19
l.Oj*
2.2^
2)00
4.4
1.42
.97
144.0
.74*
). 62*
1.64
>.4)
I.*4
• 6J4
0000
0.4)
1.71
.71
11.6
4.12
2.14*
• .72
.M*
i.yd1
.»
0100
2.69*
).2«
.))
.90
1.47
r
11.)
1.71*
1.77*
J.IJ*
0200
.296*
).l)
1.01
r
4.77*
r
69.2
1.44*
J.62-
2.V44
0»1
.127*
2.47
.72
r
1.14*
.47*
Itl.O
12.0
nV
1.81
0*00
.12 J4
I.H
.U
.67*
r
11.)
7.II4
OSOO
).76
1.91
1.22*
6.10*
1M.0
11.)
r
.17*
0400
4.10
r
11.1
i.oe
1.4**
1.0)
2000
.09
.14
.1)
.16
13.7
.16
.07
.17
.11
2100
.01
.18
.09
.11
.06
.04
.09
.01
2200
2.61
.ai
2.64
.11
r
.11
.09
.20
2)00
.0)
.10
.02
1.17
.02
.04
.02
.01
.0)
6.26
0000
.09
.06
.02
1.(7
.02
.0)
.02
.02
.01
.12
0100
.04
.10
.16
6.92
12.7
I
.06
.02
.04
.19
0200
.02
.1)
.61
r
9.11
r
.9)
.01
.01
1.66
0)00
.01
.10
.06
r
• 9.17
9.70
4.03
1.76
.46
.10
0400
.0)
19.4
.03
1.87
S.)J
r
I. )9
1.07
0M0
.04
12.)
19.0
1.96
1.11
.2)
.11
0600
14.1
r
.0)
.01
.07
.19
Sm (ootaol.. «t «oi o( Cabli. IchUwmI)
-------
TABLE 7
(coollauftd)
9MfllD| Local loa^
bvi
Cq^o^4
(at)
2
i
IS
20
19
1«
11
12
4
7
2000
r
4.47
4.4}
4.VJ4
16. +
9. }7
}.}6
26.14
}.U4
2100
r
2.02
J.1)
1.4}4
J.624
).92
10.1
10. }4
2.79
2200
it.)
9. }7
4.7}
r
6. 76
6. Md
10.24
2)00
1.174
l.ti
4.J)
2.0}
1.6 J4
6.98
2.06d
I).I-
1.49
0000
1.12
i. n
10.•
J.22
2.7T4
1.974
}.24
.U
4.42"
I.>4*
0100
l.)4d
1.7)
2.42
J.S6
l.)Jd
p
}.))
1.20
}.«d
4.72d
0200
1.2 J*
1.764
i.M
r
l.}2d
r
2.})
1.40
7.0O4
4.1)4
0300
l.l7d
1.99
• .2)
r
l.}2d
6.614
2.00
1.6)
9.6)d
4.17
0400
l.J*-
J.18
4.00
1.074
r
2.62
6. ))d
0100
).}*
1.2}4
4.7}4
2.9)
l.2}d
P
8.12 ^
0600
27.4
r
).04
10.}
9. }2d
}.24
c2ci4
2000
.
2.04
2.02
.44*
2.4J4
l.M
)• 16
16. T4
10.4
2100
1.26
2.7}
.it*
2.IT4
l.»
21.6
}.36d
1.47
2200
24.0
2.64
1.94
r
4.6)
*.•/«
). }6d
2X10
.91
4.6)
4.7}
1.0J4
l.M4
1.62
.604
7.144
.4)
0000
.)}4
.55
1.224
2.14
• 4I4
1.01d
1.7}
,44d
.H4
.794
0100
.)14
.M
1.0)
1. }6
.B94
P
1.46
.47
4.024
2.J74
0200
.)i4
- .44
1.67 >
r
. 74d
P
*<. '*
.46
}.0}4
2.20^
0)00
.)i
.*1
2.69
r
.))
2.4
-------
TMLI 7 (coatlou**)
lMflU| Local
CoaeoiflU
bar
(CST)
1
i
li
10
6j
It
11
12
4
7
v>a>
2000
1100
7.74
4.41
l.M
7.»J
9.1*
10. «*
9.24*
7.11
.41
4.97 .
J. 97
10.7*
9.74
J.O4
1.1 J*
1200
11.1
14.1
J. 19
(4.01)10*
r
9.14
10.1
10V
1300
11.)
» j.j
>.M
.41
>6200.0*
10.4*
1.79
14.0
11.4"
1.49*
0000
2.*?
4.10
4.71
t.Jl
9.72
4.14'
J. J J*
0100
J.M*
1.19
1.*}
9.29
10.9
r
11.9
l.M
9.9S*
10.0*
0100
3.SJ*
.J44
• .17
r
J. JO*
r
4.14
3.03
10.4*
9.74'
0)00
0400
4.4?-1
4.»l4
1.17
t.Jl
11.1
4.72
r
11.7
11.9
11. T4
S.17
4.17
10.7
9.1)'
9.4*'
7.17
o:oo
a.i
11.4
11.4
U.6*
1.1*
11.1
4.09*
0400
1.91
t
1.12
9.19
10.4
9.14*
3
'CgaculitllaM upuMd La it — fV. k-
k,
(,,« WP
U# ri|wi i Car iMfllai Iwxisu. ^
r • lUt k<|.
'fMk Ih1|U MtniraHii.
-------
TABLE 8. OBSERVED WEATHER CONDITIONS IN WICHITA AREA,
10-U AUCUST 1981
Hour
(CST)
Sky
Cover
(tenths)
Wind
Temperature
(Of)
Relative
Hubid icy
(X)
At Site
Mld-Contlnent
Airport
Direction
(°)
Speed
(bs-1-)
Direction
<°>
Speed
(as-1)
2200
1
70
1.8
100
2.1
69
73
2300
7
100
1.8
110
2.1
69
73
2400
10
85
1.3
40
T
2.6
69
73
0100
10
45
0.4
360
1.5
69
76
0200
10
30
0.9
20
2.1
68
79
0300
10
30
0.9
350
2.1 .
66
84
0400
10
20
0.9
40
2.1
66
84
0500
8
15
0.9
30
2.6
65
84
0600
9
20
0.9
—
0
64
84
34
-------
(laboratory
HOBILK
! ,—r-VH
:',?uirr sire
¦DUCES.
RELEASE
LINE
AMPLER LOCATIONS
after 0115 7
-------
Observed Concentration#
Table 9 gives Che observed concentrations of the tracer and the
chlorinated hydrocarbons. As shown In Table 8t the winds between 2400
and 0100 at the site were generally from the east (85°) and shifted to
the northeast during the following hour. The data collected during these
2 h show significant concentrations of CCI4 and C2HCI3. It appears that
there may have been above-normal releases of material that were detected
by the sampling network during this period. For that reason, 2 h have
been chosen for interpretation and for estimating the release rates.
They will be discussed in greater detail In the next section.
TEST OF 11-12 AUCUST 1981
Meteorological Conditions
The meteorological conditions that prevailed during this test are
summarized in Table 10. The skies were overcast during most of the period,
so the atmospheric stability was neutral. Winds at the site were generally
from the east until about 0300, when they shifted to the southeast. Winds
were from the east at the Mid-Continent Airport until about midnight,
when they shifted to southerly; later, the winds were more from the south-
east. During the last 3 h of the test period, the airport winds were
approximately south-southeast. Table 10 shows that the temperature
dropped about 3° F during this period. Relative humidity rose corres-
pondingly from 64 to 7BZ. The cloud cover throughout this period was
cirrostratus at an altitude of about 7,500 m.
Test Operations
Figure 8 shows the sampler locations and the route traveled during
the tracer release operations. The sampling array was set up to
accomodate the winds at the beginning of the period. These winds were
from the east or slightly northeast, so the array was quite suitable
for measuring lnfluencej Croa the tracer line and from the plant.
However, around 0&CC CST, she winds shifted, so most of the mar»rials
were p.-.tiocl north of r .. sampling line.
Observed Concentrations
Table 11 gives the concentrations of the various materials. Aa
noted earlier, the wind directions during the first few hours of this
period were nearly perpendicular to the roadway on which the tracer was
released. This Is the most suitable alignment for application of the
method described earlier. One of the test hours (2200-2300 CST) has
provided Interesting data that could be used to estimate emissions from
the plant.
36
-------
TMLZ 9. HKASUUD OOMCCMTtATIOHS, 10-11 AUCUST 1MI*
Hour
(CST)
b
Stapling lACatloa
18
15
14
11
12
11
10
9
a
7
t
1
4
1
"<1
2200
1.12
.21
.01
.06
r«
.01
.M
.01
.07
2100
.05
.01
.01
.02
l.U
.20
.01
.01
0000
J.16
.18
.01
.01
.01
.61
l.U
2.21
.20
.04
0100
.10
.01
.07
.27
.01
.01
.04
fC
.08
0200
.10
.01
.02
.07
.06
.06
.06
r*
r* .
