-------





1
fcl
g
H
§


S
1 .
fe
s
s
1
s
2

H
N
ft
JQ
H





•V
|HB
I
M
t^
^ 0
^suf5
Wl
— .2 IS
i

- »
«*• *
11"

w * "
5""



^ j
~ " i«
~ * •
• ^ «.
* o

« ,
I5"

o


to
tH
O

O
n
in
r-
•-t



in
CM

-t
in
m



o
00
Ci
00
in

o

r4
O
O

o
a
n
• 0

o
en
*


0%
m
o
o
r-




r-
a




o
o
f-i
e
n
o

m
o






























CN
O
i
























































































































































































































































































































































2-39

-------
ft.
m
* |^
M !*•*
S
"
•g ..
* ™" »5
i*"1
•• i
•
• *•*

- .2 i
"• 3
5
«. a
* .2
*• ^5 *
ss"
u -
tJm
$*"
«_
5 E
Ji
is\.
«• U t*4
s-


u



to
m
0

o
^
CN


O

CM


o^
^
f
o

o
o
o
ec
IN
C
C
f
V
f
c
in
a
a
r*
0


in
*N
* 0

m
m
2


o
r-t
s


,_,
in
o

o
o
•H
O
O


o
o
in
^
o
in
o
o
o
<*>
o
m
es
o

CD
f*l
"


O
CN
«


in
1*1
o


r-i
O
in
r-
r-l
-
O
in
r-
c
ao
o
o
c
«
































r-
m
n
o
tr
































































































































































4
























)
































































































































































2-40

-------
2.1.4  Precision of Emission Factors Determined by Method 4
     Although it is not possible to calculate the precision of
a guess, it is reasonable to estimate such precision.  For example,
given a specific process category, a guess cannot be any more
precise than the most precise AP-42 factor within that category.
As a practical matter, one cannot assume that a guess will be
more precise than the least precise emission factor for the given
category — it may even be less precise.  Since the exact precision
cannot be determined, the unit emission factors used in calcu-
lating emissions by Method 4 were estimated to be as precise
as the least precise AP-42 factor within that category.
2.1.5  Precision of Emission Factors Determined by Method 5
     There is no basis for stating that an emission factor developed
by, for example, a state air pollution control agency is any
more precise than the corresponding AP-42 factor.  Therefore
an emission factor developed by a state or other agency is estimated
to be as precise as the corresponding AP-42 factor.  Where no
AP-42 factor is given, but the process emits the pollutant in
question, precision of the emission factor is estimated to be
the same as that of a factor determined by guess  (Method 4).
2.2  PRECISION OF CONTROL DEVICE EFFICIENCY VALUES
     The precision of any value given for control device efficiency
depends on how the number was derived by personnel completing
the NEDS form.  Methods for determining this three-digit number
include the following:
     a) Stack measurements before and after a control device.
     b) Use of manufacturer's data.
                                 2-41

-------
     c) Use of graphs for specific control devices relating
        efficiency to unit size, pressure drop, power
        requirements, and other parameters.

     d) Estimation or guess.

     Since there is no way of telling which of these methods

led to the value entered on the NEDS data sheet, one must take

a conservative approach in determining precision of values for

control device efficiency.

     The following guidelines were established:

     1)  All numbers not ending in .5 or .0, such as 92.3
         or 76.6, have a precision of + 0.1.

     2)  Numbers ending in .5, such as 75.5, have a preci-
         sion of + 0.3.

     3)  Numbers ending in .0 but not 0.0 or 5.0, such as
         58.0, have a precision of + 0.5.

     4)  Numbers less than 80 ending in 5.0, such as 55.0,
         have a precision of + 5.0.

     5)  The numbers 85.0 and 95.0 have a precision of
         + 2.5.

     6)  Numbers ending in 0.0, such as 50.0 or 90.0,
         have a precision of + 5.0.


     Statement 1) assumes that any number in this category
     must be fairly precise since it contains three signifi-
     cant figures.  A properly conducted emission test yields
     values that are precise to the third significant figure.

     Statement 2) considers values ending in 0.5, which can
     possibly be accurate to three significant figures, but
     also can be the average of two numbers  (such as 84 and
     89).  A precision factor of + 0.3 represents a conserva-
     tive approach.

     Statement 3) suggests that these numbers, usually from
     control system manufacturer's literature, are precise
     to the nearest unit, or + 0.5.

     Statement 4) and 6) assume conservatively that the entry
     may have been a guess or that the literature source  states
     values in increments of 5 or 10.  Therefore these numbers
     are assigned a precision of + 5.0.
                                2-42

-------
     Statement 5) assumes that even though these numbers end
     in 5.0 and may have been determined by guess, entries of
     80.0 or 90.0 would have preceded guesses of 85.0 and 95.0.


2.3  PRECISION OF THRUPUT VALUES

     Instead of determining precision of a thruput value for

each SCC, we assigned one precision to each major SCC category.

Although the major categories include many different types of

processes or fuels, the precisions are applied to the entire

category .

     A literature search was made to determine how specific indus-

tries in the major categories normally report their production

or thruput.  In addition, questionnaires were completed by the

Directors and Executive Committee of the Air Pollution Control

Association.  These questionnaires provided estimates of typical

thruput or production numbers for "average" installations in

the major industries, along with the units and magnitudes in

which the thruput or production is usually reported.  With this

information and engineering judgement, we developed precision

values for thruput in each major SCC category.

     The first major SCC category encompasses generation of elec-

tric power by burning of coal, gas, and oil.  For various thruputs,

X, the following precisions, P, were determined:

                  X > 200,000        P  =  ± 0.001

         10,000 < X < 200,000        P  =
                  X <  10,000        P  -  + 0.001
                                2-43

-------
     Numbers greater than 10,000 in this category represent coal
usage in major power plants; the precisions therefore reflect
the precision of coal thruput values.  The precision of numbers
reported in this range is approximately + 200.  Thruput values
greater than 200,000 have precisions less than 0.001, but the
precision value was rounded to 0.001 since the precisions of
emission factors are given only to three decimal places .  Thruput
values less than 10,000 usually indicate gas and oil usage.
Since the reported thruputs of these fuels are quite precise,
on the order of , v QQQ* a precision of + 0.001 is specified.
                , v
     Another major SCC category is industrial boilers , whose
reported thruputs are assigned the following precision values:
                  X > 200,000         P  =  + 0.001
         50,000 < X < 200,000         P  -  + 5£°-
                —                           ™  A
          1,000 < X <  50,000         P  -  + 0.020
                  X <   1,000         P  =  + 0.001

     Again, the small thruputs are considered precise since they
represent mainly gas, oil and small coal boilers.  The consumption
of gas and oil can be known and reported very precisely.  As
the thruput values increase, coal usage begins to dominate and
the precisions reflect the change.  Thruputs in the range of
                                        500
1,000 to 50,000 have a precision of + 25 Q'Q'O" or 0.020.  Thruputs
in the range of 50,000 to 200,000 are precise within + 500.
For very large thruputs, the values should be known and reported
precisely.  Thruputs in this range are also ± 500; rounding off
                                 2-44

-------
to three decimal places gives a precision of + 0.001.  Two other
major categories, commercial/institutional boilers and in-process
fuel usage, are assigned the same thruput precision values as
industrial boilers.
     The precision of thruput values for all stationary internal
combustion sources is + 0.002.  This value is based on engineering
judgment and consideration that the thruput is likely to be
small, but not necessarily reported to the same precision as
thruput of industrial boilers .  The fuel normally used is gas
or oil and the thruput is reported with a precision of ?^-r or
     t which equals + 0.002.
     The precision of thruputs for all chemical processes is
estimated to be + 0.010.  The category of chemical manufacturing
encompasses a wide variety of types of processes and a large
range of thruputs .  Although the thruput or production in these
industries is likely to be known to a high degree of precision,
many operators are reluctant to divulge precise production figures
because of competition within the industry.  It is assumed, there-
fore, that the precision to which thruput is generally reported
is ....Q0 for small operations, 7QO 'QQQ for medium-sized processes,
and f nmlimii for large operations; these values yield an overall
         f U v U
precision of + 0.010.
     The precision of thruput reported in the food/agricultural
category is estimated to be 0.100.  Again, the category encompasses
a wide variety of types of processes and it is judged that the
thruput values reported for these processes are not as precise
                                2-45

-------
aa those of some other Industries.  Country elevators often report



their thruput with a precision of gnn' nnn bushels and terminal
                                     t UUU
elevators report thruputs of about vnfnnnfnn» bushels.  A preci-
                                   iU , UUU /UUU


sion of + 0.100 is assigned to the entire food/agricultural cate-



gory.



     Precision of thruputs for metallurgical  (primary and secondary)



processes is specified as follows:




                  X > 500,000         P  =  + 0.001




         50,000 < X < 500,000         P  =  + 522.
               - ^                           "™  A



                  X <  50,000         P  =  + 0.010





     Precision of thruput values for large metallurgical processes



is judged to be good, and for small installations, not as good.



Thruputs of the magnitude 50,000 to 500,000 are reported within



± 500.  Thruputs larger than 500,000 are still within ± 500,



but the precision for this category is rounded to three decimal



places, or 0.001.  Thruputs less than 50,000  are as precise as



those at 50,000.  That is, thruputs of 25,000 have a precision




of
     The precision of thruputs for mineral product installations



follows a similar trend:



                  X > 1,000,000       P  =  +  0.001




         50,000 < X  <  100,000       P  -  +  1>°°°




         10,000 < X  <    50,000       P  =  i  0.020




                  X  <    10,000       P  *  +  0.050
                                 2-46

-------
     The large installations know and report their thruputs to



a higher degree of precision than do the small ones.  Thruputs



greater than 50,000 are reported + 1,000; precision of thruputs



greater than 1,000,000 are rounded to 0.0001.  Thruputs between



10,000 and 50,000 are generally reported with precisions in the



range of Or .mr» or + 0.020.  Included among the small installa-
         4bD , UUU     —


tions are seasonal rock crackers, for which the accounting of



thruputs is not very precise, usually in the range of 5 OQO or



± 0.050.



     Precision of thruput values for the petroleum industry is



as follows:




                  X > 200,000         P  =  + 0.001




          1,000 < X < 200,000         P  =  +




                  X <   1,000         P  «  + 0.002





     This category includes both refinery operations and process



heaters.  Thruputs less than 1,000 represent the process heaters,



whereas the larger thruputs represent refining operations.  The



larger the refinery capacity, the more precisely the thruput



is reported.  This trend was determined from analysis of informa-



tion presented in National Petroleum News Factbook Issue.  Preci-



sions given for thruputs greater than 1,000 reflect the finding



that the refinery capacities are generally reported + 1,000 barrels



per day.  The precision values take into account conversion from



barrels per day to thousands of barrels per year.  Precisions



of thruput values for the process heaters are all on the order
                                2-47

-------
of 500 or + 0.002.



     Precisions of thruput values for the wood products industries



are:



                  X > 10,000          P  =  + 0.010



                  X < 10,000          p  =  + 0.050




This category includes both paper pulping and sawmill operations.



With a break point of 10,000, a precision of + 0.010 represents



mainly pulping operations and a precision of + 0.050 represents



various sawmill operations.  Thruputs for pulping operations



are generally reported with a precision of g-QQ''pQQ- or ± 0.010.



Thruputs for sawmill operations are generally reported in board-



feet and are then converted to tons for entry on the NEDS form.



This conversion entails some additional imprecision.  The



thruputs entered on the NEDS forms therefore are assigned a



precision of approximately I -.„ or + 0.050.
                           D , UUU    —


     The precision of thruput values for metal fabrication, leather



products, textile manufacturing and "other/not classified" cate-



gories is estimated to be + 0.050.  These are small, diverse,



and relatively low-polluting industries.  It is estimated that



the precision of thruput values reported by any of these industries



would not be better than the precision of the smallest thruput



reported in any of the other major categories.  Therefore, the



precision assigned to these categories is an average of the preci-



sions of the smallest thruputs of the other categories.



     The precision of thruput values for all point source evapora-



tion operations is + 0.010.  This category includes petroleum
                                2-48

-------
storage and surface coating  (i.e., painting) among others.  The
precision of +0.010 is valid of all thruputs in this category.
For small thruputs, the precision is r Q0^; for large thruputs
the precision is         or        oa'  The precision of all
these thruputs is + 0.010.
     The precision of thruput values for all solid waste disposal
operations is estimated as follows:
                  X > 10,000          P  =  + 0.100
          1,000 < X < 10,000          P  =  + 0.200
                  X <  1,000          P  =  + 0.500
Sources disposing of more than 10,000 tons per year usually are
municipal incinerators, whose operations maintain records of
the quantities burned.  The precision of values for the amount
burned is not very good, however, generally about 25*000 or
+ 0.100.  Industrial and commercial incinerators burn less waste,
and operators keep few, if any, records of the amounts burned.
Generally, the less waste incinerated, the less precise the
thruput.  Thruputs for industrial incinerators often range from
1,000 to 10,000, with an approximate precision of + - 'QQQ or
+ 0.200.  Thruputs of smaller units (generally commercial units
or small industrial incinerators) are often estimated when re-
ported for NEDS purposes.  The resulting precision is in the
           OCA
range of + |^- or + 0.500.
     Precisions assigned to thruput values for all categories
are summarized in Table 2.17.
                                2-49

-------
Table 2.17.  PRECISION OF THRUPUT VALUES
SCC designation
1 01 XXX XX
(Electric
generation)
1 02 XXX XX
(Industrial
boilers)
1 03 XXX XX
(Commercial/
institutional
boilers)
2 XX XXX XX
(Internal Combus-
tion engine)
3 01 XXX XX
(Chemical
manufacture)
3 02 XXX XX
(Food/agricultural)
3 03 XXX XX
3 04 XXX. XX
(Metallurgical)
3 05 XXX XX
(Mineral products)
Thruput,
X ^
10,000 2 X *
X *
X 2
50,000 « X «
1,000 i X «
X *
X =
50,000 « X «
1,000 S X «
X *
All X

All X

All X

X 2
50,000 * X *
X «
X 2
50,000 ^ X «
10,000 * X *
X «
X
200,000
200,000
10,000
200,000
200,000
50,000
1,000
200,000
200,000
50,000
1,000






500,000
500,000
50,000
1,000,000
1,000,000
50,000
10,000
Precision, P
P
P
P
P
P
P
P
P
P
P

P

P

P
P
P
P
P
P
P
= t 0.001
= t-yr-
= ± 0.001
= + 0.001
= ± 0.020
= ± 0.001
= ± 0.001
+ 500
= 1 0.020
= 1 0.001
= i 0.002

« +• 0.010

= + 0.100

= +0.001
7500
= ±0.010
* +0.001
4.1,000
* ±0.020
» ±0.050
                    2-50

-------
Table 2.17 (Continued).  PRECISION OF THRUPUT  VALUES
SCC designation
3 06 XXX XX
(Petroleum
industry)
3 07 XXX XX
(Wood products)
3 09 XXX XX
(Metal fabrication)
3 20 XXX XX
(Leather products)
3 30 XXX XX
(Textile
manufacturing )
3 90 XXX XX
(In-process fuel)

3 99 XXX XX
(Other/not
classified)
4 XX XXX XX
(Point source
evaporation)
5 XX XXX XX
(Solid waste
disposal)
Thruput, X
X = 200,000
10,000 * X « 200,000
X « 10,000
X * 10,000
X « 10,000
All X

