-------
Figure 2-13. Benzene Concentration (Rate of Decrease with Distance)
a.
d
01
C
o
CD
§
C5
C CD
Otfl
CO
O 2
^ o
OCD
X
«
eo
o
o
o
CO
o
CM
QO
aa fit
a on
Q=^ / _.
CSfiSiS
o
o
o
CO
o
o
o
O
o
o
CM
O
O
o
o
o
o
o
CO
o
o
o
CO
o
o
o
o
o
o
04
ii i i i i i In 11 i i i i
MI i in111 i i i
OJ
_c
'«
01
CD
-CD
CD
O
CO
-a
c
o
a
CM
O
o
O
(eiu/6n)
22
-------
similar (about -0.0001). While it is reasonable to assume that contaminant levels continue to
decrease with distance, it cannot be demonstrated that the slope is different from zero given the
high degree of variability in the concentration data. In other words, the rate of decrease beyond
1000 feet is very slow, and the data cannot support that it is different from zero. This is, at
least partly, a result of combining data from different fires into a single analysis. Other sources
of variability, as previously discussed, may also be important.
From a practical standpoint, it is not very useful to apply the formulas obtained from the curve
fitting to obtain estimates of concentration at a given distance. The data do not strongly support
doing so and the rate of decrease is so slow that the difference in concentration between any two
distances beyond 1000 feet is insignificant as a practical matter.
For practical purposes, a single statistic should adequately characterize contaminant levels
beyond 1000 feet. The "average" (some suitable central tendency) concentration level for all
samples collected at more than 1000 feet from the fire is representative (within some confidence
limits) of the concentration at any distance from the fire greater than 1000 feet. In addition, a
reasonably conservative estimate of the maximum expected value (e.g., maximum, second
highest value, or 90th percentile) can be obtained for any distance over 1000 feet as the likely
maximum expected value for all distances over 1000 feet.
Wind direction information was preserved in the scatter plots in order to further examine
whether concentrations appear to be dependent on recorded wind direction. The plots do not
show any clear distinction between concentration levels for downwind, missing, and variable
wind directions.
23
-------
Figure 2-14. Benzene Concentration (Rate of Decrease at > 1000 feet)
a.
CM
s
c
o
•™ CD
3
0)
C5
o i
C a,
02
OCD
Q
s
C
0)
ffl
i
OC
a a
aa
o
CO
O
a
q
d
O)
O)
TT
CM
h-
00
CO
I I I I I
o
o
o
CO
o
o
o
o
o
o
CM
o ^-.
g o
o «
CD
c
"w
o
o
o
00
o
o
o
CO
o
o
o
o
o
o
CM
O CD
11
A CO
li
01
*-•
en
Q
$
I
o
a
i
a
CT CM
CM
O
CM ,-
O
ra CM
o
O
(etu/6n)
24
-------
Figure 2-15. Toluene Concentration (Rate of Decrease at > 1000 feet)
O
">
o £
w o
^^ ®
£.
O
O
O
o
o
o
00
o
o
o
CO
o
o
o
o
o
o-
04
CM
o
n CM ,-
o
P3 CM
o
a
o>
c
'co
OJ
CO
^A CO
O ,
i >
CO
"w
b
.73
c
I
o
a
(SLU/6n) UOUBJJU30UOO
X
25
-------
Figure 2-16. Styrene Concentration (Rate of Decrease at > 1000 feet)
s
^-
05
O 5
^^* rt)
•? to
O co
OS
CD
Q
C**-
o
O o
CO
*-*
CO
CO
00
o
o
«
CO
in
oo
CD
Tf
CO*
CM
a o
! 1 II 1 1 1 1
l| 1 1 I 1 1
O)
'to
03
o
o
o
CO
o
o
o
o
o
o
CM
o ^
o *
o fi
"*" O CD
o §3
° T "5
O A Co
00 ^>
o '
c >
o
o
o
CO
o
o
o
Tf
o
o
o
CM
CO
b
5
o
Q
CM
m
o
CM
CM
o
CM ,_
o
CM
o
o
(eiu/6n)
26
-------
3.0 DISCUSSION AND CONCLUSIONS
3.1 Data Summary
Because a distributional model has not been established for the data value, non-parametric
statistics (median, confidence limits, about the median, 90th percentile, and maximum) were
selected to describe contaminant concentrations at distances less than and greater than 1000 feet.
The summary statistics derived for the 17 contaminants are presented in Tables 3-1 and 3-2 for
distances less than 1000 feet and greater than 1000 feet, respectively. Based on experience with
the distribution of the data, the median serves as a reasonable measure of central tendency.
Upper and lower 90 percent confidence limits were established to provide a range of uncertainty
about the median. These limits were calculated so that the skewed distribution was adequately
represented without using highly sophisticated methods.
Both the upper and lower confidence limits are determined by counting up and down from the
median the number of data points specified by a, where a is chosen so that the probability that
the true median falls within the confidence limits is approximately 90 percent. This is based on
the probability that a given number of points will fall to one side or the other of the median.
Since the "true" median represents the middle of the population, the probability that a given
point will fall on one side of the median versus the other is 50 percent. Thus, the problem is
identical to determining the likelihood of obtaining more than a given number of heads in N coin
tosses. The number of "heads" is one half of the sample size plus one half of the confidence
interval (N/2+a). To simplify computation, the normal approximation to the binomial
distribution is used for N greater than 12. For N less than 12, intervals are calculated based on
the binomial distribution using a table of cumulative binomial probabilities found in Wonnacott
(Wiley 1985).
While the median, with its associated confidence limits, characterizes the central tendency of the
data, maximum expected values are more important from a public health perspective. The
27
-------
Table 3-1. Summary Statistics ("Fingerprint") (Distance <1000 Feet)
Analyte
Benzene
Toluene
Styrene
Xylenes4
m,p-Xylene
o-Xylene
Methylene Chloride
Chloroform
Ethylbenzene
Trichloroethene4
1,1,2-Trichoroethane
1,1, 1-Trichoroethane
1, 1-Dichloroethane
Chlorobenzene
Trichloroethane4
Carbon Tetrachloride
Tetrachloroethene
N
101
94
86
41
30
49
39
33
57
45
33
43
26
33
17
31
28
Fires
21
21
14
9
6
10
10
9
12
11
7
12
10
11
7
10
9
Units of Atg/m3
Median
121
220
85
17
76
35
8
42
49
0
0
0
0
0
0
0
0
90%
LCL1
33
38
20
0
1
1
0
0
0
0
0
0
0
0
0
0
0
90%
UCL1
525
527
174
607
282
109
89
197
204
41
82
10
0
0
1
0
0
"a"2
17
16
15
11
9
12
10
9
12
11
9
11
8
9
7
9
9
90th
Pent3
6375
3766
2320
1424
912
336
565
533
502
425
316
39
16
2
1
0
0
Max
79693
206753
2705
3809
999
564
836
1085
1477
881
542
817
42
11
1
44
0
1 The 90 percent confidence limits lower and upper as determined for the median
2 Where a is the number of data values from the median to the upper and to the lower 90 percent confidence
limits [derived from cumulative binomial probability table in Wonnacott (Wiley 1985)]
3 The analytes in this table are arranged in order of 90th percentile (except for the o-xylene isomer)
4 Contains mixed isomers
28
-------
Table 3-2. Summary Statistics ("Fingerprint") (Distance >1000 Feet)
Analyte
Styrene
Ethylbenzene
Toluene
Benzene
Xylene4
m,p-Xylene
o-Xylene
Chlorobenzene
1, 1, 1-Trichloroethane4
Trichloroethane4
Carbon Tetrachloride
Trichloroethene4
1, 1-Dichloroethane
1,1,2-Trichloroethane
Chloroform
Methylene Chloride
Tetrachloroethene
N
45
18
45
47
20
28
38
29
30
34
8
6
7
6
3
14
8
Fires
5
5
10
10
4
3
6
5
5
4
4
4
3
2
3
3
4
Units of /tg/m3
Median
1
3
5
4
0
2
1
1
1
1
0
0
0
0
0
0
0
90%
LCL1
0
0
1
0
0
1
1
0
0
0
0
0
0
0
0
0
0
90%
UCL1
16
172
37
29
0
9
5
1
1
1
0
18
0
0
0
0
0
"a"2
11
7
11
11
7
9
10
9
9
10
4
3
3
3
1
6
4
90th
Pent3
554
172
156
67
4
14
13
1
1
1
0
0
0
0
0
0
0
Max
2705
1390
634
524
20
999
521
1
7
3
0
18
0
0
0
660
0
1 The lower and upper 90 percent confidence limits as determined for the median
2 Where a is the number of data values from the median to the 90th percentile [derived from cumulative
binomial probability table in Wonnacott (Wiley 1985)]
3 The analytes in this table are arranged in order of 90th percentile (except for the xylene isomer)
4 Contains mixed isomers
29
-------
maximum values, however, tend to be extremely high for some analytes in this data set
(especially benzene and toluene) and may not provide a reasonable estimate of the maximum
concentration likely to occur. Such extreme concentrations seem to occur very rarely, and
probably do not persist for long periods. These extreme values may not, in fact, represent
ambient air concentrations at all. Other factors such as measurement or data recording errors
might be found to be responsible where such errors might still persist in the database. While
it is important, at this stage, to retain all values in the analysis, it may be reasonable to use the
90th percentile as a surrogate for the maximum value. The 90th percentile occurs more
frequently than the extreme values and is less likely to be later identified as an erroneous value.
3.2 Data Comparisons
Of the hundreds of potential tire fire air pollutants, only the 17 analytes shown in Tables 3-1 and
3-2 were common to many of the air monitoring efforts at tire fire incidents. To estimate
concentrations for more than the 17 analytes, the relative concentrations of the 17 analytes were
used as source profiles or "fingerprints." The fingerprints included both median and 90 percent
upper confidence limits (90% UCL).