0)00
.10
.01
.11
.01
.02
2.44
.07
.04
.04
.01
MOO
.06
.16
.04
r®
.11
.07
r<
r«
0*00
.20
.11
.02
.01
.M
.01
.07
.01
.01
otoo
.01
.01
.14
.01
.04
.07
.07
.06
.02
.01
0C,4
2200
I
7.17
• lld
14.4
191
21.2
11.6
KIT*
.11
2 XX)
i.yr
.91*
4.11
1.19"
.19
l.il
.66*
10.1
114
7. II-
1.89
.11
0000
1.19
.05*
6.18
.2J4
.29
12.1
4.26
2.46
11.8
4.41d
1.21
2.69
1.11
.11
0100
1.20
21.9
4.41d
6.7
22.2
.41*
1.01
6.08
1.71d
.61
0200
.79"
. It'
.41
7.27
.69*
1.2 Jd
oioo
i.u£
1.16"
.09
.72*
l.lld
.10
MOO
2.12*
1.78"
.99
1.84d
OSOO
4.17
.66
1.09*
.21
.11*
1.41d
.91
0600
.51*
1.10*
.21
1.60
1 • isd
.17
C7"Cll
2200
llV
8.92
l.U"
8.04
S.91
1.91*
1.06
1.28d
1.44d
2)00
26.7
I.42"
J.11
4.64"
2.61
1.19
2.77*
10.4
8.42
12.9*
2.02
2.85d
0000
20.1
0.2"
11.8
2.68"
18.4
10.1
1.66
22.1
16.1
22.4J
26.7
1.79
12.8
B.69d
0100
17.2
8.72
14.l'
26.1
11.1
8.04*
2ft.0
12.7
11.6d
16.4
0200
19. ld
1.00*
1.0Sd
17.2
1.26
4.10
10.4d
C,J0
2. 79*
*.71d
11.4
1.64*
9.69d
2.90d
MOO
),K"
8.46*
10.8
1.91*
OSOO
l.U
1.79
1.79"
6.11
2. ll"
1.64*
2.74d
otoo
1.41*
1.71'
9.61
22.1
4.J0*
4.10
l«« (MIMtM at
m4 b.'
(cMtUaW)
-------
TAIU (vunl loued)
SaaptIng
b
Ucil Ion
Kour
(CST>
1J
n
l»
1 1
12
II
10
9
8
7
6
1
4
1
SCl4
2100
i as
1.61
. 19^
2. >2
1.12
.87
1.68
2.024
.11
1X0
1.50
1.86
.18
.JId
2.50
4.29
i.ia*
1.17
.w
0000
2.42
1.61
.If
2.04
1.76
• 48*1
1.67
1.01
l.ktf*
1.81
1.70
.48
0100
?. e*
J. 12
2.04d
1.1 i
1.76
.1 1*
2.08
1.48
1.09"*
1.07
0200
2.71d
\ 74
.64
2.41
1.01
l.BT*
0)00
• H
.IB
2.9 J
.14
1.24d
.48
0400
!..lid
1.1*1
2.OS
1.74d
0100
l.lld
t .40
.49
2.20
.26*
1.10*
1.19
0600
.U"1
l.0lJ
1.8*
.09
1.21d
.98
Wl
2200
J.20
i.k*
8. J9
6.10
10.0*
1.11
7.14d
6.094
no0
11.6d
6.21
9.04"
19.a
.9)
7.t4d
9.18
6.90
9.46d
1.14
1.94*
0000
4.26
4. J*'
' 7.79
B.I24
9.81
9.19
7.01d
2.14
1.99
9. Ud
9.18
1.01
8.17
1.10
0100
J<.69
6. 1?
i.tr*
8.14
8.)]
J. XT*
9.44
7.84
a .it*
7.94
0100
i.92d
6. JO4
11.6
8.22
a.71
MB*
8.61d
0100
5.OA
6.14"
a.66
11.1
8.20d
8.09*
MOO
.i»d
9.0 J4
8.18
a.W
0*00
6.01
10.1
8.41
1.10d
8.67d
0<>00
4.08
-------
TABLE 10. OBSERVED WEATHER CONDITIONS IN WICHITA AREA,
11-12 AUCUST 1981
Hour
(CST)
Sky
Ccvar
(tench*
ULnd
Temperature
(°F)
Relative
Humidity
(*)
At Site
Hid-Contlnent
Airport
Direct Ion
(°)
Speed
(ms-1)
Direction
(°)
Speed
(ms-1)
2100
10
80
2.2
70
3.1
72
64
2200
10
90
1.8
80
2.6
71
66
2300
10
100
1.8
90
3.1
70
68
2400
10
110
1.3
190
1.5
67
78
0100
10
90
0 u
140
1.5
67
73
0200
10
100
0.9
130
2.6
67
. 71
0300
10
95
0.9
130
2.1
65
78
0400
10
135
0.9
160
1.5
64
79
0500
8
155
0.9
160
1.5
64
78
0600
8
140
0.9
160
1.5
65
78
39
-------
>ah
A
¥.
/ 2 *17
/ MOBILE Jf
\laboratory /
L! ,-r-VJL
PUNT Sin
i
Figure 8. Sampling locatlona and tracer release line for
Che night of 11-12 Auguae 1981.
40
-------
TAIUK II. NU5UMD OMCOrTUT IONS, 11-13 AUGUST 1M1*
Cnfnial
SMpl lag
Uci( lM
h
liMir
(CST>
J
6
7
a
9
10
II
12
11
1
17
14
11
7100
• 04
.06
.01
.04
.01
.10
.69
ai
.11
1.14
.14
.14
1)00
.0)
.0)
.01
.01
.04
.07
1.11
1.08
1.42
.12
.79
.17
.41
0000
.30
.21
.04
.01
.04
.01
.19
1.17
1.60
1.47
1.11
1.19
0100
.M
.90
.01
.02
.04
.42
1.04
16.7
2.21
1.11
1.68
2.08
1.68
0100
.01
.04
.07
.02
.09
.61
1.17
1.12
2.70
2.21
4.81
.11
.67
ov»
.01
.04
,t>2
.01
.07
.18
.60
1.74
2.98
2.88
6.00
.11
MOO
.97
1.84
.0)
.02
.26
.04
.01
.10
.81
3.61
1.14
12.9
owe
.0*
.11
.02
.04
.04
.04
.02
.11
.09
1.11
2.01
.88
.11
0600
.71
1.12
.01
.11
.01
.01
.61
.13
.01
.81
1.97
.17
cci
1100
2.44d
14.8
1.11
1.66d
61.8
1.42*
44.7
6.92
.4*
.72
.61
7.0^
%
JVX)
1.42d
9.16
.71
.21
19.8
1.124
11.1
7.02
1.9^
12.0
i.of
0000
1.13d
10.6
.72
1.02
.ie*
4.01
1.13d
11.7
9.06
9.11
17.7
11.1
0100
74.7
.10
.60
.W
10.1
l.44d
11.0
7.2/'
S.71d
28.0
8.2^
0100
2.04d
7.12
• 61d
4.87
1.84d
1.28
11.3
4.91
11.1
6.74
4.1/1
0100
1.M54
• .Of
.92
.60
.47*
21.7
4. lld
11.9
9.81
10.0
241
4.11d
0400
6.97d
1)1
1.01
.12
.81
1.37?
. 82d
1.19
71.2
11.1
6.81*
0100
1.70"1
4.77
,86d
.57d
8.11
I.14d
l.^O-
1.14
14.1
II./4
1.1*4
04C0
1.27d
1U
.SO
.U
,12d
140
1.61d
2.14
1.76
2.6*1
.oaf1
.if
C7*Cll
210
18.4d
21.1
17.1
11.2d
22.4
28. ld
18.9
6.11
8.6^
12.8
26.6
24. £
m
1.60d
t
10.1
6.11
2.88d
1.44
6.66d
1.97d
1.61
7.9^
.01
9.16
7.9*/1
oooc
! 1.77d
1.17
9.90
9.90
2.11
6.12
6.78d
1.61
1.87
4.22
12.0
1.49
0100
b
2.91
6.92d
4.81
2.16
8.11
7.10d
2.11
1.64*
11.J-
4.22
1 .U?
0100
; S.M4
1.41
2.46d
.12
1.76d
1.19d
1.70
1.9^
49.7
1.08
i.vf
oxxj
;
1.11
1.81
1.99
l.l6d
11.9
8.40d
1.17
1.41
i.ed*
9.26
4.91-
MOO
j «.nd
2.81
4.42
7.16
J.77d
11.7
6.10d
2.11
2.69
29.9
40.8
1.44d
0100
4.68d
.67
9.90
1.97d
11.7
6.94d
1.18d
1.78'
1.14d
SO.l
.13
1.94d
SMC
1 **M<
1.17
4.44
8.91-
2.96d
60.1
1.17d
2.04
1.61
1.4*1
17.r1
47./*
Im fMlaotti • I «f tall«,
(cMt tCMfed)
-------
TABLE II (cunt tnu*d)
b
S*ra t Inf Local Ion
Hour
Cofiuftd
(CST)
5
6
7
8
9
10
11
12
11
1
17
14
15
C.Ci.
2100
1.8*
2.90
2.21
.92*
4.10
1.9*
2.05
.8*
,52d
2.78
2.10
1.91d
2 %
2 VM
1.9<*
4.04
1.85
.4*
1.26
1.2*
.81*
4.72
.65"
2.02*
6.11*
1.2)'
0000
1.72*
2.04
2.56
1.52
.15*
.67
1.0*
1.67
.11
.85
4.19
9.5"
0100
1.14
1.55
1.28
.29 '
2.15
1.47*
.49
.52d
1.61*
5.22
6.17*
0200
1.15*
2.28
.11*
1.15
.6*
.61
2.14
.81*
1.94
2.74
2.01*
0)00
1.47*
1.71
1.14
1.74
.22*
1.16
1.8*
1.60
.06
.70*
11.8
1.J8*
MOO
4.22*
7.15
2.77
1.14
.26*
1.06
1.2*
.41
.45
1.66
2.26
1.99*
OSOO
1.09*
1.72
2.74
.15*
1.18
l.l*
l.l*
.17
.11
2.88
1.20
1.01*
ouo
1.11*
1.47
1.61
.97
.26*
1.12*
1.2*
.11
.11
.19*
2.00*
1.04*
ci"jcij
2200
J.**1
5.46
5.10
i. vfl
5.05
5.1*
5.99
7.04
4.46
5.71*
5.51
8.90*
2100
s.ot*
5.56
4.21
2.15*
4.42
5. 76*
14.8
2.62
2.91*
9.85
4.86
1.65*
0000
5.86*
. 3.29
4.48
4.21
1.18
2.87
4.6V1
5.61
18.1
1.91
6.54
11.0
0100
4.92*
4.6]
1.57
1.87
1.20
7.64
6.1*
14.0
7.05
7.6*
5.61
9.11*
0200
6.02*
5.24
1.60*
8.11
1.1*
1.49
5.11
4.16*
16.1
4.29
8.91*
0100
4.86*
4.91
1.61
1.08
2.79
11.1
8.4*
5.78
16.6
7.51
8.16
5.06*
MOO
6.10*
4.14
4.00
1.80
1.76*
5.25
12.2*
6.55
1.15*
.90
4.56
6.26*
OSOO
5.97*
1.42
5.65
11.2
4.67
8.7*
10.4*
14.6
4.45
4.91*
0600
6.<6*
4.65
1.94
4.79
8.25*
8.18*
8.90^
12.9
1.61*
1.19*
6.79*
4.85*
$CoMMlrat ina «ipr««iid In v% m \
b
Sm fl|vr« 6 for lupllng location*.