All X

All X

x : 200,000
50,000 * X « 200,000
1,000 : x < 50,000
X « 1,000
All X

All X

X ± 10,000
1,000 : x « 10,000
X « 1,000
Precision, P
P
P
P
P
P
P

P

P

P
P
P
P
P

P

P
P
P
= + 0.001
* IT"
= + 0:001
= + 0.010
= + 0.050
= ± 0.050

= + 0.050

= + 0.050

= + 0.001
+ 500
~ - TT
= + 0.020
= + 0.001
= + 0.050

= + 0.010

= + 0.100
= + 0.200
= + 0.500
                         2-51

-------
2.4  PRECISION OF VALUES FOR ASH AND SULFUR CONTENT OF FUEL
2.4.1  Anthracite and Bituminous Coal
     Users of large quantities of coal periodically test the
fuel's sulfur content by a procedure developed by the American
Society for Testing Materials (ASTM).  Even though the test proce-
dure is relatively simple, results vary because of retention
of sulfate sulfur in the ash and degree of precision with which
the sample is weighed.  The ASME Power Test Code for Solid Fuels,
PTC 3.2-1954,  Par. 69 specifies that reproducibility of results
of ash and sulfur content determinations shall be as follows:
                              Permissible differences when
                              tested in different laboratories,
     Coal characteristics     absolute percent
  TEST FOR ASH
     No carbonates present                 0.3
     Carbonates present                    0.5
     More than 12% ash, containing
     carbonate and pyrite                  1.0
  TEST FOR SULFUR
     Sulfur - under 2 percent              0.10
     Sulfur - over 2 percent               0.20
Smith and Gruber report that the average ash content of coals
from the major producing districts is always less than 12 percent
and that two carbonate compounds occur frequently in coal.   It
is assumed that carbonates are present in all coals represented
on the NEDS form.  Precision of reported ash contents can then
be expressed as follows:
                                 2-52

-------
Reported ash content
< 12.0 %

> 12.0 %
Precision
°-5
- % A
- FT
     Precision of the reported sulfur contents of solid fuels
can be expressed as follows:
Reported sulfur content
< 2.0 %
> 2.0 %
Precision
+ °-1Q
+ !?-20
2.4.2  Distillate and Residual Oil
     As with solid fuels, the ASTM specifies the testing of oils
to determine sulfur content.  The ASME Power Test Code for liquid
fuels, PTC 3.1-1958  specifies reproducibility of results of
sulfur content determination as follows:
           Sulfur
           content,
           percent
      Reproducibility by different
         operator and apparatus,
(absolute deviation  from mean, percent)
           0 to 0.5
      over 0.5 to 1.0
      over 1.0 to 2.0
      over 2.0 to 3.0
      over 3.0 to 4.0
      over 4.0 to 5.0
      over 5.0
                  0.03
                  0.04
                  0.05
                  0.07
                  0.09
                  0.12
                  0.14
     The precision of reported contents of oils can be expressed
                                2-53

-------
by the following equation:
    Reported sulfur content
         S < 0.5


         0.5 < S < 1.0


         1.0 < S < 2.0


         2.0 < S < 3.0


         3.0 < S < 4.0


         4.0 < S < 5.0


         S > 5.0
                      Precision
                          0.03
                          %~~S~

                          °'04


                          0.05
                        .  0.07
                        i FT-
                        ,  0.09
                          0.12
                          0.14
2.4.3  Liquefied Petroleum Gas

     The standard method of test for total sulfur in fuel gases

is specified per ASTM Designation D1072-56 .  The absolute preci-

sion of this method as determined by ASTM is + 0.1 grain of

sulfur per cubic feet.  This is equivalent to + 0.0001 absolute

percent sulfur by weight.  The precision of any reported value

of sulfur content  (reported in percent on the NEDS form) is  there-

fore:
Precision = +
                                   0.0001
                                     T~S~
                                 Z-54

-------
        3.0  PRECISION OF AREA SOURCE EMISSION VALUES

     Emissions from area sources are calculated by use of
the appropriate AP-42 emission factors and data entered on
the NEDS area source coding form (Figure 3.1).  In this
section, the emission factors, categorized as nonautomotive
and automotive, and the NEDS input data entries are analyzed
to determine the precision values.
3.1  PRECISION OF NONAUTOMOTIVE AREA SOURCE EMISSION FACTORS
     The area source emission factors used in the NEDS
computer program are taken from EPA Report AP-42.  Where the area
source emission factor (or category) corresponds to a point
source emission factor, the precision determined for the
corresponding point source emission factor is also applied
to the area source emission factor.  Precisions of the area
source emission factors are tabulated in Table 3.1.
3.1.1  Priority Pollutants
     The priority area source pollutants (nonautomotive)
are particulate, SO , and NO  from industrial combustion of
bituminous coal, NO  from commercial and institutional
combustion of natural gas, HC from solvent evaporation and
gasoline marketing, and HC and CO from off-highway gasoline
usage.
                                 3-1

-------
ill
i!
  a?
  r»
J%d
»oe4
' ion
O.M.O.I
|>4M.
I1   ~
* r    ,. i
ii    K
                3s
                ill
        i-53d
        S*V3H
_!-!
 IW
jut
H^IH
 113
 In
 *P
                ss£-
                      *•*[='
                      r I o'
                        in
E
LE HLES

uluit Ruill
vfmtn
                                £f
                      r«Li.
                      « 0,-J«1_
                      g-lir
      » 14-1
     ^ £ ^ ^£V-J
     3 > *i ^ 1
                        n-^
                        <

                         T
EFUS


e«k
I
                                     f|
                       s
 r*fr
             l^
             33
             * =
      5"i1
                l^-^*
                2s
        •»
        =t
        5s«l|d
        II*
O m

M
ll
CWMtRCIAL M

BXux* I Onl OM
U> t»i 1 ig4
Arti

»> tot
                  (**
                        -
                      S B .31
                      - It-i.r
si ^
sfirri
 ?.r5
 t -'-s
              uifo
              SS[^
               — t r>' *
                     ill^J
                     . j-^^s	i
                     uJ £ S ^» :
                     S   ~
                     ri-fe[sd
                     s  hSH
                      1,^
                      i «,~
                      ! S1^..
                      !i^
                      " i si

                     S..HH
R«i
io '
              s ^r*«T
              * •« i n*
              i?!r.r
              3 s

              1^
                      I«r~in
                            L!^

                                -?M
          M—!

                            1
                                                        f

                                                        1
                                                        S
                       3-2

-------
 §
 H
 CO
 CO
 H
ro

 O








c
a
a
"5
a.












g



U


M
e



Q
v»

a
•^
U
~
IQ
o.
hi
XI

= 1
= 3

U





^
w
e
M

n
0






































rH
1C
0
U
01
•P
•H
U
IB
rl
^
J^
*

m
ro
o

ro
CN
o
CM
r-
•-J
o

VO
^*
•-J
o

rH
rH
CN
O
at
o
r-l
O
o
ro
0
rH




rH
id
•rl
4-1
C

ir
f
c
r»
ro
c\
o
CN
r->
•~
e

vfl
^j
«-
O

r>
r-
CN
C

r-
CN
O
o
ro
C
rn




H
IB
•H
4-1
C
0)
T3
•H
CO
(U
a:



in
ro
o
^
ro
CN
O
O
o
P*
o

*3*
in
CN
o

CN
ro
rH
o

OH
CNCN
OO
00
mro
OO
rHrH
.
03
C
M

rH
•H
U
rl
(U


o
U


r-
m
ro
o

IT
CN
O
id
o
o
f-
o
id

in
CN
O
id
CM
ro
rH
O
<*
o
CM
O
O
CN
0
rH





rH
id
•rl
rl
^J
CO
3
•O
C
H



































rH
•H
O
0)
4J
>d
rH
rH
•rl
•P
CO
•H
a
r»
m
ro
o
r-
ro
CM
o
CN
r>"
r-
O

vo

rH
O

rH
rH
•"vl
O
ro
O
in
O
o
ro
O
rH




rH
id
•rl
•P
C
0)
•a
•rl
CO
«


r»
in
ro
o
r-
ro
CN
O
CN
r~-
rH
O

VO
TT
rH
O

rn
rH
CN
O
CN
O
in
o
o
ro
O
rH
4J
CO
C
H
X
rH
id
•rl
O
rl
(U

s
0
y


r~
in
ro
0
i~
ro
rv
O
CN
t^

O

VO
^1
rH
0

rH
rH
CM
O
CM
0
in
o
o
CM
o
rH





rH
id
•H
rl
4J
CO
3
•a
c
H





































rH
•rl
O

rH
*d
3
13
•H
CO
(U

r-
m
ro
o
r-
ro
CN
O
ro
O
0
o

r-
CN
o
o

rH
rH
CM
O
CN
O
TJ*
O
O
CN
O
rH




rH
id
•H

C
(U
•a
•rl
CO
£



in
ro
0

ro
CN
O
ro
o
o
o

p*.
CN
0
O

rH
H
CN
O
CN
O

O
0
CM
O
rH
•P
m
C
H

rH
id
•rl
O

0)

g
o
U



in
«*
o

ro
CM
O
ro
0
o
o

(^
CN
O
O

^-
f.
O'
o
CM
O
*4*
O
0
CM
O
rH





rH
id
•H

4J
CO
3
T3
C
H


(U
U
M
3
O
CQ

id
CD
M
id
                                                                                                                          4J

                                                                                                                          •rl

                                                                                                                           rl

                                                                                                                           O
                                                                                                                          •rl

                                                                                                                           l-f

                                                                                                                          Oj
                                                                3-3

-------
§
H
(0
CO
•O
 O
4-1


O
U
ro

 <0







c
IV
JB
"S
o.








O
O



X


M
e

ffM
O

ft
o
'•£
rrj
O-

**
e_
a **
C3
o
u








^
w
o
oa

**
CJ










































tli
«J
O

^
«J
J.

n
2


ir
r1
C
r
(*"
CM
C
r-
oc
•-
a
VO
rH
O
H
rH
CS
a
m
o
vo
0
0
f
o
r—







r-
(C

4-i
c
V
•c

V.
a




IT
01
O
r-
»r
CM
a
p-
a
•-
o
VO
r-
a

r-
(N
O

o
VO
0
o
rO
c
r—



4.
in
C
H
X,

re
• r-
O
>H
0)
i
e
o
u




I/I
rn
a
r>
rn
CM
O
«3
H
0>
n
o
vo
rH
O
rH
rH
^|
O
CM
O
VO
o
0
CM
O
rH








r— j
ro
-rH
^
4-1
W
3

C
H




in
m
o

(T
CM
O
CM
r-
rH
O
VO
rH
O
iH
rH
ft
O
CM
O
G\
O
O
CM
O
rH
10
0)
•H
)H
O
cr
0)
4J
(0
o

rH
rH
(0

I

•g
o



in
cn
o

m
CM
O
rH
fj>
n
o
vo
rH
O
•H
rH
CM
O
CM
O
VO
o
o
CM
O
rH


•
(0
3
T3
C
H

U]
(0
O

in
in
0)
u
C

0,



























c
o
•rH
4->
m
)_
0)
c
-rH
U
C
H

a)
4J
.,.
CO
c
o

IT
fl
a
r-
rn
CM
o
CM
^
H
O
VO
rH
0
rH
rH
CM
O
CM
O
rH
O
o
CM
O
in







r-
(C
•H
4J
C
(1)
•a
•r-
V.
K



m
n
o

n
CM
o
CM
1**
rH
O
VO
rH
O
rH
rH
• CM
O
•H CM
O 0
•H rH
O O
O 0
CM CM
O O
in in








rH
(0
-H
U
M
0)
E

C



in
ro
o
«*.
CO
CM
O
CM

rH
O
VO
•H
O
rH
rH
CM
O


rH CN
O O
0 1
CM rH
O O
in •—








rH
(fj
•H
)H

U)
3
•a
c
M





































cn
c
•H
C

3


C
0)
a
o

m

0
r--
ro
CM
o
CM
f^
rH
O
VO
rH
0
rH
rH
CM
O
CM
O
CM
O
O
m
0
in







rH
10
•rH
4J
C
0)
-a
•H
m
&



in

o
r-
CO
CM
O
CM
{^
rH
O
VO
rH
O
rH
iH
CM
O
CM
O
CM
O
o
r-
m

o
r-
f>
CM
O
CM
r^
iH
O
VO
rH
O
rH
rH
CM
0
CM
O
CM
O
O
<**"> 1 en
0 0
m








rH
ro
•rH
O

0)

p
6
u



in








rH
10
•H
^
4-)
10
3
TJ
C
t-t






































C
o
•H
4-1
(0
Vl
0
a,
10
^
w


1

0
o
o
o


1


1



I
1



















) >
c
(U

rH
O
co



0)
o
M

O
U)

(0
0)
IH
(0
                                                                                                                         rH

                                                                                                                         O
                                                                                                                         cu
                                                              3-4

-------
§
H
ca
to
1

•H
4J



§











B
IV
a
«
e














O





™*



M
e
ac


o*
M


IV
u
+*
IB
X.




u,
Jo
E-o
a «*
c a
u
CJ
v>










>*
£;
QA
V
n
o















































^.^
*
•a

4J
c
o
o
**•*

c
o
•H
4J
a
M
g

^
H


1
1

n)
rH
CN
cn
•
O


1
1



1
1



1
1

















tr
c
•H
4J
<]>

J_(
1C
g

CD
c
•H
rH
O
CO
m
O




1
1




1
1



1
1



1
1


a
X




























co
TJ
(0
o
«
4-1
IH
*H
Q


1
1




|
1



1
1



1
1


Q
X
























to
C

J»
^.
0)
JN
•H
4J
^.
•rl
a


i
i




i
|



i
i



i
i


a
X
JjJ

























c
o
•H
4J
CJ
3
M
JJ
CO
C
0
O


|
1




1
1



1
|



1
1


Q
X
Z















(u
^

CO

<«

Oi
c
•H

•a
c
B

O
O


a

[2



a
X



a
X



Q

55


Q

^

























co
at
Vl
•H
PT,
4J
CO
0)

o


in
n
•
o

r^
ro
CN
•
0

p«fc
rH
•
O


|
i

00
o
(N
•
O




CN
O
CN
O
O

O
in











tr
c
•H
C
M
3
CD
A
CO
n)
rH
U)

Q
X




Q
X



Q
X



a
X
z


a
X
z
























f—\
o

4J
C
O
u
4J
CO
O

fe

a
X
z



a
X
z


Q
X
z


Q

2


O
X
z






















CO
0)
M
••H
fa

0)

3
4J
O
3

4J
CO

Q
X
Z



a
X
z


a
X
z


Q
X
Z


Q
X
^


















{y
p
•H
C

3
PQ

d)
CO
3

cu
K
tH
(0
o
u



















































^
i ca
4-> CO
•H -H
rl g
0 Cd
•H
M II
Q
X
R) z
                                                3-5

-------
CO

CO
§
M
CO
•O


 §


•H

4J



 O

 U
 0)
iH
a
 a








t .
K
*a
~o
o-







f.
E
•a
e
u
u
C/l

















e
u


o




o
z

(N
0
VI

"3
'•£
ra

"S
a




**
e
40
^*
n
^







































-P
»4_
ft
^
u
M
•r-
^

in
01
o

ro
CM
o
CM
p*
iH
o
VO
H
o
H
CM
•
O










>

It

•H
P»|
•H
g



m
ro
0
r-
00
CM
0
CM
r^
•H
O
VO
i-4
O
„
CM
•
O













i^
•H
^
•H
U



in
m
o
t-
01
CM
O
CM
r-
H
O
VO
1-1
o
rt
CM
•
O








rH
id
•r-l
O
^
(L

€
O
u


































in
r-i
0)
CO
to
0)


in
01
o

n
CM
o
CM
r^
•"I
o
VO
H
O
H
CM
•
O




r-l
id
o
o

m
•s
o
c
•H
£T
P
4-1
•H
n


r*-
in

o

n
CM
O
CM
p^
iH
0
VO
•H
O
rf
CM
•
O





•H
O

V
4J

rH
rH
•H
4J
CO
•H
Q


r-
m
1
o
r-
01
^
o
CM
r**
•H
O
VO
H
O

CM
•
O






iH
•rl
O

rH
(TJ
3
13
•H
10
(U
ft


r»
in
01
0
r-
01
CM
o
CM
r*
•-]
0
VO
3
o

CM
•
O










(U
c
•H
rH
0
(0
id
0




a
g


D
§


Q
Z

a
2

Q
X
2




1C

^.
4-
(fl
^
•a
c
H

i

0)
,*
0
u










































































































































































































































•a
 0)
                                                                                                                   (1)
                                                                                                                   •a