These fingerprints were compared to the Hagersville data set because it is the most
comprehensive of those available from tire fire incidents. Only the Hagersville analytes that
showed decreasing concentrations over distance were considered in this fingerprint match, and
only values above the detection limit were used in calculating average concentrations for
Hagersville analytes. The results of the match are shown in Figures 3-1 and 3-2 for distances
less than 1000 feet and distances greater than 1000 feet, respectively. The Hagersville data set
had five analytes (ethyl benzene, xylene, styrene, toluene, and benzene) in common with Tables
3-1 and 3-2. The m,p-xylene and o-xylene data from Tables 3-1 and 3-2 were summed to give
the xylene values shown in Figures 3-1 and 3-2.
For distances less than 1000 feet, the average Hagersville values were comparable to the median
fingerprint, with the largest discrepancy being about 2.5-fold for toluene. The 90% UCL of the
fingerprint was much higher than the maximum values of the common Hagersville analytes. A
30
-------
fingerprint match was not obvious using the Hagersville data for distances greater than 1000
feet.
In addition, data common to the Hagersville incident and the EPA simulated open burning
(Ryan, 1989 and U.S. EPA, 1989) are compared in Figures 3-3 and 3-4. These figures include
data for analytes not in the fingerprint of 17 common analytes, but which were measured at both
Hagersville and the simulated burn. The average concentrations for these simulated-burn
analytes were greater than the Hagersville data with the exception of trimethyl benzene (1MB).
"Spikes" of TMB occurred at several monitoring distances at the Hagersville incident, suggesting
that sources other than the tire fire may have contributed.
3.3 Preliminary Conclusions
The Group 1 contaminants (benzene, toluene, and styrene) measured at the most fires, are
represented by a relatively large number of measurements, and exhibit the highest overall
concentrations.
While it seems reasonable to aggregate data across different incidents, further study is needed
to resolve the impact of individual fires on the combined analysis. Additional effort is also
needed to characterize the impact of variables other than distance and wind direction on
contaminant levels. The recorded wind direction data do not seem to have a strong relationship
to concentration. This may be due to inaccuracies and variations in recording practices.
Principal component analysis may be useful for further clarifying the relationships among
analytes and the characteristics of the tire fire incidents and monitoring parameters.
The concentration distributions show that there are many more cases of low concentrations for
each analyte at distances of both less than and greater than 1000 feet. It may be reasonable to
describe the distribution as log-normal. If so, the geometric mean would be the appropriate
measure of central tendency.
31
-------
For each analyte, concentrations appear to decrease very rapidly at first, and then very slowly
with increasing distance from the fire boundary. This may be an artifact of the monitoring
distances typically chosen at tire fire incidents.
The data can be reasonably summarized by statistics for measurements taken at less than and
greater than 1000 feet from the fire boundary. At less than 1000 feet, the data are clustered at
or near the fire boundary. At greater than 1000 feet, the rate of decrease is so gradual, that,
within confidence limits, a single statistic can describe concentration levels at any given distance.
The initial efforts to compare the Hagersville data at for distances of less than 1000 feet with
the fingerprint suggests that the Hagersville data may be useful in estimating tire fire pollutant
concentrations. The data collected during the EPA simulated open burning study appears to
show much higher concentrations of various analytes than the data collected at actual tire fire
incidents. Using the Hagersville data set, it may be possible to derive analyte-specific or
analyte-group factors to convert the simulated burn data to "real world" conditions at distances
less than 1000 feet.
32
-------
Figure 3-1. Hagersville vs. Fingerprint Data (Distances < 1000 Feet)
<
Q
Q_
LJ 8
O
V
LU
CO
oo" £
Ld J/5
(/">
CE
o
<
X
//////////A
//A
LU
z
LU
M
LU
CD
LU
z
LU
O
LU
Z
LU
o:
co
LU
z
LU
>
X
M
LU
CQ
_l
X
Q
LU
o:
o.
cc
Ld
o
o
LU
co
a:
Ld
o
o
O
o
o
o
o
o
o
o
CN
o
o
33
-------
Figure 3-2. Hagersville vs. Fingerprint Data (Distance > 1000 Feet)
Q
«•»
•^
Z
C£
LU
5
g
Q
00
LJ
c/i
(£
LJ
O
^^
KXXXX)^
^^
W///M
LU
2
LU
M
LU
CD
LU
Z
LU
O
_J
o
LU
LU
<£
(/)
LU
LU
_1
X
Nl
LU
CD
a
LU .
a:
a.
a:
LU
o
1
LU
NOIJ.Vy_LN33NO3
34
-------
Figure 3-3. Hagersville vs. Simulated Burn Data (Distance < 1000 Feet)
«
1
Q
z;
en
i—
fi\ '-'-
i™J-J t""^
Q
Q f
LJ v
| ' '"" |_jj
< ^
_l |
-^ LU
(/O |
^
. CO
CO y
> ^
01
Uce
LU
dl
1 1 i
LLJ
O
f
c
c x xx x Xx,XxXXxiX.X,x,X,XX,XiXx,xXX.X,XLx X
^\\\\^\\\\\\\\\X\X\\X\\\\N\\\\\\\\\\\NSX\N\\,\XX
*" KX//xX/x/
^ S !2^!
H2 )QCX Xxxxxxxxxxxxxxxx,xxx
,1. Jv\\X\XX\\\X\\\\X\X\\X\\X's
iT) '"
S KXXXxxxxxxxxxxx^xx
— JXXXXXAXXXxkxvv"
<"" v^,,,
^ l^^\\vSivx>OvX\s'
-^
to t X X X x x x x
u> [\tf?\\N
in fc
§ ^
00
»— k x x x x x
V t^VVvS
jt ^
^ b< X X XXXXXXXX X"X
OT KX\\XW
o
d
in 7-
rj- 5x.
*h
Ln K.X X X X
1*1 r~. JO-
S *4&
CM niiVV
I I 1 1 1 ! 1 1 1 1 1 1 1 I 1 ! 1 '!
3OOOOOOOOOOOOOOOOOOO
>i^^^-^- — — «- — — — odd dddddd
LU
^
LU
~Z
LU
LU21
^1
LU
O
"-LLJ
•z.
LU
C/l
_LU
LU
CD
ROPYLBENZ.
INZ. ETHYL
ANALYTE
0-5
11
< 2
— fsj
~z. q
~M^
>-
1 1 1
•— >
3
o
o
/•y
CO
Q
LU
1
(ft
r~ i
L\N
y
, MAX(UG/M3
~;
LU
rs,
yy
O
LU
— I
—
HAGERSV
lv
rO
URN.MAX(UG/
CD
CO
l/S
PX]
•
i
(spuosnom)
n) NOIJ.Va±N3ONOO
35
-------
Figure 3-4. HagersvUle vs. Simulated Burn Data (Distance > 1000 Feet)
CN
rO
I
^1^
Q
~z.
f^s
LL.
fT~l t
o
o
Q 2
1 , 1 A
^f ^
— ' 1
^> UJ
_I — '
(/i
> i
00
III UJ
y 1
00
cr
i i
1_LJ
<£
KXXXXXXXXXXXXXXXXXXXXXXXXXXX
I»^\\\\\\\\\\XX\\X\\\\\X\\\\\\\X\\X\\\\XNXX\\\\NS
B
to
in tC/'C^C/?! XXXXXXXXXXXXXXXXXX
— [\V\\\X\XX\\\\V\'\XV\\\'\X\'N
1
IT)
CM K X^( xxxxxxXXX?t
„,,... ...,,_..
03 C^XXXx/xxxxxx
^ I^\\\\\\\.\1\\\\M
1
fo [.xx^xxxx1
u> K\\\s
ro bf1 V
2 ^
*~ ICX X X X X
•*• KsvN
ik
04 (>iXXXXXXxxXxxx
01 ^S\\\\\\N
"*" "N
m JiTx" x" x" x"
I^) [yN
f77////
I
1 ! I i 1 1 ! ! ! 111!!
UJ
z
Z 0
UJ >
Qj * "5
5 £
UJ UJ
- <
Di I
•" ^3
ooooooooo
(spuosnoqj.)
(£uj/5n)
36
-------
APPENDIX A
DATA ACQUISITION
(Intended to be useful for scientists interested in using
the database system for additional research.)
A-l
-------
APPENDIX A. DATA ACQUISITION
A.I Data Sources
The tire fire incidents with air monitoring data were identified through telephone conversations
and written correspondence with persons identified by the Air RISC Hotline the EPA Emergency
Response Office in Edison, NJ; State and local staff in air pollution and solid wastes offices in
all 50 States; and a number of other government agencies, universities and business groups.
A.2 Definition of Data Requirements
In consultation with air monitoring and dispersion modeling staff, a listing of information was
produced that would be useful in evaluating the air monitoring data. This listing of site
characteristics that may influence contaminant concentrations included monitoring distance from
the fire, fire size and duration, and topographical and meteorological conditions. During
telephone conversations with the site contacts, the type and format of information available was
listed. This site contact information was used to determine the level of detail that could be
expected. A preliminary database system was written, consisting of about 50 topics. The
preliminary database system was then revised in several iterations reviewed by three site
contacts.
A.3 Receipt of Data
Tire fire incident information typically consisted of brief reports, laboratory data sheets, and
handwritten notes compiled from telephone conversations with tire fire incident contacts.
Inventory and follow-up actions were tracked using a computer file. Follow-up telephone
conversations usually provided additional information, although many fields in a given record
could not be completed. Air monitoring data sets for 22 tire fire incidents were entered into the
database system. The original numbering system is retained throughout this report; therefore,
the numbering of the 22 tire fire incidents is not consecutive. We were not able to locate what
we believe to be a large air monitoring data set for the Denver, Colorado tire fire incident that
occurred on June 11, 1987. The tire fire incidents in the database and summary of the number
A-2
-------
of analytes and the number of contaminant measurements for each incident are shown in Table
A-l.