Cr • flat V-cg.
VhIi Kft|h; M«iur«tn(.
-------
TEST OF 12-13 AUGUST 1981
Meteorological Conditions
Table 12 summarize* Che observed weather conditions for 12-13 August
1981. This was another overcast night, but the clouds were lower on this
night than on che preceding two nights. The overcast vis altostratus
at altitudes of a few thousand meters. .During the earlier part of the
period, there was light drizzle; a 2-hour period of light rain turned to
drizzle at about 2015 CST. Because of the rain and the heavy overcast,
temperatures remained nearly constant throughout the period, and the
humidity was consistently high.
The winds during tills night were generally from the south and south-
southeast, usually within 30° of a south wind. The wlndspeeds were light,
but somewhat stronger than on the other nights, exceeding 4 m s"1 by 0600 CST.
Test Operations
figure 9 shows the location of the samplers used during this test.
It should be noted that the selection of chese sampling locations was
based on che southeast winds that prevailed at Che beginning of Che test.
As the winds became more southerly at later hours, this array proved to
be less Chan Ideally located. As the winds became more Southerly, they
also became more nearly parallel to the tracer release line shown In the
figure. This was a poor choice of tracer release route. A better choice
would have been along the east-west road to Che south of the plant.
Observed Concentrations
Table 13 summarizes che observed concentrations at various locations
on this night. Two of the hours during the period, when winds were at a
greater angle to Che tracer release line, were suitable for analysis and
for estimating the emission rates.
43
-------
TABLE 12. OBSERVED WEATHER CONDITIONS IN WICHITA AREA.
12-13 AUGUST 1981
Wind
Sky
Cover
(tenths)
At Site
Mid-Contlnent
Airport
Rela cive
Hour
(CST)
DirertIon
(°)
Speed
(as"1)
Direct Ion
(°)
Speed
(ms_1)
Temperature
<°F>
Humidity
(X)
Remarks
2100
10
150
2.2
160
3.1
72
84
Light drizzle
2200
10
160
1.3
160
2.6
72
84
2300
10
160
1.3
150
2.1
72
84
2400
10
175
1.3
170
3.1
84
0100
10
170
1.3
180
*
3.1
71
87
0200
10
170
1.8
170
3.1
71
87
0300
10
160
1.8
150
3.1
71
87
0400
10
150
2.2
170
3.6
71
84
0500
10
160
2.7
170
3.6
71
84
light r«in
0600
10
160
4.0
160
4.6
71
87
light shower,
fog
44
-------
uk.
ftf.
SAMPLER LOCATIONS !
££15—13—1
\n 5 6 *7 lSOl^ . ^>2 10
I t
HOBILK
LABORATORY
PUN!
i
TiRU
Figura 9. Sampling locations and tracer rclaate line for
th« night of 12-13 August 1981.
*5
-------
TAIIJE I). HKASUIED C0MCUT1UTI0MS, ' 1 2-11 UKUST 1961*
Sopl l«i liXtt !»•'
Uuur
ClT»Ufcg«d
ICS!)
s
9
II
i
4
7
IB
15
14
11
12
10
20
19
U4
2IOO
.66
.82
.40
r
.40
.82
2.17
1.51
1.80
.52
.10
7.61
2XM
.M
.04
.14
.78*
1.59
11.4
1.11
r6
1.12
2.04
1.71
1.57
.71
28.9
0000
1.12
17.4
.14
.11
.77
5.44
.98
.92
1.39
.71
4.48
7.11
.»
0100
27.4
8.05
41.5*
.07
1.24
1.14
r®
7.79
6.44
4.87
1.41
2.09
1.11
0100
.02
¦
.2)
.21
.21
24.4
0.1
1.74
1.71
2.82
2.91
4. 11
1.0?
1.02
1.17
0 wo
II.}
.21
21.2
2.04
1.46
1.61
2.28
1.40
r
20.7
'
1.17
.04
MOO
.0%
.55
1.17
1.25
11.5
2.51
2.46
1.59
2.89
.2»
1.92
.11
.08
.15
ovoo
.01
26.0
.44
1.41
1.41
2.11
4.70
2.05
5.22
.71
.15
.25
.71
e«4
2200
i.n
13.7
F<
5.44*
9.14
1.84
74.7
4.00*
1.15*
4.07
' 1.75*
I MO
I.*?
.59*
7.12
. 84*
1.51*
21.1
1.51
r*
6. 39*
4.25*
11.5
5.95*
84.7
1.98*
0000
.3;
10.4
I*
1.15*
7.47
.84
41.4
5.46*
1.72*
7.07
5.12*
24.1
6.44*
0100
1.11
.41*
2.40*
1.01*
8.62
1.18
r*
2.14*
.91*
10.5
8.28*
12.2*
4.55*
0300
.BJ*
9.14
.12*
J.94*
17.8
9.95
21.0
5.66*
5.52*
71.5
8.47*
1.68*
5.52*
0100
.17
It.}
.80*
1.82*
1.41
59.5
119
1*
4.86*
4.86*
7.07*
4.12*
MOO
.92
1.14*
11.8
1.71*
5.17*
142
72.1
182
60.5
4.84*
59.5
6.61*
12.5
1.81*
OWJ
1.24*
.12
.78*
1.62*
45.7
6.74
478 -
¦4
5.87*
5.65*
44.1
6.19*
1.22*
ci"cii
2200
1.1}
2.29
re
4.44*
4.11
8.11
¦ 2.12
6.27*
2.11*
11.8
.18*
1100
14.4
.25*
1.42
1.12*
5.21*
10.1
tO.4
re
4.75*
1.61*
5.12
.24*
4.04
1.49*
oooo
4.10
9.45
Tc
5.19*
9.91
7.04
1.71
7.06*
1.79*
4.94*
.M*
5.91
6.94*
0100
9.82
.10*
2.5C*
7.19*
9.91
10.1
re
4.14*
1.87*
27.0
.14*
4.27
1.97*
oioo
.02*
1. 19
1.81*
4.81*
9.14
7.48
8.04
J.88*
1.79*
5.94
9.18*
8.02
2.48*
oioo
C.9I
.14*
7.4)
2.49*
5,21*
J. 79
1.99
2.21
P
2.96*
5.94*
2,79*
MOO
9.24
.02*
9.41
1.41*
5,02*
2.02
4.04
2.28
5.21
1.61*
6.85
4.27*
1.81
5.82*
OMO
.04*
.41
1,14*
4.40*
J. W
11.7
2.88
1.17*
2.21*
6.7)
6.94*
6,44*
-------
TAJ LI. II (cunl InuH)
Siapllng locallun^
Hour
Cof iwd
(CST)
a
9
II
1
6
7
IS
1)
14
11
17
10
20 '
19
c,ci4
?;oo
1.80
1.4b
I1
2.9id
4.20
1.18
1.22
2.48J
.69d
1.71
1.74d
2 MX)
I 06
,60J
1 .IB
.t 0d
1.79d
1.28
l.U
Fc
2.ltJ
1. 2bd
4.24
2.41
. 70d
OOOO
7.70
1. 12
re
1.41d
1.61
1.12
1.11
2.20J
1. 0od
.12
2.17-
2.14
2.70d
0100
1.6)
• 4|d
¦ 9?d
l.61d
. W
l.U
rc
7.49d
. 16d
2.82*
6.21d
1.97
. 92d
o/uo
. 79d
1.56
.61d
2 .1 ld
1.81
20.9
2.16
I.S7d
• 61d
).9S
2.17d
2.11
.96d
0 100
1.58
l.46d
1.98
,71d
2.11-
1.47
1.96
1.01
~ c
• 9Bd
2.18d
116Q0d
,91d
0* 00
2.06
.lld
1.17
.;?d
2.16d
4.20
1.91
1.61
7.09
2.l)d
9.10
2 ,02d
1.41
I.12d
osoo
.8Jd
. 71
• 12d
2.61d
1.18
.01d
9.71
5.11°
1.41d
6.92
2.81d
2.18d
ei"icl»
2700
68.5
1.99
r
1.71d
0.0
1.16
1.11
2 KM
7.17
7.64
2. 74d
10. 6d
11.1
1.01
4.11
11.1
OOOO
7.00
11.8
Fc
9.14d
7.12
J.J2
4.16
6.11
'7.7
oiuo
6.SI
10.9d
9,04d
8.16
1. 18
f£
7.81
11.7
0700
7.SI •
6.19d
7.47
1.02
6.80
) .60
20.0
0 K)0
It. >
1.00J
1.19
1.01
4.46
0400
%. 51
6.18
1.90
4.71
7.19
7.17
7.91
7.48
0)00
4.42
1.27
9. 79
14.9
¦i
7.92
'Coacmtril long «>pr««««d I
b
Ttgurr % I tr adapting heat (una.
CF • (111 tMg.