                                                                                                                   4-1
                                                                                                                   (U
 g
                                                                                                                   U
                                                                                                                   a
 o
•H
 tn
 in
•H
                                                                                                                    II

                                                                                                                   Q
                                                            3-6

-------
3.1.1.1  Area Source Combustion - The priority pollutants
from area source combustion are also priority pollutants
from point sources of combustion.  Calculations determining
the precision of the area source emission factors (as in the
section on point source emission factors), are shown in Table
3.2.
3.1.1.2  Evaporation - The emission factor for solvent
evaporation is 2000 Ib/ton, which means that the amount of
solvent purchased for a given operation is equal to the
amount needed to make up evaporation losses.  Therefore, any
imprecision is entailed in the estimate of the amount of
solvent purchased, and the emission factor of 2000 Ib/ton is
considered precise.
     An AP-42 emission factor is given for HC from gasoline
marketing.  Since this is a priority area source emission
factor, it is handled in the same manner as a priority point
source emission factor.  Calculation of precision of these
factors is given in Table 3.2.
3.1.1.3  Off-Highway Gasoline Usage - Emission factors for
off-highway gasoline usage are assumed to have been derived
from factors for emissions from heavy-duty, gasoline-powered
(HDG) highway vehicles.  Using the HDG emission rates as the
source of the emission factors for off-highway gasoline
usage may affect the accuracy of the off-highway emission
factors, but does not affect the precision.  Therefore, the
precision values for HDG emission rates are used in cal-
culating the precision of off-highway gasoline emission
factors.
                                 3-7

-------
CM
  •
n

 4)
i-l
A
 <0
*.
n
O 1^
A.
•M
W*
• *5j
£ "^ '^S
18 s **T
V M
•M **
** I
• *-•*
^ M
2 5 ^
V» £
1
o •
. w *
5 .2 ^
o *»
U »

^
7* — *
C» "
•c e
— * ,„
S E
3}
t* W M
is*

~
M
a
r





»•
^
M
•*
*•
W



M
n
o

rs
f»
r-


a
•-J
t>


*f
IT
0
<
r-
i*



ea

0


N
N

Ol
*
o
00
l-l


•a

o
id
kl C
3 O
4J-H
Id 4J
c n

r-l £

••H 0
4J °
II) ID
3 a
ig cr
HI
o
0
o
o





















0
o
o
fM



' g
^^
1 •
id u


u* E
Id -~

41 n
n
•w o
C r-l
01
> C
•H e
S I
fs
0
•

f
T
rH


»
Oi
H

CN
at
0




t.
X
0
"i
U
rH O
CN r-l
"V
£>
rH



g
O"
c

±
01

u
a
E

0)
c
•r
'c
u
C
IM
c
c





















m
in
m




^





»
^,
id o
S c
£ •'
•-T'c
£ I
1 A
IW t
O
00
o
o





















o
o
m



O
0







>

4
H
3
1
1
7'

                                                                  I
                                                                  JJ

                                                                   I
                                                                   a
                                                                  JS
                                                                   o
                                                                id a

                                                                S"
                                                                41 TJ
                                                                a a
                                                                au
                                                                o o
                                                                   o
                                                                Oh
                                                                •8S
                                                                        a
                                                                        u
                                                                        a
                                                                 41 *J

                                                                AJ e
                                                                   o

                                                               ••H O

                                                              U  kl-H
          >  O
          r-4  4J

          ss
             IU

      ,JJIB
   _ J  C M  O
   O    Id    -<
u a  -4J IM  a
•H n-i u  3 o  «
     41 rH    -H
 - C O rH rH  E
 • O C  0 "I  S
>•* Q  a.n

CJ W QJ  4) 4J  Q
   -H IM  CJ Id  X
41 6 01  kl £
tr S kl  3 4J   c  u e
« OI-H    o  c
> •-*   4J-H  O
id ja TJ  c -u-rt
   ™ c -H a »
 "-I 3  O E  -H
O-H o  a3  o
   n! 14-1    n  4)

o m c 4?"  a
     o -H
^H 41 -H  h 41  4)
   U 4J  O £  £
01 C Id -H 4J  *J
•-I 41 >  kl
   ki ki  a B  c
c 4i ai    o  o
   u-i n  a
4) 4) Xt    *O  *O
ki ki O  0 4>  41
4i      ana
c O o >-i a  id
3 z z < m  n
                                                              a .a OTJ 0*4
                                                               3-8

-------
     Precision of emission factors for CO from off-highway
gasoline usage is assumed to be equal to that of the emis-
sion factor for uncontrolled  (pre-1970 model year) HDG
exhaust.  This precision value, discussed in detail in a
later section on automotive factors  (Section 3.3.2.1), is
calculated as follows:
  Precision_o =
Var (C.) 1/2            . ..
        1     = [Q.00081]^
(C.)
                             0.028
     The emission factor for HC from off-highway gasoline
usage is calculated as the sum of the HC in uncontrolled HDG
exhaust and combined HC emissions from crankcase and evap-
oration.  Thus the variance of the off-highway HC emission
rate is equal to the sum of the variances of the uncontrolled
HDG exhaust and combined crankcase and evaporation HC emis-
sion rates:
   Var  (Off-Highway HC) = Var  (C^.^) + Var  (fHDG_HC)
                        » 0.272 + 0.952
                        - 1.224
     The precision of the emission factor for HC from off-
highway gasoline usage is therefore:
  Precision
           HC
Var (Off -Highway HC)
(CHDG-HC + fHDG-HC^ 2
1/2
31
1.224
(17 + 8.2)~
                                                        1/2
                                         =  0.044
3.1.2  Non-P.riority Pollutants
     Emission factors for all other area sources are given
the precision of the corresponding point source emission
                                 3-9

-------
factor.  When there is no corresponding point source emis-
sion factor, the area source is assigned the average pre-
cision that was determined for point source emission
factors.  Precision values are given in Table 3.1.
3.2  PRECISION OF AREA SOURCE CODING FORM ENTRIES
     The purpose of this section is to define a precision
for every entry on the NEDS area source coding form, ex-
cluding the emission estimates.  The coding form and the
entries are discussed in detail in EPA's "Guide for Com-
piling a Comprehensive Emission Inventory" , published as
APTD-1135.  When used in conjunction with the precision
ratings of the emission factors, the precision values for
NEDS entries will define a quality rating of the area source
emission values in an emission inventory.
     For all entries on the area source coding form, the
method of determination is assumed to be correct.  The
precision of the entry is then dependent on the precisions
of the factors used to determine the entry.  For instance,
if an entry X is a function of factors A, B, and C according
to the equation:
          X = (A)(B)(C) then the precision of X  (P ) is a
function of the precisions of the individual factors as
follows:
     For the cases in which APTD-1135 provides several
methods for determining a single entry, precisions are
                                 3-10

-------
calculated only for the method believed to be  most.


widely used.  For a factor whose precision varies over a


wide range of values dependent on the size of the factor, a


median value is used in determining precision.


     All entries entail rounding errors.  The units of each


information field on the coding form define what rounding


off, if any, must be performed.  For example, if values to

                                                  4
be entered in a field are expressed in units of 10  gallons,


the entry should be considered precise within + 5 x 10


gallons.  In addition to the rounding errors, the precision


of the data-gathering techniques must be considered.   When


the precision of data-gathering techniques is zero, the


precision of rounding is used.  When a precision other than


zero is assigned to the data-gathering techniques, that


precision is used for the entry on the coding form.


     Precisions of entries on the NEDS area source coding


form are summarized in Table 3.3.


3.2.1  County Data


3.2.1.1  Population - County population figures are obtained
                                   g
from the 1970 Census of Population.   The Bureau of the Census


reports almost complete coverage in the 1970 census.   They

report the population to an exact number, much more precise

than the nearest thousand coded on the NEDS form.  The


precision therefore depends only on the significant figures


reported.  For this field the precision equals +500 divided


by the entry.
                                3-11

-------
Table 3.3  SUMMARY OF THE PRECISIONS OF ENTRIES



      ON THE NEDS AREA SOURCE CODING FORM

Card
1



1



1
1
1
1
2



2
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3

Columns
50-53



54-58



59-63
64-68
69-73
74-77 '
10-14



15-19
20-24
25-29
30-33
34-35
36-41
42-47
48-51
52-56
57-61
62-66
67-69
70-73
10-15
16-20
21-24
25-30
31-36
Precision
of entry
+ 0.196

+ 0.045

+ 0.196

+ 0.045

+ 0.104
+ 0.144
+ 0.054
+ 0.555
+ 0.615

+ 0.637

+ 0.294
+ 0.312
+ 0.144
+ 0.050
+ 0.269
+ 0.664
+ 1.26 x
+ 0.150
+ 0.023
+ 0.023
+ 0.127
+ 0.250
+ 0.200
+ 10.00
+ 10.00
+ 10.00
+ 10.00
+ 10.00

Comments
(for a zero in Card 1, Columns
54-58)
(for an entry or blank in Card 1,
Columns 54-58)
(for a zero in Card 1, Columns
50-53)
(for an entry or blank in Card 1,
Columns 50-53)




(for a zero in Card 2, Columns
36-41)
(for an entry or blank in Card 2,
Columns 36-Al)






102 (entry) ~'97











                        3-12

-------
Table 3.3  (continued).  SUMMARY OF THE PRECISION OF  ENTRIES



             ON THE NEDS AREA SOURCE CODING FORM
Card
3
3

3
3

3
3.
3
3
3
4
4
4
4
4
4
4
4



Columns
37-42
43-49

50-54
55-59

60-64
65-67
68-72
73-76
77
10-13
14-18
19-23
24-27
28-31
32-36
37-40
41-46



Precision
of entry Comments
+ 10.000
+ 0.045 (for an entry in Card 4, Columns
52-76) Variable G^
+ 0.071 (for no entry in Card 4, Columns
52-76) Variable GX
+ 0.045 (for an entry in Card 4, Columns
52-76) variable G2
+ 0.072 (for no entry in Card 4, Columns
52-76) variable G,
[0.022T2-.572 TP + 28.73 P2]1/2
T
where T = entry in Card 3, Columns 55-59,
and P = entry in Card 3, Columns 73-76
+0.5 Variable G3
entry
+ 0.082
+ 0.100
i SOP
entry
precise number, p = 0.000 Variables P and PR
t 0-5
entry
+ 0.5
entry
+ 0.5
entry
+ 0.030
+ 0.130
+ 0.180
+ 0.158
for the entry in Card 3, Columns 73-76
in the range :
0 to 100 	 + o 936
100 to 500 	 + n AQ A
500 to 1000 	 + 0 246
1000 to 9999 	 + 0.165
                               3-13

-------
Table 3.3 (continued).  SUMMARY OF THE PRECISION OF ENTRIES




             ON THE NEDS AREA SOURCE CODING FORM
Card
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
5
5
5
Columns
47-51
52-57
58-63
64-69
70-76
10-16
17-21
22-26
27-31
32-38
39-41
42-47
48-50
51-54
55-57
58-61
62-67
68-70
Precision
of entry
+ 0.060
+ 0.032
+0.032
+ 0.032
+ 0.032
+ 0.503
+ 0.000
+ 0.5
~ entry
+ 0.5
entry
+ 0.050
+ 0.150
+ 50 	
~ entry
+ 0.250
+ 0.5
~ entry
+ 5.0
~ entry
+ 5.0
entry
+ 0.270
precise
Comments

Variable M,
Variable M2
Variable M3
Variable M.








Position 385-388
Position 389-391


number, p = 0.000
                               3-14

-------
3.2.1.2  Density Code - The 1970 census also classifies
urban and rural populations in a county as a percentage
urban.  This percentage is precise based on the classifi-
cation specified by the Bureau of Census.  Since APTD-1135
defines a code for each percentage, no variation is intro-
duced, and the density code is a precise number.
3.2.2  Sulfur Content
     APTD-1135  specifies using a weighted average for
sulfur content of fuels consumed in a county.  This aver-
aging creates serious problems in determining precision.  To
simplify the determination, assume that fuels consumed by
different categories of users have the same sulfur content
countywide.  This is a logical assumption, since most coal
used in a given area is mined from the same general region.
3.2.2.1  Anthracite and Bituminous Coal - The precision of
tests for determining sulfur content of coal is discussed in
Section 2.4.1, which specifies that precisions of the sulfur
content of solid fuels can be expressed as follows:
Reported sulfur content
< 2.0%
> 2.0%
Precision
, 0.10
- ~%S~
. 0.20
- ~%s~
Distillate and Residual Oil - The precision o
tests for determining sulfur content of fuel oil is dis-
cussed in Section 2.4.2.  Precisions of the sulfur content
of fuel oil can be expressed as follows:
                                3-15

-------
         Reported sulfur content,%
             S  2  0.5
             0.5  «  S  =  1.0
             1.0  «  S  Z  2.0
             2.0  *  S  Z  3.0
             3.0  «  S  ;  4.0
             4.0  «  S  i  5.0
                  5.0
Precision
                                              + 0.03
                                                0.04
 ± 0.05
   % S

 ± 0.07
   % S

 ± 0.09
   % S

 ± 0.12
   % S

 ± 0.14
   % S
3.2.3  Ash Content

     Again, APTD-1135  specifies using a weighted average to

determine ash content of coals over an entire county.  In

this case, also, assume that the ash content is constant

over the county.  The ash contents for anthracite and

bituminous coal are determined in the same manner if ASTM

procedures are followed, as discussed in Section 2.4.1.

Accordingly, precision of the ash content value reported on

the NEDS form may be expressed as follows:
Reported ash content , %
<_ 12.0
> L2.0
Precision
+ ^!
- % A
. 1T0
- % A
                                3-16

-------
3.2.4  Residential Fuel
     For most fuels, residential fuel usage is determined by
apportioning statewide fuel totals to residential categories
and then apportioning the residential categories to counties
on the basis of dwelling units per county heating with the
fuel.
3.2.4.1  Anthracite Coal - Residential consumption of an-
thracite is usually determined on the basis of the number of
dwelling units heating with coal as enumerated by the
                       9
1970 Census of Housing.   The average consumption per
dwelling unit is the product of the coal heating requirement
factor , the average annual heating degree-days from Local
Climatological Data  , and a correction factor for the
median number of rooms per dwelling unit in the county
                                                          9
versus the national median of 5.0 rooms per dwelling unit.
The calculated residential usage of coal must be compared
with the statewide area source totals, determined by sub-
tracting the point source consumption of anthracite from the
total state consumption of anthracite given by the Bureau of
Mines.    If the calculated residential usage of coal should
exceed the statewide area source usage of anthracite, con-
sider the statewide usage as the residential consumption of
anthracite.  Assume that the difference between the cal-
culated residential usage of coal and the area source con-
sumption of anthracite represents the residential consump-
tion of bituminous coal.  The precision of the entry for
                                3-17

-------
residential consumption of anthracite coal will depend on

precisions of the factors used in the determination.