A.4 Sampling and Analytical Methods
The sampling and analytical methods used for determining airborne concentrations of analytes
at tire fire incidents were considered in the decision whether to enter data into the database
system. For example, analytical equipment used to measure the actual concentrations of analytes
in a given volume of sample must have acceptable detection limits and employ appropriate
standards. For most of these incidents, sufficient information pertaining to the type of sampling
and analytical equipment used at each of the tire fire incidents was available.' This information
in presented in Table A-2.
A-3
-------
i
s
a
o
.
«
n
d
U
a
H
h
*
M
fl
EH
n
4-1
fl
e
ID
09
n)
0}
2
4J
0
H
U
4-1
rH
fti
fl
I
fl
ident
u
fl
H 0)
W Q
•H
M
•H
H
+J
•r-t
U
fl —
H 03
e
OJ rt
w d
-H ~—
Cu
M
-H
H
Q
H
H
00
m
0
rH
0
CJ
rri
T!
rd
id
u
o
nj
O
1>
r-H
•H
CO
M
CO
tn
™
rH
rH
in
CJ
00
ot
S
.J
dj
Q)
W
(N
00
N
CO
CO
H
d
i}
Cn
O
S
41
2
«
rH
Ot
CO
S
3
•H
4-1
rt
U
-
m
-<*
i-i
00
4J
rj
u
d
H
JJ
rrj
0)
•H
0!
in
-j
CO
CO
o
M
0
CJ
i-H
0)
Q)
1-4
.
5
rH
at
•^
"i
^
w
m
3
-
rH
CO
$
X
0
M
Q
00
00
f.
t-
CO
<4*
<0
rH
*^H
0]
a;
m
«
M
O
U
0)
1—1
0]
0)
d
0
H
M
at
CO
a*
r5
•H
d
rt
P
rH
CO
CO
3
•H
a
^
in
CO
H
CO
3
V
0
c
0
Q,
t-l
a
00
a
0
Jj
5
u
5)
01
00
rH
in
H
H
A
00
2
-^
0]
(D
S
0
o
tM
.
O
at
ft)
nJ
d
(fl
CJ
u
a1
O
-u
f-t
r-j
ffl
S
w
rH
OJ
CO
O
H
CO
_
2
S
•H
ffl
3
m
n
in
at
00
CO
i
j
>
0
a
01
«
CO
CO
N
U
2
O
T!
H
^
•H
H
(J
CM
in
CO
CO
a
QJ
<
^:
S
0
o
rt
H
CO
(N
3
at
CO
u>
H
£
En
a,1
V
u
~"
d
V
73
d
•H
3
O
m
rH
CO
c
0
c
—
ro
S
nj
*H
IT)
tu
n
Ot
in
at
in
n)
4J
0
EH
r-H 0
ni N
d
-------
JJ
a
9
•a
•H
U
H
(D
rl
CO
rl
•H
EH
X!
id
H
JJ
a)
•O
0)
n
a
•O
0
a
-H
a
•O
(0
Dl
ifl
CO
XI
id
T?
o
jj
S irt
Q)
•H i2
ffl "-1
rH
id
rij
^ id
JJ -H
G fO
CO CO
ft
-H 01
&*H
W rH Si
_J m
Q« |±
&1 S
C ro
-H CO
r-(
Ql (U
S P*
col?
jj
G
a)
•H
G~
H 0)
0) id
l-i G
-H ~^
Cu
(D
iH
-H
=8:
Q
rH
co
jgj
•~^
CJ
i rj
-H JJ
C X!-H
O £ CO
-H TJ Oi
JJ 0) iH *C3
ft-H CO -H
iH MH i-H ^
O -H ft JJ
CQ t3 E rl
co o 53 id
•a s m o
rH 1 CO M
id jj co
S ffl id jj
H 33 rH rH
0) CJ > Ol
^
o
-rl
iH
id
JJ
a
0
-.
D
rH
-H
[>
to id
iH T3
a) id
tn c
(d (d
ffi u
rH
•a
o
jj
1
a-
H
•a s
•H J-
3 Qj
O* (fi
r} Ol
0
1 JJ
id
S O -—*
\ rl 0
U £ rH
O U w>
^
id
c
CO
EH
•a
G
id ffl
£j
rH O
S^
CJ Ol
rl G
id -H
X! rH
CJ ft
•v. S
I'm
ft rl
0
G MH
id
•H ffl
rH CO
rH X)
-H 3
O JJ
fCIJ
3
~
^j
jj
0)
)H
OJ
m
•o
t* id ^
c^ ro 03
H [^ -- CO
^s-s
i 0) JJ
m — TH rH
CO tO-H CO
rl CO MH Ol
3 XI
T3 3 co id
CO JJ JJ U
u id-H
O rH rH i-H
rl Id 3-H
ft o u ffl
CJ-H —
K iH JJ
co id )H oo
O X! (d u>
g S3""1
ffl
Ol iH
G co
•H JJ
C* rH H
ft-H
"O S MH
§lu
ffl CO
JJ
co )H id
XI 0 rH ffl
3 MH 3 CO
JJ CJ JJ
ffl -H id
i-H CO JJ rH
Id XI rl 3
0 3 id CJ
CJ JJ ft-H
rl JJ
(d rH CO in
js cu ca id
CJ Ol O ft
•••». i-H
tid 3 Ol
OH C
-H H -H
CU H CO H
-H U ft
Cm S
id ' id
-HO) ffl
H Ol CO
H Id U rl
-H JJ O O
O m > MH
s
-
a
CO
*o
Dl
o
„
i^4
0)
XI
CO
ro
O
-H -a
JJ CU C
id n m
§ "" —
rl 0 CQ
id o c
E ~- in o
co H xi -a
O H # rl CO
HO (d JJ
2 in » u id
Oi*'ffl IH co "m
C G T3 Oi C
-H m O >i O O
ffl co XI X! H XI
3 rl rl Id rl
3 id ?-. £! id
Q 'O O ~ — CJ
H 0) O •_! O
fc U in 7 ro in
^ 0 T3 o o "O
CJ rl >,Vfl 0 >,
iH
0
MH
ffl
Of to G
C CO 0
-H XI XI
ft jj id
S o
3 G 0
m O rl
XI T)
0) rl >1
G (d X!
-H CJ
S "^- ffl
H T! 3
CO 0) 0
JJ ffl -H
CO 3 rl
•a id
4J >
4J C
0 CO Ol
C i G
TJ rl H
H JJ ft
3 CQ S
0 G 3
CJ -H ffl
ix
S
~
rH
r-t
-H
y
CD
JJ
rH -H CO -H H "D ffl
3 H O S O CO 3
•a -H H 13 ^ ffl
co co 2; 3 co
o cj f- • •« fa
3 g g CJ CD Cd
id O O h U XI g
O XI G (d ^8 •- 3
-H iH -H «— CU ^ JJ 0)
jj id ffl N
i-H 0) VD CO O Id H
id (UXlrHO ftG id
G XI 3 H id CO C
rtj JJ JJ i 13 > EH (d
id co
O CO
-H O
H ffl H
•H CO 3
Ol ffl •*£) H
G — 3 H
-H CO CD JJ CO
rH Ol CO CJ iH
ft id G XI 0
S JJ -H in 13 MH
id CO S O CO —
ca ' id m X
ro 0 -H *-.
co ^- cj s e t±<
G T) -H i ta
•H 0) JJ 0 --O
S ffl id in ~ g
in 3 e X! CQ —
0) 0 CJ iH
JJ -— iH \ 0 CO — -
0) ffl BJ X ft r* ffl
T3 -^ — id id co 01
jj a > jj jj
JJ C ffl 01 H Id
O CO 0) EH O -H rH
G g XI^-H MH 3
3 3 G G U
T) rl JJ 0 Id rl -H
H JJ XI Ol 0) JJ
3 ffl H ^ iH JJ iH
0 G 0) (d 0 CD (d
U -H O) O — CO ft
g
"in
0)
JJ
ffl
CO
u
a
-H
£2
• — *
4J
'^-4
m
43
0)
G
•H
5
in
co
^j
-s^
CJ
o
ffl
0)
XI
3
JJ
g
XI
SH
id
CJ
*o
fj
(d
X
id
G
0)
EH
ft
ft
G
id
-H
rH
H
-H
o
u
3
-
CO
01
o
--I
o
rH
(D
£>
0)
to
ffl
01
iH JJ
Bo id
> *" "3
Cu CJ
Olrf -H
G CJ JJ
-H H rl
N id
S •» ft
H ffl
id G in
GOO
Id-H MH
rl (d Ol
O rl G
MH JJ -H
fl N
CO 0) >i
g O rH
•^ G id
u o c
O 0 id
0)
rl JJ
0 (d
MH H ffl
3 rl
ffl CJ CO
rl-H JJ
CO JJ rH
JJ iH -H
ffl Id MH
-H ft
G rl ffl
id 0) 0) co
u i xi jj
3 -H id
U H MH H
O O 3
> > ffl CJ
i ffl -H
OI-H id jj
G J3 rH k
-H CD id
CQ •» ft
3 ffl Oi
U G 0)
0) O-H G
H > CO-H
ft 3 H
E Oi ft
id G in S
CQ -H 0) (d
H rH ffl
X) ft ft
id E 6 >H
iH id id O
O ffl ffl MH
-"jj
l>
^
TJ
rH
U)
-H
4-1
0)
rd
CO
s
— ^
0
ffl
CO
OlXI
G 3
-H JJ
H
trH
id
o
ffl CJ
JH
co id
G X!
-H CJ
£ "^^
p T3
CO CO
JJ CO
(U 3
•d
JJ
JJ C
0 0)
c g
T3 in
H JJ
3 ffl
o c
^
£>
~
^
rH
o
1-1
o
oo
CO
\
CJ
o
rH
CO
Ol
id
CJ
-H
rH
Ol-H
C ffl
-H
rH T3
t(d
CQ X
fd
CO G
G CO
-H EH
Ei -a
0) 01
JJ ffl
JJ
JJ G
O CO
J3
73 iH CO
H JJ 0)
3 ffl XI
0 C 3
U -H JJ
Qi
,<£
^
a)
rH
I — 1
-rH
J>
ffl
0)
iJ
rd
03
en
-------
•a
o
jj
$,<,
ill
m S
CO t>
rH
id
^ id
JJ -rl
G 'O
cu cu
g S
ft
-rl 01
&-rl
W rH g
Ol g
G (o
-H CO
rH
ft C
ca ca -H
fi CU T) g
O rH 0 fi
MH ftX!