4
P««k h«tghi aMMraBtnti,
-------
SECTION 6
INTERPRETATION OF RESULTS
GENERAL APPROACH
For reasons that were alluded to earlier, not all the data were
suitable for estimating emission rates. Frequently, the alignment of
the tracer release line and the wind were not as they should•have been.
Sometimes there were too few valid samples. Flat bags, CC traces suitable
for peak, height analysis but not for peak area analysis, concentrations
too low to be reliably estimated, and uncertainties regarding background
concentrations all contributed to making many of the cases unsuitable
for complete analysis. However, it was possible to Identify five 1-hour
test periods for which data could be used to estimate emission rates.
These five periods are:
(1)
11
August
1981,
0000-0100
CST
(2)
11
August
1981,
0100-0200
CST
(3)
11
August
1981,
2200-2300
CST
(4)
12
August
1981,
2300-2400
CST
(5)
13
August
1981,
0400-0500
CST.
For each case, the concentration data (excluding those obtained by peak
height analysis) were plotted as a function of distance along the Barapling
line. Smooch curves were drawn by hand to estlaate the concentration
distribution of each of the materials along this line. The lower con-
centrations near che ends of the line were taken to represent the back-
ground concentrations. In this way. It was possible to estlaate the
concentration contribution from the plant for each material believed to
be "Tlttci irom the plant. Concentration dJ srr<>-tiont Lasted
in the same way for SF^. Estimates of the term B In Eq. (/', vae derived
from the data, using a value of 0.63 for b [from Eq. (2)j and estimates
of distance to the upwind and downwind edges of the area source were
based on the geometry of each test and the extent of the region where
emissions were believed to be probable (as shown earlier In Figure 2).
As noted earlier, the assumed area from which emissions took place was
2.5 x 10-1 n2.
The same ratios of chlorinated hydrocarbon concentration to tracer
concentration were used with Eq. (11) to obtain other estimates of the
emission rates from the plant. The appropriate value of the term K In
Eq. (11) was selected on the basis of meteorological observations.
49
-------
ESTIMATED EMISSION RATES
11 August 1981, 0000-0100 CST
Figure 10 shows the concentrations observed at the various sampling
locations for SF&, CCl^, C2CI4, and CjHClj. The measurements of C2H3CI3
were not sufficiently reliable on this day to be used to estimate the
emission rates. The curves shown in Figure 10 were used to estimate the
ratio of the peak, concentration of the various compounds (above back-
ground) to Chat of Che peak SFg. These ratios are entered In the second
column of Table 16.
Table J.4 summarizes the values used for the various terms in order
to estimate the emission rates of the three compounds for which reliable
measurements were available. The last two columns are of greatest
Interest; they give the emission rates obtained from the area-source
formulation (assuming a total area of 2.5 x 104 m2) and from the point-
source formulation. It is evident that In this case the two approaches
for estimating total emission rates are quite consistent. The emission
rates given are in g s~*; 1 g 3"* is equal to 3.6 kg h~l, so the
emission rates given In Table 14 range from about 300 g h~* to about
3 kg h-1.
11 August 1981, 0100-0200 CST
Figure 11 shows the observed concentrations for the various compounds
and the smooth curves that were drawn through the available points In
order to estimate peak values. As before, ratios were determined between
the observed peak concentrations of the chlorinated hydrocarbons and
those of the tracer. These ratios are given in Table 15, which also
summarizes the caluclatlons used to estimate the emission rates.
The last two columns in Table 15 show the estimated emission rates
for the compojnds. As wa3 true before, the two approaches for estimating
emission rates yield quite consistent results. However, this hour shows
emission rates much higher than thoi.-a estimated for the preceding hour.
It appears that emission rates of about 20 to 25 kg h~^ were observed for
C2HCI3 and CCl^. Tills is .» factor of 10 to 20 times thns»» rrtes n •
during the preceding hour. The apparent is s ion rates for C2CI4
also much higher 'uring this hour than during Che preceding hour. As
will be'seen from subsequent discussions, the lower values seem more typical,
but similarly high estimates of CCI4 emission rates were observed on at
least one other occasion.
11 August 1981, 2200-2300 CST
Figure 12 shows the observed concentrations of three of the
chlorinated hydrocarbons and that of the tracer. The concentrations of
C2HCI3 were not above background, and it can be assumed that the emissions
for this compound were quite low. The ratios determined from the graphs
In Figure 12 are shown In Table 16.
50
-------
C-HCl
m
93
<8
s
a
s
c,ci
ca
3
S
7
t
I
I
I
X
11
SAX?!.n«C UXUTTOX
Figure 10. Concentrations on 11 August 1981, 0000-0100 CST.
51
-------
TABLE 14. EMISSION RATE ESTIMATES FOR 11 AUGUST 1981, 0000-0100 CST
Compound
VC1
B In
Equat lo;J 9
<*1 *
(g m~* B~^)
sin#
In Equation 11
Estimated Emission
Kates
Equation 9
(8 «"2 «_1)
Assuming
area of
2.5 x 10 m2
(g 8_1)
Equation 11
(g s-1)
cciA
5
io"2
3.4 x 10~*
6.7 x 10~2
1.7 x 10"J
0.4
0.3
C2C1A
1
io"2
3.4 x 10~4
6.7 x 10~2
3.4 x 10"6
0.08
0.06
CjHClj
9
io-2
3.4 x 10~A
6.7 x IO-2
3.1 x 10~5
0.8
0.6
-------
B
C.1C1
a
s
*
a
C,C1
CC1
>»
s
X
0
I
t
0
»
OJ
ai
e
J
s
c
«
I
t
»
s
IWL1K LOCATION
Figure 11. ConcentratIons on 11 August 1981, 0100-0200 CST.
53
-------
TABLE 15. EMISSION RATE ESTIMATES FOR II AUCUST 1981. 0100-0200 CST
KJ*
Estimated Ealsalon
Rates
Compound
VC1
B In
Equation 9
« -?1-s
(g ¦ a )
b lnt
In Equation 11
Equation 9
(g «_1)
Assualng
area of
2.5 x 104 «2
(g »_1)
Equation 11
(g ."l)
CC1,
90
10~2
3.4 x 10~^
6.7 x 10~2
3.1 x 10"A
a
6
C CI
2 4
7.5
10"2
3.4 * 10"4
6.7 x 10~2
2.6 x 10"5
0.6
0.5
CjHCij
70
io"2
3.4 * IO-4
6.7 x IO-2
1
o
X
N
6
5
CjHjCIj
7
io"2
3.4 x IO-4
6.7 x 10"2
2.4 x 10"5
0.6
0.5
-------
C,C1
CC1
ir,
OJ
0.1
o.«
OJ
(mvlipc uxxnm
Figure 12. Conc«nCratloni on 11 August 1981, 2200-2300 CST
55
-------
TABLE 16. EMISSION RATE ESTIMATES FOR II AUGUST 1981, 220O-2300 CST
Compound
VC1
B In
Equation 9
Q1 ,
-------
The estimated emission races are shown In the last columns of the
table. Again, Che estimates made by the cwo approaches are In reasonable
agreement. The apparenc emissions for Chls hour were somevhac smaller
Chan Chose for Che last hour discussed, 0100-0200 CST. The estimated
CCl^ emissions were abouC 20 kg h-1. and Chose of Che other Cwo compounds
were abouC 0.5 kg h"1.
12 August 1981. 2300-2400 CST
Figure 13 shows Che observed concentrations of the chlorinated
hydrocarbons and Che tracer, and Che estimated distribution of these
concentrations along Che sampling line. The rados shown In the second
column of Table 17 were derived from the estimated peak concentrations
of the various compounds.
The last Cwo columns of Table 17 show the estimated emission rates
obtained by Che cwo approaches discussed In this report. The agreement
in this case Is much poorer than it was in the preceding case. Further-
more, che estimate derived frcm a polnt-sourc^ approximation Is greater
than Che estimate derived from the area-source approxiaatIon; in the
three preceding cases, the area-source approximation yielded larger
estimates. During this test, the angle of the wind to the line-source
release was only about 20° or 30°, a face Chat may have contrlbuced to
Che observed discrepancy between che Cwo approaches. As noted before, a
release more nearly normal Co che wind direction better satisfies Che
approximations made in the derivation of Che equations.
13 August 1981, 0400-0500 CST
This was the last case that was suitable for estimating emission
rates. It suffered from che same shortcomings with regard Co the angle
between the wind direction and che line source as did Che case for 12
August, 2300-2400 CST. Figure 14 shows the observed concentrations of
the chlorinated hydrocarbons and the tracer, along with the smooth
curves thac were used for estimating Che rados shown in Table 18.
As was Che case for 12 August, Che esCimates derived from Che area-
source and Che polnc-source approximations do not agree very well. (See
the last r.vc rr"' ^ns of Tables 17 and 18.) Again, the estimate based
on the . cj assumption Is greater than Che estimate based on the
area-source assumption. The Cwo results differ by about a factor of 3.
The estimate based on Eq. (11) indicates Chat Che emissions of CC14 were
about 5 kg h ; the estimate based on Eq. (9) Indicates the emissions were
about 2 kg h~^. The Eq. (11) estimates for Che other compounds range
from about 100 g h~, for CjHClj Co about 700 g h~^ for CjH^Clj. Aa
noted before, Che estimates with Eq. (1) are about one-third Chose from
the other equations.
57
-------
C,MC1
¦
8
C,C1
CC1
11
s
i
;
a b * u a
»
s
SAWLIWC LOOT 10*
Figure 13. Concentrations on 11 August 1981, 2300-2600 CST.