     Number of dwelling units heating with coal - This
                                                  n
figure is obtained from the 1970 Census of Housing  under

"House Heating Fuel".  This source gives a county-by-county

enumeration of houses heated with various types of fuel.

The figure is obtained from a 5 percent sample of all the

housing units nationwide.  The Census of Housing provides

tables for'calculation of the approximate standard error

based on the percent sampled and the number of housing units

in the area.  Since the standard error varies from county to

county, we generalize a precision value based on data from

the states of Ohio, Michigan, Kentucky, and Iowa.  See

Appendix A.

          precision = + 0.14

     Coal heating requirement factor - This factor repre-

sents the coal required to heat an average 5-room dwelling

for one degree-day.  The factor cited in APTD-1135 is 0.0012

ton coal per dwelling unit per degree-day.  Calculating by

the method in Appendix B we obtain:

          Type House          Factor

          Insulated           0.001097 ton coal/D.U.-D.D.

          Noninsulated        0.00161 ton coal/D.U.-D.D.

     Analysis based  on these values indicates that the

factor cited in APTD-1135 was derived by assuming that  80

percent of households heating with anthracite coal are

insulated.
                                 3-18

-------
     Factor - 0.8  (0.001097) + 0.2  (0.00161) * 0.0012



     Assume that 66.7 percent of the values occur in the



range from 100 percent insulated to 60 percent insulated and



that this range represents two standard deviation units.



                              Factor



          100% insulated      0.001097



           60% insulated      0.0013022




     standard deviation = °-0013022 ~2 °'0°W = 0.0001026



           -    precision^ ^01026 =± o.084




     Annual heating degree-days - Degree-days are obtained



from local meteorological data on average 24-hour temper-



atures by defining a degree-day as a day where the average



24-hour temperature is 1 degree below 65°F.  Temperature may



be measured very precisely.  Considering a 0.5-degree error



in the significant range of 10 degrees or greater below 65°,



we get a maximum error of + 0.05.



     Correction factor for number of rooms per dwelling unit



in the study area - This factor is based on a 20 percent



sample of all households and therefore should be precise



within the limits of the reported units.  Since this value



is reported to the nearest tenth, rounding off errors would



give +_ 0.05 or less.  The precision would then be +^ -£— .



The 5.0 in the denominator is a precise figure since it is



the basis for calculation of heat loss.  Therefore, precision



- + 0.01.



     Total statewide area source consumption of anthracite



coal - The precision of this factor depends on the precision
                                3-19

-------
of the state consumption totals in Minerals Yearbook   and
the precision of the state point source totals.  The state
totals are reported to the nearest ton and at least four
significant figures.  Consider this as a precise number,
since the worst case is precise within + 0.0001.  The
reported point source consumption is based on information
from each point source.  The precision depends on the number
of significant figures reported.  Most point sources report
at least three significant figures.  Assuming this is the
case, the precision equals + y^- = + 0.005.  Assuming that
                           — j.uu   —
the point source total accounts for no more than 90 percent
of the state totals, the precision of the area source totals
is + 0.045.
     Overall precisions - Two cases must be considered in
determining the overall precision of the anthracite coal
entry.
     Case 1:  Anthracite coal is the only coal used as
     residential fuel  (a zero will appear for residential
     bituminous coal consumption) .  The consumption of
     anthracite is the product of the factors developed
     above.  Overall precision  (P ) is determined as fol-
     lows :                       °
     P 2 =  (0.17)2 +  (0.084)2 +  (0.05)2 +  (0.01)2
      o
     P 2 -  0.03856
      O
     P  « + 0.196
      o   —
     Case 2 ;  Both anthracite and bituminous coals are used
     by residential sources.  In this case the residential
     anthracite consumption equals the difference between
     the Bureau of Mines figure and th$ point source totals.
     This precision,  calculated above, equals +_ 0.045.
                                3-20

-------
3.2.4.2  Bituminous Coal - Essentially the same procedures

and factors are used for calculating residential consumption

of bituminous coal as for calculating residential consump-

tion of anthracite coal; the same precisions apply.  Two

cases are considered for bituminous coal:

     Case 1;  Bituminous coal is the only coal used as
     residential fuel (a zero will appear for residential
     anthracite coal consumption.   This  total is  determined
     by the product of the factors defined in Section
     3.2.4.1.  The precision will be the same as determined
     for Case 1 in Section 3.2.4.1 and equals + 0.196.

     Case 2;  Both anthracite and bituminous coal are con-
     sumed by residential sources.  The figure is calculated
     by determining total coal usage and subtracting the
     total area source consumption of anthracite (assuming
     only residential sources).   The precision is dependent
     on the factors in Section 3.2.4.1, but since a sub-
     traction is involved, an assumption must be made as to
     the contribution from each factor.  The equation used
     to determine bituminous coal usage is:

                    B = T - A

          where     B = bituminous coal usage

                    A = anthracite coal usage

                    T = total coal usage

          From Section 3.2.4.1 the precision of T = + 0.196
          and

          the precision of A = + 0.045

          If it is assumed that anthracite consumption
          accounts for no more than 80 percent of the total
          coal consumption (A <_ 0.8T), the maximum error
          possible would be:

               + 0.196T + 0.045 (0.8T)

               or + 0.241T

          The precision of B, based on the maximum error and
          B + 0.2T, is

               ± 0-241T =
                  0.2T    - -i
                                3-21

-------
3.2.4.3  Distillate Oil - Residential consumption of dis-
tillate oil is determined by the same procedures outlined
with regard to anthracite coal.  The only factors that
differ are the number of dwelling units heating with fuel
oil and the oil heating requirement factor.
     Number of dwelling units heating with fuel oil - This
figure is obtained by the same procedure outlined for
anthracite coal calculations.  The calculation of standard
error is based on a generalization of data from Ohio,
Michigan, Kentucky, and Iowa, as in Appendix A.
               precision = + 0.06
     Oil heating requirement factor - This factor represents
the amount of oil required to heat an average 5-room dwell-
ing for 1 degree-day.  The value is given as 0.18 gallon oil
per dwelling unit per degree-day.  According to Appendix B
(ASHRAE Handbook values):
          Type House          Factor
          Insulated           0.16847 gal oil/D.U.-D.D.
          Noninsulated        0.26578 gal oil/D.U.-D.D.
     Most oil-heated homes are insulated; upon analysis, the
factor appears to represent a distribution of 88 percent
insulated and 12 percent noninsulated homes.  Since we have
no data on the actual ratio of insulated to noninsulated
homes for different counties, an assumption must be made.
If we assume that for 68 percent of the counties the range
is 75 to 100 percent insulated homes and that this range
represents two standard deviation units, we get:
                                 3-22

-------
                              factor



          100% insulated      0.16847



           75% insulated      0.1927975


     standard deviation - 0-1927975 - 0.16847 = 0>
                 0
     precision =  p     = ± 0.068



     Overall precision - Precision of the value for resi-



dential consumption of distillate oil is equal to the



product of precisions of the factors for degree-days (0.05)



and rooms-per-dwelling (0.01, both from Section 3.2.4.1) and



the factors for number of dwelling units heating with dis-



tillate oil and oil heating requirements as just calculated.



The precision is then:




     P 2 = (0.05)2 +  (0.01)2 + (0.06)2 + (0.068)2
      o




     P 2 - 0.010824
      o




     P -  - + 0.104
      o    —



3.2.4.4  Residual Oil - Residual oil is not usually burned



in residential furnaces.  If it is used within the area of



the inventory, however, contact the local fuel dealers to



determine what percentage of their customers are sold residual



oil.  The percentage is applied to the total amount of oil



consumed by residential users (with precision of 0.104,



determined above) to calculate gallons of residual oil.



Assume that the dealers are precise within 10 percent when



reporting the distribution of oil.  Precision of the value



for residual oil consumption is then determined as follows:
                                 3-23

-------
     PQ2 -  (0.104)2 +  (0.10)2





     PQ2 - 0.020816





     P   - + 0.144
      o    —



3.2.4.5  Natural Gas - Natural gas consumption by residential



users is reported by state in Reference 13.  The Bureau of



Mines states that these reports are obtained by direct



contact with dealers and others; they assert that these



contacts account for 80 percent of the gas produced and that



most of the remaining gas is accounted for in marketing



reports of the companies contacted.  The statewide totals



are apportioned to the county by the ratio of county/state



dwelling units heating with natural gas.  If we assume that



the Bureau of Mines statement implies that at least three-



fourths of the remaining 20 percent of natural gas usage is



accounted for, then the precision of the value for resi-



dential natural gas consumption per state is + 5 percent.



The precision of the number of dwelling units per county



heating with natural gas, calculated according to Appendix



A, is + 0.02.  The precision of the number of dwelling units



per state heating with natural gas is also calculated according



to Appendix A by using the average number of dwelling units per



state and consulting Reference 10.  The precision is + 0.001.



The overall precision  is then:





     P 2 «  (0.05)2 +  (0.02)2 +  (0.001)2
      o




     P~2 *  0.002901
      o




     P   -  +  0.054
      o     —
                                 3-24

-------
3.2.2.6  Wood - The procedures outlined with regard to
anthracite coal consumption are used again in determining
residential consumption of wood.  The two factors that
differ are the number of dwelling units heating with wood
and the wood requirement factor.
     Number of dwelling units heating with wood - This
figure is obtained by the same procedure used earlier for
other fuels.  The calculation of standard error is based on
a generalization of data from Ohio, Michigan, Kentucky, and
Iowa, as in Appendix A.
               precision = + 0.42
     Wood heating requirement factor - This factor repre-
sents the wood required to heat an average 5-room dwelling
for 1 degree-day.  The value given in APTD-1135 is 0.0017
ton of wood per dwelling unit per degree-day.  According to
Appendix B (ASHRAE Handbook values) :
          Type House          Factor
          Insulated           0.00189 ton wood/D.U.-D.D.
          Noninsulated        0.00286 ton wood/D.U.-D.D.
     These factors are higher than that given in APTD-1135.
If it is assumed that the APTD value is the result of two
observations, the precision may be calculated:
     a'  ~ 0.00286^0.00189  =8.6xlQ-4

     a   - 8'6 * 10"4  = 6.1 x la'4
     precision =             = ± 0.358
                                3-25

-------
     Overall precision - Precision of the value for resi-
dential consumption of wood is equal to the product of
precisions of the factors for degree-days (0.05) and rooms-
per-dwelling (0.01, both from Section 3.2.4.1) and the
factors for number of dwelling units heating with wood and
wood heating requirements, as just calculated.  The precision
is then:
     PQ2 * (0.05)2 + (0.01)2 + (0.42)2 + (0.358)2
     P 2 » 0.3072
      o
     P0  " ± °-554
3.2.5  Commercial and Institutional Fuel
     For most fuels, commercial and institutional use is
determined by obtaining statewide totals from Bureau of
Mines publications and apportioning to county on the basis
of population.
3.2.5.1  Anthracite Coal - Commercial institutional use of
anthracite coal is determined by subtracting statewide
residential use of anthracite from the total statewide area
source consumption of anthracite and apportioning the
balance to county on the basis of population.  The pre-
cisions of the values for statewide residential consumption
and total area source consumption of anthracite were deter-
mined in Section 3.2.3.4.  Also, the precisions for county
and state population figures were determined earlier, and
both are precise values.  The overall precision is then the
                                 3-26

-------
precision of the number representing the difference between
residential usage and total area source usage.  Assuming
that residential sources consume 75 percent of the anthra-
cite, the precision is determined as follows:
               T » A - R
     where     T = commercial institutional consumption
                   of anthracite
               A * total area source consumption of
                   anthracite
               R * residential consumption of anthracite
Var (T)
2
Var (A)
w.. /v*\
_ Var (A) + Va
(A - R)2
- (0.045A)2
ir (R)

     Var  (T)  m  (0.045A)2 +  (0.147A)2
       T2           (A - 0.75A)2
     Var (T)  = 0.0236A2
      TT        0.0625A2
     a  (T)  _ 0.154A
      T       0.25A
     but 0.25A * T
     therefore    a  (T)  a 0.154  (4T)

     precision = + 0.615
This precision value applies only to cases where no anthra-
cite is consumed by industrial sources.  For cases where
                                3-27

-------
anthracite is consumed by industrial sources, first subtract



statewide residential usage from total area source usage,



then assign 60 percent of that balance to the commercial



institutional category and 40 percent to the industrial



category.  Multiplying by 60 percent reduces the precision



of the entry.  Assuming that 60 percent is precise within 10



percent gives an overall precision as follows:





      o      1      60




     PQ2 - (0.615)2 +  (0.167)2





     P 2 - 0.4061
      O




     P   * + 0.637
      o    —


3.2.5.2  Bituminous Coal - The value for commercial insti-



tutional consumption of bituminous coal is the difference



between the Bureau of Mines figure for statewide consumption



of bituminous coal  (obtained from retail dealers) and the



residential consumption of bituminous coal, apportioned to



counties on the basis of population.



     The Bureau of Mines figure represents almost complete



coverage of coal producers.  The figure should be precise



within the nearest thousand tons as reported.  The error



would be + 500 tons.  Based on a median value of 193 thou-



sand tons for retail dealers distribution of bituminous coal



in the state, the precision equals 193,000 or + 0.003.  The



precision of the population figures is zero  (Section 3.2.1).



The precision of the value for residential usage of bitu-



minous coal has also been determined; assuming no anthracite
                                3-28

-------
ooal consumption, it is + 0.196.  The overall precision is

dependent on precision of the percentage of coal consumed by

residential sources.  Assuming that residential consumption

represents 60 percent of the retail dealers figure gives the

following precision:


               T - D - R

     where     T = commercial institutional consumption
                   of bituminous coal

               D = retail dealers figure from Bureau of
                   Mines

               R * residential consumption of bituminous
                   coal

               Var  (T) _ Var (D) + Var  (R)

                 T2         (D - R)2

               but R - 0.6D

               Var  (D) = (0.002D)2

               Var  (R) = (0.196)2 » (0.1176D)2

               Var  (T) m (0.002D)2 +  (0.1176D)2

                 T2              (0.4D)2

               Var  (T) m 0.01383376D2

                 T2          0.16D2

               precision - q-(T) = + 0.294
                            X      ~~

3.2.5.3  Distillate Oil - Commercial and institutional

distillate oil consumption is the difference between resi-

dential consumption and the sum of the figures given for

"distillate-type heating oils", "kerosene used for heating",

and "distillate used by the military" from Tables 5, 6, and

12 of Reference 12.  The difference is then apportioned to

county on the basis of population.
                                 3-29

-------
     The Bureau of Mines totals represent essentially complete
coverage of oil producers.  The figures should be precise
within the nearest thousand barrels for each category.  The
median values for each category are 532,000 barrels  "kero-
sene used for heating"; 3,339,000 barrels - "distillate-type
heating oils"; and 96,000 barrels - "distillate used by the
military."  Assuming the maximum error equal to the standard
deviation in each case, the precision is calculated.
            T - K + D + M.
                         d
      where T = total use of distillate for heating
            K m kerosene used for heating
            D « distillate-type heating oils
            H, - distillate used by the military
             d
      a (K) « + 500 barrels
      a (D) » + 500 barrels
      a (M) - + 500 barrels
     Var (T)  m Var  (K) + Var (D) + Var  (Md)
        T2             (K + D + M)2
     Var (T)  m 3(500)2
        T2       (T2)
     Var (T) = 750,000
     a (T)  - 866
                          866
a (T) _ .,	    _
  T   ~ (3,339,000 +96,000 + 532,000}
precision * -2J£*- * ± 0.0002
                                 3-30

-------
For most cases  the totals  are  essentially precise figures.
Imprecision  is  introduced  in determining the  difference
between the  totals and  residential  consumption.   Assuming
that residential  sources account  for  75 percent  of the
distillate used for heating, the  precision  is calculated as
follows:
     C - I » T  -  R
     C - I = commercial institutional area  source con-
             sumption of distillate oil
         T « distillate oil -  total area source  consumption
         R « residential consumption of distillate oil
     Var (C  - I)   = Var (T) +  Var (R)
       (C -  I)2          (T - R)2
     a (R) * 0.104R (Section 3.2.4.3)
     R « 0.75T
     Var (C  - I)   = (0.0002T)2 +  (0.08T)2
      (C - I)2           (0.25T)2
     Var  (C - I)  _  0.00000004 +  0.0061
        (C - I)2
     Var  (C - I)
       (C - I)2
     cr  (C - I)
  (C - I)"2"  ~        0.0625
	~—  -  0.097
 (C - IP
      -TC - I)   - °-312

Apportionment to county on the basis of population  adds  no
imprecision, therefore
          precision = + 0.312
3.2.5.4  Residual Oil - The value for commercial and  insti-
tutional usage of residual oil is the difference between the
amount of residential consumption and the sum of the  figures
                                 3-31

-------
given for "residual-type heating oils* and "residual used by



the military" from Tables 7 and 12 of Reference 12.  The



difference is then apportioned to  counties on the basis of



population.