CQ
1)
G
0
^
rH
CO
,a_
u
o
fi
0
MH
CO
CU
XI
3
jJ
rH
id
O
U
fi
id
XI
(J
3 o
ft>
Ol Ol
G G
•rl -H
i-H rH
ft ft
g g
td 10
CO CO
f3j
a.
^
D
rH
rH
-rl
J>
G
td
Q
H
co
S3
u
O
ca
^
Ol
G
-H
rH
ft
§
3
CO
fi
0
MH
CO
ffl
2
JJ
X
id
C
CD
EH
ffi
S
,,
(U
rH
rH
-rl
^>
J^
id
Q
in
H
co
.s
U
o
•o
0)
ca fi
3 O
ft
ft co
cu
>,£}
fi 3
CU JJ
JJ
JJ fi
id a>
XI 01
"* V-)
0) Q
rH
JJ T3
jj G ca
0 (d U
X ^
t3 id
CU G Ol
JJ CD G
id EH -H
3 rH
U X! ft
> -rl S
td S tn
<£
2
^
fi
(U
u
G
D
ft
CO
H
co
S3
CJ
U
ca
CJ
O
Oi
-H
rH
a
fo
[Q
^1
0
MH
CU
i-H
JJ
JJ
0
XI
T3
CD
JJ
id
3
U
id
w
S!
a
^
r^
s
0
JJ
fi
0)
XI
u
rH
(D
ffl
CO
H
• ».
CU CQ CO T3
N 13 d) 7; 0
>l CO G rH O X! CO
rH CU 3 ft CO JJ -rl
id X) ' o g cu ca
§3 -ri a 3 MH g s
jj g g 0 ca o I-H
CD O JJ CD id
O rH CQ U Ol CQ G G
4J id •o (d G - G S
\ (d to
ft Ol-rl
g G rH fi
3 -rl rH 0
fti-H -H MH
ftCD
G g ca
-H CO CO 0)
i-H CU JJ
•-H fi CQ i-H
-H 0 (d -H
O MH OlMH
M
£2
s.
iJ
a)
CO
^i
0)
o
CO
o
O]
co
S3
U
0
0)
§
X!
jj
cu
^1
*3
^
rH .
O
a
^i
cu
rH
ft
ca
cu
ca
rH fi
0 CU
> JJ
1 rH
-H -H
K MH
a5
T3
fO
C
flj
o
*,
CJ
CD
XI
3
Ol
^
(U
rH
X!
aj
^S
^J
CO
H
OJ
fi CU
id cu xi
rH G JJ
U TJ-H
O cu g cu
> JJ fi N
CU >i
Ol OlJJ rH
G G cu id
•H -rl T3 G
N ca id
s» 3 JJ
rH O 0
id G G JJ
§cu ca
AJ -a TJ cu
Id H CU XI
fi jj 3 ca 3
0 0 3 JJ
MH ca cj
CU T) fi
CO rH •» O CD
a ft CQ xi 01
\ G OlJJ (d
U ID (d CD fi
CJ w XI £ T)
CO
Ol
id
XI
V-t
id
i-H
T)
CD
EH
T)
G
id
ca
cu
XI
3
jj
^i
CD
01
id
Q
s
§
•.
td
5
m
co
G
0
-H
-•-^
U
o
tr
id
i-H
i-H
-H
ft
id
u
ca
U
O
^
Ol
G
-H
rH
ft
g
Id
ca
fi
0
MH
CO
Ol
id
XI
fi
id
rH
T)
0)
EH
g
S
X.
I >-J
0)
^
0
T)
£1
5
O]
•o
rH
3
O
U
ca
-HTJ
ca CD
^i G
rH -H
id g
G fi
id CD
JJ
IMH 0)
0 T3
T3 CU
0 XI
X!
JJ JJ
CU 0
a G
H
0
MH CU
G
CQ 0)
cu 3
52
fi T>
CU G
01 id
id
fi CU
Q G
•\ CD
ft N
1 c
a xi
CO fi
s o
0 MH
rH
rH Ol
m 01
G
CU -H
g rH
3 a
rH g
0 (d
> CO
i fi
0 0
1-3 MH
^-^
£
fi
3
XI
JJ
ca
0)
4J
- —
<£
S
.,
td
g
0
u
td
EH
00
(M
rQ
i-H
O
CJ
ca
•H -a
CQ d)
^i G
rH -H
id g
id cu
HH 0)
0 -O
T! 0)
0 XI
JJ JJ
CD 0
O
JJ
13 .
CD
ca
3
ca
fi
cu
jj
ca
-H
G
fd
u
g
u ca
id CJ
> 0
fi
CD 0)
JJ rH
*rl ft
1 — ' g
1 (Q
\D CO
f^
H
~
, — ,
g
fi
td
TT,
O
Q)
U
gj
- —
G
0)
T!
-H
a
o
-------
Lalysis Method
Used
5
<;
\ (d
JJ -rl
G "O
cu a)
ft
-H 01
CT-H
Wft™
01 €
C fo
-rl W
ft CD
P Pi
W EH
JJ
G
(U
•a
•iH
o
G *— *
H CD
S
a) id
'/H C
•H ' —
to
0)
-rl
Q
H
<4H
O
concent rat i ons
i—i
Q
EH
H
\
U
a
0
CU
rH
ra
0
jj
ca
cu
•§
jj
1
ft
3
(8
X
^
, — .
G
0
CO
3
O
K
Z,
>~*
ra
G
nj
J3
lH
-rl
(8
to
(j_|
o
e reported as
tive estimates
rl
CU
^
ra
u
g
CO
u
o
(8
JJ
-rl
I >
G
(8
sxnce an
-------
APPENDIX B
DATA SET HANDLING PROCEDURES
B-l
-------
MEMORANDUM
To: Dan Bowman Date:Monday 4 May 92
Marilyn Bulman
Ritchie Buschow
Ken Jones
Nancy Rohr
project file
From:Bill Mitchell
Subject: Tire Fire Data Handling Procedures (Updated: 4 May 92)
B-2
-------
TIRE FIRE WORK ASSIGNMENT
DATA SET HANDLING PROCEDURES
Updated: 4 May 92
20 changes!**) since 1 April 92;
5 changesf*) since 30 March 92.
(This document replaces all project modifications completed before 4 May 92.)
General Procedures:
All files and diskettes are stored in file cabinet drawer in Bill Mitchell's and Marilyn
Bulman's Quadrangle offices. When removing a file or diskette, sign and date the
check-out form.
Handwritten entry of a data set:
1. Current updates of the hard copies of the data
base system "screens" are in the project file
cabinet drawer. Use only these "screens".
2. Obtain the entire data set file, including the 2
diskettes, for a given tire fire incident.
** 3. Obtain "tire fire incident identification number
"incident name" from the Tire Fire Login
Sheet. This "Sheet" is a Lotus 1-2-3 file in
[Scratch] named "LOGIN.WK3".
4. Review the ten conversions on the incident
diskettes. These ten appear to be the most
common conversions, although additional ones
may be required. New conversions must be
checked by a second person before use. After
confirming the accuracy of the conversion, the
second person should date and initial the print-
out containing the conversions. This print-out
should then be filed in a file folder labeled
"Conversions" (there is one "Conversions"
folder for each tire fire incident). When unit
conversions are necessary, perform
conversions using appropriate conversion files
on the "primary" incident diskette. Save files
to both the "primary" and the "backup"
diskettes.
B-3
-------
5. When writing conversions on a "screen" hard
copy, make two columns and write the data
units at the top of each column. Enter the data
from the original document in the first column
and the converted data in the second column.
* 6. Use the "Phone Conversation Record" forms
and keep detailed telephone conversation
notes when talking to tire fire incident
contacts. These notes are kept in the project
files.
** 7. The complete re-entry of data sets for the
Everett, Chadbourn (C&J Tire) and the
Somerset tire fire incidents must be completed
by a second Alliance staff member.
8. All incident data sets must be handwritten and
reviewed before computer entry except for
several very large data sets such as Rhinehart
(Winchester) and Hagersville.
* 9. Questions?: Ask Bill or Marilyn or use Everett
and Level Cross incidents as examples. If a
modification to these procedures is needed,
use a "project modification" form.
Computer entry of data set into dBase system:
1. Refer to Specific Procedures...
2. Set "confirm" to on.
3. Leave field blank if no data is available.
4. Save files to both diskettes frequently.
Specific Procedures
"Tire Fire Incident identification" database:
1. Obtain tire fire incident identification #
incident name from the Tire Fire Login Sheet.
This "Sheet" is a Lotus 1-2-3 file in [Scratch]
B-4
-------
# *
2.
named "LOGIN.WK3". Some tire fire incidents
may have had two or more separate teams
monitoring air concentrations, with two or
more discrete sets of sampling data. Each
incident is given only one Incident ID#, even if
there is more than one set of data and more
than one contact organization, person, and
phone number. The additional contacts and
contact phone numbers should be entered into
the "Comments" data base. The data sets are
distinguished from each other by assigning
different MONCRIS #s for the monitoring
instruments. (Be sure to note on the Data Set
Source Sheet the Instrument ID #s to facilitate
tracking these separate data sets.)