58
-------
TABLE 1'/, EMISSION RATE ESTIMATES FOR 12 AUGUST 1981, 2300-2400 CST
Compound
Cc/Cl
B t-i
Equation 9
Q1
(g ¦ a )
slnf
In Equation 11
Estimated Emission Rates
Equation 9
(e m~2 b'1)
Assuming
area of
2.5 x 10 m2
(g s"1)
Equation 11
(g s"1)
CC1,
4
1
5.1 * 10"3
2.9 x 10"*
0.1
1.5 x 10"6
0.04
0.1
C2C14
0.1
5.1 * i0~3
2.9 x 10~*
0.1
1.5 x 10"7
0.004
0.01
c2hci3
0.4
5.1 x -0"3
2,9 x 10"*
0.1
5.9 x 10"7
0.01
0.04
C2H3C13
0.1
5.1 x iO-3
2.9 x 10"*
0.!
1.5 x 10'7
0.004
0.01
-------
t—i 1 i i 1 1—i r—i 1 1 i i i I I
o
c.aci
8
c,ei
cci
0
6
B
S
e
s
i
7
a s » d a ix
s
SAKTLIHC LOCATION
Figur* 1^. Concentration* on 13 August 1981, 0400-0500 CST.
60
-------
TABLE 18. EMISSION RATE ESTIMATES FOR 13 AUGUST 1981. 0400-0500 CST
Estimated Ealaston
Hates
Coapowd
c /c,
c 1
B la
Equ-tlon 9
(k ¦
qi
-1 -1.
e )
rQlxJ-»
sin*
In Equation 11
Equation 9
(g B~2 8~l)
4Assuming
area of
2.5 x 104 m2
(R 8~l)
Equation 11
(8 •~L>
cci4
15
5.. x 10~3
2.9
x 10~*
0.1
2.2 x 10"5
0.6
1.5
C2CI4
0.6
5.1 x 10-3
2.9
x 10~4
0.1
B.8 x 10~7
0.02
0.06
c2hci3
0.3
5. x 10~3
2.9
X
++
o
1
0.1
«.4 x 10-7
0.01
0.03
c2«3ci3
2
r*>
1
O
K
u> |
2.9
x io'4
0.1
3.4 x 10"6
0.09
0.2
-------
SECTION 7
SUMMARY AND RECOMMENDATIONS
ESTIMATED EMISSIONS
Table i9 summarizes Che estimated emissions rates described In the
preceding sections. It Is immediately apparent from the table that the
emission rates vary by more than two orders of magnitude. For the kinds
of emissions that were measured, this does not seem unreasonable.
Probably, the most frequently observed emissions would be comparable to
those in the "minimum" columns of Table 19. It appears that some sporadic
activities raise the emission rates appreciably for relatively short
periods of time. It Is not hard to Imagine activities thac could con-
tribute to such elevated emissions: Drum filling, loading and unloading
of containers, and other material transfer operations come immediately
to mind.
SUGGESTED IMPROVEMENTS IN THE METHODOLOCY
The results obtained during this Initial effort indicate that the
methodology that has been developed has conjlderable promise, but that
this first application did not always achieve that promise. It Is not
surprising that the potential of the method was not fully exploited In
this effort, and future efforts should be able to profit by the errors
made during this study. The following paragraphs present some suggestions
for future efforts, based on our experience with the data collected
during this study.
First, samples should be collected with greater spatial and temporal
resolution. The data collected during this program suggest considerable
variability In the emission rate. This means that there may be relatively
short-term "puffs" of materials Chat come from different parts of the
operation at different times during the hour-long sampling period,
closely cn*'" d uiaplers woi-Jd prov_de ;h« »patLa! .solution nece»i..,
to resolve any such small-scale emissions. The closer spacing would
also permit bett«v definition of the peaka. Frequently the data col-
lected in this ..Cudy defined a peak with only a few observations. Of
course, this problem was compounded by the difficulties that were
encountered in looe of the chemical analyses.
Finer temporal resolution; i.e., samples collected over a shorter
period of time, would provide the Information necessary to determine the
short-term peak concentrations and the variability of emission rates.
63
-------
TABLE 19. SUMMARY OF EMISSION RATE ESTIMATES
FOR VULCAN PLANT
Rates Based on
Equat ion 9
(kg h"1)
Rates Based on
Equation 11
(kg h-1)
Compound
Max Lmum
Min imam
Average
Ma x Laum
Mln Lbub
Average
cci4
30
0.1
11
20
0.4
9
C2Cl4
2
0.01
0.7
2
0.04
0. 5
C2HC13
20
0.04
6
20
0.1
5
C2H3CI3
2
0.01
0.7
2
0.04
0.7
64
-------
For a line-source length comparable to that used In this study, It would
b« possible to get valid averages for t Lme periods aa short as about
10 mln.
If the samples are collected at more closely spaced locations, then
It will be very Important to develop techniques to limit the number of
samples that must be analyzed In order to reduce costs. This could be
done In at least two different ways.
First, a screening technique could be developed. The SF^ results
were used to some extent during this study to decide which bags were to'
be analyzed for the chlorinated hydrocarbons. A rapid, inexpensive method
sensitive to chlorinated hydrocarbons might improve the screening process
even further. Another approach to reducing the number of samples would
be to limit Che extent of the sampling line itself. The results of this
study have shown that the emissions and the tracer affect only a limited
length of the total sampling line at a given hour. The problem, of course,
is to determine ahead of time just where the ^effects will be felt or to
make appropriate adjustments during the sampling operation. SRI recom-
mends that as much use as possible be made of U.S. National Weather Ser-
vice forecasts. Accurate forecasts of changes in wind direction would
allow each field rest to be planned In such a way that the samplers could
be moved during the course of a test If necessary. On-site wind-direction
observations are an obvious necessity. An attempt was made during this
study to change the sampling array following changes in wind direction,
but no plans for changes In deployment were formulated ahead of time,
baaed on forecasted wind-direction changes. Any redeployment of samplers
could probably be carried out more effectively if it were planned ahead
of time, using U.S. National Weather Service forecasts.
It should be noted that redeployment of samplers will introduce
difficulties. Some data will be lost while samplers are being moved
from place to place, but prior planning should reduce these losses.
Serious complications will arise if record* are not kept of the Simpler
locations at all times during a test.'' Careful bookkeeping with regard to
sampler location will be absolutely easential. Again, prior planning
based on weather forecasts should facilitate that bookkeeping
There may be diurnal patterns of activity in a plant, * that data
collected at night could be unren-^u .iatlve. Thei^tor^., .uturit
should include dcytlme measurements. The methodology described in this
report emphasized its application at nlghi, but the method can be applied
In the daytime as well. The best approach -jould be to apply the method-
ology only during periods of neutral atmospheric stability. Such sta-
bility conditions occur when skies are overcast, day or night. The tech-
niques and aquation* described in this report could be applied in the
daytlm" undar neutral conditions. If tests were undertaken during
periods of instability, it would be necessary to revir.e the aquations.
It is our recommendation that daytime applications of the method be
limited to conditions of overcsst skies with light winds.
65
-------
The patterns of activity nentloned In the preceding paragraph could
provide useful clue* regarding specific sources of observed chlorinated
hydrocarbon concentrations downwind of a plant. The plant studied In
this program is arranged so that It Is Impossible to observe and record
events at the plant froa outside and the lack of such Information has
proved to be unfortunate. Such Information might have provided reasons
for the observed enlsslon variability. At sites where It is possible,
It would be desirable to have an observer recording Information about:
(1) Truck arrivals and departures
(2) Movements of railroad tank cars
(3) Outdoor drum-filling operations that might be visible
(4) Changes In the visible emissions from the plant complex.
Such Information Is sometimes subjective and almost always difficult
to incorporate Into an automated data processing system, but as noted
earlier. It can be very valuable to the Interpretation of observed
concentrations.
The original Intent in developing this methodology was to release
the tracer from within the plant complex by using some sort of extended
manifold or multiple sources to simulate the areal extent of the fugitive
sources. For reasons noted in Section 1, It was not possible to do Chit.
However, SRI still believes that that approach would Introduce the
fewest uncertainties and would provide the most reliable results. Vhen-
ever It is possible to release tracers collocated with the sources of
Interest, the experiment should be conducted that way.
In summary, the methodology described in this report has been
demonstrated to be workable, although it still needs Improvement. The
method has already provided some Interesting estimates of emissions from
a chlorinated hydrocarbon plant, by using only measurements made outside
the plant area. -If the recommended'Improvements are Instituted, future
programs using this methodology could provide even better estimates of
emission rates.
66
-------
RtKiR-tNCtS
Buaaa, A.D., and J.R. Zlm«r**n. Ubitb Culd* tor th» Ci i«itolo| kli
Dispart Ion Modal. R4-73-024, U.S. Environmental Protection Agancjr,
Research Triangle Park, North Carolina, 1973. 144 pp.
Turner, D.B. Workbook of Atmospheric Dlaparalon Est lutes.
-------
18 July 1985
Rural Route #1
Benton. Kansas 67017
Morris Kay, Administrator
USEPA Regi on VII
72o Minnesota
Kansas City, Kansas 66101
re: Vulcan Materials Company
^y\jAA^o
Dear M
f tnr.-.rM
t, .jL 11 l_ <-J
I am enclosing a copy of testimony that 1 presented at a publ ic
hearing on 17 July 1985 regarding Vulcan's applications -for
repermitting of five underground injection wells at their Wichita
p1 an t.
I based my testimony on a review o-f a portion of the Vulcan files
at the Kansas Department of Health and Env ironment. The Vulcan
files are extensive and some of the information was withheld under
a C5I claim. Evidently an attempt wa = made by Vulcan to also
claim confidentiality on groundwater monitoring data but we did
gain access to a limited amount of analyses. The groundwater
monitoring information is spotty and according to the KDHE, no
organized assimilation of the data is available at this time.
Mrs. Donna Hinderliter had requested monitoring data from the EPA
but none was available and at the suggestion of someone in the UIC
branch, she contacted Mr. James Boyd of Vulcan for monitoring
data--and was refused.