     As mentioned earlier, the Bureau of Mines figures



should be precise within + 500 barrels for each category.



The median values for each category are 561,000 barrels - "re-



sidual-type heating oils" and 103,000 barrels - "residual



used by the military."  Assuming that the maximum  error is



equal to the standard deviation in each case, the  precision



is calculated:



               T - R + Mj^



     where     T = total residual used for heating



               R « "residual-type heating oils"



               M_= "residual used by the military"



     Var  (T)    Var R + Var M,,
     	2	~2  K

       T           (R + Mj^)






     Var  (T)   -   2(500)2

         2             2
        *F             T




     Var  (T) - 500,000



     o  (T) = 707



     q  (T)   	707	

       T      (561,000 + 103,000)





            •  .0011
                                 3-32

-------
Again calculation  of  the  residential  consumption  of residual
oil introduces  imprecision into  the totals  for  commercial
institutional consumption.  Assuming  that residential
sources account for 50  percent of  the residual  used for
heating, the precision  is calculated  as  follows:

     C - I - T - AR
     C - I - commercial institutional consumption of
               residual oil by area sources
         T « residual oil - total area source consumption
         A£ = residential consumption of residual oil
     Var (C - I)  _  Var  (T) + Var (AR) _
       (C - I)2           (T -

     O (Aj^) » 0.144AR   (Section  3.2.4.4}
     AH » 0.5T
     Var (C - I) = (0.0011T)2 +  (0.072T)2
      (C - I)2            (0.5T)2
     Var (C - I) _ O.OOOOQX21 +  Q.QQ5184
      (C - ir              57Z3

     Var V V*  = 0.02074
      (c - ir
                 0.144
     precision •» _+ 0.144
3.2.5.5  Natural Gas - Commercial institutional usage of
natural gas is the sum of values given for "commercial" and
                                3-33

-------
•other users" in Table 6 of Reference 13.  The total is
apportioned to counties on the basis of population.
     The Bureau of Mines figures should be precise within +
5 percent.  Apportioning to counties on the basis of popu-
lation should add no imprecision.  Therefore, the overall
precision is + 0.05.
3.2.5.6  Wood - A figure for commercial institutional con-
sumption of wood can be entered only if adequate local data
from fuel dealers or other reliable sources are available.
Quantities of wood are usually reported in units of cords.
Since cords measure volume rather than weight, any vari-
ability in density of the wood would affect the precision of
the weight reported.  Reference 22 reports that densities of
wood vary by more than 20 pounds per cubic foot.  Assuming
that the wood in a given location differs by only 10 pounds
from an average density of 40 pounds per cubic foot gives a
precision equal to -rx or +_ 0.25.  The quantity of wood
reported also entails imprecision.  Fuel dealers should give
figures precise within 10 percent.  The overall precision of
the value for commercial institutional consumption of wood
is then:
     P 2 *  (0.25)2 + (0.10)2
      o
     PQ2 - 0.0725
     P   » + 0.269
      o    —
 3.2.6   Industrial Fuel
     For  most  categories, usage of  industrial  fuel by  area
 sources is calculated  by subtracting state totals  for  point
                                 3-34

-------
source consumption from Bureau of Mines data and apportion-
ing the balance to counties according to a ratio of the
number of manufacturing employees in the county to the total
number of manufacturing employees in the state.
3.2.6.1  Anthracite Coal - The total consumption of an-
thracite coal by industrial users per state is determined in
the same manner as consumption by commercial institutional
users except that the factor for industrial usage is 40
percent rather than 60.  The total is then apportioned to
counties on the basis of the total number of manufacturing
employees per county.
     Industrial area source consumption of anthracite per
state - Precision of this figure is calculated in the same
way as the value for commercial institutional usage with the
exception of the 40 percent factor:
     P2 = (0.615)2 + (Jg)2
     P2 - 0.4407
     P  « + 0.664
     Number of manufacturing employees - This figure is
obtained for county and state from Bureau of Census data
(Reference 15).  This census encompassed all establishments
employing one or more persons during the census year (1967)
and thus includes all sources considered as industrial in
the scope of the NEDS system.  Since an industrial firm can
accurately report the number of employees, and since all
firms were contacted, the census figure should be accurate.
                                 3-35

-------
But the Census Bureau reports employees to the nearest
hundred, leading to a maximum error in the county total of +
50 employees.  If we assume that the error is random, the
state total will also have a maximum error of + 50 employees.
Because of the variability of numbers of employees in
different counties and states , we assume that the precision
of median values will apply to all cases.  The median value
for a state is 175,000 employees.  The precision of this
value is    50   or + 0.0003.  The median value for a county
         175,000                                     5Q
is 2150 employees.  The precision of this value is
or £ 0.023.  Since the factor for number of employees is
county totals divided by state totals, the precision of the
factor is:


     P2 -  (0.0003)2 +  (0.023)2

     P  m + 0.023
     Overall precision - The overall-precision figure for
industrial usage of anthracite coal is the product of
precisions of the factors just calculated for industrial
area source consumption and number of manufacturing em-
ployees.  The precision is then:
     P 2 -  (0.664)2 +  (0.023)2
      o
     P 2 * 0.4414
      o
     P   * + 0.664
      o    —
                                 3-36

-------
3.2.6.2  Bituminous Coal - First the total statewide industrial

consumption of bituminous coal is determined by consulting

the Minerals Yearbook   Table 40, pg. 357, under the category

"All Others."  To determine industrial area source consumption,

use the value for statewide industrial point source usage,

taking care that coal used for processes (i.e. coke oven

charge) is not included.  Subtract that value from the total

statewide industrial coal combustion to determine the amount

of bituminous coal burned by industrial area sources over

the state.  This figure is then apportioned to counties by

assuming that industrial fuel usage in a county is pro-

portional to the total number of manufacturing employees in

that county.  The number of manufacturing employees  for

state and county is obtained from the 1967 Census of Manu-

facturers , "Area Statistics."

     In summary, industrial area source combustion of

bituminous coal for a county is determined as follows:

                                   EC

           D • C •     B • C •    B • C •   E
     where:
          Ag c  = industrial area source combustion of bitumi-
                  nous coal per county

          IB P  = state total combustion of  bituminous coal by
            "  *    industrial users

          PB c  = point source combustion of bituminous coal
                  for the state

          E  • number of manufacturing employees  in the county

          Efl « number of manufacturing employees  in the state
                                 3-37

-------
     Industrial source consumption of bituminous coal by



state   - This figure is based on data collected quarterly



by the Bureau of Mines from coal handlers who normally sell



or produce 100,000 tons or more of coal annually.  Reports



from these coal handlers are said to cover 94 percent of all



coal shipped or produced.  The balance is estimated by



statistical methods, and the data should be reliable, at



least within the limits of the units reported in the table



(1,000 tons).    For any one state, the maximum error in



I_ _  would be + 500 tons.
 B.C.          —


     Point source combustion of bituminous coal by state -



This factor is the summation of the coal usage reported by



all industrial point sources.  Reports of the amount of coal



burned are usually in round figures.  Since a source usually



has accurate figures for the coal burned,  precision depends



on the number of significant figures reported, which is



usually two or three.  In such cases the maximum error would



be:



     + 5%   for two significant figures



     + 0.5% for three significant figures



For consistency with the units of I_ _ , consider that the
                                   D • \* •


point source usage is reported to the nearest 1,000 tons.



Assuming random error distribution, the maximum error is +^



500 tons.



     Number of manufacturing employees per county and



state   - The referenced census encompassed all establish-



ments employing one or more persons during the census year
                                3-38

-------
(1967) and this includes all sources considered as industrial
in the scope of the NEDS system.  Since an industrial firm
can accurately report the number of employees and since all
firms were contacted, the census figure should be accurate.
But the Census Bureau reports employees to the nearest
hundred, leading to a maximum error in E  of + 50 employees.
If we assume that the error is random, the state total will also
have a maximum error of + 50 employees.
     Overall precision - To determine the overall precision
for area source industrial consumption of bituminous coal
assume the maximum error in a factor equals the standard
deviation of that factor.  For:
          *B.C.
                         - P
                            B.C .
  Var
but  Var (I,
therefore: Var
                      Vard
                               -Pn ,, )  .  Var EC ,  Var Es
                                  P . L .   +   A    +   A
                          r - p    r     E/      E/
                        D.U.   O.C.        C        S
                                        Var
               *B.C.
                                  (IB.C.-PB.C.>
since
  SJ^.c.l- =
    *B.C.
               Var (I    )  + Var (P,, r )
                     a. I*.        &  t> • t- . —
                                3-39

-------
            °(AB C}
Calculating -j—^-  for several cases according to Appendix
             *B.C.
C, we obtain the following general equation to define a

precision of the area source entry for industrial bituminous

coal consumption.


     precision = q(AB.C.)  =1.26 x 102 (A^ „ )~°'97
                 —sr	                B.C.
                  A'B.C.

3.2.6.3  Coke - Consumption of coke by area sources can be

determined only when local information is available.  These

data usually come directly from the companies using the

coke, who should have reliable figures for annual consump-

tion.  Estimating the figures to be precise at least to within

+ 15 percent, the precision is + 0.15.

3.2.6.4  Distillate Oil - The industrial area source use of

distillate oil statewide is the sum of figures given for

"industrial" and "oil companies" in Tables 8 and 9 of

Reference 12 (Bureau of Mines) minus the amount consumed by

industrial point sources.  State totals are then apportioned

to counties on the number of manufacturing employees.

     The Bureau of Mines totals should be precise within the

limits of the nearest thousand barrels for each category.

The median values for each category are 662,000 barrels for

"industrial" and 74,000 barrels for "oil companies".  Assuming

that the precision of the median cases applies to all cases

gives:

               T = I + O

     where:    T - total consumption of distillate oil by
                   industrial users per state
                                3-40

-------
                I « industrial consumption of distillate
                    oil
                0 » consumption of distillate oil by oil
                    companies
      Var (T)  = Var (I)  + Var (0)
       T^           (I + O)2
      Var (I)  - (500)2
      Var (0)  = (500)2
      I + 0 »  662,000  + 74,000  = 736,000
                     SxlQ5
         T         5.41696  X 10  A
      q  (T)   -    9.6  x 10~4
        T
      precision   = +  0.00096
      Consumption of  distillate  oil by industrial point
 sources  is  reported  to  the  nearest thousand gallons.   Since
 most  companies  keep  records of  purchases  of fuel oil,  the
 figures  should  be precise within  500  gallons.   Assuming that
 the point source consumption is around 60 percent of  the
 total distillate oil consumption,  precision may be deter-
 mined for the median case.
      p  _      500
       p   0.6(736,000)42
      P  =*-.+; 2.7  x 10~5

     The precision of values derived by apportioning  to
counties on the basis of manufacturing employees was deter-
mined in Section  3.2.6.
                                 3-41

-------
PM - 0.023

The overall precision is determined as follows:
where A_ = industrial area source consumption of
       "   distillate oil

     T  = total consumption of distillate oil by
          industrial users

     P  = consumption of distillate oil by point
          sources

     E  • number of manufacturing employees per
      c   county

     E  « number of manufacturing employees per
      8   state

Var  (A  ) _ Var  (T-) + Var P     Var  (E.)  . Var  (£„)
     =— u— -       u      5   u— -r      5"^""        3~~^*~
  V          ^o-V           V        Es
Var
      cl + Yar(Esi ,  (0>
  Ec
using median values


      D
Var  (T) =  (+  5  x  105)422  gallons2
Var  (P  )  =  (+  500  gallons)2
Var  (A  )  _  8.8  x  108  x  2.5  x  105   ?
  7~^~      [(736000)42  -  18547200]^
  AD


              8.8025 x 108
            1.5288827  x  10
            5.3476  x  10'4
  AD
 precision -  -~D- -  +0.023
                                3-42

-------
 3.2.6.5  Residual Oil - The area source use of residual oil
 statewide is the sum of figures given for "industrial" and
 "oil companies" in Tables 8 and 9 of Reference 12,  minus the
 amount consumed by industrial point sources.   State totals
 are then apportioned to counties on the basis  of the number
 of manufacturing employees.
      Precisions are the same as for distillate oil  cal-
 culations,  but the median values change.   On the basis of
 950,000 barrels for "industrial" and 178,000 barrels for
 "oil companies",  the precision of the state totals  is + 4 x
 10  .   Precision of the point source totals is + 1.3 x 10~5.
 The overall precision is then:
          [Var  (TR) + Var (PR)
             
                             +  (0.023)
 1/2
where T_ » total industrial consumption of residual
       R     oil
      PR = point source consumption of residual oil
using median values
                    8
Var  (TR) = 8.82 x 10
Var  (PR) = 2.5 x 103

(TR - PR)2 =  (1.89504 x 107)2
(TR - PR)2 = 3.6 x 1014
      8.82 x 108 x 2.5 x 105
           3.6 x 10
                   14
                             + 5.29 x 10
-4
po * - 0.023
                                3-43

-------
3.2.6.6  Natural Gas - To determine total area source con-

sumption of natural gas by industrial consumers, the total

statewide industrial consumption of natural gas is deter-

mined from Table 5, p. 734 of the Minerals Yearbook under

the category "Industrial".  Next we consult the point source

inventory to obtain industrial point source consumption for

the state.  The difference between total statewide industrial

usage and statewide industrial point source usage is the

statewide industrial area source usage of natural gas.  This

figure is apportioned to counties by assuming that county

industrial fuel usage is proportional to the number of

manufacturing employees per county.  The number of manu-

facturing employees for state and county is obtained from

the 1967 Census of Manufacturers "Area Statistics."

     In summary, the determination of industrial usage of

natural gas attributable to area sources is as follows:
     where
             *N.G. " PN.G.  _s
A,, G  = industrial area source combustion
        of natural gas per county
               I  _  = total statewide usage of natural
                 " "   gas by industrial firms

               P  _  = statewide point source combustion
                N*t"   of natural gas

               E     = number of manufacturing employees
                c      in the county

               E     = number of manufacturing employees
                8      in the state

     Statewide usage of natural gas by industrial firms -

These data are based on annual surveys of the companies that
                                3-44

-------
distribute natural gas; these surveys account for  80 percent



of the natural gas produced.  Most of the remaining natural  gas



usage is covered in the purchase listings of the reporting



companies.  Assuming that the figure reported covers 95



percent of the natural gas consumed, the precision is +



0.05.