One of three entries should be used in the
"Complete?" field: "Y" (Yes), "N" (No), and
"M" (Minimal). Some incidents do not have
data for fields in the data bases or the time
needed to seek out these data is excessive. If
you think you have a "Minimal" data set, first
explain this to Bill and then determine if the
following "M" information is available:
Screen
INCIDENT IDENTIFICATION
FIRE FIGHTING DATA
ENVIRONMENTAL SITE DATA
MONITORING INSTRUMENT DATA
SAMPLED ANALYTE DATA
BIBLIOGRAPHY
Field
Complete? = M
Incident ID #
Incident Name
Total Tires
none
none
Instrument ID
Instrument's Site-to-instrument
Wind Orientation
Instrument Distance from Fire
Boundary
Analyte
Detect Flag
Air Concentration
none
B-5
-------
COMMENTS none
2. Write the name of a contact person who is
easily accessible by telephone, knowledgeable
about the air monitoring and the incident
location.
"Site Data Available for Incident" database:
1. A "y" is entered in the Information Available
field if quantitative data or specific
descriptions are available (e.g. hourly wind
speed measurements or symptom
questionnaires.
"Fire Fighting Activity Data"
** If more than one fire-fighting activity was
used, list first the one that was most used.
Also list the fire fighting activity types and the
dates they were used in the "Comments" data
base.
"Monitoring Instrument Data"
1. Because of space limitations, when assigning
a Monitoring Instrument ID, use this hierarchy:
a. try to match instrument ID with
that used in the original
document,
b. assign an arbitrary number (i.e.
1,2,3...),
c. assign an arbitrary number
followed by an abbreviation of
"chemical grouping" used in the
original document.
2. In the following order, write at least four
descriptive words separated by commas in
Monitoring Instrument Description that
address:
a. collection device
(e.g. Gillian pump)
B-6
-------
b. collection media (e.g.
Tenax tube)
c. analytical method (e.g. GCMS)
** d. quality concerns (i.e. y,n or ?). If "n" is
entered, explain in "Comments" data
base (e.g. potentially: clogged filter,
absorption tube breakthrough, lab
contaminant)
** 3. Use of the Monitoring Instrument Compass
Direction From Tire Fire Center field is
optional.
** 4. Use only "up", "down" or "variable" in the
Monitoring Instrument Site-To-lnstrument
Wind Orientation. Indicate plume color, if
known, in "Comments".
** 5. Write in first the Monitoring Instrument Site
Terrain feature that is judged by the
documentation to most significantly affect
contaminant air concentrations. If including
more than one feature, separate with commas.
Use only the following features "Depression",
"hilltop", "slope", "level", "large surface
water","buildings","trees".
** 6. If a monitoring instrument is used for sampling
on more than one day at the same location,
assign the instrument a different monitoring
instrument ID# for each day. !f the location of
a monitoring instrument is changed, assign a
new Monitoring Instrument ID #.
"Specific Sampled Analyte Data":
** 1. Review all documentation and select the most
appropriate documents containing the incident
air monitoring data set. The "data set source"
form should be attached to copies of the data
set before these materials are stored in the
project file cabinet. Enter data quality
concerns in the "Comments" data base.
B-7
-------
2. Use appropriate conversion spreadsheets on
the incident diskettes.
3. If there are two or more chemical air
concentrations for one analyte at a given
Monitoring Instrument ID # (e.g. 9 a.m. and 3
p.m. sampling times), assign an arbitrary
Sample Num for each (i.e. 1,2,3...).
** 4. Only concentrations for specific analytes
should be entered into the data base.
Documents containing data for analyte
"groupings" such as Total Suspended
Particulates (TSP), Polynuclear Aromatic
Hydrocarbons (PAHs or PNAs), Volatile
Organic Compounds (VOCs) should be entered
and filed in the Bibliography. The data
collected using non-specific field instruments
such as an Organic Vapor Analyzer (OVA) or
a Flame lonization Detector (FID) should also
be listed in the "Comments" database and
documents filed in the Bibliography.
* 5. In general, only target analytes should be
included in a data set although if only a few
non-target analytes are included in the data
summary of an original document the
following procedure should be followed: If
two or more specific analytes are a "group"
and are assigned a single air concentration,
write each analyte on a separate line and write
the air concentration divided by the number of
analytes in the "group".
6. If a specific analyte appears in a data subset
for a given monitoring instrument (i.e. same
instrument, date and time), assign a separate
Sample Num for each.
** 7. Enter only air concentration numeric values.
If an analyte air concentration is listed ND (not
detected), 8TL (below detection limit) or 0,
leave SAMAIR field blank.
B-8
-------
8.
For each tire fire, label a file folder "Data
Sets" If the data set is small, place
photocopies of data used in the tire fire
database in the file folder and attach a blank
"Data Set Source Sheet" form. If the data set
is unreasonably large to photocopy, fill out a
"Data Set Source Sheet" form and place it in
the "Data Set" folder.
"Comments":
* *
The "Comments" data base is used to record any information that may be important
in interpreting the air concentration data but does not have a specific field. Examples
include:
1. Smoke plume characteristics (e.g. color,
opaqueness, height)
2. References for selected data set
** 3. Brief description of analytical methods
4. Brief data quality summary. Summarize
information concerning quality of data entered
into the data base. Include comments about
clogged filters, absorption tube breakthrough,
lab contamination, lapsed holding times.
5. Synonyms for the tire fire incident
6. If used, the name of the air dispersion model
7. If modeling was done to estimate non-target
compounds, make a note in the "Comments"
section.
8. Listing of fire fighting activity types and the
dates they were used if they affected air
sampling data.
9. if in-vitro assays were performed during a tire
fire, make a note in the "Comments" section.
10. List additional Site Contact persons and phone
numbers.
11. List data "groupings" available such as Total
Suspended Particuiates (TSP).
"Bibliography":
1. Organize hard copies of documents as follows and enter
into "Bibliography" data base:
a. Publications
b. Stand-alone reports
* *
* *
B-9
-------
c. Series of memos, letters, workplans,
analytical reports (may also be included
as a "Stand-alone") that address a
particular issue.
** d. Newspaper articles (group all
articles into a file folder labeled
"Newspaper Articles" and staple
a "Bibliography" screen to them.
2. Since Book. Magazine, or Journal Title and
Article Title fields have a 50 character limit,
use common abbreviations and "...".
3. Attach a hard copy of the "Bibliography"
screen to the document before filing.
B-10
-------
APPENDIX C
DATABASE CONSTRUCTION
C-l
-------
APPENDIX C. DATABASE CONSTRUCTION
C.1 Background
Initially, TRC identified 31 tire fire incidents were identified in which air monitoring
data had been collected for up to 100 pollutants. None of these data were available
in machine-readable form. A database system was developed, based on the
information gathered during telephone conversations with contacts for several tire fire
incidents.
•
C.2 Purpose
The purpose of the database system for this work assignment was to facilitate the
efficient and accurate entry of data, and to provide a flexible method to summarize
and evaluate the data.
C.3 System Requirements
The following assumptions were made in selecting an adequate database system: data
entry would be performed by secretarial staff using existing hardware, data could be
accessed by EPA staff using existing software on nominal "PC" hardware, and data
could be exported to Lotus 1-2-3 or Axum for analysis.
C.4 System Design
The database software selected was dBase!V version 1.1. A database "system" was
constructed using dBaselV to accommodate both the data entry and data summary
requirements consisting of seven database files and a total of 65 fields. The files are
linked by common fields, and only one data entry is needed for a field common to two
or more of the seven database files. The fields were designed to facilitate the correct
C-2
-------
entry of data, including a limited number of characters for each field, multiple choice
fields, and a data entry enhancement so that a pollutant name needed to be entered
(and spelled correctly) only once, even though the pollutant appeared in many of the
tire fire incident data sets. In addition, a menu item was included to allow the data
entry operator to stop before completing a record and finish the record at another
time. Although dBaselll+ currently is more widely used within EPA than dBaselV,
dBaselV includes better data querying capabilities among multiple database files. A
set of the field specifications for the seven database files is included in Appendix D
of this report.
»
C.5 System Use
Data entry was performed using the system developed by TRC. Depending on how
complicated the data interpretation process was for a given tire fire incident, data
entry consisted of either writing the information on forms and subsequently
completing the computer entry, or directly entering the data into the computer
database system. Data summaries were prepared using dBase queries. The queries
producing useful information were made into new databases and exported to Lotus
1-2-3 version 3.1 or Axum version 1.0.
C.6 System Limitations
The specific procedures for correct entry of data into the database system were
documented for use by data entry staff. These data handling procedures are included
in Appendix B of this report. The database system was not designed to perform unit
conversions (e.g., ppm to fjg/m3}. Instead, Lotus 1-2-3 spreadsheets were used to
perform and document any necessary conversions.