The situation at and around the Vulcan site is very serious. I
had no idea of the magnitude of the problem until I reviewed the
Vulcan documents and have to assume that you weren't aware of it
either. The KDHE Oil Field and Geology Section has abused their
UIC authority by allowing such conditions to prevail. As you will
note, I have publicly petitioned the USEPA to conduct an
environmental audit of the area.
I still haven't acquired the art of condensing information, so I
wiH apologize in advance for asking you to personally review the
enclosed statement that I presented at last evening's hearing—but
please read i t .Ule were pleased that EPA was represented by Ted
Fri^r last night. Mr. Fritz observed the meeting but did not
e s t i f / .
b. u c e : 1 y
Sh ar t 1 yn Dien st
-------
My name is Sharilyn Dienst. I am going to begin my statement with
a request for an extension of the comment period. I appreciate
the opportunity to voice my concerns with any repermitting of the
Vulcan injection wells. I am opposed to the use of injection
wells as a means of hazardous waste disposal. I have made that
statement publicly numerous times based upon a general objection
to industry's use of our natural resources as a sewer for chemical
offage .
That argument is based upon the premise that the waste is actually
going to get to the Arbuckle. After reviewing files on Vulcan's
injection wells, I am also concerned with the effect that the
Vulcan injection wells have had on the groundwater- from the wel 1
head TP THE ARBUCKLE during the past several years.
A rsvifK of 3"=tilAble material indicates that every environmental
media has been contaminated in the Vulcan area. The extent of
water and air contamination both on and off-site indicate that
this may well be one of the major environmental problems in the
state of Kansas. I am basing that statement on groundwater
samples taken in April of 1985 and on an EPA study of air
contamination conducted in 1981.
In 1976, the Vulcan site was found to be in a highly contaminated
condition. An extensive amount of effort and money was reportedly
expended in an attempt to "clean up the site". Three years later,
Vulcan was pronounced a "CLEANED UP SITE" in an August 3, 1979
letter from Mel Gray of the KDHE to the Environmental Protection
Agency. At that time, the Vulcan site was removed from the
National Listing of Hazardous Waste sites. THAT IS AMAZING.
If the conditions that prevail today at the Vulcan site are what
the regulatory agencies deem "CLEANED UP", I can foresee some
great problems at FuMey when it's time to decide whether the
remedial work at that site is adequate.
Considering the air and water contamination frjm the Vulcan site
— and the fact that Vulcan has been identified by the USEPA as
one of tne =:tes in Region VII ¦-•hose m»nnf ar tur i no operations have
the potential to result in PI OXIN CONTAMINATION—:and that Vulcan
does produce high levels of PCBs in their perch 1oroethylene
pr oc e ss — I publicly petition the United States Environmental
Protection Agency to conduct an Environmental Audit of the Vulcan
Materials site and surroundino area to determine the exact extent
of water and air contamination and to assure the safety of the
citizens who are being exposed to unknown contamination in their
a~i r and water .
Nobody should h-"& to breathe chemically contaminated air and
nobody shni' i, De deprived of the use of their natural water supply
+ -¦ industrial gain. I feel sure that our governmental
. _> v knowingly condone such practices.
- 1 -
-------
Considering the condition that the Vulcan area is in, I don't see
how any assurance can be given that the Vulcan injection wells
have had no detrimental effect on useable groundwater.
In March of 1985, extremely high levels of chemical
contaminants—for instance 31,000 ppb of 2,4-D—were found in
on-site monitoring wells in all three aquifers. I saw no
monitoring results for PCBs or Dioxins—yet both are waste
byproducts of Vulcan operations. Groundwater appears to be
contaminated in all directions and in all water levels off-si te.
At least 15 private residence wells have been contaminated. I
understand that Vulcan has financed water line installation from
the city of Clearwater to most of these homes and installed carbon
filters at a few others.
Carbon filters may or may not reduce chemical contaminant levels
in drinking water. At best, filtered drinking water provides a
false assurance to the affected user. Bathing in chemically
contaminated water and inhalation of contaminated air reportedly
provides a greater source of exposure than actual consumption of
the water.
I would hope that the regulatory agencies do not consider such
acquisition of contaminated property as an adequate method of
dealing with ongoing water pollution from the Vulcan site. If so,
Vulcan conceivably could become the landlord of a lot of property
south of Wichita. PEOPLE ON THE OUTER EDGES OF EACH VULCAN
ACQUIRED PROPERTY SHOULD BE ALERT TO THIS FACT.
In 1981, an Air Emissions study was performed for the USEPA
outside the Vulcan site. The study revealed that Carbon
Tetrach1 oride, Trich1 oroethy1ene, Methylene Chloride and
Perch 1oroethylene were detected in all directions from the site
for a distance up to 3 miles. The four chemicals were found in a
range from Detectable to 36,200 parts of Trichloroethylene and up
to 5,600 parts of Perchloroethylene. The higher levels were
reported at a location near 55th Street between Hoover and Ridge
Roads—approximate1y one-half mile due north of Vu'can^
The study noted that chemical emissions come from on-site
operations such as LEAKS IN VALVES, INCOMPLETELY SEALED
CONTAINERS, FILLING AND EMPTYING OPERATIONS AND TRANSPORT OF
CHEMICALS FROM ONE PLACE TO ANOTHER ON SITE.
Neighbors of the site have complained of fumes resulting in
headaches and respiratory distress. One recently described fumes
at -her home as having a "sweet" odor. I was interested to read in
this morning's paper about a train accident that spilled Carbolic
Acid near El Dorado. The chemicals were on their way to the
Vulcan site and the Carbolic Acid was described as having a
"sweet, tarry odor".
WHAT FOLLOW LK Ar i . PK'S HAVE BEEN TAKEN BY THE REGULATORY AGENCIES
REGARDING i'Ht Ubcl-Vs .-i-PCRT OF CONTAMINATED AIR IN THE VULCAN AREA'
- 2 -
-------
INCINERATOR
Vulcan has proposed to meet Federal requirements -for a test burn
o-f their Uichita incinerator by submitting data -from a test burn
conducted in 1981 -from a much larger Incinerator at the Vulcan
plant in Lou i siana.
Vulcan became aware in 197? that POLYCHLORINATED BIPHENYLS
are present in their Perch 1oroethylene process waste streams. PCBS
AVERAGING 300 PPM HAVE BEEN OBSERVED IN THEIR HEX WASTE STREAM.
DDES VULCAN HAVE AN APPROVED PCB INCINERATOR"5 IT WAS MY
UNDERSTANDING THAT PCB INCINERATORS ARE HIGHLY REGULATED AND FEW
IN NUMBER.
IN THE STACK SAMPLING DATA, PCB DESTRUCTION EFFICIENCY lb RATED AT
99.99992V. BASED ON LIMITS OF DETECTION. However, those limits o-f
detection are shown at 10 parts per billion.
IS THE PCB BURN DATA FROM THE WICHITA INCINERATOR OR FROM THE
LOUISIANA INCINERATOR? IF THE DATA IS FROM THE LOUISIANA
INCINERATOR. DOES THAT MEAN THAT THE WICHITA VULCAN PLANT IS AND
HAS BEEN USING THEIR INCINERATOR FOR PCB INCINERATION BASED UPON
DATA FROM ANOTHER PLANT IN ANOTHER STATE?
HAS ANY AMBIENT AIR MONITORING BEEN CONDUCTED IN THE VULCAN
VICINITY SPECIFICALLY FOR PCBS OR FOR ANY OF THE OTHER CHEMICALS
AT THE VULCAN PLANT?
- 1 -
-------
I've been to several deepwell hearings in the past couple of
years. I've seen sketches of injection wells drawn on a
blackboard showing three protective casings—corrosion resistant
cement bonding around the casing—corrosion resistant tubing. At
a hearing three weeks ago, I listened to the KDHE staff explain
how mechanical integrity tests are conducted to assure foolproof
operation of injection wells. I've heard about the importance of
compatibility between the receiving formation and the injection
was t e s.
After reading the history of the Vulcan injection wells—I have
come to the conclusion that the KDHE and I are in great
disagreement as to what constitutes a problem. .
-I would like to ask what the requirements for Mechanical
Integrity Tests hav? been in the past for the Vulcan wells?
A note from a Vulcan meeting says:'LOGS ON UELLS SUBSTITUTE FOR
M.I.T.S". Is that correct? Are MITs required or»or to problems
or only as a reaction to problems?
Another memo from a Vulcan meeting with EPA last year notes that
new integrity testing of the casing is required. Vulcan claimed
that they had done an equivalent assurance of the integrity of the
casing and asked if this would suffice. Vulcan threatened to
appeal the requirement.
-Why does the Vulcan Draft Permit allow a 50'/ reduction in annulus
pressure when the KDHE required notification of a 25'/. reduction
by the High Plains well?
-Senate Bill #120 has been signed into law by the Governor. This
bill requires an application fee of $10,000 per well for existing
we 11s.
Has Vulcan submitted the $50,000 in application fees?
-Has a current review of disposal alternatives been provided"9 The
document ihat I saw was undated but dwelled mainly on the
financial feasibility of alternatives.
-What studies have been undertaken to protect oil and gas
production in the Vulcan area? Has the impact of lateral
displacement of brine in the injection zone been investigated
fully? Oil production exists within 3 miles of Vulcan. A well in
Florida caused pressure effects 40 miles away from the site.
-Considering the existing contamination of air and water from the
Vulcan site, how are the requirements of Section 213 of the RCRA
Reauthorization going to be met'
In the recently releistu, to Congress on Underground
Injection Wells in the United States, a statement was made that
sticks in my mind:
-------
THE COMMON PRACTICE AT A FEU OF THE FACILITIES HAS BEEN TO REWORK
AN INJECTION WELL ONLY AFTER LEAKS ARE DETECTED.
Vulcan was not among those sites visited but it is apparent that
Vulcan -falls into that category—and has been allowed to do so by
the state regulatory agency.
The history of the Vulcan wells is abysmal and a far cry from what
is presented by deepwell proponents.