     Statewide point source combustion of natural  gas by



industrial users - This value is the summation of  the



natural gas usage reported by all the industrial point



sources in the state.  It is usually reported in round



numbers in units of 10  cubic feet.  Since the amount of gas



used is a readily available, accurate figure, precision will



depend on the number of significant figures reported.  A



value reported to the nearest million cubic feet gives



          a  (PN _ ) = + 0.5 x 106 ft3
              N.G.    —


     Number of manufacturing employees per county  and per



state - The precision for E /E , determined earlier, is +
—————                      c  s                         —


0.023.



     D.  Overall Precision - For
     (I - P) (fc)
Var (A)
             Var  (I) + Var  (P)  . Var
                  (I - P)
                                   (Ec)'
 but Var (I - P)  - Var (I)  + Var (P)
     Var (A)
                                              Var
Var (I) + Var  (P)
Var  (Ec) + Var_JE )_


  (EJ2       (E.)2"
                       (I - P)'
                         (0.023)
                            {Es)
                                     Var  (EC)_
                                                  Var  (EC)
                                 3-45

-------
Assume that the precision for the median value of  T-, „
                                                   N . tj.


applies to all cases.  The median value of I.. _  from the
                                            N .(j.


Minerals Yearbook is 86,682 x 10  cubic feet.
      a  (IN-G>) = 0.05
                     <°-05
      Var (IN.G.>median
Assume that point sources account for  60 percent  of  the



natural gas consumed by industrial users.






      PN.G.  median = °'6 <86682 * Io6>



        (PN>G>)  '- 0.5 x 106



      Var (PN r  )  = 2.5 x 1011
            N .U .


                    1.878 x 10» + 25 x 1011 + (0>Q23)2


                       (3.46728 X 1011V
                    1.56 x 10"2 + 5.29 x 10~4
      precision = —^"^.G.^	 = jt 0.127






3.2.6.7  Wood - Usage of wood by industrial sources not



classified as point sources must be estimated from data



supplied by these sources.  Although a company may provide  a



good estimation of the feed rate to their boiler, error may



be introduced if the density of the wood is not constant.



Accounting for this type of error, assume that the esti-
                                 3-46

-------
nation is precise within 25 percent and precision is + 0.25.
3.2.6.8  Process Gas - Use of process gas by industrial area
sources can be determined only from local information.  The
most common users of process gas are petroleum refining
operations and natural gas production plants.  They usually
cannot provide a measurement in cubic feet but can supply
their heater or boiler capacity.  Because of the varying
heat content of process gas, the volume of gas required to
fire the heater or boiler to capacity also varies.  At best
the number provided by the user is an estimate, which should
be precise within 20 percent.
          P - + 0.20
3.2.7  On-Site Incineration
     On-site incineration refers to the disposal of solid waste
 in small incinerators, such as backyard burners, industrial incin-
erators, and incinerators at food and department stores,
hospitals, and schools.  Per-capita generation rates are
derived from the 1968 National Survey of Community Solid
Waste Practices  '   which provides a good indication of the
methods of solid waste disposal but a poor indication of the
quantities involved.  Analysis of the data gives a precision
for each category in the range of + 10.00.  Assume that this
precision applies to all categories of users, i.e., in-
dustrial, residential, and commercial institutional.
3.2.8  Open Burning
     On-site burning refers to the unconfined burning of
waste such as leaves, landscape refuse, or other rubbish.
                                 3-47

-------
Factors for per-capita generation rate were also derived
from the 1968 National Survey of Community Solid Waste
Practices  '  .  Again, determination of the quantity of
waste based on this survey is very imprecise.  Based on the
data examined for on-site incineration, a precision of +
10.00 is estimated for residential, commercial institutional,
and industrial sources of open burning.
3.2.9  Measured Vehicle Miles
     Entries are made in these fields of the NEDS form only
if a transportation study is available giving the measured
vehicle miles traveled.  The vehicle miles are categorized
by roadway types based on average speeds.  Since the data
for all roadway types are from the same source, the data for
each type should have the same precision.  The Bureau of the
Census reports that data gathered in the Census of Trans-
portation has a standard deviation of 3.2 percent.
     Assume that all transportation studies are as precise
as the Census of Transportation.  On the basis of Bureau of the
Census data, precisions of values for measured vehicle miles
on all types of roadways  (limited access, rural, suburban,
urban) is + 3.2 percent.
3.2.10  Gasoline Fuel
     Three methods are used in calculating consumption
of gasoline by on-highway vehicles.  Since two of the
three methods are interrelated, consider only two methods
for calculation of precisions.  Method 1 is applied
only when there is available a transportation study that
                                 3-48

-------
provides measured vehicle miles.  The study should classify



mileage by light- and heavy-duty gasoline vehicles and by



heavy-duty diesel vehicles.  Method 2 involves obtaining the


                                                      18
total statewide gasoline sales from Highway Statistics   and



apportioning values to counties on the basis of population.



     Off-highway gasoline usage is calculated by applying a



factor for annual consumption by farm tractors along with a



population-based factor to include other off-highway sources.



3.2.10.1  Light-Duty Vehicle - In Method 1 the value for



total light-duty vehicle miles is divided by 13.6 miles per


      19
gallon   to obtain total volume of gasoline consumed by



light-duty vehicles.  Both factors are obtained from trans-



portation studies and each should have a precision of +



0.032 (see Section 3.2.9).  The overall precision is then:



       222

     Po  - Pl  + P2





     P 2 - (0.032)2 + (0.032)2
      o




     P 2 = 0.002048
      o




     P  - -I- 0.0453
      o   —


     Method 2 involves obtaining the statewide sales of



gasoline for highway use, apportioning the gasoline to



counties on the basis of the ratio of county service station



sales to state service station sales, calculating total



gasoline vehicle miles traveled (VMT) by assuming an average



12.2 miles per gallon, assuming that light-duty vehicles
                                 3-49

-------
account for 89 percent of the total VMT, and assuming that



light-duty vehicles obtain 13.6 miles per gallon to obtain



gallons consumed by light-duty vehicles.  The precision of



each factor employed must be calculated to obtain the over-



all precision.



     Published gasoline sales totals by state - This figure


                                    18
is obtained from Highway Statistics.    it is compiled from



reports by state tax departments and should be a precise



figure since tax records are precise.



     County service station sales and state service station



sales - These figures are obtained from the 1968 Census of



Business.    The Bureau of Census reports that 80 percent of



all stations were canvassed by mail and data for the remain-



ing were obtained by consulting records of the IRS and



Social Security Administration.  Assuming that the records



are accurate, there should be no imprecision other than that



introduced by rounding off to the nearest $1,000.  Con-



sidering the magnitude of sales at service stations, an



error of + $1,000 is insignificant.  Therefore consider the



sales totals as precise numbers.



     12.2 miles per gallon average for all vehicles - This


                                         18
figure, obtained from Highway Statistics,   represents the



total vehicle miles in the nation divided by the total



gasoline consumed.  The total VMT is precise within 4-3.2



percent, and the value for gasoline consumed is a precise



number.  Therefore, the result is also precise within +



0.032.
                                 3-50

-------
     89 percent of total vehicle miles are contributed by
light-duty vehicles -  This  figure is obtained from Highway
Statistics18 by finding the ratio of light-duty vehicle
miles to heavy-duty vehicle miles.  Since each value has a
precision of 3.2 percent, the overall precision is deter- .
mined as follows:
           Po
-~\/2T.032V
            ^ * + 0.045
            o   —
     13.6 miles per gallon average for light-duty vehicles -
This value, also from Highway Statistics,   is obtained by
dividing the total light-duty VMT by the gasoline consumed
by light-duty vehicles.  Since the same precisions are
involved as before, the precision of this value is also +
0.032.
     Overall precision - Multiplying precisions of the five
factors just calculated gives the overall precision of the
value for off-highway gasoline consumption as obtained by
Method 2:
     PQ =  [(O)2 +  (0.032)2 +  (0.032)2 +  (0.045)2 + (0.032)2]1/2
     P  =  [0.005097]1/2
      O
     po * - °'071
                                3-51

-------
3.2.10.2  Heavy Vehicle - The methods for determining
gasoline consumption by heavy vehicles are exactly the same
as those for light vehicles.  The values of the factors
change slightly, but the sources are the same and the
precisions are the same.  For Method 1, with values obtained
from a transportation study, the precision is + 0.045; for
Method 2, which involves apportioning, the precision is +
0.071
3.2.10.3  Off-Highway - The total off-highway consumption of
gasoline is the sum of that used by farm tractors and the
amount consumed by other miscellaneous off-highway sources.
     Gasoline usage by farm tractors - The total gasoline
used by farm tractors is obtained by finding the total
                                                                 23
number of farm tractors in a county in the Census of Agriculture,
assuming that 60 percent are gasoline powered, and applying
a factor of 1,000 gallons per tractor per year.  Assuming
that all these factors are derived from Department of
Agriculture figures, and that only counties with 100 or more
                                                           23
farms are significant, Table C of the Census of Agriculture
may be used to estimate precisions.  Using the Table's
levels of reliability for fuels and machinery, the precisions
for both the fuel consumption factor and the percentage of
tractors powered by gasoline are +0.10 and the precision of
the number of tractors is + 0.042.  The overall precision is
then:
           Po
           PQ =~U2(0.10)2+ (0.042)2

           PQ -+ 0.148
                                 3-52

-------
     Miscellaneous off-highway gasoline consumption - This
figure is obtained by using a consumption factor of 13
gallons per capita-year and multiplying by the county
population.  APTD-1137  gives no reference for the factor of
13 gallons per capita-year; therefore we assume that the
factor is significant only to the nearest ten.  This would
give a standard error of + 5 or a precision of + 0.385.
Since the county population is a precise figure, the pre-
cision is + 0.385.
     Overall precision - The total off-highway gasoline
consumption is the sum of farm tractor usage and miscel-
laneous usage.  Therefore the overall precision is as
follows:

           T -  F + M
where      T =  total  off-highway  gasoline  consumption
           F =  gasoline  consumed  by  farm tractors
          M =  gasoline  consumed  by  miscellaneous  off-highway
               sources

          Var  (T) _  Var (F + M)
            T2      (F  + M)2

          Var  (T) _  Var F + Var  M
            T2        (F + M)2

          a(F) = 0.148  F
          a(M) = 0.385 M
          Var  (F) =  0.022 F2
          Var  (M) =  0.148 M2

          Var  (T) m  0.022 F2 + 0.148 M2
            T2        F2 + 2 FM + M2
                                 3-53

-------
but       M - 13 x population  (P)
and       F = T - M
therefore Var,(T) = 0.022 (T - 13P)2 + 0.148 (13P)2	
            T^      (T - 13P)J + 2 (13P)(T - 13P) + (13P)
          Var  (T) _ 0.022T2 - 0.572 TP + 28.730P2
          "?                 T2
          g(T)   =   [Var  (T)]1/2
           T        T   ^2   F
          precision
                        g  (T)  =   L0.022T2 = 0.572TP + 28.73P2J  1/2
3.2.11  Diesel Fuel
     This category includes on-highway  heavy vehicle  use,
off-highway use by miscellaneous  sources,  and  rail loco-
motive use.   The various methods  used for  calculating each
case are explained in the  following sections.
3.2.11.1  Heavy Vehicle -  Methods of calculation are  the
same as those for calculating gasoline  consumption.  The
first method  involves dividing the measured vehicle miles
traveled by diesel-powered vehicles by  a factor of 5.1 miles
per gallon.   Both  figures  are obtained  from transportation
 studies  and should have the precision indicated earlier,
 0.032.   The overall precision is then:


      P 2 = (0.032)2 +  (0.032)2
       o

      po  ~ ± °'045
                                 3-54

-------
     The second method involves apportionment of statewide
diesel fuel sales to counties on the basis of service
station sales.
     Statewide diesel fuel sales - This figure is given as
                                                        18
"on-highway use of special fuels" in Highway Statistics.
It is obtained from data reported by state tax agencies and
should be a precise figure.
     County service station sales and state service station
sales - These figures for sales of diesel fuel are from 1968
Census of Business and are precise numbers.
     Overall precision - The factors used in calculating
diesel fuel use are precise, but the coding form requires
rounding off to the nearest thousand gallons.  The precision
is then:
     P = ± °-5
         Mo. in field
3.2.11.2  Off-highway - Off-highway diesel fuel consumption
is the sum of that used by farm tractors, construction
equipment, and other miscellaneous sources.
     Diesel fuel usage by farm tractors - The method de-
scribed for calculation of gasoline usage by tractors
applies to calculating diesel fuel usage except that 35
percent of the tractors are assumed to be diesel-powered.
The same precisions apply.
     P « + 0.148
     Diesel fuel usage by construction equipment - The value
for diesel fuel usage by construction equipment is the
                                3-55

-------
product of the number of non-building construction employees



in the county and a factor of 5,000 gallons per employee-



year.  The number of non-building construction employees,



reported in County Business Patterns   is obtained by con-



sulting data from the Social Security Administration and, if



needed, by a special multiunit survey.  It should be a



precise number.  The fuel consumption figure should be



precise within the limits of significant figures.  Precision



is then ±|^j or + 0.10.



     Diesel fuel usage by miscellaneous sources - This



figure is calculated by applying a factor of 7.4 gallons per



capita annually.  Since no source is listed for the factor,



assume that it is significant only to the nearest unit.



Since population is a precise figure, the precision equals



Y~j or + 0.07.



     Overall precision - It is difficult to determine a



precision for all cases without deriving an equation in



terms of factors that do not appear on the NEDS area source



coding form.  Therefore, assume that the precision applying



to a median case applies to all cases.  The median value for



diesel fuel consumed by farm tractors is 792,050 gallons.



The median value for diesel fuel consumed by construction



equipment is 970,000 gallons, based on 194 non-building



construction workers per county.  The median value for



diesel fuel consumption by other miscellaneous sources is



105,235 gallons gased on a county population of 14,221.
                                 3-56

-------
     precision median »



   (0.148 x 792,050)2 +  (0.1 x 970,OOP)2 +  (0.07 x 105,235)2 1//2
                  (79,250 + 970,000 + 105,235)2
       0.082
3.2.11.3  Rail Locomotive - Sales of fuel oil to railroads by



state are given in Reference 12, Table 10.  County fuel


totals are obtained by apportionment on the basis of miles



of track per county.


     The Bureau of Mines totals may be considered precise


within the reported units of 1,000 barrels.  Assuming that


the median value of 1,231,000 barrels applies to all cases



gives a precision of v ?3l 000  or — °"4 x ^


     Values for miles of track per county are obtained from



detailed state maps or may be supplied by the railroads.


Assuming that the figure obtained is precise within 10 percent


gives an overall precision as follows:



     PQ2 « (0.10)2 + (0.4 x 10"3)2



     P 2 » 0.010 + 1.6 x 10~7
      o

     P   - + 0.10
      o    —


3.2.12  Aircraft


     This category includes all aircraft operations except


operations on dirt airstrips.  The required values for landing


and takeoff (LTO) cycles are obtained from three Federal Aviation


Administration (FAA) publications,  ~   which report the number of


"operations" performed by civil, commercial, and military aircraft


at all FAA-regulated airports and military airfields.  The number
                                 3-57

-------
of "operations" is divided by 2 to obtain the number of LTO
cycles.
3.2.12.1  Military Aircraft - The exact number of operations
performed by military aircraft is reported in Military Air
                       26                             7 ^
Traffic Activity Report   and FAA Air Traffic Activity  .
Therefore, the LTO cycles may be determined precisely.  The
units of the NEDS form field limit precision, in that the
figure must be reported to the nearest 100.  Precision is
then equal to
               No. LTO cycles
3.2.12.2  Civil Aircraft - Civil aircraft operations are
also enumerated in FAA Air Traffic Activity   and Military
Air Traffic Activity Report.    Since these reports do not
include operations at non-FAA-regulated airfields, figures
for these operations must be obtained from the appropriate
airfield officials, who should provide an exact figure.
Therefore, precision of the number of LTO cycles is limited
only by rounding to the nearest 10 as required on the NEDS
coding form.  The precision will be _ — _ .
                                    No. of LTO cycles
3.2.12.3  Commercial aircraft - The number of LTO cycles  for
commercial aircraft is detemrined in exactly the same manner
as that for civil aircraft  and has the same precision, which  is
No. ofLTO cycles
3.2.13  Vessels
     This category, which includes vessels and small craft, is
sub-categorized according to type of fuel.
3.2.13.1  Anthracite Coal - No reliable method for estimating
emissions from these vessels on a local basis is currently
known, because only nationwide fuel totals are available.
                                 3-58

-------
Thus  information on  local  fuel consumption can be obtained
only  from estimates  made by port authorities or ship operators.