C-3
-------
APPENDIX D
DATABASE SYSTEM SPECIFICATIONS
D-l
-------
DATABASE STRUCTURE:
C:\TIRE\TF DATA\INCIDENT.DBF
Number
Date of
FIELD
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
TOTAL
of records: 22
last update: 07/29/92
FIELD NAME
INCNUM
INCNAME
INCDATE
AGENCY
AGSAL
AGLAST
AGFIRST
AGMI
AGPHONE
CITY
COUNTY
ST OR PROV
COUNTRY "
TIRENUM
BURNPCT
BURNDUR
SITESIZE
FIRESIZE
PI LENT
PILECONF
TIREOIL
BURNMAT
INCFLG
DATABASE STRUCTURE:
Number
Date of
FIELD
1
2
3
TOTAL
of records: 7
TYPE WIDTH
N 3
C 50
D 8
C 50
C 4
C 20
C 15
C 1
C 10
C 20
C 20
C 3
C 20
N 10
N 3
N 5
N 8
N 8
N 3
C 15
N 10
C 30
C 1
318
DEC INDEX
Y
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
3 N
3 N
N
N
N
N
N
C:\TIRE\TF_DATA\HAGER1.DBF
last update: 04/06/92
FIELD NAME
SAMANAL
SAMAIR
MONDIS
TYPE WIDTH
C 40
N 19
N 6
66
DEC INDEX
N
13 N
N
D-2
-------
DATABASE STRUCTURE: C:\TIRE\TF_DATA\MONITOR.DBF
Number
Date of
HELD
1
2
3
4
5
6
7
8
TOTAL
of records: 341
last update: 07/29/92
FIELD NAME
INCNUM
MONCRIS
MONDES
MONDIR
MONDiS
MONHT
MONWIND
MONTER
DATABASE STRUCTURE:
Number
Date of
FIELD
1
2
3
4
5
6
7
8
9
10
11
12
TOTAL
of records: 5697
TYPE WIDTH
N 3
C 5
C 40
C 4
N 6
N 3
C 8
C 50
120
DEC INDEX
Y
Y
N
N
N
N
N
N
C:\TIRE\TF_DATA\ANALYTE.DBF
last update: 07/30/92
FIELD NAME
INCNUM
SAMANAL
SAMDATE
SAMNUM
MONCRIS
SAMFLG
SAM AIR
SAMDUR
SAMVOL
SAMWND
SAMWNS
SAMTEMP
TYPE WIDTH
N 3
C 40
D 8
N 3
C 5
C 1
N 19
N 5
N 6
C 4
N 3
N 3
101
DEC INDEX
N
Y
N
N
N
N
13 Y
2 N
4 N
N
N
N
D-3
-------
DATABASE STRUCTURE: C:\TIRE\TF_DATA\FIREFITE.DBF
Number of records: 21
Date of last update: 09/14/92
RELD
1
2
3
4
TOTAL
HELD NAME
INCNUM
FITETYP
FITEDATE
FITEDUR
TYPE WIDTH
N 3
C 30
D 8
N 5
47
DATABASE STRUCTURE: C:\TIRE\TF DATA\ENVDATA.DBF
Number
Date of
FIELD
1
2
3
4
5
6
7
TOTAL
of records:
last update:
FIELD NAME
INCNUM
DATATYPE
DATAVAIL
DATALAST
DATAFIRST
DATAMI
DATPHONE
147
04/09/93
: TYPE
N
C
C
C
C
C
C
WIDTH
3
4
1
20
15
1
10 .
55
DATABASE STRUCTURE: C:\TIRE\TF_DATA\COMMENT.DBF
Number of records: 13
Date of last update: 09/14/92
FIELD HELD NAME TYPE WIDTH DEC INDEX
1 INCNUM N 3 Y
2 iNCCOM M 10 N
TOTAL 14
D-4
-------
DATABASE STRUCTURE: C:\TIRE\TFJDATA\BIBUO.DBF
Number of records: 23
Date of last update: 04/09/93
RELD
1
2
3
4
5
6
7
8
9
TOTAL
HELD NAME
INCNUM
BIBLST
BIBFST
BIBMI
BIBTIT
BIBART
BIBVOL
BIBPAG
BIBDATE
TYPE
N
C
C
C
C
C
N
N
D
WIDTH
3
20
15
1
50
50
3
4
8
155
DATABASE STRUCTURE: C:\TIRE\TF_DATA\ANALLJST.DBF
Number of records: 154
Date of last update: 07/16/92
DEC JNDEX
Y
FIELD
1
TOTAL
FIELD NAME
SAMANAL
TYPE WIDTH
C 40
41
DATABASE STRUCTURE: C:\TIRE\TF_DATA\ENVLIST.DBF
Number of records: 7
Date of last update: 04/06/92
FIELD NAME TYPE WIDTH DEC
ENVJTPE C 4
DESC C 15
20
D-5
-------
APPENDIX E
QUALITY ASSURANCE/QUALITY CONTROL PROCEDURES
E-l
-------
APPENDIX E. QUALITY ASSURANCE/QUALITY CONTROL PROCEDURES
E.I Introduction
The data acquired for this project represent a wide variety of data collection, analysis, and
Quality Assurance/Quality Control (QA/QC) methods. Data were not entered into the database
if the site contact person or a site report author stated that the data were incorrect or misleading.
Data from non-specific monitoring methods, such as field-survey instruments, were also not
included. In addition, a QA/QC program was designed to minimize the potential for errors in
entering data from the original reports to the database system and to detect errors if they
occurred. The QA/QC program consisted of the following elements:
• Data "manifest" sheet: This sheet was used to keep track of the data sets
including: data receipt date, handwritten data entry date, computer data
entry date, and a listing of incomplete items.
• Double-entry procedure: Values from fields in the monitoring instrument
and analyte databases for three data sets were entered into separate
database files by two people. The files were then compared and
differences were evaluated. Differences resulted from ambiguities in the
original reports (i.e. different interpretations by two data entry people),
different interpretations in the data entry procedures, or incorrect data
entry.
• Calculation checks: Tire fire data required several unit conversions to be
made. An audit was done that compared the data received to the data
entered. Unit conversions were performed using a formula in a computer
spreadsheet.
The three data sets used for the double entry procedure were for the Belchertown, Spencer, and
Everett tire fires. The Belchertown data set contained 9 separate analyte records, the Spencer
data set contained 18 records, and the Everett data set contained 96 records, for a total of 123
analyte records. Each analyte record had 19 data entry fields. Three of these fields, SAMNUM
(Sample Number), MONCRIS (Monitoring Instrument Identification), and MONTER
(Monitoring Instrument Location Terrain), were not included in the double-entry procedure
because the values for these fields were arbitrary. Thus, 16 data entry fields were available for
double entry. The 16 data entry fields could be evaluated using two general categories. The
E-2
-------
first category included six of the "minimal" fields that were required for a data set to be entered
into the database system and an additional "field," record omission, to indicate when the double
entries differed by an entire record. The second category included the 10 fields of ancillary
information that were not available for many of the data sets. The total number of data entered
in these fields for the three tire fire incidents was 1,968 (123 analyte records times 16 data entry
fields). A simple QA/QC procedure was used to check all three double-entered data sets. A
separate printout was generated for each double-entered data set, the printouts were then
compared and any differences noted.
£.2 Preliminary Results and Discussion
The following is a description of data fields and types of differences found during the QA/QC
check. Only data fields with multiple differences are discussed.
• Record Omission. Record omissions occurred 31 times, 11 as data entry
and 20 as data interpretation. The data entry differences occurred when
a data entry operator missed a value in the original report. Data
interpretation differences occurred when one data entry operator entered
sample blanks and another did not. All 20 differences were in the Everett
fire data set. One of the Everett data sets was completed eight months
before the second data set and there were changes in procedure between
the two entry periods.
• Sample Flag (SAMFLG). Two data entry people had interpreted the
purpose of the sample flag field differently. The original purpose of the
sample flag field was to enter one-half the detection limit for all samples
in which a contaminant was not detected and enter "No" in SAMFLG.
The strategy was not used in the data analysis presented in this report
(i.e., we were no longer interested in values below the detection limit).
• Sample Air Concentration (SAMAIR). The differences in values were
actually errors and represented incorrect units for values at or below the
detection limit. No errors were found in values above the detection limit.
In particular, the sample air concentration data for the Everett tire fire
were checked.
Table E-l is a summary of the double-entry differences. The two categories of database fields
are listed across the top; "minimum" fields are shaded and ancillary fields are unshaded. The
E-3
-------
tire fire name and types of differences are given along the side of the table. The two types of
differences are data interpretation and data entry. Data interpretation differences occurred
because the data presentation in the original report was ambiguous. For example, monitoring
instrument height might be given in the original report as "2-3 feet." One data entry person
would enter "2 feet" while another entered "3 feet." Data entry differences occurred when one
data entry person incorrectly entered data or neglected to enter data.
There were a total of 425 differences between the double entries of the three data sets, resulting
in a 22 percent difference rate for the three data sets. The differences by data set were 2 percent
for Belchertown, 9 percent for Spencer, and 26 percent for Everett. The differences noted in
"record omissions" and the sample flag fields often do not represent true or important
differences since, in many cases, the differences involved sample blanks and analyte detections
given as "below detection limit" or "0" which were not significant for the analyses performed
in this report. The four remaining fields with any differences had a 6 percent difference rate.
Of particular concern are the 10 data entry differences in the sample air field. These differences
were only for values reported for the Everett tire fire at the detection limit but not for values
above the detection limits. The incorrect values were actually reported in mg/m3 instead of
/*g/m3.
E-4
-------
01
4J
H
O
tf
§
•H
•H
rH
&
S
1
O
H
««*- &
O M tt, b ^
S
H Q B •< J
a o a * H a a
SOS S H
a o a OHO)
2 0 a Q H oi
S O S Q W M
0) < S H W S 0.
0) < 2 3 S «
tn < S 2 S Q
W «« S > 0 J
M < S Q S OS
n ^ » it H at
cn «C S 3 D S
« < S Q < H W
H 3 O 3 D »
M id O O M Q
N
S
srtown, MA
1 Belch,
e
MM
^^
-a Interpretation
Q
n
^BH
^^
ro
G
14
a
a
a\
»
- M
I
r-l
^
^
d
0
•H
-tJ
ft)
4J
a
0)
a
H
Q
^
^
-
.a Entry
in
Q
o*
*
Ov
-w
I
O
O
-a Interpretation
IT)
a
t*
r-
n
N
^
ro
00
IT*
H
a
Si
o
-a Entry
a
a
ri
n
H
U>
U3
in
fl*
ro
CO
10
f*
T-4
O
OX
K1
^
S
Count
[I Total
t-t
>H
M
04
r*
H
t-4
M
rH
H
-
rn
u
1
*
m
rl
S
without omi»»iou
1 and akliFLa
Total
I K«eor<
-------
£.3 Preliminary Conclusions
The potential error rate was less than the 6 percent difference rate identified by comparing the
three double-entered data sets. In particular, additional QA/QC efforts .addressed the
unconverted values within the sample air field. In general, 90 to 95 percent of the differences
were readily identifiable and correctable. The results of the double-entry process suggest that
there is a greater likelihood of differences in data entered in the larger data sets. Potentially,
some of these differences could be errors. The records for the analytes used in the exploratory
data analysis were checked in the following QA/QC procedure:
• Checked data field SAMFLG against SAMAIR. If there was a
concentration given in SAMAIR, then SAMFLG should say yes, and vice
versa. Also checked for patterns in SAMAIR numbers. Looked for
unexpected values.