THE WELL CASING IS USED TO PREVENT CONTAMINATION OF UNDERGROUND
SOURCES OF DRINKING WATER BY CONFINING THE INJECTION FLUID INSIDE.
A CRUCIAL INDICATOR OF WELL FAILURE IS THE ANNULUS PRESSURE.
LEAKS IN THE CASING CAN BE DETECTED BY A DROP IN ANNULUS PRESSURE.
EVEN A VULCAN REVIEW OF THEIR WELLS NOTES THE FOLLOWING REGARDING
THE IMPORTANCE OF THE ANNULUS ZONE:
"THE ANNULUS PRESSURE ASSURES THAT. SHOULD A LEAK OCCUR IN THE
INJECTION STRING. NO WASTEWATER COULD FLOW INTO THE ANNULUS. THE
WASTEWATER WOULD REMAIN CONTAINED.'
The following is a brief summary of six of Vulcan's wells,
including the five for which they seek repermitting.
VULCAN WELL *3
3-29-77: ANNULUS PRESSURE VARIED ERRATICALLY AND DROPPED TO ZERO.
When the well was pulled, several deteriorated gaskets, thread
imperfections on casing joints and EXTENSIVE CASING CORROSION WAS
FOUND. THE CASING REVEALED DAMAGE. AND LEAKS FROM THE 500 TO 700
FOOT LEVEL.
3-31-77: ' TUBING REINSTALLED BUT COULD NOT RE-ESTABLISH ANNULAR
PRESSURE. The following is a direct quote from a letter from
Vulcan to Bryson "ME:
NEVERTHELESS. THE LARGE VOLUME OF IMPOUNDED WASTEWATER REQUIRED
THAT WE RETURN THE WELL TO SERVICE.
THE WELL WAS USED FOR 18 DAYS WITH NO ANNULUS PRESSURE and WITH
NO REPAIR. Although no firm conclusion was ever reached as to the
problem, well #3 was put back into service.
1-19-80: TUBING PULLED BECAUSE OF BRINE SLUDGE BLOCKAGE.
TWO MONTHS LATER (APRIL 1980): TUBING PULLED DUE TO BRINE SLUDGE
BLOCKAGE.
SIX MONTHS LATER (OCTOBER 1980): WELL SUFFERING REDUCED FLOW OWING
TO BOTTOM HOLE BLOCKAGE.
FOUR MONTHS LATER (FEBRUARY 1981)i i"UB;Nb .^LUttGED AND DRILLED
THROUGH HARD BLOCKAGE AT 3980'.
- 2 -
-------
LATER THAT MONTH: ANOTHER BLOCKAGE DEVELOPED IN TUBING STRING.
IN 1983, 141,900,000 GALLONS OF WASTE WAS INJECTED THROUGH THIS
WELL.
VULCAN IS SEEKING REPERMIT FOR WELL #3
WELL W4
JULY 1982: REDUCTION IN ANNULUS PRESSURE. The repair work is as
f oJ 1 ows:
-The 141st joint o-f tubing had separated leaving the remaining 57
joints o-f tubing in the bottom of the hole.
-Fibercast tubing cowered with very thick, black
sludge — apparently the result o-f chemical contamination o-f the
annulus oil be 1ow the tubing leak.
-Re-ference is made to damaged bottom hole conditions.
-Hole in joint at 2074" and at 2850' depths.
-Several thread -failures.
-Older pipe very badly deteriorated inside and at the ends.
On July 14, 1982 several new areas o-f possible casino damage were
revealed--most significant beino at 2076 and 2857' depths.
-Other areas o-f metal loss at depths o-f 2273, 2292 and 2601'.
-A blockage was noted at 3413' depth and it was noted that
PREVIOUS LOGS HAVE SHOWN SEVERE DAMAGE BELOW THIS DEPTH.
On July 1? i coring r-p s n?ta 11 at i on o-f the tubing, the tubing
stopped at' the 196th joint. "ALTHOUGH THE TUBING IS HUNG UP
DOWNHOLE, IT IS NOT ACTUALLY ANCHORED AT BOTH ENDS". After
unsuccessfully attempting to free the string, a meeting was held
with KDHE and PERMISSION WAS GRANTED TO RETURN THIS FAULTY WELL TO
SERVICE.
IN 1983, 109,600,000 GALLONS OF CHEMICAL WASTE WAS INJECTED
THROUGH WELL #4. VULCAN SEEKS REPERMITTING OF THIS WELL.
WELL *6
7-5-78: ANNULUS PRESSURE FELL. PRESSURE TEST REVEALED 1 TAKS AND
DAMAGED THREADS OR LINER. ONLY 39 OF 196 JOINTS FIT.fQR REUSE.
8-14-79: ANNULUS PRESSURE FELL. 'THE NhXT DAY huiD SERVICE WAS
TERMINATED" . I ASSUME THAT THIS MEANS THAT FOR A FULL DAY, THE
-------
ACID SERVICE WAS CONTINUED ALTHOUGH A LEAK IN THE WELL WAS
INDICATED BY THE DROP IN ANNULAR PRESSURE.
OCTOBER 197?: CASING DAMAGE WAS FOUND AT THE FOLLOWING INTERVALS:
SLIGHT
SLIGHT
COMPLETE FAILURE
SEVERE DAMAGE TO COMPLETE FAILURE
COMPLETE FAILURE
SEVERE
COMPLETE FAILURE
SEVERE TO MODERATE
MODERATE TO SLIGHT
COMPLETE FAILURE
SEVERE DAMAGE TO COMPLETE FAILURE
NO STEEL CASING PRESENT
MECHANICAL CALIPER TOTAL DEPTH
Vulcan met with KDHE's Oil Field and Geology staff and it was
decided to repair the casing damage.
-The -first cement squeeze would not hold pressure.
-They had trouble with the third squeeze.
-The inflatable Lynes plug partially deflated and moved down the
hoie for 9 feet.
-Attempts to pull or push the plug resulted in getting it stuck in
the hoie.
-While drilling out the cement, the formation cuttings at 3949'
depth changed. The author notes: "I believe that we started
drilling a new hole at this point. Circulation was lost at 3970'
and never regained. A new hole was drilled to 4072'."
3150-3175
3215-3238
3238-3254
3254-3278
3278-3295
3295-3352
3352-3357
3357-3458
3458-3590
3590-3600
3600-3745
3475-3938
3938
12-28-79: Fibercast tubing eventually landed at 3965' and
disposal of contaminated water began. The consultant noted:
"THE PROBLEM WITH USING THIS WELL FOR ACID SERVICE IS THAT
FIBERCAST TUBING FAILURE IN THE LOWER PORTION OF THE HOLE IS HARD
TO DETECT. "THERE"ARE HOURLY FLUCTUATIONS IN ANNULUS PRESSURE AND
THESE CHANGES COULD 'MASK* A TUBING FAILURE."
The well was returned to service for 'primarily* 'essentially"
neutral or basic wastewater. A profound statement was made at
this point: "DIFFICULTIES WILL ARISE ONLY AFTER THE TUBING
FAILS".
3-29-80: ANNULUS PRESSURE AND TUBING PRESSURE EQUALIZED AND THEY
THOUGHT THE TUBING WAS EITHER PINCHED OR COLLAPSED AT ABOUT 3800
FEET.
A year later a KDHE geologist drove by the Vulcan plant and
noticed that a "rig was positioned over Vulcan #6". He stopped to
confer with Vulcan officials. It is noted that Well #6 had been
down for approximately 60 days with no evident not i f i r? t i r.n - the
Depar tmen t.
- 4 -
-------
Well M6 had STOPPED TAKING FLUID IN MAY 1981. When the tubing was
retrieved, the lower part was badly twisted.
The KDHE Geologist notes that he thinks that the CASING HAS PARTED
AND IS OFFSET.
An inspection log clearly showed extensive casino damage above
190'. The 7" casino was completely pone -from a depth o-f 82' to
109/
Maintenance on Well #6 was -finally discontinued and the well was
abandoned—NEARLY TWO YEARS AFTER A CASING INSPECTION REVEALED
EVERYTHING FROM SEVERE DAMAGE TO NO CASING AT ALL NEARLY 1000 FEET
ABOVE THE DISPOSAL ZONE.
WELL W8
December 1983: LOSS IN ANNULUS PRESSURE AND MAJOR MAINTENANCE.
-COMPLETE SEPARATION AT 30TH JOINT OF TUBING STRING.
-SOME TUBING LEFT IN OPEN HOLE.
MAJOR MAINTENANCE CONTINUED THRU EARLY JANUARY 1984.
85,400,000 GALLONS OF WASTE WAS INJECTED INTO WELL #8 IN 1983.
ON MAY 16, 1985: LOSS IN ANNULUS PRESSURE. AFTER LOADING ANNULUS
SEVERAL TIMES AND CONTINUING TO LOSE PRESSURE. VULCAN CONCLUDED
THAT THEY PROBABLY HAVE A HOLE IN THE TUBING. SEVERAL ATTEMPTS
WERE MADE TO FIND THE PROBLEM AND IT WAS FINALLY DECIDED THAT THE
ANNULUS OIL HAD WATER IN IT. THE OIL WAS REPLACED AND THE WELL
RETURNED TO SERVICE ON JUNE 4. 1985—A MONTH AGO.
CEMENTING
THE MAJOR FUNCTIONS OF THE CEMENT THAT IS APPLIED BETWEEN THE
OUTER WALLC OF THE CASING AND THE BOREHOLE OR OTHER CASING ARE TO
RESTRICT MOVEMENT OF FLUIDS BETWEEN THE SURFACE AND THC SUBSURFACE
OR BETWEEN DIFFERENT STRATA IN THE SUBSURFACE, TO SUPPORT THE
CASING, TO PREVENT POLLUTION OF UNDERGROUND SOURCES OF DRINKING
WATER AND TO PREVENT CASING CORROSION.