     Anthracite coal usage should be assigned to counties
with major ports, according to tonnage handled in the ports.
These data are available from Waterborne Commerce of the
              28
United States,   published by the U.S. Army Corps of Engineers.
     The nationwide  totals are reported to the nearest
thousand tons.  The number should be precise within this
limitation; for 1969 the precision is *~ ^nn or 0.03.
                                      X / , UUv
     Waterborne Commerce of the United States reports the
tonnage handled in the ports to the nearest ton.  Since all
ships are required by law to report their tonnage, this
should be a precise  figure.  The overall precision of the
value for anthracite coal consumption by vessels is then +
0.03.
3.2.13.2  Diesel Oil - Detailed apportioning methods
for determining fuel oil consumption by vessels are given in
APTD-1135.   Assume that all data from Waterborne Commerce
of the United States are precise.
     In-port fuel consumption - The number of vessels
entering a port, obtained from Reference 28, is a precise
number.  Assume that each vessel spends 3 days in port, and
estimate that this figure is precise within 1/2 day or —j^
« 0.167.  The vessel-days in port must be distributed
between ships using distillate (diesel) fuel and ships using
residual fuel.  This is accomplished by using the ratio of
distillate oil vessel-days to residual oil vessel-days.
First, obtain the number of barrels of each fuel consumed in
                                 3-59

-------
the state from Reference 12.  Convert to vessel-days by
assuming that each vessel consumes 660 gal/day of diesel fuel
and 1900 gal/day of residual oil.  Assuming that these
figures are precise within their significant figures gives
precision of the diesel fuel factor equal to  -F^Q or + 0.008
and precision of the residual fuel factor equal to , QOQ-
or + 0.026.  Using median values for total statewide fuel
oil consumption by vessels, with a standard deviation of +
21,000 gallons gives the following precision:
     distillate oil - 8,484,000 gallons
     residual oil - 7,854,000 gallons
     distillate-oil vessel-days = 8>4^'000 = 12,855
precision =
/   21,000
(8,484,000
                                     660
                                  (0.008)
               - 4- 0.009
 residual-oil vessel-days =
                                            = 4,134
          precision =
                          21,000
                        7,854,000
                      + 0.061
                                 (0.026)
                                          1/2
      total vessel-days = 12,855 + 4,134
                        » 16,989
         precision =
                      (0.009 x 12,855)2 + (0.061 x 4,134)2
                                   (16,989)
                                           2
                                                            1/2
                     + 0.02
      percent diesel-oil days = ,g'Qor.
                                io,
          precision =
                  (0.009r + (0.02)
                 + 0.022
                                           1/2
                                  3-60

-------
      total  diesel-oil  days = percent diesel days
      x number  of vessels x 3 days/vessel

          precisio
          tion =  (0.
022)2 + o2 + (0.167)2
                                           1/2
                      + 0.009
      residual-oil vessel-days =  7'j?5f;°°°  =  4,134
                                   j. f y uu

                                                 1/2
         precision
                          21,000
                        7,854,000

                        0.061
                              +  (0.026)
     total vessel-days  =  12,855  +  4,134

                        =  16,989
    precision =


              . ± 0.02

percent diesel-oil days =
                        (0.009 x 12,855)2 +  (0.061  x  4,134)

                                    (16,989)2
                                                              1/2
         precision =   (0.009)2 +  (0.02)2

                   =  + 0.022
                                          1/2
     total diesel-oil days = percent diesel-oil days x
     number of vessels x 3 days/vessel
          precisio
             n =  (0.
 022)2  +  O2  +  (0.167)2'1/2
                        0.17
     diesel oil consumption = total diesel-oil days x  660
     gallons per day

          precision -  |(0.17)2 +  (0.008)2 1/2|

                    - + 0.17

     Underway fuel consumption - Underway diesel fuel  con-

sumption is assumed to be the difference between the total

distillate oil used by vessels (Reference 12), and the total

distillate fuel consumed in-port.  Using median values,  the
                                 3-61

-------
     precision
in-port consumption of distillate is 4,437,668 gallons out

of a total of 8,484,000 gallons.  Precision of underway

diesel fuel consumption is then:

                  ).17x 4,437,668)2 + (21,OOP)2  1/2

                     (8,484,000 - 4,437,668)2

     precision = + 0.187

     Overall precision - Precision of the value for total

diesel fuel usage by vessels equals precisions of the in-

port consumption value plus precision of the underway con-

sumption value.  Based on.median values:
     precision =
                  (0.17 X4,437,668)    +  (0.187 x 4,046,332)
                                                                1/2
                                  8,484,000'

     precision - + 0.13

3.2.13.3  Residual oil - In-port residual oil consumption is calculated

in the same manner as for diesel fuel.  From the preceding section,

precision of residual-oil vessel-days equals + 0.061, precision of

total vessel-days equals + 0.02, and precision of the residual fuel

factor equals + 0.026.
     precision =  (0.061)2 + (0.02)2
                                       1/2
                 + 0.064
     total residual-oil days = percent residual-oil days x
     number of vessels x 3 days/vessel
     precision
I.064)2  + O2
                                   (0.167)
                                             1/2
                 + 0.018
     residual oil consumption = total residual-oil days x
     1900 gallons/day
     precision =
                   (0.18)* +  (0.026)

                 * 0.18
                                   2  1/2
                                 3-62

-------
3.2.13.4  Gasoline - This category includes craft operated
by inboard, outboard, or inboard-outboard motors on lakes
and rivers.  Gasoline consumption by vessels can be esti-
mated by using a factor of 160 gallons of gasoline per boat
registration per year.  This factor was derived from in-
formation on average annual operating hours, average horse-
power per vessel, and average fuel consumption per horse-
power-hour of operation.  The estimate also approximately
accounts for any non-motorized boats that may be included in
the state total of registrations and for boats with small
outboard motors that may not be registered.  Boat regis-
tration data should be obtained from the appropriate state
agency, such as the State Department of Recreation of the
Fish and Game Department.
     The total gasoline consumption predicted by use of the
state total of boat registrations should be apportioned to
counties on the basis of inland water surface area.  County
inland water areas may be determined from the Bureau of the
                                 29
Census1 Area Measurement Reports.
     The factor of 160 gallons per boat registration should
be a good estimate of gasoline consumption.  Assuming that it
is precise within 25 gallons for each county, the precision
is j|| or ± 0.156.
     The number of boat registrations is a precise number.
     The surface area of water in a state is reported in square
miles.  The Bureau of Census reports that the figures are accurate
                                3-63

-------
within 1 percent.  Based on median values for county and state,



precision for the state figure is 0.01, but the county figure



is less precise because of rounding off:  P equals si T-Q' or °-023



     Based on median values, the overall precision of a value



for gasoline consumption by vessels is:




     PQ -  (0.156)2 +  (0.01)2 +  (0.023)2|1/2



     P  = + 0.158
      o   —




3.2.14  Evaporation



     This category includes losses of hydrocarbons and other



organic compounds by evaporation at sources not considered as



point sources.  These  sources include gasoline-handling, dry



cleaning, surface coating, and miscellaneous solvent-use opera-



tions .



3.2.14.1  Solvent-Purchased Methods of Determination - Methods



for determining evaporation losses as a function of solvent



purchases entail  (1) a questionnaire survey of all solvent



users, or  (2) the assumption that solvent use is proportional



to population and is dependent on the population density.



     Questionnaire survey of all solvent users - A survey  is



the best and most accurate method of determining solvent



use.  Most companies maintain records  of purchase orders that



show solvent use.  Precision depends on the significant figures



reported on  the  questionnaires.  Although  this  figure  varies



among the  reporting  companies,  one  can assume  that  it  is



reported within  + 50 tons.   Precision  is then:
                                 3-64

-------
                        P  «  +     50
                              " No. in field


     Assume solvent use is proportional to population and

dependent on population density - This method is used for most

counties, since surveys require considerable time and effort.

     For dry cleaning operations use factors of 2.0 solvent per

person per year in moderate climates and 2.7 Ib per person per

year for colder climates.

     For other solvent losses apply values from the following

table:
 County  population
     Solvent loss,
Ib per capita per year
     <100,000

100,000 to  500,000

500,000 to  1,000,000

     >1,000,000
           3
           8

          18
          28
     Since APTD-1135 does not cite references from which these

factors for evaporation losses were obtained, the best possible

precision value is based on an engineering estimate.

     Losses from dry cleaning operations are based on population,

climate, and factors for dry cleaning loss per person.  Since

the population value is a precise figure, all imprecision arises

from the factors dependent on climate; that is, 2.0 and 2.7 Ib/

person for moderate and cold climates, respectively.  Assuming
                                 3-65

-------
that the factors are correct, the usage should cover  a  range

from  2.0 to 2.7, since climate does not change  abruptly  from

moderate to cold.  If errors are randomly distributed,  the  range

2.0 to 2.7 should contain 99 percent of the values.   Assuming

the range is 3 standard deviations, T = + 0.2333.  Related  to

a median value of 2.35 for the use factor, the precision  equals

+ 10 percent.

     Other solvent losses are also considered in this category.

Since the population figure is precise, any imprecision results

from errors in the solvent-use factors.  Estimating the error

in each factor as + 5 and using the median value for  each range,

precision for each range is determined:

        T «  total solvent  loss

        D =  dry  cleaning  loss

        S =  miscellaneous  solvent loss

        X =  population

        a(D) =  0.1  (2.35)(X)

        o(S) =  5 (X)

        Var  (D)  =  [cr (D)]2 = 0.055225 X2

        Var  (S)  =  [o (S)]2 - 25 X2
         Since  T  =  D  +  S

          »r(T)   =  Var(E

          T2          (D + S)1
Var(T)  =  Var(D) + Var(S)
                    ,2
        Var(T)  =  25.055225 X2
         a(T)  ="\/25.055225 X2

         o(T)  =  5.00552 X
                                 3-66

-------
Range
0 - 100,000
(X « 50,000)
100,000-500,000
(X - 300,000)
500,000-1,000,000
(X - 750,000)
>1, 000, 000
(X - 1,500,000)
T
5.35 X
10,35 X
20.35 X
30.35 X
»-2$a
+ 0.936
+ 0.484
+ 0.246
+ 0.165
3.2.14.2  Gasoline Handling - This category includes only evapora-

tive losses that occur as a result of service station operations.

Among three possible data collection methods, the most common

method involves totaling the entries for light vehicle, heavy

vehicle, and off-highway consumption of gasoline.  Precision then

depends on the precisions of these entries.  For most counties,

off-highway consumption is negligible in comparison with consump-

tion by light and heavy vehicles.  Assume then that precision

depends only on the precisions of values for light and heavy

vehicle consumption:

            T * L + H

     where  T = gasoline marketed at service stations

            L = light-vehicle gasoline usage

            H = heavy-vehicle gasoline usage
     Var (T)   = Var (L) + Var (H)

       T2           (L + H)2


     but T2 = (L + H)2
     therefore

         Var  (T)
            12
Var (L) + Var (H)
       IT
                                3-67

-------
        Var  (L)   -   lo(L)]2  - (0.071L)2 (per Section 3.2.10.1)


        Var  (H)   -   [o(H)]2  - (0.072H)2 (per Section 3.2.10.2)


        Var  (T)   =   0.0050L2 + 0.0052H2

           wi              n*




Assume that  light vehicles using  13.6 miles  per  gallon .account for


89 percent of the VMT and heavy vehicles using  8.4  miles per gallon


account for  11 percent; then


     H = 0.2L



     Var (T)  _  0.0050L2 + 0.0052(0.2L)2


       T2    "            T2



     Var (T)  = 0.00521L2






     g (T)  _ Q.072L

       T        T



     but  T  = 1.2L




     q (?)    = 0.060
       T




     precision = + 0.060






3.2.15  Fugitive Dust Sources


     This section deals with precision  of values reported for


emissions of fugitive dust from sources that include dirt roads,


dirt airstrips, construction sites, rock handling sites,  and


several miscellaneous sources.


3.2.15.1  Dirt Roads - The value  for vehicle miles  of travel


on dirt roads is obtained by using the  figure given by the State
                                   3-68

-------
Highway Department for total mileage of dirt roads in each county



and by estimating the amount of traffic.



     The State Highway Department usually can supply a figure



for the mileage of dirt roads that is precise within the nearest



mile.  Estimating at least 10 miles of dirt road per county



as a median value gives a precision of j^- or + 0.05.



     Estimating the number of vehicles that travel over dirt



roads introduces more imprecision.  If at least some data are



available on which to base an estimate, the figure for number



of vehicles should be precise within 50 percent.



     The overall precision is then:




     PQ2 - (0.5)2 + (0.05)2




     P   = + 0.502
      o    -




3.2.15.2  Dirt Airstrips - Airstrip operators who are contacted



for the required information should be able to supply the exact



number of LTD cycles performed on the airstrip.



     Precision = + 0




3.2.15.3  Construction Sites - This category includes construc-



tion sites and other areas subject to severe soil erosion.  These



areas are estimated by the appropriate state officials, presumably



to the nearest 5000 acres with precision.





               precision =
                           No.in field



3.2.15.4  Rock Handling and Storage - These are sources too



small to qualify as point sources.  For rock-handling operations,



state offices should have registration forms giving thruput
                                3-69

-------
values.  Each source should report thruput precisely at least
to the nearest thousand tons.  The precision of the area source
totals then depends on the units reported.  The figure is rounded
to the nearest thousand tons, giving precision equal to:
                      ±0.5
                   No. in field
3.2.16  Miscellaneous Sources
     Sources in this category are characterized by intermittent
emissions of fugitive dust.  Techniques for collection of data
for estimating emissions generally do not yield precise values.
3.2.16.1  Forest Fires -
     Area - Data on areas involved in forest fires are obtained
from the U.S. Forest Service or from state forest aaencies,
which report acreage to the nearest acre.  It is doubtful that
these acres are measured precisely, without the use of aerial
photography, which yields the most precise value for affected
acreage but is not used by most state forest agencies.  Other
methods of measurement involve estimation of the extent of the
fire by personnel using maps or other indicators to determine
dimensions of the burned area.  These estimates will produce
a precise figure if the outer limits of the burned area are
determined carefully.  Assuming this is the case, the area value
should be precise within + 0.05.
     Quantity  (tons/acre) - A value for tons of wood per acre
is also obtained from the U.S. or state forestry services.
If data are not available, use 40 tons/acre for heavily forested
                                 3-70

-------
areas or 20 tons/acre for thinly forested areas.  Since no data



are at hand on which to base a precision value, assume that



the density of forestation is precise within + 0.15.



3.2.16.2  Slash Burning -



     Area - This category includes areas of slash burning and



agricultural field burning.  Values for slash burning areas



can be obtained from the U.S. Forest Service or the state Forestry



Department.