• To reduce record omissions, double checked analyte list versus data
entries. Counted sources and verified analyte numbers.
• Checked one SAMAIR value from each data set for proper units for
values at or below the detection limit.
E.4 QA/QC Procedure for Random Checks
QA/QC checks were performed on each of the five tire fire incidents which contained more than
200 records in the database. Table E-2 shows the type and number of data checked. A random
number table was used to determine the order for checking each of the incidents, as well as
randomly selecting which records to review. For each of the five incidents, the data submitted
with the documentation (i.e., "the raw data") for each of the incidents were checked against the
entries made in the database.
QA/QC notations were made on the database master printout. These notations were made as
follows:
• A slash mark (-) was made in the left margin next to each randomly
chosen record using red ink.
E-6
-------
The entire entry line for each randomly chosen record was highlighted
with a yellow marker.
Potential differences noted in a given field(s) for each record checked was
circled with a pencil (i.e., first pass).
For each potential difference in a given field, the original data for that
particular incident were double-checked to determine if it was an actual
difference (i.e., second pass).
If an actual difference was noted during the second pass, the corrected
value was noted in red ink on the master printout next to the error, and
an arrow was drawn from the notation to the field for which the correction
would be made.
Table E-2. Recommended QA/QC Checks
Database'"
BIBLIO
FTREFITE
ENVDATA
COMMENTS
INCIDENT
MONITOR
ANALYTE
Fields
9
4
7
2
23
8
12
65
Total
Records
23
21
147
13
22
341
5967
6534
QC Checks
Records
1
1
8
1
1
17
250
279
Total
Values
207
84
1,029
26
506
2,728
71,604
76.184
QC Checks'2'
Values
9
4
56
2(3)
23
136
3,000
3.230
(1)
(3)
Data sets (/. e., tire fire incidents) are randomly selected for QA/QC check except
for ANALYTE which will be performed on the analytes used in the exploratory
data analysis.
QC checks comprise approximately 5 % of all database data.
Additional checks for omissions are suggested in the nine data sets for which there
are no COMMENTS records.
In some instances, additional differences between the raw data and information entered into the
database were noted as a result of this procedure. In such cases, the differences noted in these
records were obvious as a result of checks performed on the randomly chosen records.
E-7
-------
Although the procedure involved checking approximately 5 percent of the entire database,
differences found in any additional records were noted.
E.5 QA/QC Procedure for the Everett, Washington Tire Fire Incident
In addition to the random QA/QC checks performed on the five tire fire incidents containing
greater than 200 analyte records, an additional QC check was performed on the Everett,
Washington tire fire incident analyte database to ensure that analyte concentrations recorded at
or below the applicable detection limit were not entered into the database. This QC check was
necessary due to a change made in the data entry procedure which occurred during the data
gathering phase of the project. In general, the procedural change required that data at one-half
and below the detection limit not be entered into the SAMAIR database field for each analyte
record contained in the database.
As such, the analyte database for the Everett, Washington tire fire incident was checked using
this procedure. The Everett database was chosen to undergo this QC check due to the
complications involved in extracting the analyte data from the documentation received on this
incident.
The results of this QC check were as follows:
• Concentrations for styrene and xylene contained in the incident documentation
were presented as a single value. As such, concentrations for both of these
analytes were entered into the database at one-half the value contained in the
documentation. As a result, double entries for styrene were noted.
• Double entry for naphthalene was noted in the database. For this double entry,
a different concentration value was entered into the database. The reason behind
this double entry is not known.
• The Everett database contained 67 records which had analyte concentrations (i.e.,
in the SAMAIR field) below the detection level. Of these 67 analyte records (or
67 fields), the detection limit value was recorded in a total of 8 analyte fields (as
opposed to leaving the field blank). The reasons for these noted differences can
be attributed to the change in procedure as previously explained. In addition, for
E-8
-------
MONCRIS #1, a concentration value for methylene chloride was recorded. This
difference was observed since there were no detectable concentrations for
methylene chloride noted for any of the samples taken.
• Of the 8 analyte fields containing values at the detection limit, 6 of the
corresponding SAMFLG fields (i.e., detectable data) were noted as yes. Two
additional SAMFLG fields which had corresponding undetectable concentrations
were entered as "yes."
• No analyte data were entered for Benzo[a]pyrene for sample station # 9
(MONCRIS #9). It should be noted that the sample # 9 data were taken by a
sampling team other than the team which took the other 8 samples during the tire
fire incident.
For all differences noted above, the appropriate changes to the database were made.
E.6 Final Conclusions
The actual error rate discovered during the final random QA/QC check of the data from the tire
fire incidents with greater than 200 records was approximately 4 percent. The QA/QC of the
Everett, Washington data also resulted in an error rate of 4 percent. The results of the checks
are presented in Table E-3.
E-9
-------
Table E-3. QA/QC Final Results
Everett, WA
Double Entry
Analyte
Level Cross, NC
Date of Sample
Wakefield, VA
Danville, PA
Hagersville, Ontario, Canada
Sample Air
Concentration
Webber, UT
TOTAL
Errors
4
1
1
0
0
8
0
14
Records
Checked
116
20
12
12
190
14
364
Error Rate
4%
5%
0%
0%
4%
0%
4%
E-10
-------
APPENDIX F
REFERENCES
F-l
-------
APPENDIX F. REFERENCES
Anderson, J. "On the Fireground at Daruk." Fire Journal - Australia. 11:2, Autumn 1987,
pp. 18-21.
Anonymous. "Energy from Wastes." Power. Special Section. March 1988.
Anonymous. "Waste Tires Burned to Fuel WTE Plant." World Wastes. February 1987, p. 32.
Best, G.A. and B.I. Brookes. "Water Pollution Resulting from a Fire at a Tyre Dump."
Environ. Pollut. fSeries B"). 2, 1981, pp. 59-67.
Butt, T. "Tyre Blaze - A Week-Long Furnace." Fire Journal - Australia. 11:2, Autumn 1987,
p. 15.
Campagnia, Phil. EPA Emergency Response Office, Edison, NJ. Personal communication with
Bill Mitchell (TRC). June 3, 1992.
Clark, C., K. Meardon, and D. Russell. "Burning Tires for Fuel and Tire Pyrolysis: Air
Implications." Office of Air Quality Planning and Standards, EPA, Research Triangle Park,
NC. EPA-450/3-91-024, December 1991.
Colin, T., G. Grigoleit, and G. Bracker. "Pyrolytic Recovery of Raw Materials from Special
Waste." Chemie-Ingenieur-Technik. 50, No. 11, November 1978, pp. 836-841 (German).
"Current Intelligence Bulletin: 2-Nitropropane." Dept. of Health, Education, and Welfare,
Public Health Service, Center for Disease Control, NIOSH, Rockville, MD. April 25, 1977.
Drabek, J. and J. Willenberg. "Measurement of Polynuclear Aromatic Hydrocarbons and Metals
from Burning Tire Chips for Supplementary Fuel." From TAPPI Proceedings, 1987
Environmental Conference, April 27-29, 1987, pp. 147-152.
Fukazawa, H., Y. Ajioka, A. Katahira, K. Nakamura, T. Nakajima, and S. Asakawa. "The
Study of the Heavy Metal Components Released from the Boiler Using Scrap Tires." Shizuoka-
ken Eisei Kankyo Senta Hokoku. Vol. 25, 1982, pp. 157-160 (Japanese).
Greene, R. "Finding Offbeat Uses for Scrap Tires." Chemical Engineering. 85:18. August 14,
1978, p. 88.
•
Hanson, K.A., J.A. Guenthoer, K.S.Mackey, and A.F. Blaisdell. "State of Washington
Department of Ecology Rubber Tire Chip Trial Burn at Holnam Incorporated Industries, Stack
Testing and Chemical Analysis, October 15-19, 1990." Volume I. Am Test. Inc.. Preston, WA.
January 23, 1991.
F-2
-------
Higgins, A.J., J.L. Suhr, M. S. Rahman, M.E. Singley, and V.S. Rajput. "Shredded Rubber
Tires as a Bulking Agent for Composing Sewage Sludge." Project Summary. Waste
Engineering Research Laboratory, EPA, Cincinnati, OH. EPA-600/S2-87-026, May 1987.
Hoglin, D.A. "What Goes Around Comes Around." Journal of Environmental Health. 51:3,
p. 174.
Illinois, State of. "Illinois Department of Energy and Natural Resources is Supporting a Test
Burn of Scrap Tires." Environmental Science and Technology. 24:12, December 1990.
* *
Illinois, State of. "Tire Eire Chemical Emissions and State Regulations REgarding the Storage
of Waste Tires." Division of Air Pollution Control, Illinois Environmental Protection Agency.
May 20, 1988.
Kearney, A.T. "Scrap Tire Use/Disposal Study." Final Report. Scrap Tire Management
Council, Washington, DC. September 11, 1990.
Koogler and Associates. "Summary of Paniculate Matter, Volatile Organic Cotnpounds, Semi-
Volatile Organic Compounds, Furans and Dioxins, Sulfur Dioxide, Nitrogen Oxides, Metals and
Visible Emission Measurements - Tire Derived Fuel Conditions" For Central Power and Lime,
Inc., Brooksville, FL. Koogler and Associates Environmental Services, Gainesville, FL.
September 18-24, 1990.
Lemieux, P.M. and D.M. DeMarini. "Mutagenicity of Emissions from the Simulated Open
Burning of Scrap Rubber Tires." EPA, Washington, DC. EPA-600/R-92-127, July 1992.
Lewis, P.M. and P.W. Chartrand. "A Scrap Tire-Fired Boiler." American Society of
Mechanical Engineers, 1976 National Waste Processing Conference, May 23-26, 1976.
National Fire Information Council. "Fighting Fire with Facts." National Fire Information
Council. Lansing, ME.