THE USEPA REPORT ON UNDERGROUND INJECTION STATES THAT IN ALL
CASES, CEMENT IS APPLIED IN AT LEAST ONE STRING, FROM THE SURFACE
TO BELOW THE BASE AND AT THE CONFINING ZONE ABOVE THE INJECTION
ZONE.
THE REPORT NOTES THAT WHEN THE WELL IS DRILLED, A CONDUIT IS
CREATED FOR COMMUNICATION BETWEEN THE DIFFERENT STRATA AND UNLESS
AM ADEQUATE CEMENTING PROGRAM IS FOLLOWED. MOVEMENT OF FLUIDS
COULD OCCUR AT THE INJECTION ZONE INTO OTHER FORMATIONS OR BH'UEfc'M'.
FORMATIONS PENETRATED BY THE WELL.
- 5 -
-------
THE EPA REPORT ERRONEOUSLY LISTED ALL .FIVE OF THE VULCAN WELLS AS
BEING CEMENTED TO THE SURFACE. WELLS #7 AND M9 ARE NOT.
WELL «7
In September 1976, KDHE granted Vulcan permission to construct
Well #7 •for use of highly acidic waste. Well #7 receives acid
waste with a pH of .1 to .7.
(The information in the USEPA Report erroneously showed this well
as receiving waste with pH of 1.5 to 13.0.)
A letter written TWO MONTHS AFTER PERMISSION TO CONSTRUCT REPORTS
THAT THE CEMENTING PROCESS ON WELL tt7 FAILED AT A DEPTH OF 2400'.
DOCUMENTS NOTE THAT CEMENT BONDING BETWEEN THE 7* CASING AND THE
FORMATION IS NOT CONTINUOUS THROUGHOUT THE ENTIRE LENGTH OF THE
HOLE. A CONSULTANT NOTES; 'even if the cement bond were to fail
and fluid from the Arbuckle comes in contact with outside of
casino. ONLY CORROSION WOULD RESULT."
MARCH 1977: PRIOR to use of Well #7, a letter from KDHE notes:
"We were quite concerned, although not surprised. to learn that
Well 87 has developed a STATIC COLUMN OF FLUID WHICH HAS A LOU PH.
THIS OF COURSE SUBSTANTIATES THAT COMMUNICATION EXISTS WITHIN THE
ARBUCKLE FORMATION BETWEEN WELL <47 AND ONE OR MORE OF THE DISPOSAL
WELLS IN CURRENT USE..A STATIC COLUMN OF ACIDIC FLUID WOULD HASTEN
PIPE CORROSION AND THEREFORE JEOPARDIZE THE INTEGRITY OF THE
CASING."
One month later <4-77) the KDHE approved Vulcan/s application to
use the well and approved the casino EVEN WITH THE LACK OF TOP TO
BOTTOM CEMENTING.
SIXTEEN MONTHS LATER; (AUGUST 1976) fHE rtED PIPING WOULD ONLY
TAKE FEED AT 250 GPM.
NINE MONTHS LATER (MAY 1979): A DECREASE IN ANNUlUS FRESS'JRE WAS
REPORTED. INSTEAD OF SHUTTING THE WELL DOWN. THE COMPANY MERELY
SWITCHED WASTE STREAMS AND CONTINUED TO USE THE WELL FOR TWO MORE
WEEKS.
WHEN THE TUBING WAS FINALLY PULLED, A DAMAGED SECTION WAS FOUND AT
1-430 FEET DEPTH.THE WELL WAS RETURNED TO SERVICE 3 DAYS LATER.
FIFTEEN MONTHS LATER (SEPTEMBER 1980): ANNULUS PRESSURE DROPPED.
As the string was removed, IT PARTED AT THE 6TH JOINT WHERE IT HAD
UNDERGONE CORROSION.
IN 1983, 151,800,000 GALLONS OF WASTE WAS DISPOSED OF THRU THIS
WELL.
APRIL 1985; LOST ANNULUS PRESSURE.
- 6 -
-------
-HOLES PRESENT IN 7" CASING FROM 3848' AND BELOW.
-A TUO-FOOT SPLIT WAS OBSERVED IN 5TH AND 8TH JOINTS FROM BOTTOM.
-FOUR MORE JOINTS SPLIT DURING PRESSURE TEST.
-LEAKS IN THREADS OF 168TH JOINT FROM BOTTOM.
-ELECTRONIC CASING CALIPER LOG INDICATED POSSIBLE DAMAGE IN BOTTOM
100 FOOT OF HOLE.
AFTER A KDHE-VULCAN MEETING, LARRY KNOCHE OF KDHE SAID IF THE WELL
COULD PASS A PRESSURE TEST, IT COULD BE RETURNED TO SERVICE. THE
WELL PASSED THE TEST AND WAS BACK IN SERVICE ON MAY 14. 1985—TWO
MONTHS AGO.
WELL #9
-------
8. pH VIOLATION CAUSING CORROSION OF WELL
ALL OF THESE THINGS HAPPEN RATHER ROUTINELY UITH THE VULCAN tJELLS
AND VET VULCAN IS LISTED AS HAVING NO PROBLEMS.
WHERE DID THE USEPA OBTAIN THE INFORMATION REGARDING THE VULCAN
DELLS?
- 8 -
-------
GROUNDWATER
As the largest generator of hazardous waste in the state of
Kansas, Vulcan manufactures ammonia, chlorine, caustic soda,
hydrogen, chloroform, carbon tetrachloride, methylene chloride,
perch 1oroethy1ene and Pentach1 orophenol. Wastes from the Vulcan
plant include all of the above and hexach1orobenzene, solvents,
PCBs, Dioxins and other assorted hazardous wastes.
UJith the limited information available from groundwater
monitoring, it isn't possible to take a comprehensive look at the
groundwater conditions. What I can see is that most of the
a-f or emen t i oned chemicals — along with several others— have been
detected in three levels of groundwater as recently as four
months ago.
Intercept wells, designed to retrieve the contaminated
groundwater, have been operated for several years on and near the
site. The highly contaminated water pumped from the ground is
then injected into the Arbuckle along with Vulcan's chemical
waste.
The Vulcan site has a 41.4 acre chemical waste landfill which has
been covered with six feet of clay and 'is monitored extensively".
Scientific studies have found that chemicals attack soils, causing
dissolution of the clay, and allowing the chemical waste to enter
the groundwater. It has become commonly known that landfills leak.
The state of Kansas found that out in 1982. Putting a top on a
leaking landfill does not preuent the migration of chemicals from
the clay sides and. bottom of the landfill. Vulcan has stated that
their concern is not with the groundwater On the site out beyond
their boundaries. It would be nice if the chemicals in the
groundwater WOULD remain at the •."^ftcwl ine, but the/ haven't, they
don't and they won't.
In March of 1985, some of the chemicals found in the groundwater
on site, in all levels of groundwater include:
ORTHO-CLOROPHENOL 3800 PPB
2,4 DICHL0R0PHEN0L 7500
2,4,6 TRICHLOROPHENOL 5300
¦2,6 DICHL0RDPHEN0L 1200
2,4-D 31,000
2,6-D 22,000
2,4,6-T 4500
BEN2ENE 384 PPB
TETRACHLOROETHYLENE 241
LHLOROPHENOL 5600
THIS IS NOT A LIST OF CHEMICALS FOUND PRIOR TO THE "CLEAN UP" OF
THE VULCAN SITE. THIS IS A MONITORING REPORT FROM APRIL OF 1985.
-------
HAVE ANY PCB ANALYSES BEEN CONDUCTED ON THE GROUNDWATER ON AND OFF
OF THE VULCAN SITE? UP TO 300 PARTS PER MILLION OF PCBS ARE FOUND
IN THE PERCHLOROETHYLENE PROCESS HASTE STREAMS.
The condition of the groundwater beyond the Vulcan boundaries is
alarming. Groundwater appears to be contaminated in all directions
in all three water levels.
SOUTHEAST OF VULCAN, the aquifers contain B of the -chem i cal s found
in on-s i te we 11s.
SOUTHWEST cf IMcan, the groundwater contains ten of the
chemicals.
Northeast of Vulcan, analyses reveal 12 of those chemicals. The
carbon tetrachloride levels in that area are nearly 8 TIMES HIGHER
THAN IN 1984 despite the use of interceptor wells.
Twenty-eight chemicals were found in shallow and deep aquifers
SOUTH of Vulcan. This well appears to be at the south end of the
landfill and would create a reasonable conclusion that the
1andfill is 1eak i ng.
At least 15 PRIVATE RESIDENCES HAVE CONTAMINATED UJELL UIATER. The
wells contain up -to NINE CHEMICAL CONTAMINANTS ranging from less
than 1 FPB to 35 parts per billion of individual chemicals—the
same chemicals in the aforementioned groundwater samples. It is
my understanding that Vulcan has purchased property with
contaminated groundwater and financed water line installation from
the city of Clearwater to supply these contaminated homes. A
memo titled Vulcan Meeting Notes states that "Carbon filters
installed on many residents' well water. Carbon filters may or
jnjay not reduce chemical contaminant levels in the drinktno water.
At the best, filtering drinking water provides a false assurance
to the affected citizens. According to the National Journal of
, Public Health, skin absorption and inhalation represent a
significant route of exposure to chemicals. In other words, if it
is indeed possible to filter • chemicals from dr i nk i no
water—bathinca in contaminated water and inhalation of
con tarninated air apparently provides a oreater source of exposure
than actually consuming the water.
While that may be a responsible action for the company to take, I
would hope that the regulatory agencies do not consider such
acquisition of contaminated property as an adequate method of
dealing with ongoing water pollution from the Vulcan site. If so,
Vulcan conceivably could become the landlord of a lot of property
south of Wichita. PEOPLE ON THE OUTER EDGES OF EACH VULCAN
act.-/_ property should be alert to this fact.
------- |