     Values for agricultural field burning area must be estimated



by the state Agricultural Department.  Precision values vary



from county to county dependent on:  1) whether burning is exten-



sive in the locale; 2) significant figures reported; 3) sources



of the data.  The best precision estimate must be based on several



assumptions to be applied to all counties.



     Where the total slash burning area in the county is less



than 100 acres, assume the number is precise within + 50 percent.



Where the total burning area is 100 acres or greater, assume



the figure is precise within 50 acres.




                   .  .           ±50
               precision  =  jj-—•   -> *.,
               r             No. in field




     Quantity (tons/acre) - The quantity of timber that is slash-



burned should be obtainable from the U.S. Forest Service or



the state Forestry Department.  If it is not, use 75 tons/acre.



     For agricultural field burning area use a state Agricultural



Department estimate or a factor of 2.5 tons/acre.



     For counties in which both slash and agricultural field
                                3-71

-------
burning are practiced, use a weighted average.

     Again, a precision value applicable to all counties is

determined on the basis of these assumptions:

     1.  Assume that all counties with a figure in
         this field have both slash and agricultural
         field burning.

     2.  Assume the same precision for values obtained
         for slash and agricultural field burning:  + 0.25.


3.2.16.3  Frost Control

     Number of orchard heaters - County or state Agricultural

Departments probably can indicate the number of orchard heaters

in use.  Local orchard operators can confirm this information

or give a better figure.  Assuming that orchard operators are

willing to supply this information, the figure should be precise

within the nearest hundred as reported on the NEDS area source

coding form.


               precision  =
                             No. of orchard heaters


     Firing days per year - Determine this figure by a repre-

sentative sampling of orchard grove operators.  Assume that

all heaters in a county are fired for the same amount of time

and that the number reported is precise at least within 10 days:


               precision  =  dayS5per year


3.2.16.4  Structure fires - Local fire control agencies can

supply the number of structural fires.  Assume that the number

of unrecorded or unreported fires is 5 or less:  precision is
                                3-72

-------
then equal to No.^f°fires .
3.2.16.5  Coal refuse burning - Coal refuse piles consist of
below-standard coal that has been discarded.  Ignition of these
piles and smoldering cause significant air pollution.
     Size of bank - Three dimensions of the bank are estimated
to obtain the volume.  The area factor should be precise with
10 percent, since maps can usually give detailed outlines of
the bank.  Estimation of depth of the bank introduces more error,
since depths vary according to the terrain on which burning
occurs.  Estimate that the depth factor is precise within 25
percent.  Precision is then calculated as follows:
     P2 = (0.10)2 + (0.25)2
     P  - + 0.27
     Number of burning coal refuse piles - The state air pollution
control agency should provide an exact number of coal refuse
burning operations:
               precision  =  +0

3.3  PRECISION OF AUTOMOTIVE EMISSION FACTORS
     As with emissions from point sources and from other area
sources, we consider emissions from automotive vehicles in terms
of priority and non-priority pollutants.  We did not make a
detailed study and determination of precision of emission factors
for non-priority pollutants from motor vehicles.  Instead, we
apply to this category the average precisions calculated for
emission factors for non-priority pollutants from point sources.
                                3-73

-------
 •These precision values were calculated in Section 2.1.3 and



 are summarized below:




                       Part.       SO      HC     CO
                          •••" •"•       "~""X     -       ——


   Average precision   0.211       0.146   0.237  0.357



 Precisions of all automotive emission factors are summarized



 in Table 3.4.



   Table 3.4.  PRECISIONS OF MOTOR VEHICLE EMISSION FACTORS
Road Type
Urban
Suburban
Rural
Limited Accc
Vehicle
classification
LDG
HDG
HDD
LDG
HDG
HDD
LDG
HDG
HDD
iss LDG
HDG
HDD
Precision
Part.
0.211
0.211
0.211
0.211
0.211
0.211
0.211
0.211
0.211
0.211
0.211
0.211
S0x
0.146
0.146
0.146
0.146
0.146
0.146
0.146
0.146
0.146
0.146
0.146
0.146
NO
A
0.028
0.028
0.035
0.028
0.028
0.035
0.028
0.028
0.035
0.028
0.028
0.035
HC
0.025
0.019
0.237
0.024
0.019
0.237
0.024
0.018
0.237
0.023
0.018
0.237
CO
0.036
0.034
0.357
0.036
0.034
0.357
0.036
0.035
0.357
0.036
0.035
0.357
      The balance of this section is devoted to calculations of pre-



 cisions of the priority pollutant emission factors as they are



,related to several categories of motor vehicles.   The priority



 pollutants for light-duty and heavy-duty gasoline-powered highway



 vehicles are NO , HC,  and CO.  For heavy-duty, diesel-powered
                X


 highway vehicles the priority pollutant is NOx.



      Because it is not possible to determine the exact process




 by which the emission factors were calculated for use in the



 NEDS program, several assumptions are necessary.  Using the
                                 3-74

-------
  data and  procedure outlined  in AP-42, we  assume that the  1-year-

  old  vehicles in  the population are  of the 1973  model year and

  apply the emission rates and deterioration factors designated

  for  low-altitude areas  (excluding California) .

  3-3-1  Light-Duty,  Gasoline-Powered  Highway Vehicles^

        Emission factors for light-duty,  gasoline-powered highway

  vehicles  (LOG)  are  calculated  by the method presented in  Section

  3.1.2  of AP-42,  in  which LOG exhaust emission factors for NO ,
                                                                        X
  HC, and CO are  expressed mathematically as:
                                 n+ I
                             enp = Z-     Cj d, itij $j
                                 i = n-12
         enp = Emission factor in grams per vehicle mile for calendar year (n), and pollutant (p)

         Cj =The 1975 Federal test procedure emission rate for pollutant (p) in g/mi for the i1*1 model year
            at low mileage


         d; =The controlled vehicle pollutant  (p) emission deterioration factor for the itn model year
            at calendar year (n)

         mj =The weighted annual travel of the i"1 model year during calendar year (n) The determination
            of this variable involves the use of the vehicle model year distribution

          Sj = The weighted speed adjustment factor for the i1'1 model year vehicles


       The factor  for evaporative and  crankcase emissions  of HC

 is calculated and added to  the exhaust HC emission  factor to

 give  the total HC emissions from LDG's.   The following expression

 for calculating  evaporative and crankcase HC emissions is  also

 from  Section 3.1.2 of AP-42:

           n+l
             n-!2
when:    fn * The combined evaporative and crankcase hydrocarbon emission factor for calendar year (n)

        hj s The combined evaporative and ccankcase emission rate for the i'" model year

        nt| • The weighted annual travel of the 1th model year during calendar year (n)
                                      3-75

-------
     In order to calculate the precision of the emission factor,

the precisions of c., d., m., s., and h. must first be determined.

3.3.1.1  Deterioration factor, d^ - Factors to account for deteri-

oration of exhaust control devices were obtained from Table

3.1.2-5 of AP-42.  To calculate the precision of the deteriora-

tion factors (and also of other terms used to calculate the

emission factor)  we consulted the references used to develop

the factors.  A report of the California Air Resources Labora-

tory   is cited with respect to calculation of deterioration

factors.  This report presents graphs of HC and CO emissions

versus mileage for California automobiles of model years 1966

through 1970.  Deterioration factors for each model year are

calculated by dividing the emissions at a certain vehicle age

(as determined by mileage levels)  by the emissions at age zero.

The relationship between age and mileage3 is given as:

 Vehicle Age (Years)     0123456

 Mileage                4.0   17.5  33.6  46.8  58.2  69.9  79.9
 (thousands of miles)

     In addition to the graphs, the report gives gross regres-

sion data by model year and manufacturer at various mileages.

These data  (value, upper and lower limits) are interpreted to

mean that the upper limit minus the value, divided by two equals

the standard deviation of the value.  We calculated a weighted

average of the standard deviation using the weighting factors

given in the reference as a basis for the graphs of emissions

versus maleage:
a) Data obtained from D. Kircher, U.S. EPA, March  9,  1973.
                                 3-76

-------
                                   Fraction of total sales

     Chrysler                              0.16

     Ford                                  0.34
     General Motors                        0.50


The weighted standard deviation was then based on this formula:


     Vmileage =  [*+ »f» *  *+ «g>*  *

where W , Wf, and W  = weighting factors for Chrysler, Ford,
       c           ^   and GM, respectively.

      S , Sf, and S  = standard deviations for Chrysler, Ford,
                   '   and GM, respectively.


The variance at X-miles was:

     Var (X-miles)  = fs  - miles "I
                     L x         j

Since the formula for the deterioration factor, d, is

                Value x-miles
      d
                Value4000
the formula for variance of the deterioration factor, Var(d), is;


     Var(d) =  (d)2 f  Var(4000)  +  Var  (x-miles)
                   ["
                   L
                     (Value4oooT2


Table 3.5 summarizes the variance of deterioration factors;

Tables 3.6 through 3.11 show the calculations performed  for  model

years 1968 through 1970.

     Reference 31 does not provide enough data for calculation

of precision of the deterioration factors for NO  emissions.

It presents a graph of 1971 model year vehicle emissions versus

mileage, but gives no gross regression data for that graph.

EPA used the 1971 California NOV data in the report in compiling
                               X

1973 nationwide (except California) NO  deterioration factors.
                                      X

We estimated that the average precisions of the HC and CO deteri-

oration factors for 1973 model cars at age 1 year would give
                                3-77

-------






y
M
4J


flj
•g
w












o
o








^»
•H N
•O -»
^** "H
M T3



-H
•o

M
n)


•H
•o


-H tN
t) ^
*^ -H
M "O


•H
•a




•a
fH
V M
•o «
O 0)

t)
o fH in o ^> r«
vc oo o o rn 04
O O l-H r-l O O
o o o o o o
ooooooooooooo
ooooooooooooo


vo oo •v m o "»•
v£ o\ ro ro in ^*
O O iH r-( O O
o o o o o o
ooooooooooooo
ooooooooooooo
inonmnooooooooo
O«HrHr-lf>JfNOOOOOOO

eg r« rj> n oo r-i
 vo (N r-i
o o o o o o
ooooooooooooo
ooooooooooooo
cooiaoooocoooooooo
i-inn^koinooooooo
i-JfHrHHi-lrHtHi-liHr-»iHlHl-l
-.
r* r*** P** r** \o ^o vo vo ^o KD VD vo vo 'O


,-t H i-l i-H
3-78

-------
o


m
CO

T3
M
0)
a
a
Value
o
w
T3
M
0)
CX
04


O
in
r-
o
o
in
rH
o
CN
rH
rH
r-
cn
o
m
CN
rH
O
in
CN
O
CN
O
rH
r-
r~
o
o
o
00
VO
T
CO
in
t-
0
•
0
in
rH
O
*!•
rH
rH
en
en
o
CO
o
0
VO
rH
o
VO
rH
rH
O
0
rH
in
CN
rH
O
in
CN
O
in
0
•
rH
0
CO
o
o
o
fN
GO
in
T
3
rl
ra

c
o
4-1
(0 -
0 0
•H 4J
rl U
(U rO
i 1 \l j
(U
Q
-a
0) 0)
4J 3
-C rH
tj1 fO
•H >
a)
S
X
rl
rt)


X
en

(U
i-H
U (U
•H CP
J5 10
0)

o
o



o
0

'""'






CN
I--
O

,00038
o
o
CN
O
0

0


a\
0
o
o
0



CO
rH

"""*






in
00
o

.00108
o
CO
n
0
o

rH


rH
in
O
0
0



CN
ro

rH






in
a\
o

,00202 ]
o
in
^3*
O
O


C
rd

0)
3
rH
(0
£>

1
(!)
cu
Q,
2-


g
•o
,,
0)
rl
0)
0
in
o
II

IT*
S
*
^1*
ro
•
o

II
>M
S

VO
i-H
0
II

U
S
C
<0
1

CN
v^
rH
. .
fN
CP
CN
3

+
CN
VM
10
CN
VM
U

1
CN L
u
10
CN
U

II

X
to

II
CN
>i^
rH
"x
*— '
M
ITJ


g
§
i-H
(0
J>

•
EH
S

O
0
o
0)
3
r-l
10
^

J
*
II
>0
•a

c
0)
•C
5


1

"g
^^
(^
nj
^

N
"~g
0)
3
rH
(0


+

O
o
0
rf
•*-•
^1
(0
>

N
O
O
O
^*
0)
3
i-H
•0
1
i


CN
10
•o
II


^18
S
(^
10
>




                                               3-79

-------
Ok
8
in o





§









'O
o
(u









rl
0)
rH
0>
J*
^
U




tr
to


e
•a
M
Q)
a
a
D
3
id
CO


E
•a

M
0)
a
a
D
(U
3
rH
id
>
u
CO


E
T3

M

(U
D^
id
0)
^ J
•H
S
0)
U (U
*H 5*
a)
CM
O
0
in
o
o
o
tN
*

in
rH
H'
CN
O
O
T
O
0

•w
^
o

o
f-
o
o
o
o
CO
o
o


r-
u
o

o\
CO

0

o
o
o
*s^



o

in
o
o
o
H
O
?
*

3
H*
in
o
0
o
i-H
O

r-
,_4

rH

r-
o
•
rH
in
00
o
o
f^
H
O


rH
4
rH

1*
a
rH

O
O
en
en
u>


m














































































































































^^
^

r4
>

C
0
-rH -0
m -
M M
o o
•H 4J
Vl U


^<
id
X
to


0)
1
U QJ
1*









O
o
,—4






0
°)
0
0.000243 J
10
rH
O
o


0
o
(M
(N
O
O
o



CO
VO
1






rH
in
rH
rH
rH
O
O
0
ro
m
o
o


in








































































































El™
-0 1
II
u>
to

•a
c
1C

^
<4H
to

^
y
to

-a
c


0)
3
i— i
fd
^>

i
M
0)
a
£}
^-^
u
6
•0
,.
0)
(U
o
in
0
H
o*
u

%
*4*
n
•
o
u
i[ [

U3
rH
o
II

0
2
•o
c
1

H
t»
II
id
•o

c
0)
x:




•«-'
(^
<



i
(U
a
rH
id
5*
••*•

H-


^^
0
o
0
^f
*~*
IH
id
>

CN
^^^
O
o
o
^"
OJ
3
rH
10
*" 1






                                          3-80

-------
oo
vo
8
•
•0
H
O
tn
Chrysler
&
(A
•d
M
0)
(X
Q,
D
Value
VH
w
T3
M
0)
a
a
D
Value
0
w
J
M
(1)
O,
O,
D
Value
Mileage
Vehicle
age
in


r-
vo











































s
k
ra

c
o
-H -a
-U
ffl -
M r-l
o o
•H 4J
n o
a) m
•M VM

Q)
*X
i-l
(0
>

X
C/l

0)
1 J
u a)
•H O>
x: 

1
k-l
0)
CU
Ch
D


II

£
-a
o
in
•
o
II

0>

«
•*•
M
•
O
II
VM
S
VO
rH
O
II

U
3
•o
C
1 rt
1
M
\
rH
f"~~ "1
CM
tr
w
CM
IT
»^
»s
+
CM
H-l
w
CM
VM
S
+
CM w
0
C/l
CM
O
t S ,

II

X
w

II

CM
\
rH



E
§
r-l

1 ^H'
1
id
•o

c
V
3




£
M
fO



^
V
3
rH
ia
>


+

o
o
0
T
^^
M
(0
>

CM
O
0
0
*r
0)
3
rH
rt)
>^
^^ 1


CM
*
•o

II


,-*,
f3
T3
*—^
U
ifl
>
                                            3-81