Niles, R.C. "Energy and Environment Practical Concerns."
NIOSH. "Hazard Evaluation and Technical Assistance, Report No. TA 76-90." Newport
Industrial Products, Firestone Tire and Rubber Co., Newport, TN. For the U.S. Dept. of
Health. Education, and Welfare, Center for Disease Control, NIOSH. January 1978.
Ohio, State of. "Air Emissions Associated with the Combustion of Scrap Tires for Energy
Recovery." Malcolm Pirnie, Inc., Columbus Ohio. Ohio Air Quality Development Authority,
May 1991.
Ohio, State of. "Results of The Ohio Edison Tire Burn Test at Ohio Edison Company, Toronto
Plant, Toronto, Ohio, May 21-25, 1990." Ohio EPA. August 1990.
F-3
-------
Pacey, M.D. "Down in the Dumps - But Waste Management May Come Out Smelling Like a
Rose." Barren's. February 24, 1975.
Radian Corporation. "Modesto Energy Company Waste Tire to Energy Facility, Westley, CA -
Final Emission Test Report." For Oxford Energy, Boston, MA. Radian Corporation, Research
Triangle Park, NC. April 25, 1988.
Radian Corporation. Recycling Research Institute. Scrap Tire News. Suffield, CT. 5, No. 5,
May 1991.
Ryan, J.V. "Characterization of Emissions from the Simulated Open Burning of Scrap Tires."
Acurex Corporation, Research Triangle Park, NC. EPA Contract No. 68-02-4701. Air and
Energy Engineering Research Laboratory, EPA, Research Triangle Park, NC. EPA-600/2-89-
054, June 1989.
Sabath, D. "Burning Waste Tires with Coal Promising." Cleveland Plain Dealer. July 6, 1990,
p. dlO.
Stoneberger, M. "Tire Fire Chemical Emissions and State Regulations Regarding the Storage
of Waste Tires." Draft. Air Toxics Unit, Permits Section, Division of Air Pollution Control,
Illinois Environmental Protection Agency. May 20, 1988.
Sunia, L. "$750,000 Tire Store Blaze in Madera, CA." American Fire Journal. 37:11,
November 1985, p. 37, 47.
Truax, H. "Built to Last-They're tough. They're durable. They're an environmental
headache." Environmental Action. March/April 1988, pp. 9-11.
U.S. EPA. "Burning Tires for Fuel and Tire Pyrolysis: Air Implications." U.S. Environmental
Protection Agency, Research Triangle Park, NC, EPA-450/3-91-024. December 1991.
U.S. EPA. "ECAO CTC-AIR RISC Tire Burning Project." Progress Report. EPA Contract
No. 68-D8-0090. January-February 1989.
U.S. EPA. "ECAO CTC-AIR RISC Tire Burning Project." Progress Report. EPA Contract
No. 68-D8-0090. March 1989.
U.S. EPA. "ECAO CTC-AIR RISC Tire Burning Project." Progress Report. EPA Contract
No. 68-D8-0090. May 1989.
Washington, State of. "Source Test, Boise Cascade Wallula Plant." Washington State
Department of Ecology, May 21, 1986.
F-4
-------
Zaharchuk, Roman, and Legatski, L. Karl. "SO2 Scrubber Passes Test at Firestone." Pollution
Engineering. April 1977, pp. 50-52.
F-5
-------
APPENDIX G
A TIRE FIRE BIBLIOGRAPHY
(from the database system)
G-l
-------
APPENDIX G. A TIRE FIRE BIBLIOGRAPHY
Anand, Raj. K. and Gordon Marker. "Modesto Whole Tire Burning Power Plant." Presented
at National Waste Processing Conference, 1988, Vol. 13, pp. 335-43.
Anonymous. "Tire Fires Can be Prevented Through Recycling Waste-to-Energy Use." PR
Newswire. March 13, 1990.
Anonymous. "Waste Tire Fluidized Bed Combustion Boiler Project." NTIS Accession No.
DE84008535, March 1984.
Bauman, B.D. "Scrap Tire Reuse Through Surface-Modification Technology." Air Products
and Chemicals, Inc., Allentown, PA, 1991, 24 pages. (EGG-M-91033, CONF-910216-1)
Bisaro, T., Makansi, J. "Proper Fuel Handling Reduces Solid Waste, Helps Control Emissions
from Woodwaste-Fired Boiler." ASME Industrial Power Conference. Pittsburgh, pp. 87-91,
October 5-8, 1986.
Boscak, V., R. Kenson, and P. Barlett. "Plan Energy Conservation in Solving Odor Problems."
Pollution Engineering. February 1978, pp. 34-38.
Brion, J., S. Carpentier, G. Chevalier, R. Delarue, J. Pradel. "Cleaning of Industrial Waste
from Incinerator Gaseous Effluents. Case of Tires." Conference Proceedings, International
Symposium on Chemical Engineering in the Service of Mankind, Paris, September 3, 1972.
(French)
Camfield, Stacey. "Tyre Fires Stimulate Debate." European Rubber Journal. 172:4, April
1990, pp. 24-27.
Humpstone, Charles C., Edward Ayres, Sam G. Keahey, Theodore Schell. "The Recycling and
Reuse Incentives." NTIS-Report PB-234 602/1WP, 1974.
Kemper, C.C. "Oregon Recovery Effort Wins Public and Private Sector Backing." Solid Waste
Management Refuse Removal Journal. September 1977, pp. 62, 84.
Kenney, R. and J.F. Joyce. "National-Standard Company Waste-Tire Fluidized-Bed-Combusion
Boiler Project, Final Report." Department of Energy Publication No. DOE/ID/ 12163-T1.
Kofoed, Jensen P. "Refuse Refineries." Conservation and Recycling. 1, No. 2, 1977, pp. 201-
208.
J^ewis, P.M. "A Scrap Tire-Fired Boiler." Presented at 1976 National Waste Processing
Conference: From Waste To Resource Through Processing, New York, 1976, pp. 301-311.
G-2
-------
Makansi, Jason. "Putting Powerplant Wastes to Work." Power. July 1983, pp. 23-31.
Mathews, Jay. "Garbage In, Power Out: A Clean Solution to a Heap of Problems."
Washington Post. November 18, 1987, p. A3.
Moseley, C.L., S.A. 1-ee, B. Hills. "Health Hazard Evaluation Report HETA 84-044-1441,
Rhinehart Tire Fire, Winchester, Virginia." NIOSH Accession No. PB85-185155/XAB.
Murphy, Michael L. "Fluidized Bed Combustion of Rubber Tire Chips: Demonstration of the
Technical and Environmental Feasibility." Energy Biomass Wastes. Vol. 11, 1988, pp. 371-
380.
Niles, Robert C. "Energy and Environment-Practical Concerns." Presented at U.S. Dept. of
Energy/AGA/NCA/EPRI 5th Energy Technology Conference, Washington, DC, February 27-
March 1, 1978, Vol. 78, pp. 889-894.
Purcell, A.H. "Tire Recycling: Research Trends and Needs." Presented at Recycling World
Congress, Basel, March 6-8, 1978, Vol. 5, pp. 3-4-1.
Rouge, James D. and John Ix»we. "Air Toxics Impacts from Resource Recovery Projects: A
Comparison of Health Risk Assessment Methodologies and Emission Factors." Proceedings of
APCA Annual Meeting, 1986, Session 79, No. 5, 18 pages.
Schneider, Keith. "Worst Tire Inferno Has Put Focus on Disposal Problem (Tire Fire at Dump
in Hagersville, Ontario)." New York Tunes. March 2, 1990, pp. A8, A10.
Shang, J.Y., J.S. Mei, J.E. Notestein. "Fluidized-Bed Combustion of Scrap Tires: Technical
Note." Department of Energy publication no. DOE/METC-86/4068, October 1981, 35 pages.
Stofferahn, Jeffrey A. and Simon Verneta. "Emergency Response to a I-arge Tire Fire:
Reducing Impacts to Public Health and the Environment." Haztech International Converence.
St. Louis, August 26-28, 1987, pp. 483-497.
Taggart, Robert H. Jr. "Shredded Tires as an Auxiliary Fuel." Conference Proceedings
Environmental Aspects Chem. Use Rubber Process, 1975, pp. 361-370.
Timmann. Hinrich. "Practical Experience from Cracking Scrap Tires and Plastic Waste in a
Fluidized Bed." Recyci Int.. Ed. Karl J. Thomme-Kozmiensky, 1984, pp. 609-614.
"Tire Fire Emits Toxics." Environ Manage News. 5, No. 1, 1990, p. 7.
U.S. EPA. "Characterization of Emissions from the Simulated Open Burning of Scrap Tires
(Final Report)." EPA-600/2-89-054, October 1989, 69 pages.
U.S. EPA. "Superfund: A Six Year Perspective." EPA/9200.5-000, October 1986. 45 pages.
G-3
-------
Wallace, J. "All Tired Out." Across the Board. 27:11, November 1990, pp. 24-30.
Zylkowski, Jerry, and Shelton Ehrlich. "Combustion of Waste Fuels in a Fluidized-Bed Boiler."
Proceedings of Am. Power Conf., 1983, pp. 263-270.
G-4
-------
APPENDIX H
ACKNOWLEDGMENTS
H-l
-------
APPENDIX H. ACKNOWLEDGMENTS
On behalf of the U.S. Environmental Protection Agency (EPA), Office of Air Quality Planning
and Standards (OAQPS), TRC Environmental Corporation acknowledges the cooperation and
efforts of all those who have contributed to the compilation of the data used in this report. In
particular, we are grateful to the EPA's Emergency Response Office in Edison, New Jersey, for
providing access to their files so that analytical data from eight tire fire incidents could be
included in our data system. We would also like to acknowledge Professor Jonathan Bamett,
of the Worchester Polytechnic Institute's Fire Safety Program, who contributed to estimating the
incidence of tire fire occurrences by providing the results of a search of the National Fire
Incident Reporting System for 1988.
H-2
-------